diff options
278 files changed, 13139 insertions, 7436 deletions
diff --git a/Documentation/DocBook/gpu.tmpl b/Documentation/DocBook/gpu.tmpl index a8669330b456..fe6b36a2fd98 100644 --- a/Documentation/DocBook/gpu.tmpl +++ b/Documentation/DocBook/gpu.tmpl @@ -2886,52 +2886,8 @@ void (*postclose) (struct drm_device *, struct drm_file *);</synopsis> </sect2> <sect2> <title>File Operations</title> - <synopsis>const struct file_operations *fops</synopsis> - <abstract>File operations for the DRM device node.</abstract> - <para> - Drivers must define the file operations structure that forms the DRM - userspace API entry point, even though most of those operations are - implemented in the DRM core. The <methodname>open</methodname>, - <methodname>release</methodname> and <methodname>ioctl</methodname> - operations are handled by - <programlisting> - .owner = THIS_MODULE, - .open = drm_open, - .release = drm_release, - .unlocked_ioctl = drm_ioctl, - #ifdef CONFIG_COMPAT - .compat_ioctl = drm_compat_ioctl, - #endif - </programlisting> - </para> - <para> - Drivers that implement private ioctls that requires 32/64bit - compatibility support must provide their own - <methodname>compat_ioctl</methodname> handler that processes private - ioctls and calls <function>drm_compat_ioctl</function> for core ioctls. - </para> - <para> - The <methodname>read</methodname> and <methodname>poll</methodname> - operations provide support for reading DRM events and polling them. They - are implemented by - <programlisting> - .poll = drm_poll, - .read = drm_read, - .llseek = no_llseek, - </programlisting> - </para> - <para> - The memory mapping implementation varies depending on how the driver - manages memory. Pre-GEM drivers will use <function>drm_mmap</function>, - while GEM-aware drivers will use <function>drm_gem_mmap</function>. See - <xref linkend="drm-gem"/>. - <programlisting> - .mmap = drm_gem_mmap, - </programlisting> - </para> - <para> - No other file operation is supported by the DRM API. - </para> +!Pdrivers/gpu/drm/drm_fops.c file operations +!Edrivers/gpu/drm/drm_fops.c </sect2> <sect2> <title>IOCTLs</title> @@ -3319,6 +3275,12 @@ int num_ioctls;</synopsis> !Pdrivers/gpu/drm/i915/intel_csr.c csr support for dmc !Idrivers/gpu/drm/i915/intel_csr.c </sect2> + <sect2> + <title>Video BIOS Table (VBT)</title> +!Pdrivers/gpu/drm/i915/intel_bios.c Video BIOS Table (VBT) +!Idrivers/gpu/drm/i915/intel_bios.c +!Idrivers/gpu/drm/i915/intel_bios.h + </sect2> </sect1> <sect1> @@ -3460,6 +3422,7 @@ int num_ioctls;</synopsis> </sect1> <sect1> <title>Public constants</title> +!Finclude/linux/vga_switcheroo.h vga_switcheroo_handler_flags_t !Finclude/linux/vga_switcheroo.h vga_switcheroo_client_id !Finclude/linux/vga_switcheroo.h vga_switcheroo_state </sect1> @@ -3488,6 +3451,10 @@ int num_ioctls;</synopsis> <title>Backlight control</title> !Pdrivers/platform/x86/apple-gmux.c Backlight control </sect2> + <sect2> + <title>Public functions</title> +!Iinclude/linux/apple-gmux.h + </sect2> </sect1> </chapter> diff --git a/Documentation/devicetree/bindings/display/rockchip/inno_hdmi-rockchip.txt b/Documentation/devicetree/bindings/display/rockchip/inno_hdmi-rockchip.txt new file mode 100644 index 000000000000..8096a29f9776 --- /dev/null +++ b/Documentation/devicetree/bindings/display/rockchip/inno_hdmi-rockchip.txt @@ -0,0 +1,50 @@ +Rockchip specific extensions to the Innosilicon HDMI +================================ + +Required properties: +- compatible: + "rockchip,rk3036-inno-hdmi"; +- reg: + Physical base address and length of the controller's registers. +- clocks, clock-names: + Phandle to hdmi controller clock, name should be "pclk" +- interrupts: + HDMI interrupt number +- ports: + Contain one port node with endpoint definitions as defined in + Documentation/devicetree/bindings/graph.txt. +- pinctrl-0, pinctrl-name: + Switch the iomux of HPD/CEC pins to HDMI function. + +Example: +hdmi: hdmi@20034000 { + compatible = "rockchip,rk3036-inno-hdmi"; + reg = <0x20034000 0x4000>; + interrupts = <GIC_SPI 45 IRQ_TYPE_LEVEL_HIGH>; + clocks = <&cru PCLK_HDMI>; + clock-names = "pclk"; + pinctrl-names = "default"; + pinctrl-0 = <&hdmi_ctl>; + status = "disabled"; + + hdmi_in: port { + #address-cells = <1>; + #size-cells = <0>; + hdmi_in_lcdc: endpoint@0 { + reg = <0>; + remote-endpoint = <&lcdc_out_hdmi>; + }; + }; +}; + +&pinctrl { + hdmi { + hdmi_ctl: hdmi-ctl { + rockchip,pins = <1 8 RK_FUNC_1 &pcfg_pull_none>, + <1 9 RK_FUNC_1 &pcfg_pull_none>, + <1 10 RK_FUNC_1 &pcfg_pull_none>, + <1 11 RK_FUNC_1 &pcfg_pull_none>; + }; + }; + +}; diff --git a/Documentation/dma-buf-sharing.txt b/Documentation/dma-buf-sharing.txt index 480c8de3c2c4..32ac32e773e1 100644 --- a/Documentation/dma-buf-sharing.txt +++ b/Documentation/dma-buf-sharing.txt @@ -257,17 +257,15 @@ Access to a dma_buf from the kernel context involves three steps: Interface: int dma_buf_begin_cpu_access(struct dma_buf *dmabuf, - size_t start, size_t len, enum dma_data_direction direction) This allows the exporter to ensure that the memory is actually available for cpu access - the exporter might need to allocate or swap-in and pin the backing storage. The exporter also needs to ensure that cpu access is - coherent for the given range and access direction. The range and access - direction can be used by the exporter to optimize the cache flushing, i.e. - access outside of the range or with a different direction (read instead of - write) might return stale or even bogus data (e.g. when the exporter needs to - copy the data to temporary storage). + coherent for the access direction. The direction can be used by the exporter + to optimize the cache flushing, i.e. access with a different direction (read + instead of write) might return stale or even bogus data (e.g. when the + exporter needs to copy the data to temporary storage). This step might fail, e.g. in oom conditions. @@ -322,14 +320,13 @@ Access to a dma_buf from the kernel context involves three steps: 3. Finish access - When the importer is done accessing the range specified in begin_cpu_access, - it needs to announce this to the exporter (to facilitate cache flushing and - unpinning of any pinned resources). The result of any dma_buf kmap calls - after end_cpu_access is undefined. + When the importer is done accessing the CPU, it needs to announce this to + the exporter (to facilitate cache flushing and unpinning of any pinned + resources). The result of any dma_buf kmap calls after end_cpu_access is + undefined. Interface: void dma_buf_end_cpu_access(struct dma_buf *dma_buf, - size_t start, size_t len, enum dma_data_direction dir); @@ -353,7 +350,26 @@ Being able to mmap an export dma-buf buffer object has 2 main use-cases: handles, too). So it's beneficial to support this in a similar fashion on dma-buf to have a good transition path for existing Android userspace. - No special interfaces, userspace simply calls mmap on the dma-buf fd. + No special interfaces, userspace simply calls mmap on the dma-buf fd, making + sure that the cache synchronization ioctl (DMA_BUF_IOCTL_SYNC) is *always* + used when the access happens. This is discussed next paragraphs. + + Some systems might need some sort of cache coherency management e.g. when + CPU and GPU domains are being accessed through dma-buf at the same time. To + circumvent this problem there are begin/end coherency markers, that forward + directly to existing dma-buf device drivers vfunc hooks. Userspace can make + use of those markers through the DMA_BUF_IOCTL_SYNC ioctl. The sequence + would be used like following: + - mmap dma-buf fd + - for each drawing/upload cycle in CPU 1. SYNC_START ioctl, 2. read/write + to mmap area 3. SYNC_END ioctl. This can be repeated as often as you + want (with the new data being consumed by the GPU or say scanout device) + - munmap once you don't need the buffer any more + + Therefore, for correctness and optimal performance, systems with the memory + cache shared by the GPU and CPU i.e. the "coherent" and also the + "incoherent" are always required to use SYNC_START and SYNC_END before and + after, respectively, when accessing the mapped address. 2. Supporting existing mmap interfaces in importers diff --git a/MAINTAINERS b/MAINTAINERS index f07592658119..5673421670a6 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -827,6 +827,12 @@ S: Maintained F: drivers/net/arcnet/ F: include/uapi/linux/if_arcnet.h +ARM HDLCD DRM DRIVER +M: Liviu Dudau <[email protected]> +S: Supported +F: drivers/gpu/drm/arm/ +F: Documentation/devicetree/bindings/display/arm,hdlcd.txt + ARM MFM AND FLOPPY DRIVERS M: Ian Molton <[email protected]> S: Maintained diff --git a/drivers/dma-buf/dma-buf.c b/drivers/dma-buf/dma-buf.c index 155c1464948e..9810d1df0691 100644 --- a/drivers/dma-buf/dma-buf.c +++ b/drivers/dma-buf/dma-buf.c @@ -34,6 +34,8 @@ #include <linux/poll.h> #include <linux/reservation.h> +#include <uapi/linux/dma-buf.h> + static inline int is_dma_buf_file(struct file *); struct dma_buf_list { @@ -251,11 +253,54 @@ out: return events; } +static long dma_buf_ioctl(struct file *file, + unsigned int cmd, unsigned long arg) +{ + struct dma_buf *dmabuf; + struct dma_buf_sync sync; + enum dma_data_direction direction; + + dmabuf = file->private_data; + + switch (cmd) { + case DMA_BUF_IOCTL_SYNC: + if (copy_from_user(&sync, (void __user *) arg, sizeof(sync))) + return -EFAULT; + + if (sync.flags & ~DMA_BUF_SYNC_VALID_FLAGS_MASK) + return -EINVAL; + + switch (sync.flags & DMA_BUF_SYNC_RW) { + case DMA_BUF_SYNC_READ: + direction = DMA_FROM_DEVICE; + break; + case DMA_BUF_SYNC_WRITE: + direction = DMA_TO_DEVICE; + break; + case DMA_BUF_SYNC_RW: + direction = DMA_BIDIRECTIONAL; + break; + default: + return -EINVAL; + } + + if (sync.flags & DMA_BUF_SYNC_END) + dma_buf_end_cpu_access(dmabuf, direction); + else + dma_buf_begin_cpu_access(dmabuf, direction); + + return 0; + default: + return -ENOTTY; + } +} + static const struct file_operations dma_buf_fops = { .release = dma_buf_release, .mmap = dma_buf_mmap_internal, .llseek = dma_buf_llseek, .poll = dma_buf_poll, + .unlocked_ioctl = dma_buf_ioctl, }; /* @@ -539,13 +584,11 @@ EXPORT_SYMBOL_GPL(dma_buf_unmap_attachment); * preparations. Coherency is only guaranteed in the specified range for the * specified access direction. * @dmabuf: [in] buffer to prepare cpu access for. - * @start: [in] start of range for cpu access. - * @len: [in] length of range for cpu access. * @direction: [in] length of range for cpu access. * * Can return negative error values, returns 0 on success. */ -int dma_buf_begin_cpu_access(struct dma_buf *dmabuf, size_t start, size_t len, +int dma_buf_begin_cpu_access(struct dma_buf *dmabuf, enum dma_data_direction direction) { int ret = 0; @@ -554,8 +597,7 @@ int dma_buf_begin_cpu_access(struct dma_buf *dmabuf, size_t start, size_t len, return -EINVAL; if (dmabuf->ops->begin_cpu_access) - ret = dmabuf->ops->begin_cpu_access(dmabuf, start, - len, direction); + ret = dmabuf->ops->begin_cpu_access(dmabuf, direction); return ret; } @@ -567,19 +609,17 @@ EXPORT_SYMBOL_GPL(dma_buf_begin_cpu_access); * actions. Coherency is only guaranteed in the specified range for the * specified access direction. * @dmabuf: [in] buffer to complete cpu access for. - * @start: [in] start of range for cpu access. - * @len: [in] length of range for cpu access. * @direction: [in] length of range for cpu access. * * This call must always succeed. */ -void dma_buf_end_cpu_access(struct dma_buf *dmabuf, size_t start, size_t len, +void dma_buf_end_cpu_access(struct dma_buf *dmabuf, enum dma_data_direction direction) { WARN_ON(!dmabuf); if (dmabuf->ops->end_cpu_access) - dmabuf->ops->end_cpu_access(dmabuf, start, len, direction); + dmabuf->ops->end_cpu_access(dmabuf, direction); } EXPORT_SYMBOL_GPL(dma_buf_end_cpu_access); diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig index 8ae7ab68cb97..f2a74d0b68ae 100644 --- a/drivers/gpu/drm/Kconfig +++ b/drivers/gpu/drm/Kconfig @@ -25,6 +25,14 @@ config DRM_MIPI_DSI bool depends on DRM +config DRM_DP_AUX_CHARDEV + bool "DRM DP AUX Interface" + depends on DRM + help + Choose this option to enable a /dev/drm_dp_auxN node that allows to + read and write values to arbitrary DPCD registers on the DP aux + channel. + config DRM_KMS_HELPER tristate depends on DRM @@ -106,6 +114,8 @@ config DRM_TDFX Choose this option if you have a 3dfx Banshee or Voodoo3 (or later), graphics card. If M is selected, the module will be called tdfx. +source "drivers/gpu/drm/arm/Kconfig" + config DRM_R128 tristate "ATI Rage 128" depends on DRM && PCI @@ -162,6 +172,8 @@ config DRM_AMDGPU source "drivers/gpu/drm/amd/amdgpu/Kconfig" source "drivers/gpu/drm/amd/powerplay/Kconfig" +source "drivers/gpu/drm/amd/acp/Kconfig" + source "drivers/gpu/drm/nouveau/Kconfig" config DRM_I810 diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile index 61766dec6a8d..6eb94fc561dc 100644 --- a/drivers/gpu/drm/Makefile +++ b/drivers/gpu/drm/Makefile @@ -22,10 +22,13 @@ drm-$(CONFIG_OF) += drm_of.o drm-$(CONFIG_AGP) += drm_agpsupport.o drm_kms_helper-y := drm_crtc_helper.o drm_dp_helper.o drm_probe_helper.o \ - drm_plane_helper.o drm_dp_mst_topology.o drm_atomic_helper.o + drm_plane_helper.o drm_dp_mst_topology.o drm_atomic_helper.o \ + drm_kms_helper_common.o + drm_kms_helper-$(CONFIG_DRM_LOAD_EDID_FIRMWARE) += drm_edid_load.o drm_kms_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fb_helper.o drm_kms_helper-$(CONFIG_DRM_KMS_CMA_HELPER) += drm_fb_cma_helper.o +drm_kms_helper-$(CONFIG_DRM_DP_AUX_CHARDEV) += drm_dp_aux_dev.o obj-$(CONFIG_DRM_KMS_HELPER) += drm_kms_helper.o @@ -33,6 +36,7 @@ CFLAGS_drm_trace_points.o := -I$(src) obj-$(CONFIG_DRM) += drm.o obj-$(CONFIG_DRM_MIPI_DSI) += drm_mipi_dsi.o +obj-$(CONFIG_DRM_ARM) += arm/ obj-$(CONFIG_DRM_TTM) += ttm/ obj-$(CONFIG_DRM_TDFX) += tdfx/ obj-$(CONFIG_DRM_R128) += r128/ diff --git a/drivers/gpu/drm/amd/acp/Kconfig b/drivers/gpu/drm/amd/acp/Kconfig new file mode 100644 index 000000000000..2b07813bceed --- /dev/null +++ b/drivers/gpu/drm/amd/acp/Kconfig @@ -0,0 +1,11 @@ +menu "ACP Configuration" + +config DRM_AMD_ACP + bool "Enable ACP IP support" + default y + select MFD_CORE + select PM_GENERIC_DOMAINS if PM + help + Choose this option to enable ACP IP support for AMD SOCs. + +endmenu diff --git a/drivers/gpu/drm/amd/acp/Makefile b/drivers/gpu/drm/amd/acp/Makefile new file mode 100644 index 000000000000..8363cb57915b --- /dev/null +++ b/drivers/gpu/drm/amd/acp/Makefile @@ -0,0 +1,8 @@ +# +# Makefile for the ACP, which is a sub-component +# of AMDSOC/AMDGPU drm driver. +# It provides the HW control for ACP related functionalities. + +subdir-ccflags-y += -I$(AMDACPPATH)/ -I$(AMDACPPATH)/include + +AMD_ACP_FILES := $(AMDACPPATH)/acp_hw.o diff --git a/drivers/gpu/drm/amd/acp/acp_hw.c b/drivers/gpu/drm/amd/acp/acp_hw.c new file mode 100644 index 000000000000..7af83f142b4b --- /dev/null +++ b/drivers/gpu/drm/amd/acp/acp_hw.c @@ -0,0 +1,50 @@ +/* + * Copyright 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + */ + +#include <linux/mm.h> +#include <linux/slab.h> +#include <linux/device.h> +#include <linux/delay.h> +#include <linux/errno.h> + +#include "acp_gfx_if.h" + +#define ACP_MODE_I2S 0 +#define ACP_MODE_AZ 1 + +#define mmACP_AZALIA_I2S_SELECT 0x51d4 + +int amd_acp_hw_init(void *cgs_device, + unsigned acp_version_major, unsigned acp_version_minor) +{ + unsigned int acp_mode = ACP_MODE_I2S; + + if ((acp_version_major == 2) && (acp_version_minor == 2)) + acp_mode = cgs_read_register(cgs_device, + mmACP_AZALIA_I2S_SELECT); + + if (acp_mode != ACP_MODE_I2S) + return -ENODEV; + + return 0; +} diff --git a/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h b/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h new file mode 100644 index 000000000000..bccf47b63899 --- /dev/null +++ b/drivers/gpu/drm/amd/acp/include/acp_gfx_if.h @@ -0,0 +1,34 @@ +/* + * Copyright 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * +*/ + +#ifndef _ACP_GFX_IF_H +#define _ACP_GFX_IF_H + +#include <linux/types.h> +#include "cgs_linux.h" +#include "cgs_common.h" + +int amd_acp_hw_init(void *cgs_device, + unsigned acp_version_major, unsigned acp_version_minor); + +#endif /* _ACP_GFX_IF_H */ diff --git a/drivers/gpu/drm/amd/amdgpu/Makefile b/drivers/gpu/drm/amd/amdgpu/Makefile index 20c9539abc36..c7fcdcedaadb 100644 --- a/drivers/gpu/drm/amd/amdgpu/Makefile +++ b/drivers/gpu/drm/amd/amdgpu/Makefile @@ -8,7 +8,8 @@ ccflags-y := -Iinclude/drm -I$(FULL_AMD_PATH)/include/asic_reg \ -I$(FULL_AMD_PATH)/include \ -I$(FULL_AMD_PATH)/amdgpu \ -I$(FULL_AMD_PATH)/scheduler \ - -I$(FULL_AMD_PATH)/powerplay/inc + -I$(FULL_AMD_PATH)/powerplay/inc \ + -I$(FULL_AMD_PATH)/acp/include amdgpu-y := amdgpu_drv.o @@ -20,7 +21,7 @@ amdgpu-y += amdgpu_device.o amdgpu_kms.o \ amdgpu_fb.o amdgpu_gem.o amdgpu_ring.o \ amdgpu_cs.o amdgpu_bios.o amdgpu_benchmark.o amdgpu_test.o \ amdgpu_pm.o atombios_dp.o amdgpu_afmt.o amdgpu_trace_points.o \ - atombios_encoders.o amdgpu_semaphore.o amdgpu_sa.o atombios_i2c.o \ + atombios_encoders.o amdgpu_sa.o atombios_i2c.o \ amdgpu_prime.o amdgpu_vm.o amdgpu_ib.o amdgpu_pll.o \ amdgpu_ucode.o amdgpu_bo_list.o amdgpu_ctx.o amdgpu_sync.o @@ -92,7 +93,17 @@ amdgpu-y += amdgpu_cgs.o amdgpu-y += \ ../scheduler/gpu_scheduler.o \ ../scheduler/sched_fence.o \ - amdgpu_sched.o + amdgpu_job.o + +# ACP componet +ifneq ($(CONFIG_DRM_AMD_ACP),) +amdgpu-y += amdgpu_acp.o + +AMDACPPATH := ../acp +include $(FULL_AMD_PATH)/acp/Makefile + +amdgpu-y += $(AMD_ACP_FILES) +endif amdgpu-$(CONFIG_COMPAT) += amdgpu_ioc32.o amdgpu-$(CONFIG_VGA_SWITCHEROO) += amdgpu_atpx_handler.o diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index 82edf95b7740..f5bac97a438b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -53,6 +53,7 @@ #include "amdgpu_ucode.h" #include "amdgpu_gds.h" #include "amd_powerplay.h" +#include "amdgpu_acp.h" #include "gpu_scheduler.h" @@ -74,7 +75,6 @@ extern int amdgpu_dpm; extern int amdgpu_smc_load_fw; extern int amdgpu_aspm; extern int amdgpu_runtime_pm; -extern int amdgpu_hard_reset; extern unsigned amdgpu_ip_block_mask; extern int amdgpu_bapm; extern int amdgpu_deep_color; @@ -82,10 +82,8 @@ extern int amdgpu_vm_size; extern int amdgpu_vm_block_size; extern int amdgpu_vm_fault_stop; extern int amdgpu_vm_debug; -extern int amdgpu_enable_scheduler; extern int amdgpu_sched_jobs; extern int amdgpu_sched_hw_submission; -extern int amdgpu_enable_semaphores; extern int amdgpu_powerplay; #define AMDGPU_WAIT_IDLE_TIMEOUT_IN_MS 3000 @@ -106,9 +104,6 @@ extern int amdgpu_powerplay; /* max number of IP instances */ #define AMDGPU_MAX_SDMA_INSTANCES 2 -/* number of hw syncs before falling back on blocking */ -#define AMDGPU_NUM_SYNCS 4 - /* hardcode that limit for now */ #define AMDGPU_VA_RESERVED_SIZE (8 << 20) @@ -189,7 +184,6 @@ struct amdgpu_fence; struct amdgpu_ib; struct amdgpu_vm; struct amdgpu_ring; -struct amdgpu_semaphore; struct amdgpu_cs_parser; struct amdgpu_job; struct amdgpu_irq_src; @@ -287,7 +281,7 @@ struct amdgpu_vm_pte_funcs { unsigned count); /* write pte one entry at a time with addr mapping */ void (*write_pte)(struct amdgpu_ib *ib, - uint64_t pe, + const dma_addr_t *pages_addr, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags); /* for linear pte/pde updates without addr mapping */ @@ -295,8 +289,6 @@ struct amdgpu_vm_pte_funcs { uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags); - /* pad the indirect buffer to the necessary number of dw */ - void (*pad_ib)(struct amdgpu_ib *ib); }; /* provided by the gmc block */ @@ -334,9 +326,6 @@ struct amdgpu_ring_funcs { struct amdgpu_ib *ib); void (*emit_fence)(struct amdgpu_ring *ring, uint64_t addr, uint64_t seq, unsigned flags); - bool (*emit_semaphore)(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait); void (*emit_vm_flush)(struct amdgpu_ring *ring, unsigned vm_id, uint64_t pd_addr); void (*emit_hdp_flush)(struct amdgpu_ring *ring); @@ -349,6 +338,8 @@ struct amdgpu_ring_funcs { int (*test_ib)(struct amdgpu_ring *ring); /* insert NOP packets */ void (*insert_nop)(struct amdgpu_ring *ring, uint32_t count); + /* pad the indirect buffer to the necessary number of dw */ + void (*pad_ib)(struct amdgpu_ring *ring, struct amdgpu_ib *ib); }; /* @@ -394,7 +385,7 @@ struct amdgpu_fence_driver { uint64_t gpu_addr; volatile uint32_t *cpu_addr; /* sync_seq is protected by ring emission lock */ - uint64_t sync_seq[AMDGPU_MAX_RINGS]; + uint64_t sync_seq; atomic64_t last_seq; bool initialized; struct amdgpu_irq_src *irq_src; @@ -447,11 +438,6 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring); int amdgpu_fence_wait_empty(struct amdgpu_ring *ring); unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring); -bool amdgpu_fence_need_sync(struct amdgpu_fence *fence, - struct amdgpu_ring *ring); -void amdgpu_fence_note_sync(struct amdgpu_fence *fence, - struct amdgpu_ring *ring); - /* * TTM. */ @@ -470,6 +456,8 @@ struct amdgpu_mman { /* buffer handling */ const struct amdgpu_buffer_funcs *buffer_funcs; struct amdgpu_ring *buffer_funcs_ring; + /* Scheduler entity for buffer moves */ + struct amd_sched_entity entity; }; int amdgpu_copy_buffer(struct amdgpu_ring *ring, @@ -484,8 +472,6 @@ struct amdgpu_bo_list_entry { struct amdgpu_bo *robj; struct ttm_validate_buffer tv; struct amdgpu_bo_va *bo_va; - unsigned prefered_domains; - unsigned allowed_domains; uint32_t priority; }; @@ -522,7 +508,8 @@ struct amdgpu_bo { /* Protected by gem.mutex */ struct list_head list; /* Protected by tbo.reserved */ - u32 initial_domain; + u32 prefered_domains; + u32 allowed_domains; struct ttm_place placements[AMDGPU_GEM_DOMAIN_MAX + 1]; struct ttm_placement placement; struct ttm_buffer_object tbo; @@ -544,7 +531,6 @@ struct amdgpu_bo { struct amdgpu_bo *parent; struct ttm_bo_kmap_obj dma_buf_vmap; - pid_t pid; struct amdgpu_mn *mn; struct list_head mn_list; }; @@ -621,13 +607,7 @@ struct amdgpu_sa_bo { /* * GEM objects. */ -struct amdgpu_gem { - struct mutex mutex; - struct list_head objects; -}; - -int amdgpu_gem_init(struct amdgpu_device *adev); -void amdgpu_gem_fini(struct amdgpu_device *adev); +void amdgpu_gem_force_release(struct amdgpu_device *adev); int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size, int alignment, u32 initial_domain, u64 flags, bool kernel, @@ -639,32 +619,10 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, int amdgpu_mode_dumb_mmap(struct drm_file *filp, struct drm_device *dev, uint32_t handle, uint64_t *offset_p); - -/* - * Semaphores. - */ -struct amdgpu_semaphore { - struct amdgpu_sa_bo *sa_bo; - signed waiters; - uint64_t gpu_addr; -}; - -int amdgpu_semaphore_create(struct amdgpu_device *adev, - struct amdgpu_semaphore **semaphore); -bool amdgpu_semaphore_emit_signal(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore); -bool amdgpu_semaphore_emit_wait(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore); -void amdgpu_semaphore_free(struct amdgpu_device *adev, - struct amdgpu_semaphore **semaphore, - struct fence *fence); - /* * Synchronization */ struct amdgpu_sync { - struct amdgpu_semaphore *semaphores[AMDGPU_NUM_SYNCS]; - struct fence *sync_to[AMDGPU_MAX_RINGS]; DECLARE_HASHTABLE(fences, 4); struct fence *last_vm_update; }; @@ -676,12 +634,9 @@ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct reservation_object *resv, void *owner); -int amdgpu_sync_rings(struct amdgpu_sync *sync, - struct amdgpu_ring *ring); struct fence *amdgpu_sync_get_fence(struct amdgpu_sync *sync); int amdgpu_sync_wait(struct amdgpu_sync *sync); -void amdgpu_sync_free(struct amdgpu_device *adev, struct amdgpu_sync *sync, - struct fence *fence); +void amdgpu_sync_free(struct amdgpu_sync *sync); /* * GART structures, functions & helpers @@ -799,6 +754,7 @@ struct amdgpu_flip_work { struct fence *excl; unsigned shared_count; struct fence **shared; + struct fence_cb cb; }; @@ -811,12 +767,11 @@ struct amdgpu_ib { uint32_t length_dw; uint64_t gpu_addr; uint32_t *ptr; - struct amdgpu_ring *ring; struct amdgpu_fence *fence; struct amdgpu_user_fence *user; + bool grabbed_vmid; struct amdgpu_vm *vm; struct amdgpu_ctx *ctx; - struct amdgpu_sync sync; uint32_t gds_base, gds_size; uint32_t gws_base, gws_size; uint32_t oa_base, oa_size; @@ -835,13 +790,14 @@ enum amdgpu_ring_type { extern struct amd_sched_backend_ops amdgpu_sched_ops; -int amdgpu_sched_ib_submit_kernel_helper(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct amdgpu_ib *ibs, - unsigned num_ibs, - int (*free_job)(struct amdgpu_job *), - void *owner, - struct fence **fence); +int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, + struct amdgpu_job **job); +int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size, + struct amdgpu_job **job); +void amdgpu_job_free(struct amdgpu_job *job); +int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring, + struct amd_sched_entity *entity, void *owner, + struct fence **f); struct amdgpu_ring { struct amdgpu_device *adev; @@ -850,7 +806,6 @@ struct amdgpu_ring { struct amd_gpu_scheduler sched; spinlock_t fence_lock; - struct mutex *ring_lock; struct amdgpu_bo *ring_obj; volatile uint32_t *ring; unsigned rptr_offs; @@ -859,7 +814,7 @@ struct amdgpu_ring { unsigned wptr; unsigned wptr_old; unsigned ring_size; - unsigned ring_free_dw; + unsigned max_dw; int count_dw; uint64_t gpu_addr; uint32_t align_mask; @@ -867,8 +822,6 @@ struct amdgpu_ring { bool ready; u32 nop; u32 idx; - u64 last_semaphore_signal_addr; - u64 last_semaphore_wait_addr; u32 me; u32 pipe; u32 queue; @@ -881,7 +834,6 @@ struct amdgpu_ring { struct amdgpu_ctx *current_ctx; enum amdgpu_ring_type type; char name[16]; - bool is_pte_ring; }; /* @@ -932,6 +884,8 @@ struct amdgpu_vm_id { }; struct amdgpu_vm { + /* tree of virtual addresses mapped */ + spinlock_t it_lock; struct rb_root va; /* protecting invalidated */ @@ -956,30 +910,40 @@ struct amdgpu_vm { /* for id and flush management per ring */ struct amdgpu_vm_id ids[AMDGPU_MAX_RINGS]; - /* for interval tree */ - spinlock_t it_lock; + /* protecting freed */ spinlock_t freed_lock; + + /* Scheduler entity for page table updates */ + struct amd_sched_entity entity; +}; + +struct amdgpu_vm_manager_id { + struct list_head list; + struct fence *active; + atomic_long_t owner; }; struct amdgpu_vm_manager { - struct { - struct fence *active; - atomic_long_t owner; - } ids[AMDGPU_NUM_VM]; + /* Handling of VMIDs */ + struct mutex lock; + unsigned num_ids; + struct list_head ids_lru; + struct amdgpu_vm_manager_id ids[AMDGPU_NUM_VM]; uint32_t max_pfn; - /* number of VMIDs */ - unsigned nvm; /* vram base address for page table entry */ u64 vram_base_offset; /* is vm enabled? */ bool enabled; /* vm pte handling */ const struct amdgpu_vm_pte_funcs *vm_pte_funcs; - struct amdgpu_ring *vm_pte_funcs_ring; + struct amdgpu_ring *vm_pte_rings[AMDGPU_MAX_RINGS]; + unsigned vm_pte_num_rings; + atomic_t vm_pte_next_ring; }; +void amdgpu_vm_manager_init(struct amdgpu_device *adev); void amdgpu_vm_manager_fini(struct amdgpu_device *adev); int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm); void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm); @@ -990,14 +954,11 @@ void amdgpu_vm_get_pt_bos(struct amdgpu_vm *vm, struct list_head *duplicates); void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, struct amdgpu_vm *vm); int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, - struct amdgpu_sync *sync); + struct amdgpu_sync *sync, struct fence *fence); void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_vm *vm, struct fence *updates); -void amdgpu_vm_fence(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct fence *fence); -uint64_t amdgpu_vm_map_gart(struct amdgpu_device *adev, uint64_t addr); +uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr); int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, struct amdgpu_vm *vm); int amdgpu_vm_clear_freed(struct amdgpu_device *adev, @@ -1023,7 +984,6 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev, uint64_t addr); void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, struct amdgpu_bo_va *bo_va); -int amdgpu_vm_free_job(struct amdgpu_job *job); /* * context related structures @@ -1051,10 +1011,6 @@ struct amdgpu_ctx_mgr { struct idr ctx_handles; }; -int amdgpu_ctx_init(struct amdgpu_device *adev, enum amd_sched_priority pri, - struct amdgpu_ctx *ctx); -void amdgpu_ctx_fini(struct amdgpu_ctx *ctx); - struct amdgpu_ctx *amdgpu_ctx_get(struct amdgpu_fpriv *fpriv, uint32_t id); int amdgpu_ctx_put(struct amdgpu_ctx *ctx); @@ -1096,6 +1052,8 @@ struct amdgpu_bo_list { struct amdgpu_bo_list * amdgpu_bo_list_get(struct amdgpu_fpriv *fpriv, int id); +void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list, + struct list_head *validated); void amdgpu_bo_list_put(struct amdgpu_bo_list *list); void amdgpu_bo_list_free(struct amdgpu_bo_list *list); @@ -1169,6 +1127,7 @@ struct amdgpu_gca_config { unsigned multi_gpu_tile_size; unsigned mc_arb_ramcfg; unsigned gb_addr_config; + unsigned num_rbs; uint32_t tile_mode_array[32]; uint32_t macrotile_mode_array[16]; @@ -1211,23 +1170,21 @@ struct amdgpu_gfx { unsigned ce_ram_size; }; -int amdgpu_ib_get(struct amdgpu_ring *ring, struct amdgpu_vm *vm, +int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm, unsigned size, struct amdgpu_ib *ib); void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib); -int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, - struct amdgpu_ib *ib, void *owner); +int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, + struct amdgpu_ib *ib, void *owner, + struct fence *last_vm_update, + struct fence **f); int amdgpu_ib_pool_init(struct amdgpu_device *adev); void amdgpu_ib_pool_fini(struct amdgpu_device *adev); int amdgpu_ib_ring_tests(struct amdgpu_device *adev); -/* Ring access between begin & end cannot sleep */ -void amdgpu_ring_free_size(struct amdgpu_ring *ring); int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw); -int amdgpu_ring_lock(struct amdgpu_ring *ring, unsigned ndw); void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count); +void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib); void amdgpu_ring_commit(struct amdgpu_ring *ring); -void amdgpu_ring_unlock_commit(struct amdgpu_ring *ring); void amdgpu_ring_undo(struct amdgpu_ring *ring); -void amdgpu_ring_unlock_undo(struct amdgpu_ring *ring); unsigned amdgpu_ring_backup(struct amdgpu_ring *ring, uint32_t **data); int amdgpu_ring_restore(struct amdgpu_ring *ring, @@ -1246,47 +1203,57 @@ struct amdgpu_cs_chunk { uint32_t chunk_id; uint32_t length_dw; uint32_t *kdata; - void __user *user_ptr; }; struct amdgpu_cs_parser { struct amdgpu_device *adev; struct drm_file *filp; struct amdgpu_ctx *ctx; - struct amdgpu_bo_list *bo_list; + /* chunks */ unsigned nchunks; struct amdgpu_cs_chunk *chunks; - /* relocations */ - struct amdgpu_bo_list_entry vm_pd; - struct list_head validated; - struct fence *fence; - struct amdgpu_ib *ibs; - uint32_t num_ibs; + /* scheduler job object */ + struct amdgpu_job *job; - struct ww_acquire_ctx ticket; + /* buffer objects */ + struct ww_acquire_ctx ticket; + struct amdgpu_bo_list *bo_list; + struct amdgpu_bo_list_entry vm_pd; + struct list_head validated; + struct fence *fence; + uint64_t bytes_moved_threshold; + uint64_t bytes_moved; /* user fence */ - struct amdgpu_user_fence uf; struct amdgpu_bo_list_entry uf_entry; }; struct amdgpu_job { struct amd_sched_job base; struct amdgpu_device *adev; + struct amdgpu_ring *ring; + struct amdgpu_sync sync; struct amdgpu_ib *ibs; uint32_t num_ibs; void *owner; struct amdgpu_user_fence uf; - int (*free_job)(struct amdgpu_job *job); }; #define to_amdgpu_job(sched_job) \ container_of((sched_job), struct amdgpu_job, base) -static inline u32 amdgpu_get_ib_value(struct amdgpu_cs_parser *p, uint32_t ib_idx, int idx) +static inline u32 amdgpu_get_ib_value(struct amdgpu_cs_parser *p, + uint32_t ib_idx, int idx) { - return p->ibs[ib_idx].ptr[idx]; + return p->job->ibs[ib_idx].ptr[idx]; +} + +static inline void amdgpu_set_ib_value(struct amdgpu_cs_parser *p, + uint32_t ib_idx, int idx, + uint32_t value) +{ + p->job->ibs[ib_idx].ptr[idx] = value; } /* @@ -1538,6 +1505,7 @@ enum amdgpu_dpm_forced_level { AMDGPU_DPM_FORCED_LEVEL_AUTO = 0, AMDGPU_DPM_FORCED_LEVEL_LOW = 1, AMDGPU_DPM_FORCED_LEVEL_HIGH = 2, + AMDGPU_DPM_FORCED_LEVEL_MANUAL = 3, }; struct amdgpu_vce_state { @@ -1667,6 +1635,7 @@ struct amdgpu_uvd { struct amdgpu_ring ring; struct amdgpu_irq_src irq; bool address_64_bit; + struct amd_sched_entity entity; }; /* @@ -1691,6 +1660,7 @@ struct amdgpu_vce { struct amdgpu_ring ring[AMDGPU_MAX_VCE_RINGS]; struct amdgpu_irq_src irq; unsigned harvest_config; + struct amd_sched_entity entity; }; /* @@ -1925,6 +1895,18 @@ void amdgpu_cgs_destroy_device(void *cgs_device); /* + * CGS + */ +void *amdgpu_cgs_create_device(struct amdgpu_device *adev); +void amdgpu_cgs_destroy_device(void *cgs_device); + + +/* GPU virtualization */ +struct amdgpu_virtualization { + bool supports_sr_iov; +}; + +/* * Core structure, functions and helpers. */ typedef uint32_t (*amdgpu_rreg_t)(struct amdgpu_device*, uint32_t); @@ -1944,6 +1926,10 @@ struct amdgpu_device { struct drm_device *ddev; struct pci_dev *pdev; +#ifdef CONFIG_DRM_AMD_ACP + struct amdgpu_acp acp; +#endif + /* ASIC */ enum amd_asic_type asic_type; uint32_t family; @@ -2020,7 +2006,6 @@ struct amdgpu_device { /* memory management */ struct amdgpu_mman mman; - struct amdgpu_gem gem; struct amdgpu_vram_scratch vram_scratch; struct amdgpu_wb wb; atomic64_t vram_usage; @@ -2038,7 +2023,6 @@ struct amdgpu_device { /* rings */ unsigned fence_context; - struct mutex ring_lock; unsigned num_rings; struct amdgpu_ring *rings[AMDGPU_MAX_RINGS]; bool ib_pool_ready; @@ -2050,6 +2034,7 @@ struct amdgpu_device { /* powerplay */ struct amd_powerplay powerplay; bool pp_enabled; + bool pp_force_state_enabled; /* dpm */ struct amdgpu_pm pm; @@ -2091,8 +2076,7 @@ struct amdgpu_device { /* amdkfd interface */ struct kfd_dev *kfd; - /* kernel conext for IB submission */ - struct amdgpu_ctx kernel_ctx; + struct amdgpu_virtualization virtualization; }; bool amdgpu_device_is_px(struct drm_device *dev); @@ -2197,7 +2181,6 @@ static inline void amdgpu_ring_write(struct amdgpu_ring *ring, uint32_t v) ring->ring[ring->wptr++] = v; ring->wptr &= ring->ptr_mask; ring->count_dw--; - ring->ring_free_dw--; } static inline struct amdgpu_sdma_instance * @@ -2233,9 +2216,8 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_gart_flush_gpu_tlb(adev, vmid) (adev)->gart.gart_funcs->flush_gpu_tlb((adev), (vmid)) #define amdgpu_gart_set_pte_pde(adev, pt, idx, addr, flags) (adev)->gart.gart_funcs->set_pte_pde((adev), (pt), (idx), (addr), (flags)) #define amdgpu_vm_copy_pte(adev, ib, pe, src, count) ((adev)->vm_manager.vm_pte_funcs->copy_pte((ib), (pe), (src), (count))) -#define amdgpu_vm_write_pte(adev, ib, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->write_pte((ib), (pe), (addr), (count), (incr), (flags))) +#define amdgpu_vm_write_pte(adev, ib, pa, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->write_pte((ib), (pa), (pe), (addr), (count), (incr), (flags))) #define amdgpu_vm_set_pte_pde(adev, ib, pe, addr, count, incr, flags) ((adev)->vm_manager.vm_pte_funcs->set_pte_pde((ib), (pe), (addr), (count), (incr), (flags))) -#define amdgpu_vm_pad_ib(adev, ib) ((adev)->vm_manager.vm_pte_funcs->pad_ib((ib))) #define amdgpu_ring_parse_cs(r, p, ib) ((r)->funcs->parse_cs((p), (ib))) #define amdgpu_ring_test_ring(r) (r)->funcs->test_ring((r)) #define amdgpu_ring_test_ib(r) (r)->funcs->test_ib((r)) @@ -2245,9 +2227,9 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_ring_emit_ib(r, ib) (r)->funcs->emit_ib((r), (ib)) #define amdgpu_ring_emit_vm_flush(r, vmid, addr) (r)->funcs->emit_vm_flush((r), (vmid), (addr)) #define amdgpu_ring_emit_fence(r, addr, seq, flags) (r)->funcs->emit_fence((r), (addr), (seq), (flags)) -#define amdgpu_ring_emit_semaphore(r, semaphore, emit_wait) (r)->funcs->emit_semaphore((r), (semaphore), (emit_wait)) #define amdgpu_ring_emit_gds_switch(r, v, db, ds, wb, ws, ab, as) (r)->funcs->emit_gds_switch((r), (v), (db), (ds), (wb), (ws), (ab), (as)) #define amdgpu_ring_emit_hdp_flush(r) (r)->funcs->emit_hdp_flush((r)) +#define amdgpu_ring_pad_ib(r, ib) ((r)->funcs->pad_ib((r), (ib))) #define amdgpu_ih_get_wptr(adev) (adev)->irq.ih_funcs->get_wptr((adev)) #define amdgpu_ih_decode_iv(adev, iv) (adev)->irq.ih_funcs->decode_iv((adev), (iv)) #define amdgpu_ih_set_rptr(adev) (adev)->irq.ih_funcs->set_rptr((adev)) @@ -2339,6 +2321,21 @@ amdgpu_get_sdma_instance(struct amdgpu_ring *ring) #define amdgpu_dpm_get_performance_level(adev) \ (adev)->powerplay.pp_funcs->get_performance_level((adev)->powerplay.pp_handle) +#define amdgpu_dpm_get_pp_num_states(adev, data) \ + (adev)->powerplay.pp_funcs->get_pp_num_states((adev)->powerplay.pp_handle, data) + +#define amdgpu_dpm_get_pp_table(adev, table) \ + (adev)->powerplay.pp_funcs->get_pp_table((adev)->powerplay.pp_handle, table) + +#define amdgpu_dpm_set_pp_table(adev, buf, size) \ + (adev)->powerplay.pp_funcs->set_pp_table((adev)->powerplay.pp_handle, buf, size) + +#define amdgpu_dpm_print_clock_levels(adev, type, buf) \ + (adev)->powerplay.pp_funcs->print_clock_levels((adev)->powerplay.pp_handle, type, buf) + +#define amdgpu_dpm_force_clock_level(adev, type, level) \ + (adev)->powerplay.pp_funcs->force_clock_level((adev)->powerplay.pp_handle, type, level) + #define amdgpu_dpm_dispatch_task(adev, event_id, input, output) \ (adev)->powerplay.pp_funcs->dispatch_tasks((adev)->powerplay.pp_handle, (event_id), (input), (output)) @@ -2349,7 +2346,6 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev); void amdgpu_pci_config_reset(struct amdgpu_device *adev); bool amdgpu_card_posted(struct amdgpu_device *adev); void amdgpu_update_display_priority(struct amdgpu_device *adev); -bool amdgpu_boot_test_post_card(struct amdgpu_device *adev); int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data); int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type, @@ -2359,7 +2355,9 @@ void amdgpu_ttm_placement_from_domain(struct amdgpu_bo *rbo, u32 domain); bool amdgpu_ttm_bo_is_amdgpu_bo(struct ttm_buffer_object *bo); int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr, uint32_t flags); -bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm); +struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm); +bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, + unsigned long end); bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm); uint32_t amdgpu_ttm_tt_pte_flags(struct amdgpu_device *adev, struct ttm_tt *ttm, struct ttm_mem_reg *mem); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c new file mode 100644 index 000000000000..9f8cfaab3004 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.c @@ -0,0 +1,502 @@ +/* + * Copyright 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#include <linux/irqdomain.h> +#include <linux/pm_domain.h> +#include <linux/platform_device.h> +#include <sound/designware_i2s.h> +#include <sound/pcm.h> + +#include "amdgpu.h" +#include "atom.h" +#include "amdgpu_acp.h" + +#include "acp_gfx_if.h" + +#define ACP_TILE_ON_MASK 0x03 +#define ACP_TILE_OFF_MASK 0x02 +#define ACP_TILE_ON_RETAIN_REG_MASK 0x1f +#define ACP_TILE_OFF_RETAIN_REG_MASK 0x20 + +#define ACP_TILE_P1_MASK 0x3e +#define ACP_TILE_P2_MASK 0x3d +#define ACP_TILE_DSP0_MASK 0x3b +#define ACP_TILE_DSP1_MASK 0x37 + +#define ACP_TILE_DSP2_MASK 0x2f + +#define ACP_DMA_REGS_END 0x146c0 +#define ACP_I2S_PLAY_REGS_START 0x14840 +#define ACP_I2S_PLAY_REGS_END 0x148b4 +#define ACP_I2S_CAP_REGS_START 0x148b8 +#define ACP_I2S_CAP_REGS_END 0x1496c + +#define ACP_I2S_COMP1_CAP_REG_OFFSET 0xac +#define ACP_I2S_COMP2_CAP_REG_OFFSET 0xa8 +#define ACP_I2S_COMP1_PLAY_REG_OFFSET 0x6c +#define ACP_I2S_COMP2_PLAY_REG_OFFSET 0x68 + +#define mmACP_PGFSM_RETAIN_REG 0x51c9 +#define mmACP_PGFSM_CONFIG_REG 0x51ca +#define mmACP_PGFSM_READ_REG_0 0x51cc + +#define mmACP_MEM_SHUT_DOWN_REQ_LO 0x51f8 +#define mmACP_MEM_SHUT_DOWN_REQ_HI 0x51f9 +#define mmACP_MEM_SHUT_DOWN_STS_LO 0x51fa +#define mmACP_MEM_SHUT_DOWN_STS_HI 0x51fb + +#define ACP_TIMEOUT_LOOP 0x000000FF +#define ACP_DEVS 3 +#define ACP_SRC_ID 162 + +enum { + ACP_TILE_P1 = 0, + ACP_TILE_P2, + ACP_TILE_DSP0, + ACP_TILE_DSP1, + ACP_TILE_DSP2, +}; + +static int acp_sw_init(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + adev->acp.parent = adev->dev; + + adev->acp.cgs_device = + amdgpu_cgs_create_device(adev); + if (!adev->acp.cgs_device) + return -EINVAL; + + return 0; +} + +static int acp_sw_fini(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + if (adev->acp.cgs_device) + amdgpu_cgs_destroy_device(adev->acp.cgs_device); + + return 0; +} + +/* power off a tile/block within ACP */ +static int acp_suspend_tile(void *cgs_dev, int tile) +{ + u32 val = 0; + u32 count = 0; + + if ((tile < ACP_TILE_P1) || (tile > ACP_TILE_DSP2)) { + pr_err("Invalid ACP tile : %d to suspend\n", tile); + return -1; + } + + val = cgs_read_register(cgs_dev, mmACP_PGFSM_READ_REG_0 + tile); + val &= ACP_TILE_ON_MASK; + + if (val == 0x0) { + val = cgs_read_register(cgs_dev, mmACP_PGFSM_RETAIN_REG); + val = val | (1 << tile); + cgs_write_register(cgs_dev, mmACP_PGFSM_RETAIN_REG, val); + cgs_write_register(cgs_dev, mmACP_PGFSM_CONFIG_REG, + 0x500 + tile); + + count = ACP_TIMEOUT_LOOP; + while (true) { + val = cgs_read_register(cgs_dev, mmACP_PGFSM_READ_REG_0 + + tile); + val = val & ACP_TILE_ON_MASK; + if (val == ACP_TILE_OFF_MASK) + break; + if (--count == 0) { + pr_err("Timeout reading ACP PGFSM status\n"); + return -ETIMEDOUT; + } + udelay(100); + } + + val = cgs_read_register(cgs_dev, mmACP_PGFSM_RETAIN_REG); + + val |= ACP_TILE_OFF_RETAIN_REG_MASK; + cgs_write_register(cgs_dev, mmACP_PGFSM_RETAIN_REG, val); + } + return 0; +} + +/* power on a tile/block within ACP */ +static int acp_resume_tile(void *cgs_dev, int tile) +{ + u32 val = 0; + u32 count = 0; + + if ((tile < ACP_TILE_P1) || (tile > ACP_TILE_DSP2)) { + pr_err("Invalid ACP tile to resume\n"); + return -1; + } + + val = cgs_read_register(cgs_dev, mmACP_PGFSM_READ_REG_0 + tile); + val = val & ACP_TILE_ON_MASK; + + if (val != 0x0) { + cgs_write_register(cgs_dev, mmACP_PGFSM_CONFIG_REG, + 0x600 + tile); + count = ACP_TIMEOUT_LOOP; + while (true) { + val = cgs_read_register(cgs_dev, mmACP_PGFSM_READ_REG_0 + + tile); + val = val & ACP_TILE_ON_MASK; + if (val == 0x0) + break; + if (--count == 0) { + pr_err("Timeout reading ACP PGFSM status\n"); + return -ETIMEDOUT; + } + udelay(100); + } + val = cgs_read_register(cgs_dev, mmACP_PGFSM_RETAIN_REG); + if (tile == ACP_TILE_P1) + val = val & (ACP_TILE_P1_MASK); + else if (tile == ACP_TILE_P2) + val = val & (ACP_TILE_P2_MASK); + + cgs_write_register(cgs_dev, mmACP_PGFSM_RETAIN_REG, val); + } + return 0; +} + +struct acp_pm_domain { + void *cgs_dev; + struct generic_pm_domain gpd; +}; + +static int acp_poweroff(struct generic_pm_domain *genpd) +{ + int i, ret; + struct acp_pm_domain *apd; + + apd = container_of(genpd, struct acp_pm_domain, gpd); + if (apd != NULL) { + /* Donot return abruptly if any of power tile fails to suspend. + * Log it and continue powering off other tile + */ + for (i = 4; i >= 0 ; i--) { + ret = acp_suspend_tile(apd->cgs_dev, ACP_TILE_P1 + i); + if (ret) + pr_err("ACP tile %d tile suspend failed\n", i); + } + } + return 0; +} + +static int acp_poweron(struct generic_pm_domain *genpd) +{ + int i, ret; + struct acp_pm_domain *apd; + + apd = container_of(genpd, struct acp_pm_domain, gpd); + if (apd != NULL) { + for (i = 0; i < 2; i++) { + ret = acp_resume_tile(apd->cgs_dev, ACP_TILE_P1 + i); + if (ret) { + pr_err("ACP tile %d resume failed\n", i); + break; + } + } + + /* Disable DSPs which are not going to be used */ + for (i = 0; i < 3; i++) { + ret = acp_suspend_tile(apd->cgs_dev, ACP_TILE_DSP0 + i); + /* Continue suspending other DSP, even if one fails */ + if (ret) + pr_err("ACP DSP %d suspend failed\n", i); + } + } + return 0; +} + +static struct device *get_mfd_cell_dev(const char *device_name, int r) +{ + char auto_dev_name[25]; + char buf[8]; + struct device *dev; + + sprintf(buf, ".%d.auto", r); + strcpy(auto_dev_name, device_name); + strcat(auto_dev_name, buf); + dev = bus_find_device_by_name(&platform_bus_type, NULL, auto_dev_name); + dev_info(dev, "device %s added to pm domain\n", auto_dev_name); + + return dev; +} + +/** + * acp_hw_init - start and test ACP block + * + * @adev: amdgpu_device pointer + * + */ +static int acp_hw_init(void *handle) +{ + int r, i; + uint64_t acp_base; + struct device *dev; + struct i2s_platform_data *i2s_pdata; + + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + const struct amdgpu_ip_block_version *ip_version = + amdgpu_get_ip_block(adev, AMD_IP_BLOCK_TYPE_ACP); + + if (!ip_version) + return -EINVAL; + + r = amd_acp_hw_init(adev->acp.cgs_device, + ip_version->major, ip_version->minor); + /* -ENODEV means board uses AZ rather than ACP */ + if (r == -ENODEV) + return 0; + else if (r) + return r; + + r = cgs_get_pci_resource(adev->acp.cgs_device, CGS_RESOURCE_TYPE_MMIO, + 0x5289, 0, &acp_base); + if (r == -ENODEV) + return 0; + else if (r) + return r; + + adev->acp.acp_genpd = kzalloc(sizeof(struct acp_pm_domain), GFP_KERNEL); + if (adev->acp.acp_genpd == NULL) + return -ENOMEM; + + adev->acp.acp_genpd->gpd.name = "ACP_AUDIO"; + adev->acp.acp_genpd->gpd.power_off = acp_poweroff; + adev->acp.acp_genpd->gpd.power_on = acp_poweron; + + + adev->acp.acp_genpd->cgs_dev = adev->acp.cgs_device; + + pm_genpd_init(&adev->acp.acp_genpd->gpd, NULL, false); + + adev->acp.acp_cell = kzalloc(sizeof(struct mfd_cell) * ACP_DEVS, + GFP_KERNEL); + + if (adev->acp.acp_cell == NULL) + return -ENOMEM; + + adev->acp.acp_res = kzalloc(sizeof(struct resource) * 4, GFP_KERNEL); + + if (adev->acp.acp_res == NULL) { + kfree(adev->acp.acp_cell); + return -ENOMEM; + } + + i2s_pdata = kzalloc(sizeof(struct i2s_platform_data) * 2, GFP_KERNEL); + if (i2s_pdata == NULL) { + kfree(adev->acp.acp_res); + kfree(adev->acp.acp_cell); + return -ENOMEM; + } + + i2s_pdata[0].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET; + i2s_pdata[0].cap = DWC_I2S_PLAY; + i2s_pdata[0].snd_rates = SNDRV_PCM_RATE_8000_96000; + i2s_pdata[0].i2s_reg_comp1 = ACP_I2S_COMP1_PLAY_REG_OFFSET; + i2s_pdata[0].i2s_reg_comp2 = ACP_I2S_COMP2_PLAY_REG_OFFSET; + + i2s_pdata[1].quirks = DW_I2S_QUIRK_COMP_REG_OFFSET | + DW_I2S_QUIRK_COMP_PARAM1; + i2s_pdata[1].cap = DWC_I2S_RECORD; + i2s_pdata[1].snd_rates = SNDRV_PCM_RATE_8000_96000; + i2s_pdata[1].i2s_reg_comp1 = ACP_I2S_COMP1_CAP_REG_OFFSET; + i2s_pdata[1].i2s_reg_comp2 = ACP_I2S_COMP2_CAP_REG_OFFSET; + + adev->acp.acp_res[0].name = "acp2x_dma"; + adev->acp.acp_res[0].flags = IORESOURCE_MEM; + adev->acp.acp_res[0].start = acp_base; + adev->acp.acp_res[0].end = acp_base + ACP_DMA_REGS_END; + + adev->acp.acp_res[1].name = "acp2x_dw_i2s_play"; + adev->acp.acp_res[1].flags = IORESOURCE_MEM; + adev->acp.acp_res[1].start = acp_base + ACP_I2S_PLAY_REGS_START; + adev->acp.acp_res[1].end = acp_base + ACP_I2S_PLAY_REGS_END; + + adev->acp.acp_res[2].name = "acp2x_dw_i2s_cap"; + adev->acp.acp_res[2].flags = IORESOURCE_MEM; + adev->acp.acp_res[2].start = acp_base + ACP_I2S_CAP_REGS_START; + adev->acp.acp_res[2].end = acp_base + ACP_I2S_CAP_REGS_END; + + adev->acp.acp_res[3].name = "acp2x_dma_irq"; + adev->acp.acp_res[3].flags = IORESOURCE_IRQ; + adev->acp.acp_res[3].start = amdgpu_irq_create_mapping(adev, 162); + adev->acp.acp_res[3].end = adev->acp.acp_res[3].start; + + adev->acp.acp_cell[0].name = "acp_audio_dma"; + adev->acp.acp_cell[0].num_resources = 4; + adev->acp.acp_cell[0].resources = &adev->acp.acp_res[0]; + + adev->acp.acp_cell[1].name = "designware-i2s"; + adev->acp.acp_cell[1].num_resources = 1; + adev->acp.acp_cell[1].resources = &adev->acp.acp_res[1]; + adev->acp.acp_cell[1].platform_data = &i2s_pdata[0]; + adev->acp.acp_cell[1].pdata_size = sizeof(struct i2s_platform_data); + + adev->acp.acp_cell[2].name = "designware-i2s"; + adev->acp.acp_cell[2].num_resources = 1; + adev->acp.acp_cell[2].resources = &adev->acp.acp_res[2]; + adev->acp.acp_cell[2].platform_data = &i2s_pdata[1]; + adev->acp.acp_cell[2].pdata_size = sizeof(struct i2s_platform_data); + + r = mfd_add_hotplug_devices(adev->acp.parent, adev->acp.acp_cell, + ACP_DEVS); + if (r) + return r; + + for (i = 0; i < ACP_DEVS ; i++) { + dev = get_mfd_cell_dev(adev->acp.acp_cell[i].name, i); + r = pm_genpd_add_device(&adev->acp.acp_genpd->gpd, dev); + if (r) { + dev_err(dev, "Failed to add dev to genpd\n"); + return r; + } + } + + return 0; +} + +/** + * acp_hw_fini - stop the hardware block + * + * @adev: amdgpu_device pointer + * + */ +static int acp_hw_fini(void *handle) +{ + int i, ret; + struct device *dev; + + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + for (i = 0; i < ACP_DEVS ; i++) { + dev = get_mfd_cell_dev(adev->acp.acp_cell[i].name, i); + ret = pm_genpd_remove_device(&adev->acp.acp_genpd->gpd, dev); + /* If removal fails, dont giveup and try rest */ + if (ret) + dev_err(dev, "remove dev from genpd failed\n"); + } + + mfd_remove_devices(adev->acp.parent); + kfree(adev->acp.acp_res); + kfree(adev->acp.acp_genpd); + kfree(adev->acp.acp_cell); + + return 0; +} + +static int acp_suspend(void *handle) +{ + return 0; +} + +static int acp_resume(void *handle) +{ + int i, ret; + struct acp_pm_domain *apd; + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + /* SMU block will power on ACP irrespective of ACP runtime status. + * Power off explicitly based on genpd ACP runtime status so that ACP + * hw and ACP-genpd status are in sync. + * 'suspend_power_off' represents "Power status before system suspend" + */ + if (adev->acp.acp_genpd->gpd.suspend_power_off == true) { + apd = container_of(&adev->acp.acp_genpd->gpd, + struct acp_pm_domain, gpd); + + for (i = 4; i >= 0 ; i--) { + ret = acp_suspend_tile(apd->cgs_dev, ACP_TILE_P1 + i); + if (ret) + pr_err("ACP tile %d tile suspend failed\n", i); + } + } + return 0; +} + +static int acp_early_init(void *handle) +{ + return 0; +} + +static bool acp_is_idle(void *handle) +{ + return true; +} + +static int acp_wait_for_idle(void *handle) +{ + return 0; +} + +static int acp_soft_reset(void *handle) +{ + return 0; +} + +static void acp_print_status(void *handle) +{ + struct amdgpu_device *adev = (struct amdgpu_device *)handle; + + dev_info(adev->dev, "ACP STATUS\n"); +} + +static int acp_set_clockgating_state(void *handle, + enum amd_clockgating_state state) +{ + return 0; +} + +static int acp_set_powergating_state(void *handle, + enum amd_powergating_state state) +{ + return 0; +} + +const struct amd_ip_funcs acp_ip_funcs = { + .early_init = acp_early_init, + .late_init = NULL, + .sw_init = acp_sw_init, + .sw_fini = acp_sw_fini, + .hw_init = acp_hw_init, + .hw_fini = acp_hw_fini, + .suspend = acp_suspend, + .resume = acp_resume, + .is_idle = acp_is_idle, + .wait_for_idle = acp_wait_for_idle, + .soft_reset = acp_soft_reset, + .print_status = acp_print_status, + .set_clockgating_state = acp_set_clockgating_state, + .set_powergating_state = acp_set_powergating_state, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h new file mode 100644 index 000000000000..f6e32a639107 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_acp.h @@ -0,0 +1,42 @@ +/* + * Copyright 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * Authors: AMD + * + */ + +#ifndef __AMDGPU_ACP_H__ +#define __AMDGPU_ACP_H__ + +#include <linux/mfd/core.h> + +struct amdgpu_acp { + struct device *parent; + void *cgs_device; + struct amd_acp_private *private; + struct mfd_cell *acp_cell; + struct resource *acp_res; + struct acp_pm_domain *acp_genpd; +}; + +extern const struct amd_ip_funcs acp_ip_funcs; + +#endif /* __AMDGPU_ACP_H__ */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c index 9416e0f5c1db..84b0ce39ee14 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.c @@ -1514,6 +1514,19 @@ int amdgpu_atombios_init_mc_reg_table(struct amdgpu_device *adev, return -EINVAL; } +bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev) +{ + int index = GetIndexIntoMasterTable(DATA, GPUVirtualizationInfo); + u8 frev, crev; + u16 data_offset, size; + + if (amdgpu_atom_parse_data_header(adev->mode_info.atom_context, index, &size, + &frev, &crev, &data_offset)) + return true; + + return false; +} + void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock) { uint32_t bios_6_scratch; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h index 0ebb959ea435..9e1442053fe4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atombios.h @@ -196,6 +196,8 @@ int amdgpu_atombios_init_mc_reg_table(struct amdgpu_device *adev, u8 module_index, struct atom_mc_reg_table *reg_table); +bool amdgpu_atombios_has_gpu_virtualization_table(struct amdgpu_device *adev); + void amdgpu_atombios_scratch_regs_lock(struct amdgpu_device *adev, bool lock); void amdgpu_atombios_scratch_regs_init(struct amdgpu_device *adev); void amdgpu_atombios_scratch_regs_save(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c index 3c895863fcf5..fa948dcbdd5d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_atpx_handler.c @@ -552,13 +552,14 @@ static bool amdgpu_atpx_detect(void) void amdgpu_register_atpx_handler(void) { bool r; + enum vga_switcheroo_handler_flags_t handler_flags = 0; /* detect if we have any ATPX + 2 VGA in the system */ r = amdgpu_atpx_detect(); if (!r) return; - vga_switcheroo_register_handler(&amdgpu_atpx_handler); + vga_switcheroo_register_handler(&amdgpu_atpx_handler, handler_flags); } /** diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c index f82a2dd83874..90d6fc1618aa 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_bo_list.c @@ -32,6 +32,9 @@ #include "amdgpu.h" #include "amdgpu_trace.h" +#define AMDGPU_BO_LIST_MAX_PRIORITY 32u +#define AMDGPU_BO_LIST_NUM_BUCKETS (AMDGPU_BO_LIST_MAX_PRIORITY + 1) + static int amdgpu_bo_list_create(struct amdgpu_fpriv *fpriv, struct amdgpu_bo_list **result, int *id) @@ -90,6 +93,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, bool has_userptr = false; unsigned i; + int r; array = drm_malloc_ab(num_entries, sizeof(struct amdgpu_bo_list_entry)); if (!array) @@ -99,31 +103,34 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, for (i = 0; i < num_entries; ++i) { struct amdgpu_bo_list_entry *entry = &array[i]; struct drm_gem_object *gobj; + struct mm_struct *usermm; gobj = drm_gem_object_lookup(adev->ddev, filp, info[i].bo_handle); - if (!gobj) + if (!gobj) { + r = -ENOENT; goto error_free; + } entry->robj = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); drm_gem_object_unreference_unlocked(gobj); - entry->priority = info[i].bo_priority; - entry->prefered_domains = entry->robj->initial_domain; - entry->allowed_domains = entry->prefered_domains; - if (entry->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM) - entry->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT; - if (amdgpu_ttm_tt_has_userptr(entry->robj->tbo.ttm)) { + entry->priority = min(info[i].bo_priority, + AMDGPU_BO_LIST_MAX_PRIORITY); + usermm = amdgpu_ttm_tt_get_usermm(entry->robj->tbo.ttm); + if (usermm) { + if (usermm != current->mm) { + r = -EPERM; + goto error_free; + } has_userptr = true; - entry->prefered_domains = AMDGPU_GEM_DOMAIN_GTT; - entry->allowed_domains = AMDGPU_GEM_DOMAIN_GTT; } entry->tv.bo = &entry->robj->tbo; entry->tv.shared = true; - if (entry->prefered_domains == AMDGPU_GEM_DOMAIN_GDS) + if (entry->robj->prefered_domains == AMDGPU_GEM_DOMAIN_GDS) gds_obj = entry->robj; - if (entry->prefered_domains == AMDGPU_GEM_DOMAIN_GWS) + if (entry->robj->prefered_domains == AMDGPU_GEM_DOMAIN_GWS) gws_obj = entry->robj; - if (entry->prefered_domains == AMDGPU_GEM_DOMAIN_OA) + if (entry->robj->prefered_domains == AMDGPU_GEM_DOMAIN_OA) oa_obj = entry->robj; trace_amdgpu_bo_list_set(list, entry->robj); @@ -145,7 +152,7 @@ static int amdgpu_bo_list_set(struct amdgpu_device *adev, error_free: drm_free_large(array); - return -ENOENT; + return r; } struct amdgpu_bo_list * @@ -161,6 +168,36 @@ amdgpu_bo_list_get(struct amdgpu_fpriv *fpriv, int id) return result; } +void amdgpu_bo_list_get_list(struct amdgpu_bo_list *list, + struct list_head *validated) +{ + /* This is based on the bucket sort with O(n) time complexity. + * An item with priority "i" is added to bucket[i]. The lists are then + * concatenated in descending order. + */ + struct list_head bucket[AMDGPU_BO_LIST_NUM_BUCKETS]; + unsigned i; + + for (i = 0; i < AMDGPU_BO_LIST_NUM_BUCKETS; i++) + INIT_LIST_HEAD(&bucket[i]); + + /* Since buffers which appear sooner in the relocation list are + * likely to be used more often than buffers which appear later + * in the list, the sort mustn't change the ordering of buffers + * with the same priority, i.e. it must be stable. + */ + for (i = 0; i < list->num_entries; i++) { + unsigned priority = list->array[i].priority; + + list_add_tail(&list->array[i].tv.head, + &bucket[priority]); + } + + /* Connect the sorted buckets in the output list. */ + for (i = 0; i < AMDGPU_BO_LIST_NUM_BUCKETS; i++) + list_splice(&bucket[i], validated); +} + void amdgpu_bo_list_put(struct amdgpu_bo_list *list) { mutex_unlock(&list->lock); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index b882e8175615..52c3eb96b199 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -30,47 +30,6 @@ #include "amdgpu.h" #include "amdgpu_trace.h" -#define AMDGPU_CS_MAX_PRIORITY 32u -#define AMDGPU_CS_NUM_BUCKETS (AMDGPU_CS_MAX_PRIORITY + 1) - -/* This is based on the bucket sort with O(n) time complexity. - * An item with priority "i" is added to bucket[i]. The lists are then - * concatenated in descending order. - */ -struct amdgpu_cs_buckets { - struct list_head bucket[AMDGPU_CS_NUM_BUCKETS]; -}; - -static void amdgpu_cs_buckets_init(struct amdgpu_cs_buckets *b) -{ - unsigned i; - - for (i = 0; i < AMDGPU_CS_NUM_BUCKETS; i++) - INIT_LIST_HEAD(&b->bucket[i]); -} - -static void amdgpu_cs_buckets_add(struct amdgpu_cs_buckets *b, - struct list_head *item, unsigned priority) -{ - /* Since buffers which appear sooner in the relocation list are - * likely to be used more often than buffers which appear later - * in the list, the sort mustn't change the ordering of buffers - * with the same priority, i.e. it must be stable. - */ - list_add_tail(item, &b->bucket[min(priority, AMDGPU_CS_MAX_PRIORITY)]); -} - -static void amdgpu_cs_buckets_get_list(struct amdgpu_cs_buckets *b, - struct list_head *out_list) -{ - unsigned i; - - /* Connect the sorted buckets in the output list. */ - for (i = 0; i < AMDGPU_CS_NUM_BUCKETS; i++) { - list_splice(&b->bucket[i], out_list); - } -} - int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type, u32 ip_instance, u32 ring, struct amdgpu_ring **out_ring) @@ -128,6 +87,7 @@ int amdgpu_cs_get_ring(struct amdgpu_device *adev, u32 ip_type, } static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, + struct amdgpu_user_fence *uf, struct drm_amdgpu_cs_chunk_fence *fence_data) { struct drm_gem_object *gobj; @@ -139,17 +99,15 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, if (gobj == NULL) return -EINVAL; - p->uf.bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); - p->uf.offset = fence_data->offset; + uf->bo = amdgpu_bo_ref(gem_to_amdgpu_bo(gobj)); + uf->offset = fence_data->offset; - if (amdgpu_ttm_tt_has_userptr(p->uf.bo->tbo.ttm)) { + if (amdgpu_ttm_tt_get_usermm(uf->bo->tbo.ttm)) { drm_gem_object_unreference_unlocked(gobj); return -EINVAL; } - p->uf_entry.robj = amdgpu_bo_ref(p->uf.bo); - p->uf_entry.prefered_domains = AMDGPU_GEM_DOMAIN_GTT; - p->uf_entry.allowed_domains = AMDGPU_GEM_DOMAIN_GTT; + p->uf_entry.robj = amdgpu_bo_ref(uf->bo); p->uf_entry.priority = 0; p->uf_entry.tv.bo = &p->uf_entry.robj->tbo; p->uf_entry.tv.shared = true; @@ -160,11 +118,12 @@ static int amdgpu_cs_user_fence_chunk(struct amdgpu_cs_parser *p, int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) { + struct amdgpu_fpriv *fpriv = p->filp->driver_priv; union drm_amdgpu_cs *cs = data; uint64_t *chunk_array_user; uint64_t *chunk_array; - struct amdgpu_fpriv *fpriv = p->filp->driver_priv; - unsigned size; + struct amdgpu_user_fence uf = {}; + unsigned size, num_ibs = 0; int i; int ret; @@ -181,15 +140,12 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) goto free_chunk; } - p->bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle); - /* get chunks */ - INIT_LIST_HEAD(&p->validated); chunk_array_user = (uint64_t __user *)(unsigned long)(cs->in.chunks); if (copy_from_user(chunk_array, chunk_array_user, sizeof(uint64_t)*cs->in.num_chunks)) { ret = -EFAULT; - goto put_bo_list; + goto put_ctx; } p->nchunks = cs->in.num_chunks; @@ -197,7 +153,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) GFP_KERNEL); if (!p->chunks) { ret = -ENOMEM; - goto put_bo_list; + goto put_ctx; } for (i = 0; i < p->nchunks; i++) { @@ -217,7 +173,6 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) size = p->chunks[i].length_dw; cdata = (void __user *)(unsigned long)user_chunk.chunk_data; - p->chunks[i].user_ptr = cdata; p->chunks[i].kdata = drm_malloc_ab(size, sizeof(uint32_t)); if (p->chunks[i].kdata == NULL) { @@ -233,7 +188,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) switch (p->chunks[i].chunk_id) { case AMDGPU_CHUNK_ID_IB: - p->num_ibs++; + ++num_ibs; break; case AMDGPU_CHUNK_ID_FENCE: @@ -243,7 +198,7 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) goto free_partial_kdata; } - ret = amdgpu_cs_user_fence_chunk(p, (void *)p->chunks[i].kdata); + ret = amdgpu_cs_user_fence_chunk(p, &uf, (void *)p->chunks[i].kdata); if (ret) goto free_partial_kdata; @@ -258,12 +213,11 @@ int amdgpu_cs_parser_init(struct amdgpu_cs_parser *p, void *data) } } - - p->ibs = kcalloc(p->num_ibs, sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!p->ibs) { - ret = -ENOMEM; + ret = amdgpu_job_alloc(p->adev, num_ibs, &p->job); + if (ret) goto free_all_kdata; - } + + p->job->uf = uf; kfree(chunk_array); return 0; @@ -274,9 +228,7 @@ free_partial_kdata: for (; i >= 0; i--) drm_free_large(p->chunks[i].kdata); kfree(p->chunks); -put_bo_list: - if (p->bo_list) - amdgpu_bo_list_put(p->bo_list); +put_ctx: amdgpu_ctx_put(p->ctx); free_chunk: kfree(chunk_array); @@ -336,80 +288,76 @@ static u64 amdgpu_cs_get_threshold_for_moves(struct amdgpu_device *adev) return max(bytes_moved_threshold, 1024*1024ull); } -int amdgpu_cs_list_validate(struct amdgpu_device *adev, - struct amdgpu_vm *vm, +int amdgpu_cs_list_validate(struct amdgpu_cs_parser *p, struct list_head *validated) { struct amdgpu_bo_list_entry *lobj; - struct amdgpu_bo *bo; - u64 bytes_moved = 0, initial_bytes_moved; - u64 bytes_moved_threshold = amdgpu_cs_get_threshold_for_moves(adev); + u64 initial_bytes_moved; int r; list_for_each_entry(lobj, validated, tv.head) { - bo = lobj->robj; - if (!bo->pin_count) { - u32 domain = lobj->prefered_domains; - u32 current_domain = - amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type); - - /* Check if this buffer will be moved and don't move it - * if we have moved too many buffers for this IB already. - * - * Note that this allows moving at least one buffer of - * any size, because it doesn't take the current "bo" - * into account. We don't want to disallow buffer moves - * completely. - */ - if ((lobj->allowed_domains & current_domain) != 0 && - (domain & current_domain) == 0 && /* will be moved */ - bytes_moved > bytes_moved_threshold) { - /* don't move it */ - domain = current_domain; - } + struct amdgpu_bo *bo = lobj->robj; + struct mm_struct *usermm; + uint32_t domain; - retry: - amdgpu_ttm_placement_from_domain(bo, domain); - initial_bytes_moved = atomic64_read(&adev->num_bytes_moved); - r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); - bytes_moved += atomic64_read(&adev->num_bytes_moved) - - initial_bytes_moved; - - if (unlikely(r)) { - if (r != -ERESTARTSYS && domain != lobj->allowed_domains) { - domain = lobj->allowed_domains; - goto retry; - } - return r; + usermm = amdgpu_ttm_tt_get_usermm(bo->tbo.ttm); + if (usermm && usermm != current->mm) + return -EPERM; + + if (bo->pin_count) + continue; + + /* Avoid moving this one if we have moved too many buffers + * for this IB already. + * + * Note that this allows moving at least one buffer of + * any size, because it doesn't take the current "bo" + * into account. We don't want to disallow buffer moves + * completely. + */ + if (p->bytes_moved <= p->bytes_moved_threshold) + domain = bo->prefered_domains; + else + domain = bo->allowed_domains; + + retry: + amdgpu_ttm_placement_from_domain(bo, domain); + initial_bytes_moved = atomic64_read(&bo->adev->num_bytes_moved); + r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); + p->bytes_moved += atomic64_read(&bo->adev->num_bytes_moved) - + initial_bytes_moved; + + if (unlikely(r)) { + if (r != -ERESTARTSYS && domain != bo->allowed_domains) { + domain = bo->allowed_domains; + goto retry; } + return r; } - lobj->bo_va = amdgpu_vm_bo_find(vm, bo); } return 0; } -static int amdgpu_cs_parser_relocs(struct amdgpu_cs_parser *p) +static int amdgpu_cs_parser_bos(struct amdgpu_cs_parser *p, + union drm_amdgpu_cs *cs) { struct amdgpu_fpriv *fpriv = p->filp->driver_priv; - struct amdgpu_cs_buckets buckets; struct list_head duplicates; bool need_mmap_lock = false; - int i, r; + int r; + INIT_LIST_HEAD(&p->validated); + + p->bo_list = amdgpu_bo_list_get(fpriv, cs->in.bo_list_handle); if (p->bo_list) { need_mmap_lock = p->bo_list->has_userptr; - amdgpu_cs_buckets_init(&buckets); - for (i = 0; i < p->bo_list->num_entries; i++) - amdgpu_cs_buckets_add(&buckets, &p->bo_list->array[i].tv.head, - p->bo_list->array[i].priority); - - amdgpu_cs_buckets_get_list(&buckets, &p->validated); + amdgpu_bo_list_get_list(p->bo_list, &p->validated); } INIT_LIST_HEAD(&duplicates); amdgpu_vm_get_pd_bo(&fpriv->vm, &p->validated, &p->vm_pd); - if (p->uf.bo) + if (p->job->uf.bo) list_add(&p->uf_entry.tv.head, &p->validated); if (need_mmap_lock) @@ -421,11 +369,27 @@ static int amdgpu_cs_parser_relocs(struct amdgpu_cs_parser *p) amdgpu_vm_get_pt_bos(&fpriv->vm, &duplicates); - r = amdgpu_cs_list_validate(p->adev, &fpriv->vm, &duplicates); + p->bytes_moved_threshold = amdgpu_cs_get_threshold_for_moves(p->adev); + p->bytes_moved = 0; + + r = amdgpu_cs_list_validate(p, &duplicates); + if (r) + goto error_validate; + + r = amdgpu_cs_list_validate(p, &p->validated); if (r) goto error_validate; - r = amdgpu_cs_list_validate(p->adev, &fpriv->vm, &p->validated); + if (p->bo_list) { + struct amdgpu_vm *vm = &fpriv->vm; + unsigned i; + + for (i = 0; i < p->bo_list->num_entries; i++) { + struct amdgpu_bo *bo = p->bo_list->array[i].robj; + + p->bo_list->array[i].bo_va = amdgpu_vm_bo_find(vm, bo); + } + } error_validate: if (r) { @@ -447,7 +411,7 @@ static int amdgpu_cs_sync_rings(struct amdgpu_cs_parser *p) list_for_each_entry(e, &p->validated, tv.head) { struct reservation_object *resv = e->robj->tbo.resv; - r = amdgpu_sync_resv(p->adev, &p->ibs[0].sync, resv, p->filp); + r = amdgpu_sync_resv(p->adev, &p->job->sync, resv, p->filp); if (r) return r; @@ -510,11 +474,8 @@ static void amdgpu_cs_parser_fini(struct amdgpu_cs_parser *parser, int error, bo for (i = 0; i < parser->nchunks; i++) drm_free_large(parser->chunks[i].kdata); kfree(parser->chunks); - if (parser->ibs) - for (i = 0; i < parser->num_ibs; i++) - amdgpu_ib_free(parser->adev, &parser->ibs[i]); - kfree(parser->ibs); - amdgpu_bo_unref(&parser->uf.bo); + if (parser->job) + amdgpu_job_free(parser->job); amdgpu_bo_unref(&parser->uf_entry.robj); } @@ -530,7 +491,7 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p, if (r) return r; - r = amdgpu_sync_fence(adev, &p->ibs[0].sync, vm->page_directory_fence); + r = amdgpu_sync_fence(adev, &p->job->sync, vm->page_directory_fence); if (r) return r; @@ -556,14 +517,14 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p, return r; f = bo_va->last_pt_update; - r = amdgpu_sync_fence(adev, &p->ibs[0].sync, f); + r = amdgpu_sync_fence(adev, &p->job->sync, f); if (r) return r; } } - r = amdgpu_vm_clear_invalids(adev, vm, &p->ibs[0].sync); + r = amdgpu_vm_clear_invalids(adev, vm, &p->job->sync); if (amdgpu_vm_debug && p->bo_list) { /* Invalidate all BOs to test for userspace bugs */ @@ -581,29 +542,25 @@ static int amdgpu_bo_vm_update_pte(struct amdgpu_cs_parser *p, } static int amdgpu_cs_ib_vm_chunk(struct amdgpu_device *adev, - struct amdgpu_cs_parser *parser) + struct amdgpu_cs_parser *p) { - struct amdgpu_fpriv *fpriv = parser->filp->driver_priv; + struct amdgpu_fpriv *fpriv = p->filp->driver_priv; struct amdgpu_vm *vm = &fpriv->vm; - struct amdgpu_ring *ring; + struct amdgpu_ring *ring = p->job->ring; int i, r; - if (parser->num_ibs == 0) - return 0; - /* Only for UVD/VCE VM emulation */ - for (i = 0; i < parser->num_ibs; i++) { - ring = parser->ibs[i].ring; - if (ring->funcs->parse_cs) { - r = amdgpu_ring_parse_cs(ring, parser, i); + if (ring->funcs->parse_cs) { + for (i = 0; i < p->job->num_ibs; i++) { + r = amdgpu_ring_parse_cs(ring, p, i); if (r) return r; } } - r = amdgpu_bo_vm_update_pte(parser, vm); + r = amdgpu_bo_vm_update_pte(p, vm); if (!r) - amdgpu_cs_sync_rings(parser); + amdgpu_cs_sync_rings(p); return r; } @@ -626,14 +583,14 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, int i, j; int r; - for (i = 0, j = 0; i < parser->nchunks && j < parser->num_ibs; i++) { + for (i = 0, j = 0; i < parser->nchunks && j < parser->job->num_ibs; i++) { struct amdgpu_cs_chunk *chunk; struct amdgpu_ib *ib; struct drm_amdgpu_cs_chunk_ib *chunk_ib; struct amdgpu_ring *ring; chunk = &parser->chunks[i]; - ib = &parser->ibs[j]; + ib = &parser->job->ibs[j]; chunk_ib = (struct drm_amdgpu_cs_chunk_ib *)chunk->kdata; if (chunk->chunk_id != AMDGPU_CHUNK_ID_IB) @@ -645,6 +602,11 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, if (r) return r; + if (parser->job->ring && parser->job->ring != ring) + return -EINVAL; + + parser->job->ring = ring; + if (ring->funcs->parse_cs) { struct amdgpu_bo_va_mapping *m; struct amdgpu_bo *aobj = NULL; @@ -673,7 +635,7 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, offset = ((uint64_t)m->it.start) * AMDGPU_GPU_PAGE_SIZE; kptr += chunk_ib->va_start - offset; - r = amdgpu_ib_get(ring, NULL, chunk_ib->ib_bytes, ib); + r = amdgpu_ib_get(adev, NULL, chunk_ib->ib_bytes, ib); if (r) { DRM_ERROR("Failed to get ib !\n"); return r; @@ -682,7 +644,7 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, memcpy(ib->ptr, kptr, chunk_ib->ib_bytes); amdgpu_bo_kunmap(aobj); } else { - r = amdgpu_ib_get(ring, vm, 0, ib); + r = amdgpu_ib_get(adev, vm, 0, ib); if (r) { DRM_ERROR("Failed to get ib !\n"); return r; @@ -697,15 +659,12 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, j++; } - if (!parser->num_ibs) - return 0; - /* add GDS resources to first IB */ if (parser->bo_list) { struct amdgpu_bo *gds = parser->bo_list->gds_obj; struct amdgpu_bo *gws = parser->bo_list->gws_obj; struct amdgpu_bo *oa = parser->bo_list->oa_obj; - struct amdgpu_ib *ib = &parser->ibs[0]; + struct amdgpu_ib *ib = &parser->job->ibs[0]; if (gds) { ib->gds_base = amdgpu_bo_gpu_offset(gds); @@ -721,15 +680,15 @@ static int amdgpu_cs_ib_fill(struct amdgpu_device *adev, } } /* wrap the last IB with user fence */ - if (parser->uf.bo) { - struct amdgpu_ib *ib = &parser->ibs[parser->num_ibs - 1]; + if (parser->job->uf.bo) { + struct amdgpu_ib *ib = &parser->job->ibs[parser->job->num_ibs - 1]; /* UVD & VCE fw doesn't support user fences */ - if (ib->ring->type == AMDGPU_RING_TYPE_UVD || - ib->ring->type == AMDGPU_RING_TYPE_VCE) + if (parser->job->ring->type == AMDGPU_RING_TYPE_UVD || + parser->job->ring->type == AMDGPU_RING_TYPE_VCE) return -EINVAL; - ib->user = &parser->uf; + ib->user = &parser->job->uf; } return 0; @@ -739,14 +698,8 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev, struct amdgpu_cs_parser *p) { struct amdgpu_fpriv *fpriv = p->filp->driver_priv; - struct amdgpu_ib *ib; int i, j, r; - if (!p->num_ibs) - return 0; - - /* Add dependencies to first IB */ - ib = &p->ibs[0]; for (i = 0; i < p->nchunks; ++i) { struct drm_amdgpu_cs_chunk_dep *deps; struct amdgpu_cs_chunk *chunk; @@ -784,7 +737,8 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev, return r; } else if (fence) { - r = amdgpu_sync_fence(adev, &ib->sync, fence); + r = amdgpu_sync_fence(adev, &p->job->sync, + fence); fence_put(fence); amdgpu_ctx_put(ctx); if (r) @@ -796,15 +750,36 @@ static int amdgpu_cs_dependencies(struct amdgpu_device *adev, return 0; } -static int amdgpu_cs_free_job(struct amdgpu_job *job) +static int amdgpu_cs_submit(struct amdgpu_cs_parser *p, + union drm_amdgpu_cs *cs) { - int i; - if (job->ibs) - for (i = 0; i < job->num_ibs; i++) - amdgpu_ib_free(job->adev, &job->ibs[i]); - kfree(job->ibs); - if (job->uf.bo) - amdgpu_bo_unref(&job->uf.bo); + struct amdgpu_ring *ring = p->job->ring; + struct amd_sched_fence *fence; + struct amdgpu_job *job; + + job = p->job; + p->job = NULL; + + job->base.sched = &ring->sched; + job->base.s_entity = &p->ctx->rings[ring->idx].entity; + job->owner = p->filp; + + fence = amd_sched_fence_create(job->base.s_entity, p->filp); + if (!fence) { + amdgpu_job_free(job); + return -ENOMEM; + } + + job->base.s_fence = fence; + p->fence = fence_get(&fence->base); + + cs->out.handle = amdgpu_ctx_add_fence(p->ctx, ring, + &fence->base); + job->ibs[job->num_ibs - 1].sequence = cs->out.handle; + + trace_amdgpu_cs_ioctl(job); + amd_sched_entity_push_job(&job->base); + return 0; } @@ -829,7 +804,7 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) r = amdgpu_cs_handle_lockup(adev, r); return r; } - r = amdgpu_cs_parser_relocs(&parser); + r = amdgpu_cs_parser_bos(&parser, data); if (r == -ENOMEM) DRM_ERROR("Not enough memory for command submission!\n"); else if (r && r != -ERESTARTSYS) @@ -848,68 +823,14 @@ int amdgpu_cs_ioctl(struct drm_device *dev, void *data, struct drm_file *filp) if (r) goto out; - for (i = 0; i < parser.num_ibs; i++) + for (i = 0; i < parser.job->num_ibs; i++) trace_amdgpu_cs(&parser, i); r = amdgpu_cs_ib_vm_chunk(adev, &parser); if (r) goto out; - if (amdgpu_enable_scheduler && parser.num_ibs) { - struct amdgpu_ring * ring = parser.ibs->ring; - struct amd_sched_fence *fence; - struct amdgpu_job *job; - - job = kzalloc(sizeof(struct amdgpu_job), GFP_KERNEL); - if (!job) { - r = -ENOMEM; - goto out; - } - - job->base.sched = &ring->sched; - job->base.s_entity = &parser.ctx->rings[ring->idx].entity; - job->adev = parser.adev; - job->owner = parser.filp; - job->free_job = amdgpu_cs_free_job; - - job->ibs = parser.ibs; - job->num_ibs = parser.num_ibs; - parser.ibs = NULL; - parser.num_ibs = 0; - - if (job->ibs[job->num_ibs - 1].user) { - job->uf = parser.uf; - job->ibs[job->num_ibs - 1].user = &job->uf; - parser.uf.bo = NULL; - } - - fence = amd_sched_fence_create(job->base.s_entity, - parser.filp); - if (!fence) { - r = -ENOMEM; - amdgpu_cs_free_job(job); - kfree(job); - goto out; - } - job->base.s_fence = fence; - parser.fence = fence_get(&fence->base); - - cs->out.handle = amdgpu_ctx_add_fence(parser.ctx, ring, - &fence->base); - job->ibs[job->num_ibs - 1].sequence = cs->out.handle; - - trace_amdgpu_cs_ioctl(job); - amd_sched_entity_push_job(&job->base); - - } else { - struct amdgpu_fence *fence; - - r = amdgpu_ib_schedule(adev, parser.num_ibs, parser.ibs, - parser.filp); - fence = parser.ibs[parser.num_ibs - 1].fence; - parser.fence = fence_get(&fence->base); - cs->out.handle = parser.ibs[parser.num_ibs - 1].sequence; - } + r = amdgpu_cs_submit(&parser, cs); out: amdgpu_cs_parser_fini(&parser, r, reserved_buffers); @@ -980,30 +901,36 @@ struct amdgpu_bo_va_mapping * amdgpu_cs_find_mapping(struct amdgpu_cs_parser *parser, uint64_t addr, struct amdgpu_bo **bo) { - struct amdgpu_bo_list_entry *reloc; struct amdgpu_bo_va_mapping *mapping; + unsigned i; + + if (!parser->bo_list) + return NULL; addr /= AMDGPU_GPU_PAGE_SIZE; - list_for_each_entry(reloc, &parser->validated, tv.head) { - if (!reloc->bo_va) + for (i = 0; i < parser->bo_list->num_entries; i++) { + struct amdgpu_bo_list_entry *lobj; + + lobj = &parser->bo_list->array[i]; + if (!lobj->bo_va) continue; - list_for_each_entry(mapping, &reloc->bo_va->valids, list) { + list_for_each_entry(mapping, &lobj->bo_va->valids, list) { if (mapping->it.start > addr || addr > mapping->it.last) continue; - *bo = reloc->bo_va->bo; + *bo = lobj->bo_va->bo; return mapping; } - list_for_each_entry(mapping, &reloc->bo_va->invalids, list) { + list_for_each_entry(mapping, &lobj->bo_va->invalids, list) { if (mapping->it.start > addr || addr > mapping->it.last) continue; - *bo = reloc->bo_va->bo; + *bo = lobj->bo_va->bo; return mapping; } } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c index 17d1fb12128a..17e13621fae9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ctx.c @@ -25,8 +25,7 @@ #include <drm/drmP.h> #include "amdgpu.h" -int amdgpu_ctx_init(struct amdgpu_device *adev, enum amd_sched_priority pri, - struct amdgpu_ctx *ctx) +static int amdgpu_ctx_init(struct amdgpu_device *adev, struct amdgpu_ctx *ctx) { unsigned i, j; int r; @@ -35,44 +34,38 @@ int amdgpu_ctx_init(struct amdgpu_device *adev, enum amd_sched_priority pri, ctx->adev = adev; kref_init(&ctx->refcount); spin_lock_init(&ctx->ring_lock); - ctx->fences = kzalloc(sizeof(struct fence *) * amdgpu_sched_jobs * - AMDGPU_MAX_RINGS, GFP_KERNEL); + ctx->fences = kcalloc(amdgpu_sched_jobs * AMDGPU_MAX_RINGS, + sizeof(struct fence*), GFP_KERNEL); if (!ctx->fences) return -ENOMEM; for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { ctx->rings[i].sequence = 1; - ctx->rings[i].fences = (void *)ctx->fences + sizeof(struct fence *) * - amdgpu_sched_jobs * i; + ctx->rings[i].fences = &ctx->fences[amdgpu_sched_jobs * i]; } - if (amdgpu_enable_scheduler) { - /* create context entity for each ring */ - for (i = 0; i < adev->num_rings; i++) { - struct amd_sched_rq *rq; - if (pri >= AMD_SCHED_MAX_PRIORITY) { - kfree(ctx->fences); - return -EINVAL; - } - rq = &adev->rings[i]->sched.sched_rq[pri]; - r = amd_sched_entity_init(&adev->rings[i]->sched, - &ctx->rings[i].entity, - rq, amdgpu_sched_jobs); - if (r) - break; - } - - if (i < adev->num_rings) { - for (j = 0; j < i; j++) - amd_sched_entity_fini(&adev->rings[j]->sched, - &ctx->rings[j].entity); - kfree(ctx->fences); - return r; - } + /* create context entity for each ring */ + for (i = 0; i < adev->num_rings; i++) { + struct amdgpu_ring *ring = adev->rings[i]; + struct amd_sched_rq *rq; + + rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; + r = amd_sched_entity_init(&ring->sched, &ctx->rings[i].entity, + rq, amdgpu_sched_jobs); + if (r) + break; + } + + if (i < adev->num_rings) { + for (j = 0; j < i; j++) + amd_sched_entity_fini(&adev->rings[j]->sched, + &ctx->rings[j].entity); + kfree(ctx->fences); + return r; } return 0; } -void amdgpu_ctx_fini(struct amdgpu_ctx *ctx) +static void amdgpu_ctx_fini(struct amdgpu_ctx *ctx) { struct amdgpu_device *adev = ctx->adev; unsigned i, j; @@ -85,11 +78,9 @@ void amdgpu_ctx_fini(struct amdgpu_ctx *ctx) fence_put(ctx->rings[i].fences[j]); kfree(ctx->fences); - if (amdgpu_enable_scheduler) { - for (i = 0; i < adev->num_rings; i++) - amd_sched_entity_fini(&adev->rings[i]->sched, - &ctx->rings[i].entity); - } + for (i = 0; i < adev->num_rings; i++) + amd_sched_entity_fini(&adev->rings[i]->sched, + &ctx->rings[i].entity); } static int amdgpu_ctx_alloc(struct amdgpu_device *adev, @@ -112,7 +103,7 @@ static int amdgpu_ctx_alloc(struct amdgpu_device *adev, return r; } *id = (uint32_t)r; - r = amdgpu_ctx_init(adev, AMD_SCHED_PRIORITY_NORMAL, ctx); + r = amdgpu_ctx_init(adev, ctx); if (r) { idr_remove(&mgr->ctx_handles, *id); *id = 0; @@ -200,18 +191,18 @@ int amdgpu_ctx_ioctl(struct drm_device *dev, void *data, id = args->in.ctx_id; switch (args->in.op) { - case AMDGPU_CTX_OP_ALLOC_CTX: - r = amdgpu_ctx_alloc(adev, fpriv, &id); - args->out.alloc.ctx_id = id; - break; - case AMDGPU_CTX_OP_FREE_CTX: - r = amdgpu_ctx_free(fpriv, id); - break; - case AMDGPU_CTX_OP_QUERY_STATE: - r = amdgpu_ctx_query(adev, fpriv, id, &args->out); - break; - default: - return -EINVAL; + case AMDGPU_CTX_OP_ALLOC_CTX: + r = amdgpu_ctx_alloc(adev, fpriv, &id); + args->out.alloc.ctx_id = id; + break; + case AMDGPU_CTX_OP_FREE_CTX: + r = amdgpu_ctx_free(fpriv, id); + break; + case AMDGPU_CTX_OP_QUERY_STATE: + r = amdgpu_ctx_query(adev, fpriv, id, &args->out); + break; + default: + return -EINVAL; } return r; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 65531463f88e..db20d2783def 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -636,31 +636,6 @@ bool amdgpu_card_posted(struct amdgpu_device *adev) } /** - * amdgpu_boot_test_post_card - check and possibly initialize the hw - * - * @adev: amdgpu_device pointer - * - * Check if the asic is initialized and if not, attempt to initialize - * it (all asics). - * Returns true if initialized or false if not. - */ -bool amdgpu_boot_test_post_card(struct amdgpu_device *adev) -{ - if (amdgpu_card_posted(adev)) - return true; - - if (adev->bios) { - DRM_INFO("GPU not posted. posting now...\n"); - if (adev->is_atom_bios) - amdgpu_atom_asic_init(adev->mode_info.atom_context); - return true; - } else { - dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n"); - return false; - } -} - -/** * amdgpu_dummy_page_init - init dummy page used by the driver * * @adev: amdgpu_device pointer @@ -959,12 +934,6 @@ static void amdgpu_check_arguments(struct amdgpu_device *adev) amdgpu_sched_jobs); amdgpu_sched_jobs = roundup_pow_of_two(amdgpu_sched_jobs); } - /* vramlimit must be a power of two */ - if (!amdgpu_check_pot_argument(amdgpu_vram_limit)) { - dev_warn(adev->dev, "vram limit (%d) must be a power of 2\n", - amdgpu_vram_limit); - amdgpu_vram_limit = 0; - } if (amdgpu_gart_size != -1) { /* gtt size must be power of two and greater or equal to 32M */ @@ -1434,7 +1403,7 @@ int amdgpu_device_init(struct amdgpu_device *adev, adev->mman.buffer_funcs = NULL; adev->mman.buffer_funcs_ring = NULL; adev->vm_manager.vm_pte_funcs = NULL; - adev->vm_manager.vm_pte_funcs_ring = NULL; + adev->vm_manager.vm_pte_num_rings = 0; adev->gart.gart_funcs = NULL; adev->fence_context = fence_context_alloc(AMDGPU_MAX_RINGS); @@ -1455,9 +1424,8 @@ int amdgpu_device_init(struct amdgpu_device *adev, /* mutex initialization are all done here so we * can recall function without having locking issues */ - mutex_init(&adev->ring_lock); + mutex_init(&adev->vm_manager.lock); atomic_set(&adev->irq.ih.lock, 0); - mutex_init(&adev->gem.mutex); mutex_init(&adev->pm.mutex); mutex_init(&adev->gfx.gpu_clock_mutex); mutex_init(&adev->srbm_mutex); @@ -1531,8 +1499,13 @@ int amdgpu_device_init(struct amdgpu_device *adev, return r; } + /* See if the asic supports SR-IOV */ + adev->virtualization.supports_sr_iov = + amdgpu_atombios_has_gpu_virtualization_table(adev); + /* Post card if necessary */ - if (!amdgpu_card_posted(adev)) { + if (!amdgpu_card_posted(adev) || + adev->virtualization.supports_sr_iov) { if (!adev->bios) { dev_err(adev->dev, "Card not posted and no BIOS - ignoring\n"); return -EINVAL; @@ -1577,11 +1550,6 @@ int amdgpu_device_init(struct amdgpu_device *adev, return r; } - r = amdgpu_ctx_init(adev, AMD_SCHED_PRIORITY_KERNEL, &adev->kernel_ctx); - if (r) { - dev_err(adev->dev, "failed to create kernel context (%d).\n", r); - return r; - } r = amdgpu_ib_ring_tests(adev); if (r) DRM_ERROR("ib ring test failed (%d).\n", r); @@ -1645,7 +1613,6 @@ void amdgpu_device_fini(struct amdgpu_device *adev) adev->shutdown = true; /* evict vram memory */ amdgpu_bo_evict_vram(adev); - amdgpu_ctx_fini(&adev->kernel_ctx); amdgpu_ib_pool_fini(adev); amdgpu_fence_driver_fini(adev); amdgpu_fbdev_fini(adev); @@ -1889,6 +1856,9 @@ int amdgpu_gpu_reset(struct amdgpu_device *adev) retry: r = amdgpu_asic_reset(adev); + /* post card */ + amdgpu_atom_asic_init(adev->mode_info.atom_context); + if (!r) { dev_info(adev->dev, "GPU reset succeeded, trying to resume\n"); r = amdgpu_resume(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index acd066d0a805..2cb53c24dec0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -35,32 +35,30 @@ #include <drm/drm_crtc_helper.h> #include <drm/drm_edid.h> -static void amdgpu_flip_wait_fence(struct amdgpu_device *adev, - struct fence **f) +static void amdgpu_flip_callback(struct fence *f, struct fence_cb *cb) { - struct amdgpu_fence *fence; - long r; + struct amdgpu_flip_work *work = + container_of(cb, struct amdgpu_flip_work, cb); - if (*f == NULL) - return; + fence_put(f); + schedule_work(&work->flip_work); +} - fence = to_amdgpu_fence(*f); - if (fence) { - r = fence_wait(&fence->base, false); - if (r == -EDEADLK) - r = amdgpu_gpu_reset(adev); - } else - r = fence_wait(*f, false); +static bool amdgpu_flip_handle_fence(struct amdgpu_flip_work *work, + struct fence **f) +{ + struct fence *fence= *f; - if (r) - DRM_ERROR("failed to wait on page flip fence (%ld)!\n", r); + if (fence == NULL) + return false; - /* We continue with the page flip even if we failed to wait on - * the fence, otherwise the DRM core and userspace will be - * confused about which BO the CRTC is scanning out - */ - fence_put(*f); *f = NULL; + + if (!fence_add_callback(fence, &work->cb, amdgpu_flip_callback)) + return true; + + fence_put(*f); + return false; } static void amdgpu_flip_work_func(struct work_struct *__work) @@ -76,9 +74,12 @@ static void amdgpu_flip_work_func(struct work_struct *__work) int vpos, hpos, stat, min_udelay; struct drm_vblank_crtc *vblank = &crtc->dev->vblank[work->crtc_id]; - amdgpu_flip_wait_fence(adev, &work->excl); + if (amdgpu_flip_handle_fence(work, &work->excl)) + return; + for (i = 0; i < work->shared_count; ++i) - amdgpu_flip_wait_fence(adev, &work->shared[i]); + if (amdgpu_flip_handle_fence(work, &work->shared[i])) + return; /* We borrow the event spin lock for protecting flip_status */ spin_lock_irqsave(&crtc->dev->event_lock, flags); @@ -118,12 +119,12 @@ static void amdgpu_flip_work_func(struct work_struct *__work) spin_lock_irqsave(&crtc->dev->event_lock, flags); }; - /* do the flip (mmio) */ - adev->mode_info.funcs->page_flip(adev, work->crtc_id, work->base); /* set the flip status */ amdgpuCrtc->pflip_status = AMDGPU_FLIP_SUBMITTED; - spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + + /* Do the flip (mmio) */ + adev->mode_info.funcs->page_flip(adev, work->crtc_id, work->base); } /* @@ -242,7 +243,7 @@ int amdgpu_crtc_page_flip(struct drm_crtc *crtc, /* update crtc fb */ crtc->primary->fb = fb; spin_unlock_irqrestore(&crtc->dev->event_lock, flags); - queue_work(amdgpu_crtc->pflip_queue, &work->flip_work); + amdgpu_flip_work_func(&work->flip_work); return 0; vblank_cleanup: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 9c1af8976bef..ce79a8b605a0 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -69,7 +69,6 @@ int amdgpu_dpm = -1; int amdgpu_smc_load_fw = 1; int amdgpu_aspm = -1; int amdgpu_runtime_pm = -1; -int amdgpu_hard_reset = 0; unsigned amdgpu_ip_block_mask = 0xffffffff; int amdgpu_bapm = -1; int amdgpu_deep_color = 0; @@ -78,10 +77,8 @@ int amdgpu_vm_block_size = -1; int amdgpu_vm_fault_stop = 0; int amdgpu_vm_debug = 0; int amdgpu_exp_hw_support = 0; -int amdgpu_enable_scheduler = 1; int amdgpu_sched_jobs = 32; int amdgpu_sched_hw_submission = 2; -int amdgpu_enable_semaphores = 0; int amdgpu_powerplay = -1; MODULE_PARM_DESC(vramlimit, "Restrict VRAM for testing, in megabytes"); @@ -126,9 +123,6 @@ module_param_named(aspm, amdgpu_aspm, int, 0444); MODULE_PARM_DESC(runpm, "PX runtime pm (1 = force enable, 0 = disable, -1 = PX only default)"); module_param_named(runpm, amdgpu_runtime_pm, int, 0444); -MODULE_PARM_DESC(hard_reset, "PCI config reset (1 = force enable, 0 = disable (default))"); -module_param_named(hard_reset, amdgpu_hard_reset, int, 0444); - MODULE_PARM_DESC(ip_block_mask, "IP Block Mask (all blocks enabled (default))"); module_param_named(ip_block_mask, amdgpu_ip_block_mask, uint, 0444); @@ -153,18 +147,12 @@ module_param_named(vm_debug, amdgpu_vm_debug, int, 0644); MODULE_PARM_DESC(exp_hw_support, "experimental hw support (1 = enable, 0 = disable (default))"); module_param_named(exp_hw_support, amdgpu_exp_hw_support, int, 0444); -MODULE_PARM_DESC(enable_scheduler, "enable SW GPU scheduler (1 = enable (default), 0 = disable)"); -module_param_named(enable_scheduler, amdgpu_enable_scheduler, int, 0444); - MODULE_PARM_DESC(sched_jobs, "the max number of jobs supported in the sw queue (default 32)"); module_param_named(sched_jobs, amdgpu_sched_jobs, int, 0444); MODULE_PARM_DESC(sched_hw_submission, "the max number of HW submissions (default 2)"); module_param_named(sched_hw_submission, amdgpu_sched_hw_submission, int, 0444); -MODULE_PARM_DESC(enable_semaphores, "Enable semaphores (1 = enable, 0 = disable (default))"); -module_param_named(enable_semaphores, amdgpu_enable_semaphores, int, 0644); - #ifdef CONFIG_DRM_AMD_POWERPLAY MODULE_PARM_DESC(powerplay, "Powerplay component (1 = enable, 0 = disable, -1 = auto (default))"); module_param_named(powerplay, amdgpu_powerplay, int, 0444); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c index 3671f9f220bd..97db196dc6f8 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c @@ -107,7 +107,7 @@ int amdgpu_fence_emit(struct amdgpu_ring *ring, void *owner, if ((*fence) == NULL) { return -ENOMEM; } - (*fence)->seq = ++ring->fence_drv.sync_seq[ring->idx]; + (*fence)->seq = ++ring->fence_drv.sync_seq; (*fence)->ring = ring; (*fence)->owner = owner; fence_init(&(*fence)->base, &amdgpu_fence_ops, @@ -171,7 +171,7 @@ static bool amdgpu_fence_activity(struct amdgpu_ring *ring) */ last_seq = atomic64_read(&ring->fence_drv.last_seq); do { - last_emitted = ring->fence_drv.sync_seq[ring->idx]; + last_emitted = ring->fence_drv.sync_seq; seq = amdgpu_fence_read(ring); seq |= last_seq & 0xffffffff00000000LL; if (seq < last_seq) { @@ -260,34 +260,28 @@ static bool amdgpu_fence_seq_signaled(struct amdgpu_ring *ring, u64 seq) } /* - * amdgpu_ring_wait_seq_timeout - wait for seq of the specific ring to signal + * amdgpu_ring_wait_seq - wait for seq of the specific ring to signal * @ring: ring to wait on for the seq number * @seq: seq number wait for * * return value: * 0: seq signaled, and gpu not hang - * -EDEADL: GPU hang detected * -EINVAL: some paramter is not valid */ static int amdgpu_fence_ring_wait_seq(struct amdgpu_ring *ring, uint64_t seq) { - bool signaled = false; - BUG_ON(!ring); - if (seq > ring->fence_drv.sync_seq[ring->idx]) + if (seq > ring->fence_drv.sync_seq) return -EINVAL; if (atomic64_read(&ring->fence_drv.last_seq) >= seq) return 0; amdgpu_fence_schedule_fallback(ring); - wait_event(ring->fence_drv.fence_queue, ( - (signaled = amdgpu_fence_seq_signaled(ring, seq)))); + wait_event(ring->fence_drv.fence_queue, + amdgpu_fence_seq_signaled(ring, seq)); - if (signaled) - return 0; - else - return -EDEADLK; + return 0; } /** @@ -304,7 +298,7 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring) { uint64_t seq = atomic64_read(&ring->fence_drv.last_seq) + 1ULL; - if (seq >= ring->fence_drv.sync_seq[ring->idx]) + if (seq >= ring->fence_drv.sync_seq) return -ENOENT; return amdgpu_fence_ring_wait_seq(ring, seq); @@ -322,7 +316,7 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring) */ int amdgpu_fence_wait_empty(struct amdgpu_ring *ring) { - uint64_t seq = ring->fence_drv.sync_seq[ring->idx]; + uint64_t seq = ring->fence_drv.sync_seq; if (!seq) return 0; @@ -347,7 +341,7 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring) * but it's ok to report slightly wrong fence count here. */ amdgpu_fence_process(ring); - emitted = ring->fence_drv.sync_seq[ring->idx] + emitted = ring->fence_drv.sync_seq - atomic64_read(&ring->fence_drv.last_seq); /* to avoid 32bits warp around */ if (emitted > 0x10000000) @@ -357,68 +351,6 @@ unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring) } /** - * amdgpu_fence_need_sync - do we need a semaphore - * - * @fence: amdgpu fence object - * @dst_ring: which ring to check against - * - * Check if the fence needs to be synced against another ring - * (all asics). If so, we need to emit a semaphore. - * Returns true if we need to sync with another ring, false if - * not. - */ -bool amdgpu_fence_need_sync(struct amdgpu_fence *fence, - struct amdgpu_ring *dst_ring) -{ - struct amdgpu_fence_driver *fdrv; - - if (!fence) - return false; - - if (fence->ring == dst_ring) - return false; - - /* we are protected by the ring mutex */ - fdrv = &dst_ring->fence_drv; - if (fence->seq <= fdrv->sync_seq[fence->ring->idx]) - return false; - - return true; -} - -/** - * amdgpu_fence_note_sync - record the sync point - * - * @fence: amdgpu fence object - * @dst_ring: which ring to check against - * - * Note the sequence number at which point the fence will - * be synced with the requested ring (all asics). - */ -void amdgpu_fence_note_sync(struct amdgpu_fence *fence, - struct amdgpu_ring *dst_ring) -{ - struct amdgpu_fence_driver *dst, *src; - unsigned i; - - if (!fence) - return; - - if (fence->ring == dst_ring) - return; - - /* we are protected by the ring mutex */ - src = &fence->ring->fence_drv; - dst = &dst_ring->fence_drv; - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - if (i == dst_ring->idx) - continue; - - dst->sync_seq[i] = max(dst->sync_seq[i], src->sync_seq[i]); - } -} - -/** * amdgpu_fence_driver_start_ring - make the fence driver * ready for use on the requested ring. * @@ -471,13 +403,12 @@ int amdgpu_fence_driver_start_ring(struct amdgpu_ring *ring, */ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring) { - int i, r; + long timeout; + int r; ring->fence_drv.cpu_addr = NULL; ring->fence_drv.gpu_addr = 0; - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) - ring->fence_drv.sync_seq[i] = 0; - + ring->fence_drv.sync_seq = 0; atomic64_set(&ring->fence_drv.last_seq, 0); ring->fence_drv.initialized = false; @@ -486,26 +417,24 @@ int amdgpu_fence_driver_init_ring(struct amdgpu_ring *ring) init_waitqueue_head(&ring->fence_drv.fence_queue); - if (amdgpu_enable_scheduler) { - long timeout = msecs_to_jiffies(amdgpu_lockup_timeout); - if (timeout == 0) { - /* - * FIXME: - * Delayed workqueue cannot use it directly, - * so the scheduler will not use delayed workqueue if - * MAX_SCHEDULE_TIMEOUT is set. - * Currently keep it simple and silly. - */ - timeout = MAX_SCHEDULE_TIMEOUT; - } - r = amd_sched_init(&ring->sched, &amdgpu_sched_ops, - amdgpu_sched_hw_submission, - timeout, ring->name); - if (r) { - DRM_ERROR("Failed to create scheduler on ring %s.\n", - ring->name); - return r; - } + timeout = msecs_to_jiffies(amdgpu_lockup_timeout); + if (timeout == 0) { + /* + * FIXME: + * Delayed workqueue cannot use it directly, + * so the scheduler will not use delayed workqueue if + * MAX_SCHEDULE_TIMEOUT is set. + * Currently keep it simple and silly. + */ + timeout = MAX_SCHEDULE_TIMEOUT; + } + r = amd_sched_init(&ring->sched, &amdgpu_sched_ops, + amdgpu_sched_hw_submission, + timeout, ring->name); + if (r) { + DRM_ERROR("Failed to create scheduler on ring %s.\n", + ring->name); + return r; } return 0; @@ -552,7 +481,6 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev) if (atomic_dec_and_test(&amdgpu_fence_slab_ref)) kmem_cache_destroy(amdgpu_fence_slab); - mutex_lock(&adev->ring_lock); for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; @@ -570,7 +498,6 @@ void amdgpu_fence_driver_fini(struct amdgpu_device *adev) del_timer_sync(&ring->fence_drv.fallback_timer); ring->fence_drv.initialized = false; } - mutex_unlock(&adev->ring_lock); } /** @@ -585,7 +512,6 @@ void amdgpu_fence_driver_suspend(struct amdgpu_device *adev) { int i, r; - mutex_lock(&adev->ring_lock); for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; if (!ring || !ring->fence_drv.initialized) @@ -602,7 +528,6 @@ void amdgpu_fence_driver_suspend(struct amdgpu_device *adev) amdgpu_irq_put(adev, ring->fence_drv.irq_src, ring->fence_drv.irq_type); } - mutex_unlock(&adev->ring_lock); } /** @@ -621,7 +546,6 @@ void amdgpu_fence_driver_resume(struct amdgpu_device *adev) { int i; - mutex_lock(&adev->ring_lock); for (i = 0; i < AMDGPU_MAX_RINGS; i++) { struct amdgpu_ring *ring = adev->rings[i]; if (!ring || !ring->fence_drv.initialized) @@ -631,7 +555,6 @@ void amdgpu_fence_driver_resume(struct amdgpu_device *adev) amdgpu_irq_get(adev, ring->fence_drv.irq_src, ring->fence_drv.irq_type); } - mutex_unlock(&adev->ring_lock); } /** @@ -651,7 +574,7 @@ void amdgpu_fence_driver_force_completion(struct amdgpu_device *adev) if (!ring || !ring->fence_drv.initialized) continue; - amdgpu_fence_write(ring, ring->fence_drv.sync_seq[i]); + amdgpu_fence_write(ring, ring->fence_drv.sync_seq); } } @@ -781,7 +704,7 @@ static int amdgpu_debugfs_fence_info(struct seq_file *m, void *data) struct drm_info_node *node = (struct drm_info_node *)m->private; struct drm_device *dev = node->minor->dev; struct amdgpu_device *adev = dev->dev_private; - int i, j; + int i; for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { struct amdgpu_ring *ring = adev->rings[i]; @@ -794,28 +717,38 @@ static int amdgpu_debugfs_fence_info(struct seq_file *m, void *data) seq_printf(m, "Last signaled fence 0x%016llx\n", (unsigned long long)atomic64_read(&ring->fence_drv.last_seq)); seq_printf(m, "Last emitted 0x%016llx\n", - ring->fence_drv.sync_seq[i]); - - for (j = 0; j < AMDGPU_MAX_RINGS; ++j) { - struct amdgpu_ring *other = adev->rings[j]; - if (i != j && other && other->fence_drv.initialized && - ring->fence_drv.sync_seq[j]) - seq_printf(m, "Last sync to ring %d 0x%016llx\n", - j, ring->fence_drv.sync_seq[j]); - } + ring->fence_drv.sync_seq); } return 0; } +/** + * amdgpu_debugfs_gpu_reset - manually trigger a gpu reset + * + * Manually trigger a gpu reset at the next fence wait. + */ +static int amdgpu_debugfs_gpu_reset(struct seq_file *m, void *data) +{ + struct drm_info_node *node = (struct drm_info_node *) m->private; + struct drm_device *dev = node->minor->dev; + struct amdgpu_device *adev = dev->dev_private; + + seq_printf(m, "gpu reset\n"); + amdgpu_gpu_reset(adev); + + return 0; +} + static struct drm_info_list amdgpu_debugfs_fence_list[] = { {"amdgpu_fence_info", &amdgpu_debugfs_fence_info, 0, NULL}, + {"amdgpu_gpu_reset", &amdgpu_debugfs_gpu_reset, 0, NULL} }; #endif int amdgpu_debugfs_fence_init(struct amdgpu_device *adev) { #if defined(CONFIG_DEBUG_FS) - return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list, 1); + return amdgpu_debugfs_add_files(adev, amdgpu_debugfs_fence_list, 2); #else return 0; #endif diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c index 7380f782cd14..2e26a517f2d6 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c @@ -83,24 +83,32 @@ retry: return r; } *obj = &robj->gem_base; - robj->pid = task_pid_nr(current); - - mutex_lock(&adev->gem.mutex); - list_add_tail(&robj->list, &adev->gem.objects); - mutex_unlock(&adev->gem.mutex); return 0; } -int amdgpu_gem_init(struct amdgpu_device *adev) +void amdgpu_gem_force_release(struct amdgpu_device *adev) { - INIT_LIST_HEAD(&adev->gem.objects); - return 0; -} + struct drm_device *ddev = adev->ddev; + struct drm_file *file; -void amdgpu_gem_fini(struct amdgpu_device *adev) -{ - amdgpu_bo_force_delete(adev); + mutex_lock(&ddev->struct_mutex); + + list_for_each_entry(file, &ddev->filelist, lhead) { + struct drm_gem_object *gobj; + int handle; + + WARN_ONCE(1, "Still active user space clients!\n"); + spin_lock(&file->table_lock); + idr_for_each_entry(&file->object_idr, gobj, handle) { + WARN_ONCE(1, "And also active allocations!\n"); + drm_gem_object_unreference(gobj); + } + idr_destroy(&file->object_idr); + spin_unlock(&file->table_lock); + } + + mutex_unlock(&ddev->struct_mutex); } /* @@ -252,6 +260,8 @@ int amdgpu_gem_userptr_ioctl(struct drm_device *dev, void *data, goto handle_lockup; bo = gem_to_amdgpu_bo(gobj); + bo->prefered_domains = AMDGPU_GEM_DOMAIN_GTT; + bo->allowed_domains = AMDGPU_GEM_DOMAIN_GTT; r = amdgpu_ttm_tt_set_userptr(bo->tbo.ttm, args->addr, args->flags); if (r) goto release_object; @@ -308,7 +318,7 @@ int amdgpu_mode_dumb_mmap(struct drm_file *filp, return -ENOENT; } robj = gem_to_amdgpu_bo(gobj); - if (amdgpu_ttm_tt_has_userptr(robj->tbo.ttm) || + if (amdgpu_ttm_tt_get_usermm(robj->tbo.ttm) || (robj->flags & AMDGPU_GEM_CREATE_NO_CPU_ACCESS)) { drm_gem_object_unreference_unlocked(gobj); return -EPERM; @@ -628,7 +638,7 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data, info.bo_size = robj->gem_base.size; info.alignment = robj->tbo.mem.page_alignment << PAGE_SHIFT; - info.domains = robj->initial_domain; + info.domains = robj->prefered_domains; info.domain_flags = robj->flags; amdgpu_bo_unreserve(robj); if (copy_to_user(out, &info, sizeof(info))) @@ -636,14 +646,18 @@ int amdgpu_gem_op_ioctl(struct drm_device *dev, void *data, break; } case AMDGPU_GEM_OP_SET_PLACEMENT: - if (amdgpu_ttm_tt_has_userptr(robj->tbo.ttm)) { + if (amdgpu_ttm_tt_get_usermm(robj->tbo.ttm)) { r = -EPERM; amdgpu_bo_unreserve(robj); break; } - robj->initial_domain = args->value & (AMDGPU_GEM_DOMAIN_VRAM | - AMDGPU_GEM_DOMAIN_GTT | - AMDGPU_GEM_DOMAIN_CPU); + robj->prefered_domains = args->value & (AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT | + AMDGPU_GEM_DOMAIN_CPU); + robj->allowed_domains = robj->prefered_domains; + if (robj->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM) + robj->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT; + amdgpu_bo_unreserve(robj); break; default: @@ -688,38 +702,73 @@ int amdgpu_mode_dumb_create(struct drm_file *file_priv, } #if defined(CONFIG_DEBUG_FS) +static int amdgpu_debugfs_gem_bo_info(int id, void *ptr, void *data) +{ + struct drm_gem_object *gobj = ptr; + struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj); + struct seq_file *m = data; + + unsigned domain; + const char *placement; + unsigned pin_count; + + domain = amdgpu_mem_type_to_domain(bo->tbo.mem.mem_type); + switch (domain) { + case AMDGPU_GEM_DOMAIN_VRAM: + placement = "VRAM"; + break; + case AMDGPU_GEM_DOMAIN_GTT: + placement = " GTT"; + break; + case AMDGPU_GEM_DOMAIN_CPU: + default: + placement = " CPU"; + break; + } + seq_printf(m, "\t0x%08x: %12ld byte %s @ 0x%010Lx", + id, amdgpu_bo_size(bo), placement, + amdgpu_bo_gpu_offset(bo)); + + pin_count = ACCESS_ONCE(bo->pin_count); + if (pin_count) + seq_printf(m, " pin count %d", pin_count); + seq_printf(m, "\n"); + + return 0; +} + static int amdgpu_debugfs_gem_info(struct seq_file *m, void *data) { struct drm_info_node *node = (struct drm_info_node *)m->private; struct drm_device *dev = node->minor->dev; - struct amdgpu_device *adev = dev->dev_private; - struct amdgpu_bo *rbo; - unsigned i = 0; + struct drm_file *file; + int r; - mutex_lock(&adev->gem.mutex); - list_for_each_entry(rbo, &adev->gem.objects, list) { - unsigned domain; - const char *placement; + r = mutex_lock_interruptible(&dev->struct_mutex); + if (r) + return r; - domain = amdgpu_mem_type_to_domain(rbo->tbo.mem.mem_type); - switch (domain) { - case AMDGPU_GEM_DOMAIN_VRAM: - placement = "VRAM"; - break; - case AMDGPU_GEM_DOMAIN_GTT: - placement = " GTT"; - break; - case AMDGPU_GEM_DOMAIN_CPU: - default: - placement = " CPU"; - break; - } - seq_printf(m, "bo[0x%08x] %8ldkB %8ldMB %s pid %8ld\n", - i, amdgpu_bo_size(rbo) >> 10, amdgpu_bo_size(rbo) >> 20, - placement, (unsigned long)rbo->pid); - i++; + list_for_each_entry(file, &dev->filelist, lhead) { + struct task_struct *task; + + /* + * Although we have a valid reference on file->pid, that does + * not guarantee that the task_struct who called get_pid() is + * still alive (e.g. get_pid(current) => fork() => exit()). + * Therefore, we need to protect this ->comm access using RCU. + */ + rcu_read_lock(); + task = pid_task(file->pid, PIDTYPE_PID); + seq_printf(m, "pid %8d command %s:\n", pid_nr(file->pid), + task ? task->comm : "<unknown>"); + rcu_read_unlock(); + + spin_lock(&file->table_lock); + idr_for_each(&file->object_idr, amdgpu_debugfs_gem_bo_info, m); + spin_unlock(&file->table_lock); } - mutex_unlock(&adev->gem.mutex); + + mutex_unlock(&dev->struct_mutex); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c index 9e25edafa721..b5bdd5d59b58 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ib.c @@ -55,10 +55,9 @@ static int amdgpu_debugfs_sa_init(struct amdgpu_device *adev); * suballocator. * Returns 0 on success, error on failure. */ -int amdgpu_ib_get(struct amdgpu_ring *ring, struct amdgpu_vm *vm, +int amdgpu_ib_get(struct amdgpu_device *adev, struct amdgpu_vm *vm, unsigned size, struct amdgpu_ib *ib) { - struct amdgpu_device *adev = ring->adev; int r; if (size) { @@ -75,9 +74,6 @@ int amdgpu_ib_get(struct amdgpu_ring *ring, struct amdgpu_vm *vm, ib->gpu_addr = amdgpu_sa_bo_gpu_addr(ib->sa_bo); } - amdgpu_sync_create(&ib->sync); - - ib->ring = ring; ib->vm = vm; return 0; @@ -93,7 +89,6 @@ int amdgpu_ib_get(struct amdgpu_ring *ring, struct amdgpu_vm *vm, */ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib) { - amdgpu_sync_free(adev, &ib->sync, &ib->fence->base); amdgpu_sa_bo_free(adev, &ib->sa_bo, &ib->fence->base); if (ib->fence) fence_put(&ib->fence->base); @@ -106,6 +101,7 @@ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib) * @num_ibs: number of IBs to schedule * @ibs: IB objects to schedule * @owner: owner for creating the fences + * @f: fence created during this submission * * Schedule an IB on the associated ring (all asics). * Returns 0 on success, error on failure. @@ -120,11 +116,13 @@ void amdgpu_ib_free(struct amdgpu_device *adev, struct amdgpu_ib *ib) * a CONST_IB), it will be put on the ring prior to the DE IB. Prior * to SI there was just a DE IB. */ -int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, - struct amdgpu_ib *ibs, void *owner) +int amdgpu_ib_schedule(struct amdgpu_ring *ring, unsigned num_ibs, + struct amdgpu_ib *ibs, void *owner, + struct fence *last_vm_update, + struct fence **f) { + struct amdgpu_device *adev = ring->adev; struct amdgpu_ib *ib = &ibs[0]; - struct amdgpu_ring *ring; struct amdgpu_ctx *ctx, *old_ctx; struct amdgpu_vm *vm; unsigned i; @@ -133,7 +131,6 @@ int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, if (num_ibs == 0) return -EINVAL; - ring = ibs->ring; ctx = ibs->ctx; vm = ibs->vm; @@ -141,36 +138,21 @@ int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, dev_err(adev->dev, "couldn't schedule ib\n"); return -EINVAL; } - r = amdgpu_sync_wait(&ibs->sync); - if (r) { - dev_err(adev->dev, "IB sync failed (%d).\n", r); - return r; - } - r = amdgpu_ring_lock(ring, (256 + AMDGPU_NUM_SYNCS * 8) * num_ibs); - if (r) { - dev_err(adev->dev, "scheduling IB failed (%d).\n", r); - return r; - } - if (vm) { - /* grab a vm id if necessary */ - r = amdgpu_vm_grab_id(ibs->vm, ibs->ring, &ibs->sync); - if (r) { - amdgpu_ring_unlock_undo(ring); - return r; - } + if (vm && !ibs->grabbed_vmid) { + dev_err(adev->dev, "VM IB without ID\n"); + return -EINVAL; } - r = amdgpu_sync_rings(&ibs->sync, ring); + r = amdgpu_ring_alloc(ring, 256 * num_ibs); if (r) { - amdgpu_ring_unlock_undo(ring); - dev_err(adev->dev, "failed to sync rings (%d)\n", r); + dev_err(adev->dev, "scheduling IB failed (%d).\n", r); return r; } if (vm) { /* do context switch */ - amdgpu_vm_flush(ring, vm, ib->sync.last_vm_update); + amdgpu_vm_flush(ring, vm, last_vm_update); if (ring->funcs->emit_gds_switch) amdgpu_ring_emit_gds_switch(ring, ib->vm->ids[ring->idx].id, @@ -186,9 +168,9 @@ int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, for (i = 0; i < num_ibs; ++i) { ib = &ibs[i]; - if (ib->ring != ring || ib->ctx != ctx || ib->vm != vm) { + if (ib->ctx != ctx || ib->vm != vm) { ring->current_ctx = old_ctx; - amdgpu_ring_unlock_undo(ring); + amdgpu_ring_undo(ring); return -EINVAL; } amdgpu_ring_emit_ib(ring, ib); @@ -199,14 +181,10 @@ int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, if (r) { dev_err(adev->dev, "failed to emit fence (%d)\n", r); ring->current_ctx = old_ctx; - amdgpu_ring_unlock_undo(ring); + amdgpu_ring_undo(ring); return r; } - if (!amdgpu_enable_scheduler && ib->ctx) - ib->sequence = amdgpu_ctx_add_fence(ib->ctx, ring, - &ib->fence->base); - /* wrap the last IB with fence */ if (ib->user) { uint64_t addr = amdgpu_bo_gpu_offset(ib->user->bo); @@ -215,10 +193,10 @@ int amdgpu_ib_schedule(struct amdgpu_device *adev, unsigned num_ibs, AMDGPU_FENCE_FLAG_64BIT); } - if (ib->vm) - amdgpu_vm_fence(adev, ib->vm, &ib->fence->base); + if (f) + *f = fence_get(&ib->fence->base); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c new file mode 100644 index 000000000000..f29bbb96a881 --- /dev/null +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_job.c @@ -0,0 +1,159 @@ +/* + * Copyright 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR + * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, + * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR + * OTHER DEALINGS IN THE SOFTWARE. + * + * + */ +#include <linux/kthread.h> +#include <linux/wait.h> +#include <linux/sched.h> +#include <drm/drmP.h> +#include "amdgpu.h" +#include "amdgpu_trace.h" + +int amdgpu_job_alloc(struct amdgpu_device *adev, unsigned num_ibs, + struct amdgpu_job **job) +{ + size_t size = sizeof(struct amdgpu_job); + + if (num_ibs == 0) + return -EINVAL; + + size += sizeof(struct amdgpu_ib) * num_ibs; + + *job = kzalloc(size, GFP_KERNEL); + if (!*job) + return -ENOMEM; + + (*job)->adev = adev; + (*job)->ibs = (void *)&(*job)[1]; + (*job)->num_ibs = num_ibs; + + amdgpu_sync_create(&(*job)->sync); + + return 0; +} + +int amdgpu_job_alloc_with_ib(struct amdgpu_device *adev, unsigned size, + struct amdgpu_job **job) +{ + int r; + + r = amdgpu_job_alloc(adev, 1, job); + if (r) + return r; + + r = amdgpu_ib_get(adev, NULL, size, &(*job)->ibs[0]); + if (r) + kfree(*job); + + return r; +} + +void amdgpu_job_free(struct amdgpu_job *job) +{ + unsigned i; + + for (i = 0; i < job->num_ibs; ++i) + amdgpu_ib_free(job->adev, &job->ibs[i]); + + amdgpu_bo_unref(&job->uf.bo); + amdgpu_sync_free(&job->sync); + kfree(job); +} + +int amdgpu_job_submit(struct amdgpu_job *job, struct amdgpu_ring *ring, + struct amd_sched_entity *entity, void *owner, + struct fence **f) +{ + job->ring = ring; + job->base.sched = &ring->sched; + job->base.s_entity = entity; + job->base.s_fence = amd_sched_fence_create(job->base.s_entity, owner); + if (!job->base.s_fence) + return -ENOMEM; + + *f = fence_get(&job->base.s_fence->base); + + job->owner = owner; + amd_sched_entity_push_job(&job->base); + + return 0; +} + +static struct fence *amdgpu_job_dependency(struct amd_sched_job *sched_job) +{ + struct amdgpu_job *job = to_amdgpu_job(sched_job); + struct amdgpu_vm *vm = job->ibs->vm; + + struct fence *fence = amdgpu_sync_get_fence(&job->sync); + + if (fence == NULL && vm && !job->ibs->grabbed_vmid) { + struct amdgpu_ring *ring = job->ring; + int r; + + r = amdgpu_vm_grab_id(vm, ring, &job->sync, + &job->base.s_fence->base); + if (r) + DRM_ERROR("Error getting VM ID (%d)\n", r); + else + job->ibs->grabbed_vmid = true; + + fence = amdgpu_sync_get_fence(&job->sync); + } + + return fence; +} + +static struct fence *amdgpu_job_run(struct amd_sched_job *sched_job) +{ + struct fence *fence = NULL; + struct amdgpu_job *job; + int r; + + if (!sched_job) { + DRM_ERROR("job is null\n"); + return NULL; + } + job = to_amdgpu_job(sched_job); + + r = amdgpu_sync_wait(&job->sync); + if (r) { + DRM_ERROR("failed to sync wait (%d)\n", r); + return NULL; + } + + trace_amdgpu_sched_run_job(job); + r = amdgpu_ib_schedule(job->ring, job->num_ibs, job->ibs, job->owner, + job->sync.last_vm_update, &fence); + if (r) { + DRM_ERROR("Error scheduling IBs (%d)\n", r); + goto err; + } + +err: + amdgpu_job_free(job); + return fence; +} + +struct amd_sched_backend_ops amdgpu_sched_ops = { + .dependency = amdgpu_job_dependency, + .run_job = amdgpu_job_run, +}; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c index e23843f4d877..7805a8706af7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c @@ -447,8 +447,7 @@ static int amdgpu_info_ioctl(struct drm_device *dev, void *data, struct drm_file dev_info.max_memory_clock = adev->pm.default_mclk * 10; } dev_info.enabled_rb_pipes_mask = adev->gfx.config.backend_enable_mask; - dev_info.num_rb_pipes = adev->gfx.config.max_backends_per_se * - adev->gfx.config.max_shader_engines; + dev_info.num_rb_pipes = adev->gfx.config.num_rbs; dev_info.num_hw_gfx_contexts = adev->gfx.config.max_hw_contexts; dev_info._pad = 0; dev_info.ids_flags = 0; @@ -727,6 +726,12 @@ int amdgpu_get_vblank_timestamp_kms(struct drm_device *dev, unsigned int pipe, /* Get associated drm_crtc: */ crtc = &adev->mode_info.crtcs[pipe]->base; + if (!crtc) { + /* This can occur on driver load if some component fails to + * initialize completely and driver is unloaded */ + DRM_ERROR("Uninitialized crtc %d\n", pipe); + return -EINVAL; + } /* Helper routine in DRM core does all the work: */ return drm_calc_vbltimestamp_from_scanoutpos(dev, pipe, max_error, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c index b1969f2b2038..d7ec9bd6755f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mn.c @@ -48,8 +48,7 @@ struct amdgpu_mn { /* protected by adev->mn_lock */ struct hlist_node node; - /* objects protected by lock */ - struct mutex lock; + /* objects protected by mm->mmap_sem */ struct rb_root objects; }; @@ -73,21 +72,19 @@ static void amdgpu_mn_destroy(struct work_struct *work) struct amdgpu_bo *bo, *next_bo; mutex_lock(&adev->mn_lock); - mutex_lock(&rmn->lock); + down_write(&rmn->mm->mmap_sem); hash_del(&rmn->node); rbtree_postorder_for_each_entry_safe(node, next_node, &rmn->objects, it.rb) { - - interval_tree_remove(&node->it, &rmn->objects); list_for_each_entry_safe(bo, next_bo, &node->bos, mn_list) { bo->mn = NULL; list_del_init(&bo->mn_list); } kfree(node); } - mutex_unlock(&rmn->lock); + up_write(&rmn->mm->mmap_sem); mutex_unlock(&adev->mn_lock); - mmu_notifier_unregister(&rmn->mn, rmn->mm); + mmu_notifier_unregister_no_release(&rmn->mn, rmn->mm); kfree(rmn); } @@ -129,8 +126,6 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, /* notification is exclusive, but interval is inclusive */ end -= 1; - mutex_lock(&rmn->lock); - it = interval_tree_iter_first(&rmn->objects, start, end); while (it) { struct amdgpu_mn_node *node; @@ -142,7 +137,8 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, list_for_each_entry(bo, &node->bos, mn_list) { - if (!bo->tbo.ttm || bo->tbo.ttm->state != tt_bound) + if (!amdgpu_ttm_tt_affect_userptr(bo->tbo.ttm, start, + end)) continue; r = amdgpu_bo_reserve(bo, true); @@ -164,8 +160,6 @@ static void amdgpu_mn_invalidate_range_start(struct mmu_notifier *mn, amdgpu_bo_unreserve(bo); } } - - mutex_unlock(&rmn->lock); } static const struct mmu_notifier_ops amdgpu_mn_ops = { @@ -186,8 +180,8 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) struct amdgpu_mn *rmn; int r; - down_write(&mm->mmap_sem); mutex_lock(&adev->mn_lock); + down_write(&mm->mmap_sem); hash_for_each_possible(adev->mn_hash, rmn, node, (unsigned long)mm) if (rmn->mm == mm) @@ -202,7 +196,6 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) rmn->adev = adev; rmn->mm = mm; rmn->mn.ops = &amdgpu_mn_ops; - mutex_init(&rmn->lock); rmn->objects = RB_ROOT; r = __mmu_notifier_register(&rmn->mn, mm); @@ -212,14 +205,14 @@ static struct amdgpu_mn *amdgpu_mn_get(struct amdgpu_device *adev) hash_add(adev->mn_hash, &rmn->node, (unsigned long)mm); release_locks: - mutex_unlock(&adev->mn_lock); up_write(&mm->mmap_sem); + mutex_unlock(&adev->mn_lock); return rmn; free_rmn: - mutex_unlock(&adev->mn_lock); up_write(&mm->mmap_sem); + mutex_unlock(&adev->mn_lock); kfree(rmn); return ERR_PTR(r); @@ -249,7 +242,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) INIT_LIST_HEAD(&bos); - mutex_lock(&rmn->lock); + down_write(&rmn->mm->mmap_sem); while ((it = interval_tree_iter_first(&rmn->objects, addr, end))) { kfree(node); @@ -263,7 +256,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) if (!node) { node = kmalloc(sizeof(struct amdgpu_mn_node), GFP_KERNEL); if (!node) { - mutex_unlock(&rmn->lock); + up_write(&rmn->mm->mmap_sem); return -ENOMEM; } } @@ -278,7 +271,7 @@ int amdgpu_mn_register(struct amdgpu_bo *bo, unsigned long addr) interval_tree_insert(&node->it, &rmn->objects); - mutex_unlock(&rmn->lock); + up_write(&rmn->mm->mmap_sem); return 0; } @@ -297,13 +290,15 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo) struct list_head *head; mutex_lock(&adev->mn_lock); + rmn = bo->mn; if (rmn == NULL) { mutex_unlock(&adev->mn_lock); return; } - mutex_lock(&rmn->lock); + down_write(&rmn->mm->mmap_sem); + /* save the next list entry for later */ head = bo->mn_list.next; @@ -317,6 +312,6 @@ void amdgpu_mn_unregister(struct amdgpu_bo *bo) kfree(node); } - mutex_unlock(&rmn->lock); + up_write(&rmn->mm->mmap_sem); mutex_unlock(&adev->mn_lock); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h index fdc1be8550da..8d432e6901af 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mode.h @@ -390,7 +390,6 @@ struct amdgpu_crtc { struct drm_display_mode native_mode; u32 pll_id; /* page flipping */ - struct workqueue_struct *pflip_queue; struct amdgpu_flip_work *pflip_works; enum amdgpu_flip_status pflip_status; int deferred_flip_completion; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index b8fbbd7699e4..9a025a77958d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -97,9 +97,6 @@ static void amdgpu_ttm_bo_destroy(struct ttm_buffer_object *tbo) amdgpu_update_memory_usage(bo->adev, &bo->tbo.mem, NULL); - mutex_lock(&bo->adev->gem.mutex); - list_del_init(&bo->list); - mutex_unlock(&bo->adev->gem.mutex); drm_gem_object_release(&bo->gem_base); amdgpu_bo_unref(&bo->parent); kfree(bo->metadata); @@ -254,12 +251,15 @@ int amdgpu_bo_create_restricted(struct amdgpu_device *adev, bo->adev = adev; INIT_LIST_HEAD(&bo->list); INIT_LIST_HEAD(&bo->va); - bo->initial_domain = domain & (AMDGPU_GEM_DOMAIN_VRAM | - AMDGPU_GEM_DOMAIN_GTT | - AMDGPU_GEM_DOMAIN_CPU | - AMDGPU_GEM_DOMAIN_GDS | - AMDGPU_GEM_DOMAIN_GWS | - AMDGPU_GEM_DOMAIN_OA); + bo->prefered_domains = domain & (AMDGPU_GEM_DOMAIN_VRAM | + AMDGPU_GEM_DOMAIN_GTT | + AMDGPU_GEM_DOMAIN_CPU | + AMDGPU_GEM_DOMAIN_GDS | + AMDGPU_GEM_DOMAIN_GWS | + AMDGPU_GEM_DOMAIN_OA); + bo->allowed_domains = bo->prefered_domains; + if (!kernel && bo->allowed_domains == AMDGPU_GEM_DOMAIN_VRAM) + bo->allowed_domains |= AMDGPU_GEM_DOMAIN_GTT; bo->flags = flags; @@ -367,7 +367,7 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, int r, i; unsigned fpfn, lpfn; - if (amdgpu_ttm_tt_has_userptr(bo->tbo.ttm)) + if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) return -EPERM; if (WARN_ON_ONCE(min_offset > max_offset)) @@ -470,26 +470,6 @@ int amdgpu_bo_evict_vram(struct amdgpu_device *adev) return ttm_bo_evict_mm(&adev->mman.bdev, TTM_PL_VRAM); } -void amdgpu_bo_force_delete(struct amdgpu_device *adev) -{ - struct amdgpu_bo *bo, *n; - - if (list_empty(&adev->gem.objects)) { - return; - } - dev_err(adev->dev, "Userspace still has active objects !\n"); - list_for_each_entry_safe(bo, n, &adev->gem.objects, list) { - dev_err(adev->dev, "%p %p %lu %lu force free\n", - &bo->gem_base, bo, (unsigned long)bo->gem_base.size, - *((unsigned long *)&bo->gem_base.refcount)); - mutex_lock(&bo->adev->gem.mutex); - list_del_init(&bo->list); - mutex_unlock(&bo->adev->gem.mutex); - /* this should unref the ttm bo */ - drm_gem_object_unreference_unlocked(&bo->gem_base); - } -} - int amdgpu_bo_init(struct amdgpu_device *adev) { /* Add an MTRR for the VRAM */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h index 5107fb291bdb..acc08018c6cc 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.h @@ -149,7 +149,6 @@ int amdgpu_bo_pin_restricted(struct amdgpu_bo *bo, u32 domain, u64 *gpu_addr); int amdgpu_bo_unpin(struct amdgpu_bo *bo); int amdgpu_bo_evict_vram(struct amdgpu_device *adev); -void amdgpu_bo_force_delete(struct amdgpu_device *adev); int amdgpu_bo_init(struct amdgpu_device *adev); void amdgpu_bo_fini(struct amdgpu_device *adev); int amdgpu_bo_fbdev_mmap(struct amdgpu_bo *bo, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c index 7d8d84eaea4a..d77b2bdbe800 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_pm.c @@ -119,7 +119,9 @@ static ssize_t amdgpu_get_dpm_forced_performance_level(struct device *dev, level = amdgpu_dpm_get_performance_level(adev); return snprintf(buf, PAGE_SIZE, "%s\n", (level == AMD_DPM_FORCED_LEVEL_AUTO) ? "auto" : - (level == AMD_DPM_FORCED_LEVEL_LOW) ? "low" : "high"); + (level == AMD_DPM_FORCED_LEVEL_LOW) ? "low" : + (level == AMD_DPM_FORCED_LEVEL_HIGH) ? "high" : + (level == AMD_DPM_FORCED_LEVEL_MANUAL) ? "manual" : "unknown"); } else { enum amdgpu_dpm_forced_level level; @@ -146,6 +148,8 @@ static ssize_t amdgpu_set_dpm_forced_performance_level(struct device *dev, level = AMDGPU_DPM_FORCED_LEVEL_HIGH; } else if (strncmp("auto", buf, strlen("auto")) == 0) { level = AMDGPU_DPM_FORCED_LEVEL_AUTO; + } else if (strncmp("manual", buf, strlen("manual")) == 0) { + level = AMDGPU_DPM_FORCED_LEVEL_MANUAL; } else { count = -EINVAL; goto fail; @@ -172,10 +176,293 @@ fail: return count; } +static ssize_t amdgpu_get_pp_num_states(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + struct pp_states_info data; + int i, buf_len; + + if (adev->pp_enabled) + amdgpu_dpm_get_pp_num_states(adev, &data); + + buf_len = snprintf(buf, PAGE_SIZE, "states: %d\n", data.nums); + for (i = 0; i < data.nums; i++) + buf_len += snprintf(buf + buf_len, PAGE_SIZE, "%d %s\n", i, + (data.states[i] == POWER_STATE_TYPE_INTERNAL_BOOT) ? "boot" : + (data.states[i] == POWER_STATE_TYPE_BATTERY) ? "battery" : + (data.states[i] == POWER_STATE_TYPE_BALANCED) ? "balanced" : + (data.states[i] == POWER_STATE_TYPE_PERFORMANCE) ? "performance" : "default"); + + return buf_len; +} + +static ssize_t amdgpu_get_pp_cur_state(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + struct pp_states_info data; + enum amd_pm_state_type pm = 0; + int i = 0; + + if (adev->pp_enabled) { + + pm = amdgpu_dpm_get_current_power_state(adev); + amdgpu_dpm_get_pp_num_states(adev, &data); + + for (i = 0; i < data.nums; i++) { + if (pm == data.states[i]) + break; + } + + if (i == data.nums) + i = -EINVAL; + } + + return snprintf(buf, PAGE_SIZE, "%d\n", i); +} + +static ssize_t amdgpu_get_pp_force_state(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + struct pp_states_info data; + enum amd_pm_state_type pm = 0; + int i; + + if (adev->pp_force_state_enabled && adev->pp_enabled) { + pm = amdgpu_dpm_get_current_power_state(adev); + amdgpu_dpm_get_pp_num_states(adev, &data); + + for (i = 0; i < data.nums; i++) { + if (pm == data.states[i]) + break; + } + + if (i == data.nums) + i = -EINVAL; + + return snprintf(buf, PAGE_SIZE, "%d\n", i); + + } else + return snprintf(buf, PAGE_SIZE, "\n"); +} + +static ssize_t amdgpu_set_pp_force_state(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + enum amd_pm_state_type state = 0; + long idx; + int ret; + + if (strlen(buf) == 1) + adev->pp_force_state_enabled = false; + else { + ret = kstrtol(buf, 0, &idx); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) { + struct pp_states_info data; + amdgpu_dpm_get_pp_num_states(adev, &data); + state = data.states[idx]; + /* only set user selected power states */ + if (state != POWER_STATE_TYPE_INTERNAL_BOOT && + state != POWER_STATE_TYPE_DEFAULT) { + amdgpu_dpm_dispatch_task(adev, + AMD_PP_EVENT_ENABLE_USER_STATE, &state, NULL); + adev->pp_force_state_enabled = true; + } + } + } +fail: + return count; +} + +static ssize_t amdgpu_get_pp_table(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + char *table = NULL; + int size, i; + + if (adev->pp_enabled) + size = amdgpu_dpm_get_pp_table(adev, &table); + else + return 0; + + if (size >= PAGE_SIZE) + size = PAGE_SIZE - 1; + + for (i = 0; i < size; i++) { + sprintf(buf + i, "%02x", table[i]); + } + sprintf(buf + i, "\n"); + + return size; +} + +static ssize_t amdgpu_set_pp_table(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + + if (adev->pp_enabled) + amdgpu_dpm_set_pp_table(adev, buf, count); + + return count; +} + +static ssize_t amdgpu_get_pp_dpm_sclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + ssize_t size = 0; + + if (adev->pp_enabled) + size = amdgpu_dpm_print_clock_levels(adev, PP_SCLK, buf); + + return size; +} + +static ssize_t amdgpu_set_pp_dpm_sclk(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + long level; + + ret = kstrtol(buf, 0, &level); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) + amdgpu_dpm_force_clock_level(adev, PP_SCLK, level); +fail: + return count; +} + +static ssize_t amdgpu_get_pp_dpm_mclk(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + ssize_t size = 0; + + if (adev->pp_enabled) + size = amdgpu_dpm_print_clock_levels(adev, PP_MCLK, buf); + + return size; +} + +static ssize_t amdgpu_set_pp_dpm_mclk(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + long level; + + ret = kstrtol(buf, 0, &level); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) + amdgpu_dpm_force_clock_level(adev, PP_MCLK, level); +fail: + return count; +} + +static ssize_t amdgpu_get_pp_dpm_pcie(struct device *dev, + struct device_attribute *attr, + char *buf) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + ssize_t size = 0; + + if (adev->pp_enabled) + size = amdgpu_dpm_print_clock_levels(adev, PP_PCIE, buf); + + return size; +} + +static ssize_t amdgpu_set_pp_dpm_pcie(struct device *dev, + struct device_attribute *attr, + const char *buf, + size_t count) +{ + struct drm_device *ddev = dev_get_drvdata(dev); + struct amdgpu_device *adev = ddev->dev_private; + int ret; + long level; + + ret = kstrtol(buf, 0, &level); + + if (ret) { + count = -EINVAL; + goto fail; + } + + if (adev->pp_enabled) + amdgpu_dpm_force_clock_level(adev, PP_PCIE, level); +fail: + return count; +} + static DEVICE_ATTR(power_dpm_state, S_IRUGO | S_IWUSR, amdgpu_get_dpm_state, amdgpu_set_dpm_state); static DEVICE_ATTR(power_dpm_force_performance_level, S_IRUGO | S_IWUSR, amdgpu_get_dpm_forced_performance_level, amdgpu_set_dpm_forced_performance_level); +static DEVICE_ATTR(pp_num_states, S_IRUGO, amdgpu_get_pp_num_states, NULL); +static DEVICE_ATTR(pp_cur_state, S_IRUGO, amdgpu_get_pp_cur_state, NULL); +static DEVICE_ATTR(pp_force_state, S_IRUGO | S_IWUSR, + amdgpu_get_pp_force_state, + amdgpu_set_pp_force_state); +static DEVICE_ATTR(pp_table, S_IRUGO | S_IWUSR, + amdgpu_get_pp_table, + amdgpu_set_pp_table); +static DEVICE_ATTR(pp_dpm_sclk, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_sclk, + amdgpu_set_pp_dpm_sclk); +static DEVICE_ATTR(pp_dpm_mclk, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_mclk, + amdgpu_set_pp_dpm_mclk); +static DEVICE_ATTR(pp_dpm_pcie, S_IRUGO | S_IWUSR, + amdgpu_get_pp_dpm_pcie, + amdgpu_set_pp_dpm_pcie); static ssize_t amdgpu_hwmon_show_temp(struct device *dev, struct device_attribute *attr, @@ -623,14 +910,12 @@ force: amdgpu_dpm_print_power_state(adev, adev->pm.dpm.requested_ps); } - mutex_lock(&adev->ring_lock); - /* update whether vce is active */ ps->vce_active = adev->pm.dpm.vce_active; ret = amdgpu_dpm_pre_set_power_state(adev); if (ret) - goto done; + return; /* update display watermarks based on new power state */ amdgpu_display_bandwidth_update(adev); @@ -667,9 +952,6 @@ force: amdgpu_dpm_force_performance_level(adev, adev->pm.dpm.forced_level); } } - -done: - mutex_unlock(&adev->ring_lock); } void amdgpu_dpm_enable_uvd(struct amdgpu_device *adev, bool enable) @@ -770,6 +1052,44 @@ int amdgpu_pm_sysfs_init(struct amdgpu_device *adev) DRM_ERROR("failed to create device file for dpm state\n"); return ret; } + + if (adev->pp_enabled) { + ret = device_create_file(adev->dev, &dev_attr_pp_num_states); + if (ret) { + DRM_ERROR("failed to create device file pp_num_states\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_cur_state); + if (ret) { + DRM_ERROR("failed to create device file pp_cur_state\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_force_state); + if (ret) { + DRM_ERROR("failed to create device file pp_force_state\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_table); + if (ret) { + DRM_ERROR("failed to create device file pp_table\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_sclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_sclk\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_mclk); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_mclk\n"); + return ret; + } + ret = device_create_file(adev->dev, &dev_attr_pp_dpm_pcie); + if (ret) { + DRM_ERROR("failed to create device file pp_dpm_pcie\n"); + return ret; + } + } ret = amdgpu_debugfs_pm_init(adev); if (ret) { DRM_ERROR("Failed to register debugfs file for dpm!\n"); @@ -787,6 +1107,15 @@ void amdgpu_pm_sysfs_fini(struct amdgpu_device *adev) hwmon_device_unregister(adev->pm.int_hwmon_dev); device_remove_file(adev->dev, &dev_attr_power_dpm_state); device_remove_file(adev->dev, &dev_attr_power_dpm_force_performance_level); + if (adev->pp_enabled) { + device_remove_file(adev->dev, &dev_attr_pp_num_states); + device_remove_file(adev->dev, &dev_attr_pp_cur_state); + device_remove_file(adev->dev, &dev_attr_pp_force_state); + device_remove_file(adev->dev, &dev_attr_pp_table); + device_remove_file(adev->dev, &dev_attr_pp_dpm_sclk); + device_remove_file(adev->dev, &dev_attr_pp_dpm_mclk); + device_remove_file(adev->dev, &dev_attr_pp_dpm_pcie); + } } void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) @@ -802,13 +1131,11 @@ void amdgpu_pm_compute_clocks(struct amdgpu_device *adev) int i = 0; amdgpu_display_bandwidth_update(adev); - mutex_lock(&adev->ring_lock); - for (i = 0; i < AMDGPU_MAX_RINGS; i++) { - struct amdgpu_ring *ring = adev->rings[i]; - if (ring && ring->ready) - amdgpu_fence_wait_empty(ring); - } - mutex_unlock(&adev->ring_lock); + for (i = 0; i < AMDGPU_MAX_RINGS; i++) { + struct amdgpu_ring *ring = adev->rings[i]; + if (ring && ring->ready) + amdgpu_fence_wait_empty(ring); + } amdgpu_dpm_dispatch_task(adev, AMD_PP_EVENT_DISPLAY_CONFIG_CHANGE, NULL, NULL); } else { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c index 59f735a933a9..be6388f73ba2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_prime.c @@ -73,10 +73,6 @@ struct drm_gem_object *amdgpu_gem_prime_import_sg_table(struct drm_device *dev, if (ret) return ERR_PTR(ret); - mutex_lock(&adev->gem.mutex); - list_add_tail(&bo->list, &adev->gem.objects); - mutex_unlock(&adev->gem.mutex); - return &bo->gem_base; } @@ -121,7 +117,7 @@ struct dma_buf *amdgpu_gem_prime_export(struct drm_device *dev, { struct amdgpu_bo *bo = gem_to_amdgpu_bo(gobj); - if (amdgpu_ttm_tt_has_userptr(bo->tbo.ttm)) + if (amdgpu_ttm_tt_get_usermm(bo->tbo.ttm)) return ERR_PTR(-EPERM); return drm_gem_prime_export(dev, gobj, flags); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c index d1f234dd2126..56c07e3fdb33 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c @@ -49,28 +49,6 @@ static int amdgpu_debugfs_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring); /** - * amdgpu_ring_free_size - update the free size - * - * @adev: amdgpu_device pointer - * @ring: amdgpu_ring structure holding ring information - * - * Update the free dw slots in the ring buffer (all asics). - */ -void amdgpu_ring_free_size(struct amdgpu_ring *ring) -{ - uint32_t rptr = amdgpu_ring_get_rptr(ring); - - /* This works because ring_size is a power of 2 */ - ring->ring_free_dw = rptr + (ring->ring_size / 4); - ring->ring_free_dw -= ring->wptr; - ring->ring_free_dw &= ring->ptr_mask; - if (!ring->ring_free_dw) { - /* this is an empty ring */ - ring->ring_free_dw = ring->ring_size / 4; - } -} - -/** * amdgpu_ring_alloc - allocate space on the ring buffer * * @adev: amdgpu_device pointer @@ -82,50 +60,18 @@ void amdgpu_ring_free_size(struct amdgpu_ring *ring) */ int amdgpu_ring_alloc(struct amdgpu_ring *ring, unsigned ndw) { - int r; - - /* make sure we aren't trying to allocate more space than there is on the ring */ - if (ndw > (ring->ring_size / 4)) - return -ENOMEM; /* Align requested size with padding so unlock_commit can * pad safely */ - amdgpu_ring_free_size(ring); ndw = (ndw + ring->align_mask) & ~ring->align_mask; - while (ndw > (ring->ring_free_dw - 1)) { - amdgpu_ring_free_size(ring); - if (ndw < ring->ring_free_dw) { - break; - } - r = amdgpu_fence_wait_next(ring); - if (r) - return r; - } - ring->count_dw = ndw; - ring->wptr_old = ring->wptr; - return 0; -} -/** - * amdgpu_ring_lock - lock the ring and allocate space on it - * - * @adev: amdgpu_device pointer - * @ring: amdgpu_ring structure holding ring information - * @ndw: number of dwords to allocate in the ring buffer - * - * Lock the ring and allocate @ndw dwords in the ring buffer - * (all asics). - * Returns 0 on success, error on failure. - */ -int amdgpu_ring_lock(struct amdgpu_ring *ring, unsigned ndw) -{ - int r; + /* Make sure we aren't trying to allocate more space + * than the maximum for one submission + */ + if (WARN_ON_ONCE(ndw > ring->max_dw)) + return -ENOMEM; - mutex_lock(ring->ring_lock); - r = amdgpu_ring_alloc(ring, ndw); - if (r) { - mutex_unlock(ring->ring_lock); - return r; - } + ring->count_dw = ndw; + ring->wptr_old = ring->wptr; return 0; } @@ -144,6 +90,19 @@ void amdgpu_ring_insert_nop(struct amdgpu_ring *ring, uint32_t count) amdgpu_ring_write(ring, ring->nop); } +/** amdgpu_ring_generic_pad_ib - pad IB with NOP packets + * + * @ring: amdgpu_ring structure holding ring information + * @ib: IB to add NOP packets to + * + * This is the generic pad_ib function for rings except SDMA + */ +void amdgpu_ring_generic_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib) +{ + while (ib->length_dw & ring->align_mask) + ib->ptr[ib->length_dw++] = ring->nop; +} + /** * amdgpu_ring_commit - tell the GPU to execute the new * commands on the ring buffer @@ -168,20 +127,6 @@ void amdgpu_ring_commit(struct amdgpu_ring *ring) } /** - * amdgpu_ring_unlock_commit - tell the GPU to execute the new - * commands on the ring buffer and unlock it - * - * @ring: amdgpu_ring structure holding ring information - * - * Call amdgpu_ring_commit() then unlock the ring (all asics). - */ -void amdgpu_ring_unlock_commit(struct amdgpu_ring *ring) -{ - amdgpu_ring_commit(ring); - mutex_unlock(ring->ring_lock); -} - -/** * amdgpu_ring_undo - reset the wptr * * @ring: amdgpu_ring structure holding ring information @@ -194,19 +139,6 @@ void amdgpu_ring_undo(struct amdgpu_ring *ring) } /** - * amdgpu_ring_unlock_undo - reset the wptr and unlock the ring - * - * @ring: amdgpu_ring structure holding ring information - * - * Call amdgpu_ring_undo() then unlock the ring (all asics). - */ -void amdgpu_ring_unlock_undo(struct amdgpu_ring *ring) -{ - amdgpu_ring_undo(ring); - mutex_unlock(ring->ring_lock); -} - -/** * amdgpu_ring_backup - Back up the content of a ring * * @ring: the ring we want to back up @@ -218,43 +150,32 @@ unsigned amdgpu_ring_backup(struct amdgpu_ring *ring, { unsigned size, ptr, i; - /* just in case lock the ring */ - mutex_lock(ring->ring_lock); *data = NULL; - if (ring->ring_obj == NULL) { - mutex_unlock(ring->ring_lock); + if (ring->ring_obj == NULL) return 0; - } /* it doesn't make sense to save anything if all fences are signaled */ - if (!amdgpu_fence_count_emitted(ring)) { - mutex_unlock(ring->ring_lock); + if (!amdgpu_fence_count_emitted(ring)) return 0; - } ptr = le32_to_cpu(*ring->next_rptr_cpu_addr); size = ring->wptr + (ring->ring_size / 4); size -= ptr; size &= ring->ptr_mask; - if (size == 0) { - mutex_unlock(ring->ring_lock); + if (size == 0) return 0; - } /* and then save the content of the ring */ *data = kmalloc_array(size, sizeof(uint32_t), GFP_KERNEL); - if (!*data) { - mutex_unlock(ring->ring_lock); + if (!*data) return 0; - } for (i = 0; i < size; ++i) { (*data)[i] = ring->ring[ptr++]; ptr &= ring->ptr_mask; } - mutex_unlock(ring->ring_lock); return size; } @@ -276,7 +197,7 @@ int amdgpu_ring_restore(struct amdgpu_ring *ring, return 0; /* restore the saved ring content */ - r = amdgpu_ring_lock(ring, size); + r = amdgpu_ring_alloc(ring, size); if (r) return r; @@ -284,7 +205,7 @@ int amdgpu_ring_restore(struct amdgpu_ring *ring, amdgpu_ring_write(ring, data[i]); } - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); kfree(data); return 0; } @@ -352,7 +273,6 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, return r; } - ring->ring_lock = &adev->ring_lock; /* Align ring size */ rb_bufsz = order_base_2(ring_size / 8); ring_size = (1 << (rb_bufsz + 1)) * 4; @@ -389,7 +309,8 @@ int amdgpu_ring_init(struct amdgpu_device *adev, struct amdgpu_ring *ring, } } ring->ptr_mask = (ring->ring_size / 4) - 1; - ring->ring_free_dw = ring->ring_size / 4; + ring->max_dw = DIV_ROUND_UP(ring->ring_size / 4, + amdgpu_sched_hw_submission); if (amdgpu_debugfs_ring_init(adev, ring)) { DRM_ERROR("Failed to register debugfs file for rings !\n"); @@ -410,15 +331,10 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring) int r; struct amdgpu_bo *ring_obj; - if (ring->ring_lock == NULL) - return; - - mutex_lock(ring->ring_lock); ring_obj = ring->ring_obj; ring->ready = false; ring->ring = NULL; ring->ring_obj = NULL; - mutex_unlock(ring->ring_lock); amdgpu_wb_free(ring->adev, ring->fence_offs); amdgpu_wb_free(ring->adev, ring->rptr_offs); @@ -474,29 +390,18 @@ static int amdgpu_debugfs_ring_info(struct seq_file *m, void *data) struct amdgpu_ring *ring = (void *)(((uint8_t*)adev) + roffset); uint32_t rptr, wptr, rptr_next; - unsigned count, i, j; - - amdgpu_ring_free_size(ring); - count = (ring->ring_size / 4) - ring->ring_free_dw; + unsigned i; wptr = amdgpu_ring_get_wptr(ring); - seq_printf(m, "wptr: 0x%08x [%5d]\n", - wptr, wptr); + seq_printf(m, "wptr: 0x%08x [%5d]\n", wptr, wptr); rptr = amdgpu_ring_get_rptr(ring); - seq_printf(m, "rptr: 0x%08x [%5d]\n", - rptr, rptr); - rptr_next = le32_to_cpu(*ring->next_rptr_cpu_addr); + seq_printf(m, "rptr: 0x%08x [%5d]\n", rptr, rptr); + seq_printf(m, "driver's copy of the wptr: 0x%08x [%5d]\n", ring->wptr, ring->wptr); - seq_printf(m, "last semaphore signal addr : 0x%016llx\n", - ring->last_semaphore_signal_addr); - seq_printf(m, "last semaphore wait addr : 0x%016llx\n", - ring->last_semaphore_wait_addr); - seq_printf(m, "%u free dwords in ring\n", ring->ring_free_dw); - seq_printf(m, "%u dwords in ring\n", count); if (!ring->ready) return 0; @@ -505,11 +410,20 @@ static int amdgpu_debugfs_ring_info(struct seq_file *m, void *data) * packet that is the root issue */ i = (rptr + ring->ptr_mask + 1 - 32) & ring->ptr_mask; - for (j = 0; j <= (count + 32); j++) { + while (i != rptr) { + seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]); + if (i == rptr) + seq_puts(m, " *"); + if (i == rptr_next) + seq_puts(m, " #"); + seq_puts(m, "\n"); + i = (i + 1) & ring->ptr_mask; + } + while (i != wptr) { seq_printf(m, "r[%5d]=0x%08x", i, ring->ring[i]); - if (rptr == i) + if (i == rptr) seq_puts(m, " *"); - if (rptr_next == i) + if (i == rptr_next) seq_puts(m, " #"); seq_puts(m, "\n"); i = (i + 1) & ring->ptr_mask; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c index 8b88edb0434b..7d8f8f1e3f7f 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c @@ -321,8 +321,11 @@ int amdgpu_sa_bo_new(struct amdgpu_sa_manager *sa_manager, int i, r; signed long t; - BUG_ON(align > sa_manager->align); - BUG_ON(size > sa_manager->size); + if (WARN_ON_ONCE(align > sa_manager->align)) + return -EINVAL; + + if (WARN_ON_ONCE(size > sa_manager->size)) + return -EINVAL; *sa_bo = kmalloc(sizeof(struct amdgpu_sa_bo), GFP_KERNEL); if ((*sa_bo) == NULL) { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c deleted file mode 100644 index 438c05254695..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sched.c +++ /dev/null @@ -1,108 +0,0 @@ -/* - * Copyright 2015 Advanced Micro Devices, Inc. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the "Software"), - * to deal in the Software without restriction, including without limitation - * the rights to use, copy, modify, merge, publish, distribute, sublicense, - * and/or sell copies of the Software, and to permit persons to whom the - * Software is furnished to do so, subject to the following conditions: - * - * The above copyright notice and this permission notice shall be included in - * all copies or substantial portions of the Software. - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR - * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, - * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR - * OTHER DEALINGS IN THE SOFTWARE. - * - * - */ -#include <linux/kthread.h> -#include <linux/wait.h> -#include <linux/sched.h> -#include <drm/drmP.h> -#include "amdgpu.h" -#include "amdgpu_trace.h" - -static struct fence *amdgpu_sched_dependency(struct amd_sched_job *sched_job) -{ - struct amdgpu_job *job = to_amdgpu_job(sched_job); - return amdgpu_sync_get_fence(&job->ibs->sync); -} - -static struct fence *amdgpu_sched_run_job(struct amd_sched_job *sched_job) -{ - struct amdgpu_fence *fence = NULL; - struct amdgpu_job *job; - int r; - - if (!sched_job) { - DRM_ERROR("job is null\n"); - return NULL; - } - job = to_amdgpu_job(sched_job); - trace_amdgpu_sched_run_job(job); - r = amdgpu_ib_schedule(job->adev, job->num_ibs, job->ibs, job->owner); - if (r) { - DRM_ERROR("Error scheduling IBs (%d)\n", r); - goto err; - } - - fence = job->ibs[job->num_ibs - 1].fence; - fence_get(&fence->base); - -err: - if (job->free_job) - job->free_job(job); - - kfree(job); - return fence ? &fence->base : NULL; -} - -struct amd_sched_backend_ops amdgpu_sched_ops = { - .dependency = amdgpu_sched_dependency, - .run_job = amdgpu_sched_run_job, -}; - -int amdgpu_sched_ib_submit_kernel_helper(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct amdgpu_ib *ibs, - unsigned num_ibs, - int (*free_job)(struct amdgpu_job *), - void *owner, - struct fence **f) -{ - int r = 0; - if (amdgpu_enable_scheduler) { - struct amdgpu_job *job = - kzalloc(sizeof(struct amdgpu_job), GFP_KERNEL); - if (!job) - return -ENOMEM; - job->base.sched = &ring->sched; - job->base.s_entity = &adev->kernel_ctx.rings[ring->idx].entity; - job->base.s_fence = amd_sched_fence_create(job->base.s_entity, owner); - if (!job->base.s_fence) { - kfree(job); - return -ENOMEM; - } - *f = fence_get(&job->base.s_fence->base); - - job->adev = adev; - job->ibs = ibs; - job->num_ibs = num_ibs; - job->owner = owner; - job->free_job = free_job; - amd_sched_entity_push_job(&job->base); - } else { - r = amdgpu_ib_schedule(adev, num_ibs, ibs, owner); - if (r) - return r; - *f = fence_get(&ibs[num_ibs - 1].fence->base); - } - - return 0; -} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_semaphore.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_semaphore.c deleted file mode 100644 index 1caaf201b708..000000000000 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_semaphore.c +++ /dev/null @@ -1,102 +0,0 @@ -/* - * Copyright 2011 Christian König. - * All Rights Reserved. - * - * Permission is hereby granted, free of charge, to any person obtaining a - * copy of this software and associated documentation files (the - * "Software"), to deal in the Software without restriction, including - * without limitation the rights to use, copy, modify, merge, publish, - * distribute, sub license, and/or sell copies of the Software, and to - * permit persons to whom the Software is furnished to do so, subject to - * the following conditions: - * - * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR - * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL - * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, - * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR - * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE - * USE OR OTHER DEALINGS IN THE SOFTWARE. - * - * The above copyright notice and this permission notice (including the - * next paragraph) shall be included in all copies or substantial portions - * of the Software. - * - */ -/* - * Authors: - * Christian König <[email protected]> - */ -#include <drm/drmP.h> -#include "amdgpu.h" -#include "amdgpu_trace.h" - -int amdgpu_semaphore_create(struct amdgpu_device *adev, - struct amdgpu_semaphore **semaphore) -{ - int r; - - *semaphore = kmalloc(sizeof(struct amdgpu_semaphore), GFP_KERNEL); - if (*semaphore == NULL) { - return -ENOMEM; - } - r = amdgpu_sa_bo_new(&adev->ring_tmp_bo, - &(*semaphore)->sa_bo, 8, 8); - if (r) { - kfree(*semaphore); - *semaphore = NULL; - return r; - } - (*semaphore)->waiters = 0; - (*semaphore)->gpu_addr = amdgpu_sa_bo_gpu_addr((*semaphore)->sa_bo); - - *((uint64_t *)amdgpu_sa_bo_cpu_addr((*semaphore)->sa_bo)) = 0; - - return 0; -} - -bool amdgpu_semaphore_emit_signal(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore) -{ - trace_amdgpu_semaphore_signale(ring->idx, semaphore); - - if (amdgpu_ring_emit_semaphore(ring, semaphore, false)) { - --semaphore->waiters; - - /* for debugging lockup only, used by sysfs debug files */ - ring->last_semaphore_signal_addr = semaphore->gpu_addr; - return true; - } - return false; -} - -bool amdgpu_semaphore_emit_wait(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore) -{ - trace_amdgpu_semaphore_wait(ring->idx, semaphore); - - if (amdgpu_ring_emit_semaphore(ring, semaphore, true)) { - ++semaphore->waiters; - - /* for debugging lockup only, used by sysfs debug files */ - ring->last_semaphore_wait_addr = semaphore->gpu_addr; - return true; - } - return false; -} - -void amdgpu_semaphore_free(struct amdgpu_device *adev, - struct amdgpu_semaphore **semaphore, - struct fence *fence) -{ - if (semaphore == NULL || *semaphore == NULL) { - return; - } - if ((*semaphore)->waiters > 0) { - dev_err(adev->dev, "semaphore %p has more waiters than signalers," - " hardware lockup imminent!\n", *semaphore); - } - amdgpu_sa_bo_free(adev, &(*semaphore)->sa_bo, fence); - kfree(*semaphore); - *semaphore = NULL; -} diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c index 181ce39ef5e5..c15be00de904 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sync.c @@ -46,14 +46,6 @@ struct amdgpu_sync_entry { */ void amdgpu_sync_create(struct amdgpu_sync *sync) { - unsigned i; - - for (i = 0; i < AMDGPU_NUM_SYNCS; ++i) - sync->semaphores[i] = NULL; - - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) - sync->sync_to[i] = NULL; - hash_init(sync->fences); sync->last_vm_update = NULL; } @@ -107,7 +99,6 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, struct fence *f) { struct amdgpu_sync_entry *e; - struct amdgpu_fence *fence; if (!f) return 0; @@ -116,27 +107,20 @@ int amdgpu_sync_fence(struct amdgpu_device *adev, struct amdgpu_sync *sync, amdgpu_sync_test_owner(f, AMDGPU_FENCE_OWNER_VM)) amdgpu_sync_keep_later(&sync->last_vm_update, f); - fence = to_amdgpu_fence(f); - if (!fence || fence->ring->adev != adev) { - hash_for_each_possible(sync->fences, e, node, f->context) { - if (unlikely(e->fence->context != f->context)) - continue; - - amdgpu_sync_keep_later(&e->fence, f); - return 0; - } - - e = kmalloc(sizeof(struct amdgpu_sync_entry), GFP_KERNEL); - if (!e) - return -ENOMEM; + hash_for_each_possible(sync->fences, e, node, f->context) { + if (unlikely(e->fence->context != f->context)) + continue; - hash_add(sync->fences, &e->node, f->context); - e->fence = fence_get(f); + amdgpu_sync_keep_later(&e->fence, f); return 0; } - amdgpu_sync_keep_later(&sync->sync_to[fence->ring->idx], f); + e = kmalloc(sizeof(struct amdgpu_sync_entry), GFP_KERNEL); + if (!e) + return -ENOMEM; + hash_add(sync->fences, &e->node, f->context); + e->fence = fence_get(f); return 0; } @@ -153,13 +137,13 @@ static void *amdgpu_sync_get_owner(struct fence *f) } /** - * amdgpu_sync_resv - use the semaphores to sync to a reservation object + * amdgpu_sync_resv - sync to a reservation object * * @sync: sync object to add fences from reservation object to * @resv: reservation object with embedded fence * @shared: true if we should only sync to the exclusive fence * - * Sync to the fence using the semaphore objects + * Sync to the fence */ int amdgpu_sync_resv(struct amdgpu_device *adev, struct amdgpu_sync *sync, @@ -250,123 +234,17 @@ int amdgpu_sync_wait(struct amdgpu_sync *sync) kfree(e); } - if (amdgpu_enable_semaphores) - return 0; - - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - struct fence *fence = sync->sync_to[i]; - if (!fence) - continue; - - r = fence_wait(fence, false); - if (r) - return r; - } - - return 0; -} - -/** - * amdgpu_sync_rings - sync ring to all registered fences - * - * @sync: sync object to use - * @ring: ring that needs sync - * - * Ensure that all registered fences are signaled before letting - * the ring continue. The caller must hold the ring lock. - */ -int amdgpu_sync_rings(struct amdgpu_sync *sync, - struct amdgpu_ring *ring) -{ - struct amdgpu_device *adev = ring->adev; - unsigned count = 0; - int i, r; - - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { - struct amdgpu_ring *other = adev->rings[i]; - struct amdgpu_semaphore *semaphore; - struct amdgpu_fence *fence; - - if (!sync->sync_to[i]) - continue; - - fence = to_amdgpu_fence(sync->sync_to[i]); - - /* check if we really need to sync */ - if (!amdgpu_enable_scheduler && - !amdgpu_fence_need_sync(fence, ring)) - continue; - - /* prevent GPU deadlocks */ - if (!other->ready) { - dev_err(adev->dev, "Syncing to a disabled ring!"); - return -EINVAL; - } - - if (amdgpu_enable_scheduler || !amdgpu_enable_semaphores) { - r = fence_wait(sync->sync_to[i], true); - if (r) - return r; - continue; - } - - if (count >= AMDGPU_NUM_SYNCS) { - /* not enough room, wait manually */ - r = fence_wait(&fence->base, false); - if (r) - return r; - continue; - } - r = amdgpu_semaphore_create(adev, &semaphore); - if (r) - return r; - - sync->semaphores[count++] = semaphore; - - /* allocate enough space for sync command */ - r = amdgpu_ring_alloc(other, 16); - if (r) - return r; - - /* emit the signal semaphore */ - if (!amdgpu_semaphore_emit_signal(other, semaphore)) { - /* signaling wasn't successful wait manually */ - amdgpu_ring_undo(other); - r = fence_wait(&fence->base, false); - if (r) - return r; - continue; - } - - /* we assume caller has already allocated space on waiters ring */ - if (!amdgpu_semaphore_emit_wait(ring, semaphore)) { - /* waiting wasn't successful wait manually */ - amdgpu_ring_undo(other); - r = fence_wait(&fence->base, false); - if (r) - return r; - continue; - } - - amdgpu_ring_commit(other); - amdgpu_fence_note_sync(fence, ring); - } - return 0; } /** * amdgpu_sync_free - free the sync object * - * @adev: amdgpu_device pointer * @sync: sync object to use - * @fence: fence to use for the free * - * Free the sync object by freeing all semaphores in it. + * Free the sync object. */ -void amdgpu_sync_free(struct amdgpu_device *adev, - struct amdgpu_sync *sync, - struct fence *fence) +void amdgpu_sync_free(struct amdgpu_sync *sync) { struct amdgpu_sync_entry *e; struct hlist_node *tmp; @@ -378,11 +256,5 @@ void amdgpu_sync_free(struct amdgpu_device *adev, kfree(e); } - for (i = 0; i < AMDGPU_NUM_SYNCS; ++i) - amdgpu_semaphore_free(adev, &sync->semaphores[i], fence); - - for (i = 0; i < AMDGPU_MAX_RINGS; ++i) - fence_put(sync->sync_to[i]); - fence_put(sync->last_vm_update); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c index 4865615e9c06..05a53f4fc334 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_test.c @@ -238,144 +238,10 @@ void amdgpu_test_moves(struct amdgpu_device *adev) amdgpu_do_test_moves(adev); } -static int amdgpu_test_create_and_emit_fence(struct amdgpu_device *adev, - struct amdgpu_ring *ring, - struct fence **fence) -{ - uint32_t handle = ring->idx ^ 0xdeafbeef; - int r; - - if (ring == &adev->uvd.ring) { - r = amdgpu_uvd_get_create_msg(ring, handle, NULL); - if (r) { - DRM_ERROR("Failed to get dummy create msg\n"); - return r; - } - - r = amdgpu_uvd_get_destroy_msg(ring, handle, fence); - if (r) { - DRM_ERROR("Failed to get dummy destroy msg\n"); - return r; - } - - } else if (ring == &adev->vce.ring[0] || - ring == &adev->vce.ring[1]) { - r = amdgpu_vce_get_create_msg(ring, handle, NULL); - if (r) { - DRM_ERROR("Failed to get dummy create msg\n"); - return r; - } - - r = amdgpu_vce_get_destroy_msg(ring, handle, fence); - if (r) { - DRM_ERROR("Failed to get dummy destroy msg\n"); - return r; - } - } else { - struct amdgpu_fence *a_fence = NULL; - r = amdgpu_ring_lock(ring, 64); - if (r) { - DRM_ERROR("Failed to lock ring A %d\n", ring->idx); - return r; - } - amdgpu_fence_emit(ring, AMDGPU_FENCE_OWNER_UNDEFINED, &a_fence); - amdgpu_ring_unlock_commit(ring); - *fence = &a_fence->base; - } - return 0; -} - void amdgpu_test_ring_sync(struct amdgpu_device *adev, struct amdgpu_ring *ringA, struct amdgpu_ring *ringB) { - struct fence *fence1 = NULL, *fence2 = NULL; - struct amdgpu_semaphore *semaphore = NULL; - int r; - - r = amdgpu_semaphore_create(adev, &semaphore); - if (r) { - DRM_ERROR("Failed to create semaphore\n"); - goto out_cleanup; - } - - r = amdgpu_ring_lock(ringA, 64); - if (r) { - DRM_ERROR("Failed to lock ring A %d\n", ringA->idx); - goto out_cleanup; - } - amdgpu_semaphore_emit_wait(ringA, semaphore); - amdgpu_ring_unlock_commit(ringA); - - r = amdgpu_test_create_and_emit_fence(adev, ringA, &fence1); - if (r) - goto out_cleanup; - - r = amdgpu_ring_lock(ringA, 64); - if (r) { - DRM_ERROR("Failed to lock ring A %d\n", ringA->idx); - goto out_cleanup; - } - amdgpu_semaphore_emit_wait(ringA, semaphore); - amdgpu_ring_unlock_commit(ringA); - - r = amdgpu_test_create_and_emit_fence(adev, ringA, &fence2); - if (r) - goto out_cleanup; - - mdelay(1000); - - if (fence_is_signaled(fence1)) { - DRM_ERROR("Fence 1 signaled without waiting for semaphore.\n"); - goto out_cleanup; - } - - r = amdgpu_ring_lock(ringB, 64); - if (r) { - DRM_ERROR("Failed to lock ring B %p\n", ringB); - goto out_cleanup; - } - amdgpu_semaphore_emit_signal(ringB, semaphore); - amdgpu_ring_unlock_commit(ringB); - - r = fence_wait(fence1, false); - if (r) { - DRM_ERROR("Failed to wait for sync fence 1\n"); - goto out_cleanup; - } - - mdelay(1000); - - if (fence_is_signaled(fence2)) { - DRM_ERROR("Fence 2 signaled without waiting for semaphore.\n"); - goto out_cleanup; - } - - r = amdgpu_ring_lock(ringB, 64); - if (r) { - DRM_ERROR("Failed to lock ring B %p\n", ringB); - goto out_cleanup; - } - amdgpu_semaphore_emit_signal(ringB, semaphore); - amdgpu_ring_unlock_commit(ringB); - - r = fence_wait(fence2, false); - if (r) { - DRM_ERROR("Failed to wait for sync fence 1\n"); - goto out_cleanup; - } - -out_cleanup: - amdgpu_semaphore_free(adev, &semaphore, NULL); - - if (fence1) - fence_put(fence1); - - if (fence2) - fence_put(fence2); - - if (r) - printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); } static void amdgpu_test_ring_sync2(struct amdgpu_device *adev, @@ -383,109 +249,6 @@ static void amdgpu_test_ring_sync2(struct amdgpu_device *adev, struct amdgpu_ring *ringB, struct amdgpu_ring *ringC) { - struct fence *fenceA = NULL, *fenceB = NULL; - struct amdgpu_semaphore *semaphore = NULL; - bool sigA, sigB; - int i, r; - - r = amdgpu_semaphore_create(adev, &semaphore); - if (r) { - DRM_ERROR("Failed to create semaphore\n"); - goto out_cleanup; - } - - r = amdgpu_ring_lock(ringA, 64); - if (r) { - DRM_ERROR("Failed to lock ring A %d\n", ringA->idx); - goto out_cleanup; - } - amdgpu_semaphore_emit_wait(ringA, semaphore); - amdgpu_ring_unlock_commit(ringA); - - r = amdgpu_test_create_and_emit_fence(adev, ringA, &fenceA); - if (r) - goto out_cleanup; - - r = amdgpu_ring_lock(ringB, 64); - if (r) { - DRM_ERROR("Failed to lock ring B %d\n", ringB->idx); - goto out_cleanup; - } - amdgpu_semaphore_emit_wait(ringB, semaphore); - amdgpu_ring_unlock_commit(ringB); - r = amdgpu_test_create_and_emit_fence(adev, ringB, &fenceB); - if (r) - goto out_cleanup; - - mdelay(1000); - - if (fence_is_signaled(fenceA)) { - DRM_ERROR("Fence A signaled without waiting for semaphore.\n"); - goto out_cleanup; - } - if (fence_is_signaled(fenceB)) { - DRM_ERROR("Fence B signaled without waiting for semaphore.\n"); - goto out_cleanup; - } - - r = amdgpu_ring_lock(ringC, 64); - if (r) { - DRM_ERROR("Failed to lock ring B %p\n", ringC); - goto out_cleanup; - } - amdgpu_semaphore_emit_signal(ringC, semaphore); - amdgpu_ring_unlock_commit(ringC); - - for (i = 0; i < 30; ++i) { - mdelay(100); - sigA = fence_is_signaled(fenceA); - sigB = fence_is_signaled(fenceB); - if (sigA || sigB) - break; - } - - if (!sigA && !sigB) { - DRM_ERROR("Neither fence A nor B has been signaled\n"); - goto out_cleanup; - } else if (sigA && sigB) { - DRM_ERROR("Both fence A and B has been signaled\n"); - goto out_cleanup; - } - - DRM_INFO("Fence %c was first signaled\n", sigA ? 'A' : 'B'); - - r = amdgpu_ring_lock(ringC, 64); - if (r) { - DRM_ERROR("Failed to lock ring B %p\n", ringC); - goto out_cleanup; - } - amdgpu_semaphore_emit_signal(ringC, semaphore); - amdgpu_ring_unlock_commit(ringC); - - mdelay(1000); - - r = fence_wait(fenceA, false); - if (r) { - DRM_ERROR("Failed to wait for sync fence A\n"); - goto out_cleanup; - } - r = fence_wait(fenceB, false); - if (r) { - DRM_ERROR("Failed to wait for sync fence B\n"); - goto out_cleanup; - } - -out_cleanup: - amdgpu_semaphore_free(adev, &semaphore, NULL); - - if (fenceA) - fence_put(fenceA); - - if (fenceB) - fence_put(fenceB); - - if (r) - printk(KERN_WARNING "Error while testing ring sync (%d).\n", r); } static bool amdgpu_test_sync_possible(struct amdgpu_ring *ringA, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h index 8f9834ab1bd5..9ca3735c563c 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_trace.h @@ -38,10 +38,10 @@ TRACE_EVENT(amdgpu_cs, TP_fast_assign( __entry->bo_list = p->bo_list; - __entry->ring = p->ibs[i].ring->idx; - __entry->dw = p->ibs[i].length_dw; + __entry->ring = p->job->ring->idx; + __entry->dw = p->job->ibs[i].length_dw; __entry->fences = amdgpu_fence_count_emitted( - p->ibs[i].ring); + p->job->ring); ), TP_printk("bo_list=%p, ring=%u, dw=%u, fences=%u", __entry->bo_list, __entry->ring, __entry->dw, @@ -65,7 +65,7 @@ TRACE_EVENT(amdgpu_cs_ioctl, __entry->sched_job = &job->base; __entry->ib = job->ibs; __entry->fence = &job->base.s_fence->base; - __entry->ring_name = job->ibs[0].ring->name; + __entry->ring_name = job->ring->name; __entry->num_ibs = job->num_ibs; ), TP_printk("adev=%p, sched_job=%p, first ib=%p, sched fence=%p, ring name:%s, num_ibs:%u", @@ -90,7 +90,7 @@ TRACE_EVENT(amdgpu_sched_run_job, __entry->sched_job = &job->base; __entry->ib = job->ibs; __entry->fence = &job->base.s_fence->base; - __entry->ring_name = job->ibs[0].ring->name; + __entry->ring_name = job->ring->name; __entry->num_ibs = job->num_ibs; ), TP_printk("adev=%p, sched_job=%p, first ib=%p, sched fence=%p, ring name:%s, num_ibs:%u", @@ -100,18 +100,21 @@ TRACE_EVENT(amdgpu_sched_run_job, TRACE_EVENT(amdgpu_vm_grab_id, - TP_PROTO(unsigned vmid, int ring), - TP_ARGS(vmid, ring), + TP_PROTO(struct amdgpu_vm *vm, unsigned vmid, int ring), + TP_ARGS(vm, vmid, ring), TP_STRUCT__entry( + __field(struct amdgpu_vm *, vm) __field(u32, vmid) __field(u32, ring) ), TP_fast_assign( + __entry->vm = vm; __entry->vmid = vmid; __entry->ring = ring; ), - TP_printk("vmid=%u, ring=%u", __entry->vmid, __entry->ring) + TP_printk("vm=%p, id=%u, ring=%u", __entry->vm, __entry->vmid, + __entry->ring) ); TRACE_EVENT(amdgpu_vm_bo_map, @@ -247,42 +250,6 @@ TRACE_EVENT(amdgpu_bo_list_set, TP_printk("list=%p, bo=%p", __entry->list, __entry->bo) ); -DECLARE_EVENT_CLASS(amdgpu_semaphore_request, - - TP_PROTO(int ring, struct amdgpu_semaphore *sem), - - TP_ARGS(ring, sem), - - TP_STRUCT__entry( - __field(int, ring) - __field(signed, waiters) - __field(uint64_t, gpu_addr) - ), - - TP_fast_assign( - __entry->ring = ring; - __entry->waiters = sem->waiters; - __entry->gpu_addr = sem->gpu_addr; - ), - - TP_printk("ring=%u, waiters=%d, addr=%010Lx", __entry->ring, - __entry->waiters, __entry->gpu_addr) -); - -DEFINE_EVENT(amdgpu_semaphore_request, amdgpu_semaphore_signale, - - TP_PROTO(int ring, struct amdgpu_semaphore *sem), - - TP_ARGS(ring, sem) -); - -DEFINE_EVENT(amdgpu_semaphore_request, amdgpu_semaphore_wait, - - TP_PROTO(int ring, struct amdgpu_semaphore *sem), - - TP_ARGS(ring, sem) -); - #endif /* This part must be outside protection */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 55cf05e1c81c..e52fc641edfb 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -77,6 +77,8 @@ static void amdgpu_ttm_mem_global_release(struct drm_global_reference *ref) static int amdgpu_ttm_global_init(struct amdgpu_device *adev) { struct drm_global_reference *global_ref; + struct amdgpu_ring *ring; + struct amd_sched_rq *rq; int r; adev->mman.mem_global_referenced = false; @@ -106,13 +108,27 @@ static int amdgpu_ttm_global_init(struct amdgpu_device *adev) return r; } + ring = adev->mman.buffer_funcs_ring; + rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL]; + r = amd_sched_entity_init(&ring->sched, &adev->mman.entity, + rq, amdgpu_sched_jobs); + if (r != 0) { + DRM_ERROR("Failed setting up TTM BO move run queue.\n"); + drm_global_item_unref(&adev->mman.mem_global_ref); + drm_global_item_unref(&adev->mman.bo_global_ref.ref); + return r; + } + adev->mman.mem_global_referenced = true; + return 0; } static void amdgpu_ttm_global_fini(struct amdgpu_device *adev) { if (adev->mman.mem_global_referenced) { + amd_sched_entity_fini(adev->mman.entity.sched, + &adev->mman.entity); drm_global_item_unref(&adev->mman.bo_global_ref.ref); drm_global_item_unref(&adev->mman.mem_global_ref); adev->mman.mem_global_referenced = false; @@ -499,9 +515,6 @@ static int amdgpu_ttm_tt_pin_userptr(struct ttm_tt *ttm) enum dma_data_direction direction = write ? DMA_BIDIRECTIONAL : DMA_TO_DEVICE; - if (current->mm != gtt->usermm) - return -EPERM; - if (gtt->userflags & AMDGPU_GEM_USERPTR_ANONONLY) { /* check that we only pin down anonymous memory to prevent problems with writeback */ @@ -773,14 +786,33 @@ int amdgpu_ttm_tt_set_userptr(struct ttm_tt *ttm, uint64_t addr, return 0; } -bool amdgpu_ttm_tt_has_userptr(struct ttm_tt *ttm) +struct mm_struct *amdgpu_ttm_tt_get_usermm(struct ttm_tt *ttm) +{ + struct amdgpu_ttm_tt *gtt = (void *)ttm; + + if (gtt == NULL) + return NULL; + + return gtt->usermm; +} + +bool amdgpu_ttm_tt_affect_userptr(struct ttm_tt *ttm, unsigned long start, + unsigned long end) { struct amdgpu_ttm_tt *gtt = (void *)ttm; + unsigned long size; if (gtt == NULL) return false; - return !!gtt->userptr; + if (gtt->ttm.ttm.state != tt_bound || !gtt->userptr) + return false; + + size = (unsigned long)gtt->ttm.ttm.num_pages * PAGE_SIZE; + if (gtt->userptr > end || gtt->userptr + size <= start) + return false; + + return true; } bool amdgpu_ttm_tt_is_readonly(struct ttm_tt *ttm) @@ -996,9 +1028,10 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, struct fence **fence) { struct amdgpu_device *adev = ring->adev; + struct amdgpu_job *job; + uint32_t max_bytes; unsigned num_loops, num_dw; - struct amdgpu_ib *ib; unsigned i; int r; @@ -1010,20 +1043,12 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, while (num_dw & 0x7) num_dw++; - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) - return -ENOMEM; - - r = amdgpu_ib_get(ring, NULL, num_dw * 4, ib); - if (r) { - kfree(ib); + r = amdgpu_job_alloc_with_ib(adev, num_dw * 4, &job); + if (r) return r; - } - - ib->length_dw = 0; if (resv) { - r = amdgpu_sync_resv(adev, &ib->sync, resv, + r = amdgpu_sync_resv(adev, &job->sync, resv, AMDGPU_FENCE_OWNER_UNDEFINED); if (r) { DRM_ERROR("sync failed (%d).\n", r); @@ -1034,31 +1059,25 @@ int amdgpu_copy_buffer(struct amdgpu_ring *ring, for (i = 0; i < num_loops; i++) { uint32_t cur_size_in_bytes = min(byte_count, max_bytes); - amdgpu_emit_copy_buffer(adev, ib, src_offset, dst_offset, - cur_size_in_bytes); + amdgpu_emit_copy_buffer(adev, &job->ibs[0], src_offset, + dst_offset, cur_size_in_bytes); src_offset += cur_size_in_bytes; dst_offset += cur_size_in_bytes; byte_count -= cur_size_in_bytes; } - amdgpu_vm_pad_ib(adev, ib); - WARN_ON(ib->length_dw > num_dw); - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vm_free_job, - AMDGPU_FENCE_OWNER_UNDEFINED, - fence); + amdgpu_ring_pad_ib(ring, &job->ibs[0]); + WARN_ON(job->ibs[0].length_dw > num_dw); + r = amdgpu_job_submit(job, ring, &adev->mman.entity, + AMDGPU_FENCE_OWNER_UNDEFINED, fence); if (r) goto error_free; - if (!amdgpu_enable_scheduler) { - amdgpu_ib_free(adev, ib); - kfree(ib); - } return 0; + error_free: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); return r; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index 53f987aeeacf..1de82bf4fc79 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -91,6 +91,8 @@ static void amdgpu_uvd_idle_work_handler(struct work_struct *work); int amdgpu_uvd_sw_init(struct amdgpu_device *adev) { + struct amdgpu_ring *ring; + struct amd_sched_rq *rq; unsigned long bo_size; const char *fw_name; const struct common_firmware_header *hdr; @@ -191,6 +193,15 @@ int amdgpu_uvd_sw_init(struct amdgpu_device *adev) amdgpu_bo_unreserve(adev->uvd.vcpu_bo); + ring = &adev->uvd.ring; + rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; + r = amd_sched_entity_init(&ring->sched, &adev->uvd.entity, + rq, amdgpu_sched_jobs); + if (r != 0) { + DRM_ERROR("Failed setting up UVD run queue.\n"); + return r; + } + for (i = 0; i < AMDGPU_MAX_UVD_HANDLES; ++i) { atomic_set(&adev->uvd.handles[i], 0); adev->uvd.filp[i] = NULL; @@ -210,6 +221,8 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) if (adev->uvd.vcpu_bo == NULL) return 0; + amd_sched_entity_fini(&adev->uvd.ring.sched, &adev->uvd.entity); + r = amdgpu_bo_reserve(adev->uvd.vcpu_bo, false); if (!r) { amdgpu_bo_kunmap(adev->uvd.vcpu_bo); @@ -241,7 +254,7 @@ int amdgpu_uvd_suspend(struct amdgpu_device *adev) amdgpu_uvd_note_usage(adev); - r = amdgpu_uvd_get_destroy_msg(ring, handle, &fence); + r = amdgpu_uvd_get_destroy_msg(ring, handle, false, &fence); if (r) { DRM_ERROR("Error destroying UVD (%d)!\n", r); continue; @@ -295,7 +308,8 @@ void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp) amdgpu_uvd_note_usage(adev); - r = amdgpu_uvd_get_destroy_msg(ring, handle, &fence); + r = amdgpu_uvd_get_destroy_msg(ring, handle, + false, &fence); if (r) { DRM_ERROR("Error destroying UVD (%d)!\n", r); continue; @@ -616,7 +630,6 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) { struct amdgpu_bo_va_mapping *mapping; struct amdgpu_bo *bo; - struct amdgpu_ib *ib; uint32_t cmd, lo, hi; uint64_t start, end; uint64_t addr; @@ -638,9 +651,10 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) addr -= ((uint64_t)mapping->it.start) * AMDGPU_GPU_PAGE_SIZE; start += addr; - ib = &ctx->parser->ibs[ctx->ib_idx]; - ib->ptr[ctx->data0] = start & 0xFFFFFFFF; - ib->ptr[ctx->data1] = start >> 32; + amdgpu_set_ib_value(ctx->parser, ctx->ib_idx, ctx->data0, + lower_32_bits(start)); + amdgpu_set_ib_value(ctx->parser, ctx->ib_idx, ctx->data1, + upper_32_bits(start)); cmd = amdgpu_get_ib_value(ctx->parser, ctx->ib_idx, ctx->idx) >> 1; if (cmd < 0x4) { @@ -702,7 +716,7 @@ static int amdgpu_uvd_cs_pass2(struct amdgpu_uvd_cs_ctx *ctx) static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx, int (*cb)(struct amdgpu_uvd_cs_ctx *ctx)) { - struct amdgpu_ib *ib = &ctx->parser->ibs[ctx->ib_idx]; + struct amdgpu_ib *ib = &ctx->parser->job->ibs[ctx->ib_idx]; int i, r; ctx->idx++; @@ -748,7 +762,7 @@ static int amdgpu_uvd_cs_reg(struct amdgpu_uvd_cs_ctx *ctx, static int amdgpu_uvd_cs_packets(struct amdgpu_uvd_cs_ctx *ctx, int (*cb)(struct amdgpu_uvd_cs_ctx *ctx)) { - struct amdgpu_ib *ib = &ctx->parser->ibs[ctx->ib_idx]; + struct amdgpu_ib *ib = &ctx->parser->job->ibs[ctx->ib_idx]; int r; for (ctx->idx = 0 ; ctx->idx < ib->length_dw; ) { @@ -790,7 +804,7 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx) [0x00000003] = 2048, [0x00000004] = 0xFFFFFFFF, }; - struct amdgpu_ib *ib = &parser->ibs[ib_idx]; + struct amdgpu_ib *ib = &parser->job->ibs[ib_idx]; int r; if (ib->length_dw % 16) { @@ -823,22 +837,14 @@ int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx) return 0; } -static int amdgpu_uvd_free_job( - struct amdgpu_job *job) -{ - amdgpu_ib_free(job->adev, job->ibs); - kfree(job->ibs); - return 0; -} - -static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, - struct amdgpu_bo *bo, - struct fence **fence) +static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, struct amdgpu_bo *bo, + bool direct, struct fence **fence) { struct ttm_validate_buffer tv; struct ww_acquire_ctx ticket; struct list_head head; - struct amdgpu_ib *ib = NULL; + struct amdgpu_job *job; + struct amdgpu_ib *ib; struct fence *f = NULL; struct amdgpu_device *adev = ring->adev; uint64_t addr; @@ -862,15 +868,12 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, r = ttm_bo_validate(&bo->tbo, &bo->placement, true, false); if (r) goto err; - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) { - r = -ENOMEM; - goto err; - } - r = amdgpu_ib_get(ring, NULL, 64, ib); + + r = amdgpu_job_alloc_with_ib(adev, 64, &job); if (r) - goto err1; + goto err; + ib = &job->ibs[0]; addr = amdgpu_bo_gpu_offset(bo); ib->ptr[0] = PACKET0(mmUVD_GPCOM_VCPU_DATA0, 0); ib->ptr[1] = addr; @@ -882,12 +885,19 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, ib->ptr[i] = PACKET2(0); ib->length_dw = 16; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_uvd_free_job, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); - if (r) - goto err2; + if (direct) { + r = amdgpu_ib_schedule(ring, 1, ib, + AMDGPU_FENCE_OWNER_UNDEFINED, NULL, &f); + if (r) + goto err_free; + + amdgpu_job_free(job); + } else { + r = amdgpu_job_submit(job, ring, &adev->uvd.entity, + AMDGPU_FENCE_OWNER_UNDEFINED, &f); + if (r) + goto err_free; + } ttm_eu_fence_buffer_objects(&ticket, &head, f); @@ -895,16 +905,12 @@ static int amdgpu_uvd_send_msg(struct amdgpu_ring *ring, *fence = fence_get(f); amdgpu_bo_unref(&bo); fence_put(f); - if (amdgpu_enable_scheduler) - return 0; - amdgpu_ib_free(ring->adev, ib); - kfree(ib); return 0; -err2: - amdgpu_ib_free(ring->adev, ib); -err1: - kfree(ib); + +err_free: + amdgpu_job_free(job); + err: ttm_eu_backoff_reservation(&ticket, &head); return r; @@ -959,11 +965,11 @@ int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, amdgpu_bo_kunmap(bo); amdgpu_bo_unreserve(bo); - return amdgpu_uvd_send_msg(ring, bo, fence); + return amdgpu_uvd_send_msg(ring, bo, true, fence); } int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct fence **fence) + bool direct, struct fence **fence) { struct amdgpu_device *adev = ring->adev; struct amdgpu_bo *bo; @@ -1001,7 +1007,7 @@ int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, amdgpu_bo_kunmap(bo); amdgpu_bo_unreserve(bo); - return amdgpu_uvd_send_msg(ring, bo, fence); + return amdgpu_uvd_send_msg(ring, bo, direct, fence); } static void amdgpu_uvd_idle_work_handler(struct work_struct *work) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h index 1724c2c86151..9a3b449081a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.h @@ -31,7 +31,7 @@ int amdgpu_uvd_resume(struct amdgpu_device *adev); int amdgpu_uvd_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct fence **fence); int amdgpu_uvd_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct fence **fence); + bool direct, struct fence **fence); void amdgpu_uvd_free_handles(struct amdgpu_device *adev, struct drm_file *filp); int amdgpu_uvd_ring_parse_cs(struct amdgpu_cs_parser *parser, uint32_t ib_idx); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index a745eeeb5d82..39c3aa60381a 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -74,6 +74,8 @@ static void amdgpu_vce_idle_work_handler(struct work_struct *work); */ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) { + struct amdgpu_ring *ring; + struct amd_sched_rq *rq; const char *fw_name; const struct common_firmware_header *hdr; unsigned ucode_version, version_major, version_minor, binary_id; @@ -170,6 +172,16 @@ int amdgpu_vce_sw_init(struct amdgpu_device *adev, unsigned long size) return r; } + + ring = &adev->vce.ring[0]; + rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_NORMAL]; + r = amd_sched_entity_init(&ring->sched, &adev->vce.entity, + rq, amdgpu_sched_jobs); + if (r != 0) { + DRM_ERROR("Failed setting up VCE run queue.\n"); + return r; + } + for (i = 0; i < AMDGPU_MAX_VCE_HANDLES; ++i) { atomic_set(&adev->vce.handles[i], 0); adev->vce.filp[i] = NULL; @@ -190,6 +202,8 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev) if (adev->vce.vcpu_bo == NULL) return 0; + amd_sched_entity_fini(&adev->vce.ring[0].sched, &adev->vce.entity); + amdgpu_bo_unref(&adev->vce.vcpu_bo); amdgpu_ring_fini(&adev->vce.ring[0]); @@ -337,7 +351,7 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) amdgpu_vce_note_usage(adev); - r = amdgpu_vce_get_destroy_msg(ring, handle, NULL); + r = amdgpu_vce_get_destroy_msg(ring, handle, false, NULL); if (r) DRM_ERROR("Error destroying VCE handle (%d)!\n", r); @@ -346,14 +360,6 @@ void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp) } } -static int amdgpu_vce_free_job( - struct amdgpu_job *job) -{ - amdgpu_ib_free(job->adev, job->ibs); - kfree(job->ibs); - return 0; -} - /** * amdgpu_vce_get_create_msg - generate a VCE create msg * @@ -368,21 +374,17 @@ int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct fence **fence) { const unsigned ib_size_dw = 1024; - struct amdgpu_ib *ib = NULL; + struct amdgpu_job *job; + struct amdgpu_ib *ib; struct fence *f = NULL; - struct amdgpu_device *adev = ring->adev; uint64_t dummy; int i, r; - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) - return -ENOMEM; - r = amdgpu_ib_get(ring, NULL, ib_size_dw * 4, ib); - if (r) { - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); - kfree(ib); + r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job); + if (r) return r; - } + + ib = &job->ibs[0]; dummy = ib->gpu_addr + 1024; @@ -423,20 +425,19 @@ int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vce_free_job, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err; + + amdgpu_job_free(job); if (fence) *fence = fence_get(f); fence_put(f); - if (amdgpu_enable_scheduler) - return 0; + return 0; + err: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); return r; } @@ -451,26 +452,20 @@ err: * Close up a stream for HW test or if userspace failed to do so */ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct fence **fence) + bool direct, struct fence **fence) { const unsigned ib_size_dw = 1024; - struct amdgpu_ib *ib = NULL; + struct amdgpu_job *job; + struct amdgpu_ib *ib; struct fence *f = NULL; - struct amdgpu_device *adev = ring->adev; uint64_t dummy; int i, r; - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) - return -ENOMEM; - - r = amdgpu_ib_get(ring, NULL, ib_size_dw * 4, ib); - if (r) { - kfree(ib); - DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); + r = amdgpu_job_alloc_with_ib(ring->adev, ib_size_dw * 4, &job); + if (r) return r; - } + ib = &job->ibs[0]; dummy = ib->gpu_addr + 1024; /* stitch together an VCE destroy msg */ @@ -490,20 +485,29 @@ int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, for (i = ib->length_dw; i < ib_size_dw; ++i) ib->ptr[i] = 0x0; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vce_free_job, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); - if (r) - goto err; + + if (direct) { + r = amdgpu_ib_schedule(ring, 1, ib, + AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); + if (r) + goto err; + + amdgpu_job_free(job); + } else { + r = amdgpu_job_submit(job, ring, &ring->adev->vce.entity, + AMDGPU_FENCE_OWNER_UNDEFINED, &f); + if (r) + goto err; + } + if (fence) *fence = fence_get(f); fence_put(f); - if (amdgpu_enable_scheduler) - return 0; + return 0; + err: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); return r; } @@ -521,7 +525,6 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx, int lo, int hi, unsigned size, uint32_t index) { struct amdgpu_bo_va_mapping *mapping; - struct amdgpu_ib *ib = &p->ibs[ib_idx]; struct amdgpu_bo *bo; uint64_t addr; @@ -550,8 +553,8 @@ static int amdgpu_vce_cs_reloc(struct amdgpu_cs_parser *p, uint32_t ib_idx, addr += amdgpu_bo_gpu_offset(bo); addr -= ((uint64_t)size) * ((uint64_t)index); - ib->ptr[lo] = addr & 0xFFFFFFFF; - ib->ptr[hi] = addr >> 32; + amdgpu_set_ib_value(p, ib_idx, lo, lower_32_bits(addr)); + amdgpu_set_ib_value(p, ib_idx, hi, upper_32_bits(addr)); return 0; } @@ -606,7 +609,7 @@ static int amdgpu_vce_validate_handle(struct amdgpu_cs_parser *p, */ int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx) { - struct amdgpu_ib *ib = &p->ibs[ib_idx]; + struct amdgpu_ib *ib = &p->job->ibs[ib_idx]; unsigned fb_idx = 0, bs_idx = 0; int session_idx = -1; bool destroyed = false; @@ -743,30 +746,6 @@ out: } /** - * amdgpu_vce_ring_emit_semaphore - emit a semaphore command - * - * @ring: engine to use - * @semaphore: address of semaphore - * @emit_wait: true=emit wait, false=emit signal - * - */ -bool amdgpu_vce_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - amdgpu_ring_write(ring, VCE_CMD_SEMAPHORE); - amdgpu_ring_write(ring, (addr >> 3) & 0x000FFFFF); - amdgpu_ring_write(ring, (addr >> 23) & 0x000FFFFF); - amdgpu_ring_write(ring, 0x01003000 | (emit_wait ? 1 : 0)); - if (!emit_wait) - amdgpu_ring_write(ring, VCE_CMD_END); - - return true; -} - -/** * amdgpu_vce_ring_emit_ib - execute indirect buffer * * @ring: engine to use @@ -814,14 +793,14 @@ int amdgpu_vce_ring_test_ring(struct amdgpu_ring *ring) unsigned i; int r; - r = amdgpu_ring_lock(ring, 16); + r = amdgpu_ring_alloc(ring, 16); if (r) { DRM_ERROR("amdgpu: vce failed to lock ring %d (%d).\n", ring->idx, r); return r; } amdgpu_ring_write(ring, VCE_CMD_END); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { if (amdgpu_ring_get_rptr(ring) != rptr) @@ -862,7 +841,7 @@ int amdgpu_vce_ring_test_ib(struct amdgpu_ring *ring) goto error; } - r = amdgpu_vce_get_destroy_msg(ring, 1, &fence); + r = amdgpu_vce_get_destroy_msg(ring, 1, true, &fence); if (r) { DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); goto error; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h index ba2da8ee5906..ef99d2370182 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.h @@ -31,12 +31,9 @@ int amdgpu_vce_resume(struct amdgpu_device *adev); int amdgpu_vce_get_create_msg(struct amdgpu_ring *ring, uint32_t handle, struct fence **fence); int amdgpu_vce_get_destroy_msg(struct amdgpu_ring *ring, uint32_t handle, - struct fence **fence); + bool direct, struct fence **fence); void amdgpu_vce_free_handles(struct amdgpu_device *adev, struct drm_file *filp); int amdgpu_vce_ring_parse_cs(struct amdgpu_cs_parser *p, uint32_t ib_idx); -bool amdgpu_vce_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait); void amdgpu_vce_ring_emit_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib); void amdgpu_vce_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq, unsigned flags); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index 9599f7559b3d..264c5968a1d3 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -55,7 +55,7 @@ * * @adev: amdgpu_device pointer * - * Calculate the number of page directory entries (cayman+). + * Calculate the number of page directory entries. */ static unsigned amdgpu_vm_num_pdes(struct amdgpu_device *adev) { @@ -67,7 +67,7 @@ static unsigned amdgpu_vm_num_pdes(struct amdgpu_device *adev) * * @adev: amdgpu_device pointer * - * Calculate the size of the page directory in bytes (cayman+). + * Calculate the size of the page directory in bytes. */ static unsigned amdgpu_vm_directory_size(struct amdgpu_device *adev) { @@ -89,8 +89,6 @@ void amdgpu_vm_get_pd_bo(struct amdgpu_vm *vm, struct amdgpu_bo_list_entry *entry) { entry->robj = vm->page_directory; - entry->prefered_domains = AMDGPU_GEM_DOMAIN_VRAM; - entry->allowed_domains = AMDGPU_GEM_DOMAIN_VRAM; entry->priority = 0; entry->tv.bo = &vm->page_directory->tbo; entry->tv.shared = true; @@ -154,29 +152,34 @@ void amdgpu_vm_move_pt_bos_in_lru(struct amdgpu_device *adev, * @vm: vm to allocate id for * @ring: ring we want to submit job to * @sync: sync object where we add dependencies + * @fence: fence protecting ID from reuse * * Allocate an id for the vm, adding fences to the sync obj as necessary. - * - * Global mutex must be locked! */ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, - struct amdgpu_sync *sync) + struct amdgpu_sync *sync, struct fence *fence) { - struct fence *best[AMDGPU_MAX_RINGS] = {}; struct amdgpu_vm_id *vm_id = &vm->ids[ring->idx]; struct amdgpu_device *adev = ring->adev; + struct amdgpu_vm_manager_id *id; + int r; - unsigned choices[2] = {}; - unsigned i; + mutex_lock(&adev->vm_manager.lock); /* check if the id is still valid */ if (vm_id->id) { - unsigned id = vm_id->id; long owner; - owner = atomic_long_read(&adev->vm_manager.ids[id].owner); + id = &adev->vm_manager.ids[vm_id->id]; + owner = atomic_long_read(&id->owner); if (owner == (long)vm) { - trace_amdgpu_vm_grab_id(vm_id->id, ring->idx); + list_move_tail(&id->list, &adev->vm_manager.ids_lru); + trace_amdgpu_vm_grab_id(vm, vm_id->id, ring->idx); + + fence_put(id->active); + id->active = fence_get(fence); + + mutex_unlock(&adev->vm_manager.lock); return 0; } } @@ -184,41 +187,24 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, /* we definately need to flush */ vm_id->pd_gpu_addr = ~0ll; - /* skip over VMID 0, since it is the system VM */ - for (i = 1; i < adev->vm_manager.nvm; ++i) { - struct fence *fence = adev->vm_manager.ids[i].active; - struct amdgpu_ring *fring; - - if (fence == NULL) { - /* found a free one */ - vm_id->id = i; - trace_amdgpu_vm_grab_id(i, ring->idx); - return 0; - } - - fring = amdgpu_ring_from_fence(fence); - if (best[fring->idx] == NULL || - fence_is_later(best[fring->idx], fence)) { - best[fring->idx] = fence; - choices[fring == ring ? 0 : 1] = i; - } - } + id = list_first_entry(&adev->vm_manager.ids_lru, + struct amdgpu_vm_manager_id, + list); + list_move_tail(&id->list, &adev->vm_manager.ids_lru); + atomic_long_set(&id->owner, (long)vm); - for (i = 0; i < 2; ++i) { - if (choices[i]) { - struct fence *fence; + vm_id->id = id - adev->vm_manager.ids; + trace_amdgpu_vm_grab_id(vm, vm_id->id, ring->idx); - fence = adev->vm_manager.ids[choices[i]].active; - vm_id->id = choices[i]; + r = amdgpu_sync_fence(ring->adev, sync, id->active); - trace_amdgpu_vm_grab_id(choices[i], ring->idx); - return amdgpu_sync_fence(ring->adev, sync, fence); - } + if (!r) { + fence_put(id->active); + id->active = fence_get(fence); } - /* should never happen */ - BUG(); - return -EINVAL; + mutex_unlock(&adev->vm_manager.lock); + return r; } /** @@ -228,9 +214,7 @@ int amdgpu_vm_grab_id(struct amdgpu_vm *vm, struct amdgpu_ring *ring, * @vm: vm we want to flush * @updates: last vm update that we waited for * - * Flush the vm (cayman+). - * - * Global and local mutex must be locked! + * Flush the vm. */ void amdgpu_vm_flush(struct amdgpu_ring *ring, struct amdgpu_vm *vm, @@ -260,36 +244,12 @@ void amdgpu_vm_flush(struct amdgpu_ring *ring, } /** - * amdgpu_vm_fence - remember fence for vm - * - * @adev: amdgpu_device pointer - * @vm: vm we want to fence - * @fence: fence to remember - * - * Fence the vm (cayman+). - * Set the fence used to protect page table and id. - * - * Global and local mutex must be locked! - */ -void amdgpu_vm_fence(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct fence *fence) -{ - struct amdgpu_ring *ring = amdgpu_ring_from_fence(fence); - unsigned vm_id = vm->ids[ring->idx].id; - - fence_put(adev->vm_manager.ids[vm_id].active); - adev->vm_manager.ids[vm_id].active = fence_get(fence); - atomic_long_set(&adev->vm_manager.ids[vm_id].owner, (long)vm); -} - -/** * amdgpu_vm_bo_find - find the bo_va for a specific vm & bo * * @vm: requested vm * @bo: requested buffer object * - * Find @bo inside the requested vm (cayman+). + * Find @bo inside the requested vm. * Search inside the @bos vm list for the requested vm * Returns the found bo_va or NULL if none is found * @@ -312,32 +272,40 @@ struct amdgpu_bo_va *amdgpu_vm_bo_find(struct amdgpu_vm *vm, * amdgpu_vm_update_pages - helper to call the right asic function * * @adev: amdgpu_device pointer + * @gtt: GART instance to use for mapping + * @gtt_flags: GTT hw access flags * @ib: indirect buffer to fill with commands * @pe: addr of the page entry * @addr: dst addr to write into pe * @count: number of page entries to update * @incr: increase next addr by incr bytes * @flags: hw access flags - * @gtt_flags: GTT hw access flags * * Traces the parameters and calls the right asic functions * to setup the page table using the DMA. */ static void amdgpu_vm_update_pages(struct amdgpu_device *adev, + struct amdgpu_gart *gtt, + uint32_t gtt_flags, struct amdgpu_ib *ib, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, - uint32_t flags, uint32_t gtt_flags) + uint32_t flags) { trace_amdgpu_vm_set_page(pe, addr, count, incr, flags); - if ((flags & AMDGPU_PTE_SYSTEM) && (flags == gtt_flags)) { - uint64_t src = adev->gart.table_addr + (addr >> 12) * 8; + if ((gtt == &adev->gart) && (flags == gtt_flags)) { + uint64_t src = gtt->table_addr + (addr >> 12) * 8; amdgpu_vm_copy_pte(adev, ib, pe, src, count); - } else if ((flags & AMDGPU_PTE_SYSTEM) || (count < 3)) { - amdgpu_vm_write_pte(adev, ib, pe, addr, - count, incr, flags); + } else if (gtt) { + dma_addr_t *pages_addr = gtt->pages_addr; + amdgpu_vm_write_pte(adev, ib, pages_addr, pe, addr, + count, incr, flags); + + } else if (count < 3) { + amdgpu_vm_write_pte(adev, ib, NULL, pe, addr, + count, incr, flags); } else { amdgpu_vm_set_pte_pde(adev, ib, pe, addr, @@ -345,15 +313,6 @@ static void amdgpu_vm_update_pages(struct amdgpu_device *adev, } } -int amdgpu_vm_free_job(struct amdgpu_job *job) -{ - int i; - for (i = 0; i < job->num_ibs; i++) - amdgpu_ib_free(job->adev, &job->ibs[i]); - kfree(job->ibs); - return 0; -} - /** * amdgpu_vm_clear_bo - initially clear the page dir/table * @@ -363,15 +322,18 @@ int amdgpu_vm_free_job(struct amdgpu_job *job) * need to reserve bo first before calling it. */ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, + struct amdgpu_vm *vm, struct amdgpu_bo *bo) { - struct amdgpu_ring *ring = adev->vm_manager.vm_pte_funcs_ring; + struct amdgpu_ring *ring; struct fence *fence = NULL; - struct amdgpu_ib *ib; + struct amdgpu_job *job; unsigned entries; uint64_t addr; int r; + ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); + r = reservation_object_reserve_shared(bo->tbo.resv); if (r) return r; @@ -383,56 +345,57 @@ static int amdgpu_vm_clear_bo(struct amdgpu_device *adev, addr = amdgpu_bo_gpu_offset(bo); entries = amdgpu_bo_size(bo) / 8; - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) + r = amdgpu_job_alloc_with_ib(adev, 64, &job); + if (r) goto error; - r = amdgpu_ib_get(ring, NULL, entries * 2 + 64, ib); + amdgpu_vm_update_pages(adev, NULL, 0, &job->ibs[0], addr, 0, entries, + 0, 0); + amdgpu_ring_pad_ib(ring, &job->ibs[0]); + + WARN_ON(job->ibs[0].length_dw > 64); + r = amdgpu_job_submit(job, ring, &vm->entity, + AMDGPU_FENCE_OWNER_VM, &fence); if (r) goto error_free; - ib->length_dw = 0; - - amdgpu_vm_update_pages(adev, ib, addr, 0, entries, 0, 0, 0); - amdgpu_vm_pad_ib(adev, ib); - WARN_ON(ib->length_dw > 64); - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vm_free_job, - AMDGPU_FENCE_OWNER_VM, - &fence); - if (!r) - amdgpu_bo_fence(bo, fence, true); + amdgpu_bo_fence(bo, fence, true); fence_put(fence); - if (amdgpu_enable_scheduler) - return 0; + return 0; error_free: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); error: return r; } /** - * amdgpu_vm_map_gart - get the physical address of a gart page + * amdgpu_vm_map_gart - Resolve gart mapping of addr * - * @adev: amdgpu_device pointer + * @pages_addr: optional DMA address to use for lookup * @addr: the unmapped addr * * Look up the physical address of the page that the pte resolves - * to (cayman+). - * Returns the physical address of the page. + * to and return the pointer for the page table entry. */ -uint64_t amdgpu_vm_map_gart(struct amdgpu_device *adev, uint64_t addr) +uint64_t amdgpu_vm_map_gart(const dma_addr_t *pages_addr, uint64_t addr) { uint64_t result; - /* page table offset */ - result = adev->gart.pages_addr[addr >> PAGE_SHIFT]; + if (pages_addr) { + /* page table offset */ + result = pages_addr[addr >> PAGE_SHIFT]; - /* in case cpu page size != gpu page size*/ - result |= addr & (~PAGE_MASK); + /* in case cpu page size != gpu page size*/ + result |= addr & (~PAGE_MASK); + + } else { + /* No mapping required */ + result = addr; + } + + result &= 0xFFFFFFFFFFFFF000ULL; return result; } @@ -446,45 +409,37 @@ uint64_t amdgpu_vm_map_gart(struct amdgpu_device *adev, uint64_t addr) * @end: end of GPU address range * * Allocates new page tables if necessary - * and updates the page directory (cayman+). + * and updates the page directory. * Returns 0 for success, error for failure. - * - * Global and local mutex must be locked! */ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, struct amdgpu_vm *vm) { - struct amdgpu_ring *ring = adev->vm_manager.vm_pte_funcs_ring; + struct amdgpu_ring *ring; struct amdgpu_bo *pd = vm->page_directory; uint64_t pd_addr = amdgpu_bo_gpu_offset(pd); uint32_t incr = AMDGPU_VM_PTE_COUNT * 8; uint64_t last_pde = ~0, last_pt = ~0; unsigned count = 0, pt_idx, ndw; + struct amdgpu_job *job; struct amdgpu_ib *ib; struct fence *fence = NULL; int r; + ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); + /* padding, etc. */ ndw = 64; /* assume the worst case */ ndw += vm->max_pde_used * 6; - /* update too big for an IB */ - if (ndw > 0xfffff) - return -ENOMEM; - - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) - return -ENOMEM; - - r = amdgpu_ib_get(ring, NULL, ndw * 4, ib); - if (r) { - kfree(ib); + r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job); + if (r) return r; - } - ib->length_dw = 0; + + ib = &job->ibs[0]; /* walk over the address space and update the page directory */ for (pt_idx = 0; pt_idx <= vm->max_pde_used; ++pt_idx) { @@ -504,9 +459,10 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, ((last_pt + incr * count) != pt)) { if (count) { - amdgpu_vm_update_pages(adev, ib, last_pde, - last_pt, count, incr, - AMDGPU_PTE_VALID, 0); + amdgpu_vm_update_pages(adev, NULL, 0, ib, + last_pde, last_pt, + count, incr, + AMDGPU_PTE_VALID); } count = 1; @@ -518,17 +474,16 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, } if (count) - amdgpu_vm_update_pages(adev, ib, last_pde, last_pt, count, - incr, AMDGPU_PTE_VALID, 0); + amdgpu_vm_update_pages(adev, NULL, 0, ib, last_pde, last_pt, + count, incr, AMDGPU_PTE_VALID); if (ib->length_dw != 0) { - amdgpu_vm_pad_ib(adev, ib); - amdgpu_sync_resv(adev, &ib->sync, pd->tbo.resv, AMDGPU_FENCE_OWNER_VM); + amdgpu_ring_pad_ib(ring, ib); + amdgpu_sync_resv(adev, &job->sync, pd->tbo.resv, + AMDGPU_FENCE_OWNER_VM); WARN_ON(ib->length_dw > ndw); - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vm_free_job, - AMDGPU_FENCE_OWNER_VM, - &fence); + r = amdgpu_job_submit(job, ring, &vm->entity, + AMDGPU_FENCE_OWNER_VM, &fence); if (r) goto error_free; @@ -536,18 +491,15 @@ int amdgpu_vm_update_page_directory(struct amdgpu_device *adev, fence_put(vm->page_directory_fence); vm->page_directory_fence = fence_get(fence); fence_put(fence); - } - if (!amdgpu_enable_scheduler || ib->length_dw == 0) { - amdgpu_ib_free(adev, ib); - kfree(ib); + } else { + amdgpu_job_free(job); } return 0; error_free: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); return r; } @@ -555,20 +507,20 @@ error_free: * amdgpu_vm_frag_ptes - add fragment information to PTEs * * @adev: amdgpu_device pointer + * @gtt: GART instance to use for mapping + * @gtt_flags: GTT hw mapping flags * @ib: IB for the update * @pe_start: first PTE to handle * @pe_end: last PTE to handle * @addr: addr those PTEs should point to * @flags: hw mapping flags - * @gtt_flags: GTT hw mapping flags - * - * Global and local mutex must be locked! */ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev, + struct amdgpu_gart *gtt, + uint32_t gtt_flags, struct amdgpu_ib *ib, uint64_t pe_start, uint64_t pe_end, - uint64_t addr, uint32_t flags, - uint32_t gtt_flags) + uint64_t addr, uint32_t flags) { /** * The MC L1 TLB supports variable sized pages, based on a fragment @@ -598,36 +550,39 @@ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev, unsigned count; + /* Abort early if there isn't anything to do */ + if (pe_start == pe_end) + return; + /* system pages are non continuously */ - if ((flags & AMDGPU_PTE_SYSTEM) || !(flags & AMDGPU_PTE_VALID) || - (frag_start >= frag_end)) { + if (gtt || !(flags & AMDGPU_PTE_VALID) || (frag_start >= frag_end)) { count = (pe_end - pe_start) / 8; - amdgpu_vm_update_pages(adev, ib, pe_start, addr, count, - AMDGPU_GPU_PAGE_SIZE, flags, gtt_flags); + amdgpu_vm_update_pages(adev, gtt, gtt_flags, ib, pe_start, + addr, count, AMDGPU_GPU_PAGE_SIZE, + flags); return; } /* handle the 4K area at the beginning */ if (pe_start != frag_start) { count = (frag_start - pe_start) / 8; - amdgpu_vm_update_pages(adev, ib, pe_start, addr, count, - AMDGPU_GPU_PAGE_SIZE, flags, gtt_flags); + amdgpu_vm_update_pages(adev, NULL, 0, ib, pe_start, addr, + count, AMDGPU_GPU_PAGE_SIZE, flags); addr += AMDGPU_GPU_PAGE_SIZE * count; } /* handle the area in the middle */ count = (frag_end - frag_start) / 8; - amdgpu_vm_update_pages(adev, ib, frag_start, addr, count, - AMDGPU_GPU_PAGE_SIZE, flags | frag_flags, - gtt_flags); + amdgpu_vm_update_pages(adev, NULL, 0, ib, frag_start, addr, count, + AMDGPU_GPU_PAGE_SIZE, flags | frag_flags); /* handle the 4K area at the end */ if (frag_end != pe_end) { addr += AMDGPU_GPU_PAGE_SIZE * count; count = (pe_end - frag_end) / 8; - amdgpu_vm_update_pages(adev, ib, frag_end, addr, count, - AMDGPU_GPU_PAGE_SIZE, flags, gtt_flags); + amdgpu_vm_update_pages(adev, NULL, 0, ib, frag_end, addr, + count, AMDGPU_GPU_PAGE_SIZE, flags); } } @@ -635,122 +590,105 @@ static void amdgpu_vm_frag_ptes(struct amdgpu_device *adev, * amdgpu_vm_update_ptes - make sure that page tables are valid * * @adev: amdgpu_device pointer + * @gtt: GART instance to use for mapping + * @gtt_flags: GTT hw mapping flags * @vm: requested vm * @start: start of GPU address range * @end: end of GPU address range * @dst: destination address to map to * @flags: mapping flags * - * Update the page tables in the range @start - @end (cayman+). - * - * Global and local mutex must be locked! + * Update the page tables in the range @start - @end. */ -static int amdgpu_vm_update_ptes(struct amdgpu_device *adev, - struct amdgpu_vm *vm, - struct amdgpu_ib *ib, - uint64_t start, uint64_t end, - uint64_t dst, uint32_t flags, - uint32_t gtt_flags) +static void amdgpu_vm_update_ptes(struct amdgpu_device *adev, + struct amdgpu_gart *gtt, + uint32_t gtt_flags, + struct amdgpu_vm *vm, + struct amdgpu_ib *ib, + uint64_t start, uint64_t end, + uint64_t dst, uint32_t flags) { - uint64_t mask = AMDGPU_VM_PTE_COUNT - 1; - uint64_t last_pte = ~0, last_dst = ~0; - void *owner = AMDGPU_FENCE_OWNER_VM; - unsigned count = 0; - uint64_t addr; + const uint64_t mask = AMDGPU_VM_PTE_COUNT - 1; - /* sync to everything on unmapping */ - if (!(flags & AMDGPU_PTE_VALID)) - owner = AMDGPU_FENCE_OWNER_UNDEFINED; + uint64_t last_pe_start = ~0, last_pe_end = ~0, last_dst = ~0; + uint64_t addr; /* walk over the address space and update the page tables */ for (addr = start; addr < end; ) { uint64_t pt_idx = addr >> amdgpu_vm_block_size; struct amdgpu_bo *pt = vm->page_tables[pt_idx].entry.robj; unsigned nptes; - uint64_t pte; - int r; - - amdgpu_sync_resv(adev, &ib->sync, pt->tbo.resv, owner); - r = reservation_object_reserve_shared(pt->tbo.resv); - if (r) - return r; + uint64_t pe_start; if ((addr & ~mask) == (end & ~mask)) nptes = end - addr; else nptes = AMDGPU_VM_PTE_COUNT - (addr & mask); - pte = amdgpu_bo_gpu_offset(pt); - pte += (addr & mask) * 8; + pe_start = amdgpu_bo_gpu_offset(pt); + pe_start += (addr & mask) * 8; - if ((last_pte + 8 * count) != pte) { + if (last_pe_end != pe_start) { - if (count) { - amdgpu_vm_frag_ptes(adev, ib, last_pte, - last_pte + 8 * count, - last_dst, flags, - gtt_flags); - } + amdgpu_vm_frag_ptes(adev, gtt, gtt_flags, ib, + last_pe_start, last_pe_end, + last_dst, flags); - count = nptes; - last_pte = pte; + last_pe_start = pe_start; + last_pe_end = pe_start + 8 * nptes; last_dst = dst; } else { - count += nptes; + last_pe_end += 8 * nptes; } addr += nptes; dst += nptes * AMDGPU_GPU_PAGE_SIZE; } - if (count) { - amdgpu_vm_frag_ptes(adev, ib, last_pte, - last_pte + 8 * count, - last_dst, flags, gtt_flags); - } - - return 0; + amdgpu_vm_frag_ptes(adev, gtt, gtt_flags, ib, + last_pe_start, last_pe_end, + last_dst, flags); } /** * amdgpu_vm_bo_update_mapping - update a mapping in the vm page table * * @adev: amdgpu_device pointer + * @gtt: GART instance to use for mapping + * @gtt_flags: flags as they are used for GTT * @vm: requested vm - * @mapping: mapped range and flags to use for the update + * @start: start of mapped range + * @last: last mapped entry + * @flags: flags for the entries * @addr: addr to set the area to - * @gtt_flags: flags as they are used for GTT * @fence: optional resulting fence * - * Fill in the page table entries for @mapping. + * Fill in the page table entries between @start and @last. * Returns 0 for success, -EINVAL for failure. - * - * Object have to be reserved and mutex must be locked! */ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, + struct amdgpu_gart *gtt, + uint32_t gtt_flags, struct amdgpu_vm *vm, - struct amdgpu_bo_va_mapping *mapping, - uint64_t addr, uint32_t gtt_flags, + uint64_t start, uint64_t last, + uint32_t flags, uint64_t addr, struct fence **fence) { - struct amdgpu_ring *ring = adev->vm_manager.vm_pte_funcs_ring; + struct amdgpu_ring *ring; + void *owner = AMDGPU_FENCE_OWNER_VM; unsigned nptes, ncmds, ndw; - uint32_t flags = gtt_flags; + struct amdgpu_job *job; struct amdgpu_ib *ib; struct fence *f = NULL; int r; - /* normally,bo_va->flags only contians READABLE and WIRTEABLE bit go here - * but in case of something, we filter the flags in first place - */ - if (!(mapping->flags & AMDGPU_PTE_READABLE)) - flags &= ~AMDGPU_PTE_READABLE; - if (!(mapping->flags & AMDGPU_PTE_WRITEABLE)) - flags &= ~AMDGPU_PTE_WRITEABLE; + ring = container_of(vm->entity.sched, struct amdgpu_ring, sched); - trace_amdgpu_vm_bo_update(mapping); + /* sync to everything on unmapping */ + if (!(flags & AMDGPU_PTE_VALID)) + owner = AMDGPU_FENCE_OWNER_UNDEFINED; - nptes = mapping->it.last - mapping->it.start + 1; + nptes = last - start + 1; /* * reserve space for one command every (1 << BLOCK_SIZE) @@ -761,11 +699,11 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, /* padding, etc. */ ndw = 64; - if ((flags & AMDGPU_PTE_SYSTEM) && (flags == gtt_flags)) { + if ((gtt == &adev->gart) && (flags == gtt_flags)) { /* only copy commands needed */ ndw += ncmds * 7; - } else if (flags & AMDGPU_PTE_SYSTEM) { + } else if (gtt) { /* header for write data commands */ ndw += ncmds * 4; @@ -780,38 +718,28 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, ndw += 2 * 10; } - /* update too big for an IB */ - if (ndw > 0xfffff) - return -ENOMEM; - - ib = kzalloc(sizeof(struct amdgpu_ib), GFP_KERNEL); - if (!ib) - return -ENOMEM; - - r = amdgpu_ib_get(ring, NULL, ndw * 4, ib); - if (r) { - kfree(ib); + r = amdgpu_job_alloc_with_ib(adev, ndw * 4, &job); + if (r) return r; - } - ib->length_dw = 0; + ib = &job->ibs[0]; - r = amdgpu_vm_update_ptes(adev, vm, ib, mapping->it.start, - mapping->it.last + 1, addr + mapping->offset, - flags, gtt_flags); + r = amdgpu_sync_resv(adev, &job->sync, vm->page_directory->tbo.resv, + owner); + if (r) + goto error_free; - if (r) { - amdgpu_ib_free(adev, ib); - kfree(ib); - return r; - } + r = reservation_object_reserve_shared(vm->page_directory->tbo.resv); + if (r) + goto error_free; + + amdgpu_vm_update_ptes(adev, gtt, gtt_flags, vm, ib, start, last + 1, + addr, flags); - amdgpu_vm_pad_ib(adev, ib); + amdgpu_ring_pad_ib(ring, ib); WARN_ON(ib->length_dw > ndw); - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, ib, 1, - &amdgpu_vm_free_job, - AMDGPU_FENCE_OWNER_VM, - &f); + r = amdgpu_job_submit(job, ring, &vm->entity, + AMDGPU_FENCE_OWNER_VM, &f); if (r) goto error_free; @@ -821,19 +749,76 @@ static int amdgpu_vm_bo_update_mapping(struct amdgpu_device *adev, *fence = fence_get(f); } fence_put(f); - if (!amdgpu_enable_scheduler) { - amdgpu_ib_free(adev, ib); - kfree(ib); - } return 0; error_free: - amdgpu_ib_free(adev, ib); - kfree(ib); + amdgpu_job_free(job); return r; } /** + * amdgpu_vm_bo_split_mapping - split a mapping into smaller chunks + * + * @adev: amdgpu_device pointer + * @gtt: GART instance to use for mapping + * @vm: requested vm + * @mapping: mapped range and flags to use for the update + * @addr: addr to set the area to + * @gtt_flags: flags as they are used for GTT + * @fence: optional resulting fence + * + * Split the mapping into smaller chunks so that each update fits + * into a SDMA IB. + * Returns 0 for success, -EINVAL for failure. + */ +static int amdgpu_vm_bo_split_mapping(struct amdgpu_device *adev, + struct amdgpu_gart *gtt, + uint32_t gtt_flags, + struct amdgpu_vm *vm, + struct amdgpu_bo_va_mapping *mapping, + uint64_t addr, struct fence **fence) +{ + const uint64_t max_size = 64ULL * 1024ULL * 1024ULL / AMDGPU_GPU_PAGE_SIZE; + + uint64_t start = mapping->it.start; + uint32_t flags = gtt_flags; + int r; + + /* normally,bo_va->flags only contians READABLE and WIRTEABLE bit go here + * but in case of something, we filter the flags in first place + */ + if (!(mapping->flags & AMDGPU_PTE_READABLE)) + flags &= ~AMDGPU_PTE_READABLE; + if (!(mapping->flags & AMDGPU_PTE_WRITEABLE)) + flags &= ~AMDGPU_PTE_WRITEABLE; + + trace_amdgpu_vm_bo_update(mapping); + + addr += mapping->offset; + + if (!gtt || ((gtt == &adev->gart) && (flags == gtt_flags))) + return amdgpu_vm_bo_update_mapping(adev, gtt, gtt_flags, vm, + start, mapping->it.last, + flags, addr, fence); + + while (start != mapping->it.last + 1) { + uint64_t last; + + last = min((uint64_t)mapping->it.last, start + max_size); + r = amdgpu_vm_bo_update_mapping(adev, gtt, gtt_flags, vm, + start, last, flags, addr, + fence); + if (r) + return r; + + start = last + 1; + addr += max_size; + } + + return 0; +} + +/** * amdgpu_vm_bo_update - update all BO mappings in the vm page table * * @adev: amdgpu_device pointer @@ -851,14 +836,25 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, { struct amdgpu_vm *vm = bo_va->vm; struct amdgpu_bo_va_mapping *mapping; + struct amdgpu_gart *gtt = NULL; uint32_t flags; uint64_t addr; int r; if (mem) { addr = (u64)mem->start << PAGE_SHIFT; - if (mem->mem_type != TTM_PL_TT) + switch (mem->mem_type) { + case TTM_PL_TT: + gtt = &bo_va->bo->adev->gart; + break; + + case TTM_PL_VRAM: addr += adev->vm_manager.vram_base_offset; + break; + + default: + break; + } } else { addr = 0; } @@ -871,8 +867,8 @@ int amdgpu_vm_bo_update(struct amdgpu_device *adev, spin_unlock(&vm->status_lock); list_for_each_entry(mapping, &bo_va->invalids, list) { - r = amdgpu_vm_bo_update_mapping(adev, vm, mapping, addr, - flags, &bo_va->last_pt_update); + r = amdgpu_vm_bo_split_mapping(adev, gtt, flags, vm, mapping, addr, + &bo_va->last_pt_update); if (r) return r; } @@ -918,7 +914,8 @@ int amdgpu_vm_clear_freed(struct amdgpu_device *adev, struct amdgpu_bo_va_mapping, list); list_del(&mapping->list); spin_unlock(&vm->freed_lock); - r = amdgpu_vm_bo_update_mapping(adev, vm, mapping, 0, 0, NULL); + r = amdgpu_vm_bo_split_mapping(adev, NULL, 0, vm, mapping, + 0, NULL); kfree(mapping); if (r) return r; @@ -976,7 +973,7 @@ int amdgpu_vm_clear_invalids(struct amdgpu_device *adev, * @vm: requested vm * @bo: amdgpu buffer object * - * Add @bo into the requested vm (cayman+). + * Add @bo into the requested vm. * Add @bo to the list of bos associated with the vm * Returns newly added bo_va or NULL for failure * @@ -1117,15 +1114,13 @@ int amdgpu_vm_bo_map(struct amdgpu_device *adev, */ pt->parent = amdgpu_bo_ref(vm->page_directory); - r = amdgpu_vm_clear_bo(adev, pt); + r = amdgpu_vm_clear_bo(adev, vm, pt); if (r) { amdgpu_bo_unref(&pt); goto error_free; } entry->robj = pt; - entry->prefered_domains = AMDGPU_GEM_DOMAIN_VRAM; - entry->allowed_domains = AMDGPU_GEM_DOMAIN_VRAM; entry->priority = 0; entry->tv.bo = &entry->robj->tbo; entry->tv.shared = true; @@ -1210,7 +1205,7 @@ int amdgpu_vm_bo_unmap(struct amdgpu_device *adev, * @adev: amdgpu_device pointer * @bo_va: requested bo_va * - * Remove @bo_va->bo from the requested vm (cayman+). + * Remove @bo_va->bo from the requested vm. * * Object have to be reserved! */ @@ -1255,7 +1250,7 @@ void amdgpu_vm_bo_rmv(struct amdgpu_device *adev, * @vm: requested vm * @bo: amdgpu buffer object * - * Mark @bo as invalid (cayman+). + * Mark @bo as invalid. */ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, struct amdgpu_bo *bo) @@ -1276,13 +1271,16 @@ void amdgpu_vm_bo_invalidate(struct amdgpu_device *adev, * @adev: amdgpu_device pointer * @vm: requested vm * - * Init @vm fields (cayman+). + * Init @vm fields. */ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) { const unsigned align = min(AMDGPU_VM_PTB_ALIGN_SIZE, AMDGPU_VM_PTE_COUNT * 8); unsigned pd_size, pd_entries; + unsigned ring_instance; + struct amdgpu_ring *ring; + struct amd_sched_rq *rq; int i, r; for (i = 0; i < AMDGPU_MAX_RINGS; ++i) { @@ -1306,6 +1304,17 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) return -ENOMEM; } + /* create scheduler entity for page table updates */ + + ring_instance = atomic_inc_return(&adev->vm_manager.vm_pte_next_ring); + ring_instance %= adev->vm_manager.vm_pte_num_rings; + ring = adev->vm_manager.vm_pte_rings[ring_instance]; + rq = &ring->sched.sched_rq[AMD_SCHED_PRIORITY_KERNEL]; + r = amd_sched_entity_init(&ring->sched, &vm->entity, + rq, amdgpu_sched_jobs); + if (r) + return r; + vm->page_directory_fence = NULL; r = amdgpu_bo_create(adev, pd_size, align, true, @@ -1313,22 +1322,27 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) AMDGPU_GEM_CREATE_NO_CPU_ACCESS, NULL, NULL, &vm->page_directory); if (r) - return r; + goto error_free_sched_entity; + r = amdgpu_bo_reserve(vm->page_directory, false); - if (r) { - amdgpu_bo_unref(&vm->page_directory); - vm->page_directory = NULL; - return r; - } - r = amdgpu_vm_clear_bo(adev, vm->page_directory); + if (r) + goto error_free_page_directory; + + r = amdgpu_vm_clear_bo(adev, vm, vm->page_directory); amdgpu_bo_unreserve(vm->page_directory); - if (r) { - amdgpu_bo_unref(&vm->page_directory); - vm->page_directory = NULL; - return r; - } + if (r) + goto error_free_page_directory; return 0; + +error_free_page_directory: + amdgpu_bo_unref(&vm->page_directory); + vm->page_directory = NULL; + +error_free_sched_entity: + amd_sched_entity_fini(&ring->sched, &vm->entity); + + return r; } /** @@ -1337,7 +1351,7 @@ int amdgpu_vm_init(struct amdgpu_device *adev, struct amdgpu_vm *vm) * @adev: amdgpu_device pointer * @vm: requested vm * - * Tear down @vm (cayman+). + * Tear down @vm. * Unbind the VM and remove all bos from the vm bo list */ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) @@ -1345,6 +1359,8 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) struct amdgpu_bo_va_mapping *mapping, *tmp; int i; + amd_sched_entity_fini(vm->entity.sched, &vm->entity); + if (!RB_EMPTY_ROOT(&vm->va)) { dev_err(adev->dev, "still active bo inside vm\n"); } @@ -1375,6 +1391,27 @@ void amdgpu_vm_fini(struct amdgpu_device *adev, struct amdgpu_vm *vm) } /** + * amdgpu_vm_manager_init - init the VM manager + * + * @adev: amdgpu_device pointer + * + * Initialize the VM manager structures + */ +void amdgpu_vm_manager_init(struct amdgpu_device *adev) +{ + unsigned i; + + INIT_LIST_HEAD(&adev->vm_manager.ids_lru); + + /* skip over VMID 0, since it is the system VM */ + for (i = 1; i < adev->vm_manager.num_ids; ++i) + list_add_tail(&adev->vm_manager.ids[i].list, + &adev->vm_manager.ids_lru); + + atomic_set(&adev->vm_manager.vm_pte_next_ring, 0); +} + +/** * amdgpu_vm_manager_fini - cleanup VM manager * * @adev: amdgpu_device pointer diff --git a/drivers/gpu/drm/amd/amdgpu/cik.c b/drivers/gpu/drm/amd/amdgpu/cik.c index fd9c9588ef46..6b1f0539ce9d 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik.c +++ b/drivers/gpu/drm/amd/amdgpu/cik.c @@ -1059,257 +1059,6 @@ static int cik_read_register(struct amdgpu_device *adev, u32 se_num, return -EINVAL; } -static void cik_print_gpu_status_regs(struct amdgpu_device *adev) -{ - dev_info(adev->dev, " GRBM_STATUS=0x%08X\n", - RREG32(mmGRBM_STATUS)); - dev_info(adev->dev, " GRBM_STATUS2=0x%08X\n", - RREG32(mmGRBM_STATUS2)); - dev_info(adev->dev, " GRBM_STATUS_SE0=0x%08X\n", - RREG32(mmGRBM_STATUS_SE0)); - dev_info(adev->dev, " GRBM_STATUS_SE1=0x%08X\n", - RREG32(mmGRBM_STATUS_SE1)); - dev_info(adev->dev, " GRBM_STATUS_SE2=0x%08X\n", - RREG32(mmGRBM_STATUS_SE2)); - dev_info(adev->dev, " GRBM_STATUS_SE3=0x%08X\n", - RREG32(mmGRBM_STATUS_SE3)); - dev_info(adev->dev, " SRBM_STATUS=0x%08X\n", - RREG32(mmSRBM_STATUS)); - dev_info(adev->dev, " SRBM_STATUS2=0x%08X\n", - RREG32(mmSRBM_STATUS2)); - dev_info(adev->dev, " SDMA0_STATUS_REG = 0x%08X\n", - RREG32(mmSDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET)); - dev_info(adev->dev, " SDMA1_STATUS_REG = 0x%08X\n", - RREG32(mmSDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET)); - dev_info(adev->dev, " CP_STAT = 0x%08x\n", RREG32(mmCP_STAT)); - dev_info(adev->dev, " CP_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT1)); - dev_info(adev->dev, " CP_STALLED_STAT2 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT2)); - dev_info(adev->dev, " CP_STALLED_STAT3 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT3)); - dev_info(adev->dev, " CP_CPF_BUSY_STAT = 0x%08x\n", - RREG32(mmCP_CPF_BUSY_STAT)); - dev_info(adev->dev, " CP_CPF_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_CPF_STALLED_STAT1)); - dev_info(adev->dev, " CP_CPF_STATUS = 0x%08x\n", RREG32(mmCP_CPF_STATUS)); - dev_info(adev->dev, " CP_CPC_BUSY_STAT = 0x%08x\n", RREG32(mmCP_CPC_BUSY_STAT)); - dev_info(adev->dev, " CP_CPC_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_CPC_STALLED_STAT1)); - dev_info(adev->dev, " CP_CPC_STATUS = 0x%08x\n", RREG32(mmCP_CPC_STATUS)); -} - -/** - * cik_gpu_check_soft_reset - check which blocks are busy - * - * @adev: amdgpu_device pointer - * - * Check which blocks are busy and return the relevant reset - * mask to be used by cik_gpu_soft_reset(). - * Returns a mask of the blocks to be reset. - */ -u32 amdgpu_cik_gpu_check_soft_reset(struct amdgpu_device *adev) -{ - u32 reset_mask = 0; - u32 tmp; - - /* GRBM_STATUS */ - tmp = RREG32(mmGRBM_STATUS); - if (tmp & (GRBM_STATUS__PA_BUSY_MASK | GRBM_STATUS__SC_BUSY_MASK | - GRBM_STATUS__BCI_BUSY_MASK | GRBM_STATUS__SX_BUSY_MASK | - GRBM_STATUS__TA_BUSY_MASK | GRBM_STATUS__VGT_BUSY_MASK | - GRBM_STATUS__DB_BUSY_MASK | GRBM_STATUS__CB_BUSY_MASK | - GRBM_STATUS__GDS_BUSY_MASK | GRBM_STATUS__SPI_BUSY_MASK | - GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK)) - reset_mask |= AMDGPU_RESET_GFX; - - if (tmp & (GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_CP; - - /* GRBM_STATUS2 */ - tmp = RREG32(mmGRBM_STATUS2); - if (tmp & GRBM_STATUS2__RLC_BUSY_MASK) - reset_mask |= AMDGPU_RESET_RLC; - - /* SDMA0_STATUS_REG */ - tmp = RREG32(mmSDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET); - if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK)) - reset_mask |= AMDGPU_RESET_DMA; - - /* SDMA1_STATUS_REG */ - tmp = RREG32(mmSDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET); - if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK)) - reset_mask |= AMDGPU_RESET_DMA1; - - /* SRBM_STATUS2 */ - tmp = RREG32(mmSRBM_STATUS2); - if (tmp & SRBM_STATUS2__SDMA_BUSY_MASK) - reset_mask |= AMDGPU_RESET_DMA; - - if (tmp & SRBM_STATUS2__SDMA1_BUSY_MASK) - reset_mask |= AMDGPU_RESET_DMA1; - - /* SRBM_STATUS */ - tmp = RREG32(mmSRBM_STATUS); - - if (tmp & SRBM_STATUS__IH_BUSY_MASK) - reset_mask |= AMDGPU_RESET_IH; - - if (tmp & SRBM_STATUS__SEM_BUSY_MASK) - reset_mask |= AMDGPU_RESET_SEM; - - if (tmp & SRBM_STATUS__GRBM_RQ_PENDING_MASK) - reset_mask |= AMDGPU_RESET_GRBM; - - if (tmp & SRBM_STATUS__VMC_BUSY_MASK) - reset_mask |= AMDGPU_RESET_VMC; - - if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK | - SRBM_STATUS__MCC_BUSY_MASK | SRBM_STATUS__MCD_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_MC; - - if (amdgpu_display_is_display_hung(adev)) - reset_mask |= AMDGPU_RESET_DISPLAY; - - /* Skip MC reset as it's mostly likely not hung, just busy */ - if (reset_mask & AMDGPU_RESET_MC) { - DRM_DEBUG("MC busy: 0x%08X, clearing.\n", reset_mask); - reset_mask &= ~AMDGPU_RESET_MC; - } - - return reset_mask; -} - -/** - * cik_gpu_soft_reset - soft reset GPU - * - * @adev: amdgpu_device pointer - * @reset_mask: mask of which blocks to reset - * - * Soft reset the blocks specified in @reset_mask. - */ -static void cik_gpu_soft_reset(struct amdgpu_device *adev, u32 reset_mask) -{ - struct amdgpu_mode_mc_save save; - u32 grbm_soft_reset = 0, srbm_soft_reset = 0; - u32 tmp; - - if (reset_mask == 0) - return; - - dev_info(adev->dev, "GPU softreset: 0x%08X\n", reset_mask); - - cik_print_gpu_status_regs(adev); - dev_info(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR)); - dev_info(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS)); - - /* disable CG/PG */ - - /* stop the rlc */ - gfx_v7_0_rlc_stop(adev); - - /* Disable GFX parsing/prefetching */ - WREG32(mmCP_ME_CNTL, CP_ME_CNTL__ME_HALT_MASK | CP_ME_CNTL__PFP_HALT_MASK | CP_ME_CNTL__CE_HALT_MASK); - - /* Disable MEC parsing/prefetching */ - WREG32(mmCP_MEC_CNTL, CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK); - - if (reset_mask & AMDGPU_RESET_DMA) { - /* sdma0 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); - } - if (reset_mask & AMDGPU_RESET_DMA1) { - /* sdma1 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); - } - - gmc_v7_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { - dev_warn(adev->dev, "Wait for MC idle timedout !\n"); - } - - if (reset_mask & (AMDGPU_RESET_GFX | AMDGPU_RESET_COMPUTE | AMDGPU_RESET_CP)) - grbm_soft_reset = GRBM_SOFT_RESET__SOFT_RESET_CP_MASK | - GRBM_SOFT_RESET__SOFT_RESET_GFX_MASK; - - if (reset_mask & AMDGPU_RESET_CP) { - grbm_soft_reset |= GRBM_SOFT_RESET__SOFT_RESET_CP_MASK; - - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_GRBM_MASK; - } - - if (reset_mask & AMDGPU_RESET_DMA) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA_MASK; - - if (reset_mask & AMDGPU_RESET_DMA1) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SDMA1_MASK; - - if (reset_mask & AMDGPU_RESET_DISPLAY) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_DC_MASK; - - if (reset_mask & AMDGPU_RESET_RLC) - grbm_soft_reset |= GRBM_SOFT_RESET__SOFT_RESET_RLC_MASK; - - if (reset_mask & AMDGPU_RESET_SEM) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_SEM_MASK; - - if (reset_mask & AMDGPU_RESET_IH) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_IH_MASK; - - if (reset_mask & AMDGPU_RESET_GRBM) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_GRBM_MASK; - - if (reset_mask & AMDGPU_RESET_VMC) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_VMC_MASK; - - if (!(adev->flags & AMD_IS_APU)) { - if (reset_mask & AMDGPU_RESET_MC) - srbm_soft_reset |= SRBM_SOFT_RESET__SOFT_RESET_MC_MASK; - } - - if (grbm_soft_reset) { - tmp = RREG32(mmGRBM_SOFT_RESET); - tmp |= grbm_soft_reset; - dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(mmGRBM_SOFT_RESET, tmp); - tmp = RREG32(mmGRBM_SOFT_RESET); - - udelay(50); - - tmp &= ~grbm_soft_reset; - WREG32(mmGRBM_SOFT_RESET, tmp); - tmp = RREG32(mmGRBM_SOFT_RESET); - } - - if (srbm_soft_reset) { - tmp = RREG32(mmSRBM_SOFT_RESET); - tmp |= srbm_soft_reset; - dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(mmSRBM_SOFT_RESET, tmp); - tmp = RREG32(mmSRBM_SOFT_RESET); - - udelay(50); - - tmp &= ~srbm_soft_reset; - WREG32(mmSRBM_SOFT_RESET, tmp); - tmp = RREG32(mmSRBM_SOFT_RESET); - } - - /* Wait a little for things to settle down */ - udelay(50); - - gmc_v7_0_mc_resume(adev, &save); - udelay(50); - - cik_print_gpu_status_regs(adev); -} - struct kv_reset_save_regs { u32 gmcon_reng_execute; u32 gmcon_misc; @@ -1405,45 +1154,11 @@ static void kv_restore_regs_for_reset(struct amdgpu_device *adev, static void cik_gpu_pci_config_reset(struct amdgpu_device *adev) { - struct amdgpu_mode_mc_save save; struct kv_reset_save_regs kv_save = { 0 }; - u32 tmp, i; + u32 i; dev_info(adev->dev, "GPU pci config reset\n"); - /* disable dpm? */ - - /* disable cg/pg */ - - /* Disable GFX parsing/prefetching */ - WREG32(mmCP_ME_CNTL, CP_ME_CNTL__ME_HALT_MASK | - CP_ME_CNTL__PFP_HALT_MASK | CP_ME_CNTL__CE_HALT_MASK); - - /* Disable MEC parsing/prefetching */ - WREG32(mmCP_MEC_CNTL, - CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK); - - /* sdma0 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); - /* sdma1 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); - tmp |= SDMA0_F32_CNTL__HALT_MASK; - WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); - /* XXX other engines? */ - - /* halt the rlc, disable cp internal ints */ - gfx_v7_0_rlc_stop(adev); - - udelay(50); - - /* disable mem access */ - gmc_v7_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { - dev_warn(adev->dev, "Wait for MC idle timed out !\n"); - } - if (adev->flags & AMD_IS_APU) kv_save_regs_for_reset(adev, &kv_save); @@ -1489,26 +1204,11 @@ static void cik_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hu */ static int cik_asic_reset(struct amdgpu_device *adev) { - u32 reset_mask; - - reset_mask = amdgpu_cik_gpu_check_soft_reset(adev); - - if (reset_mask) - cik_set_bios_scratch_engine_hung(adev, true); - - /* try soft reset */ - cik_gpu_soft_reset(adev, reset_mask); - - reset_mask = amdgpu_cik_gpu_check_soft_reset(adev); - - /* try pci config reset */ - if (reset_mask && amdgpu_hard_reset) - cik_gpu_pci_config_reset(adev); + cik_set_bios_scratch_engine_hung(adev, true); - reset_mask = amdgpu_cik_gpu_check_soft_reset(adev); + cik_gpu_pci_config_reset(adev); - if (!reset_mask) - cik_set_bios_scratch_engine_hung(adev, false); + cik_set_bios_scratch_engine_hung(adev, false); return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c index 5f712ceddf08..675f34916aab 100644 --- a/drivers/gpu/drm/amd/amdgpu/cik_sdma.c +++ b/drivers/gpu/drm/amd/amdgpu/cik_sdma.c @@ -295,30 +295,6 @@ static void cik_sdma_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** - * cik_sdma_ring_emit_semaphore - emit a semaphore on the dma ring - * - * @ring: amdgpu_ring structure holding ring information - * @semaphore: amdgpu semaphore object - * @emit_wait: wait or signal semaphore - * - * Add a DMA semaphore packet to the ring wait on or signal - * other rings (CIK). - */ -static bool cik_sdma_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - u64 addr = semaphore->gpu_addr; - u32 extra_bits = emit_wait ? 0 : SDMA_SEMAPHORE_EXTRA_S; - - amdgpu_ring_write(ring, SDMA_PACKET(SDMA_OPCODE_SEMAPHORE, 0, extra_bits)); - amdgpu_ring_write(ring, addr & 0xfffffff8); - amdgpu_ring_write(ring, upper_32_bits(addr) & 0xffffffff); - - return true; -} - -/** * cik_sdma_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer @@ -417,6 +393,9 @@ static int cik_sdma_gfx_resume(struct amdgpu_device *adev) cik_srbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); + WREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i], + adev->gfx.config.gb_addr_config & 0x70); + WREG32(mmSDMA0_SEM_INCOMPLETE_TIMER_CNTL + sdma_offsets[i], 0); WREG32(mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); @@ -584,7 +563,7 @@ static int cik_sdma_ring_test_ring(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ring_lock(ring, 5); + r = amdgpu_ring_alloc(ring, 5); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); amdgpu_wb_free(adev, index); @@ -595,7 +574,7 @@ static int cik_sdma_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, upper_32_bits(gpu_addr)); amdgpu_ring_write(ring, 1); /* number of DWs to follow */ amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = le32_to_cpu(adev->wb.wb[index]); @@ -645,7 +624,7 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err0; @@ -657,9 +636,8 @@ static int cik_sdma_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[3] = 1; ib.ptr[4] = 0xDEADBEEF; ib.length_dw = 5; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err1; @@ -738,7 +716,7 @@ static void cik_sdma_vm_copy_pte(struct amdgpu_ib *ib, * Update PTEs by writing them manually using sDMA (CIK). */ static void cik_sdma_vm_write_pte(struct amdgpu_ib *ib, - uint64_t pe, + const dma_addr_t *pages_addr, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags) { @@ -757,14 +735,7 @@ static void cik_sdma_vm_write_pte(struct amdgpu_ib *ib, ib->ptr[ib->length_dw++] = upper_32_bits(pe); ib->ptr[ib->length_dw++] = ndw; for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & AMDGPU_PTE_SYSTEM) { - value = amdgpu_vm_map_gart(ib->ring->adev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & AMDGPU_PTE_VALID) { - value = addr; - } else { - value = 0; - } + value = amdgpu_vm_map_gart(pages_addr, addr); addr += incr; value |= flags; ib->ptr[ib->length_dw++] = value; @@ -827,9 +798,9 @@ static void cik_sdma_vm_set_pte_pde(struct amdgpu_ib *ib, * @ib: indirect buffer to fill with padding * */ -static void cik_sdma_vm_pad_ib(struct amdgpu_ib *ib) +static void cik_sdma_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib) { - struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ib->ring); + struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ring); u32 pad_count; int i; @@ -1097,6 +1068,8 @@ static void cik_sdma_print_status(void *handle) i, RREG32(mmSDMA0_GFX_RB_BASE + sdma_offsets[i])); dev_info(adev->dev, " SDMA%d_GFX_RB_BASE_HI=0x%08X\n", i, RREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i])); + dev_info(adev->dev, " SDMA%d_TILING_CONFIG=0x%08X\n", + i, RREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i])); mutex_lock(&adev->srbm_mutex); for (j = 0; j < 16; j++) { cik_srbm_select(adev, 0, 0, 0, j); @@ -1297,12 +1270,12 @@ static const struct amdgpu_ring_funcs cik_sdma_ring_funcs = { .parse_cs = NULL, .emit_ib = cik_sdma_ring_emit_ib, .emit_fence = cik_sdma_ring_emit_fence, - .emit_semaphore = cik_sdma_ring_emit_semaphore, .emit_vm_flush = cik_sdma_ring_emit_vm_flush, .emit_hdp_flush = cik_sdma_ring_emit_hdp_flush, .test_ring = cik_sdma_ring_test_ring, .test_ib = cik_sdma_ring_test_ib, .insert_nop = cik_sdma_ring_insert_nop, + .pad_ib = cik_sdma_ring_pad_ib, }; static void cik_sdma_set_ring_funcs(struct amdgpu_device *adev) @@ -1399,14 +1372,18 @@ static const struct amdgpu_vm_pte_funcs cik_sdma_vm_pte_funcs = { .copy_pte = cik_sdma_vm_copy_pte, .write_pte = cik_sdma_vm_write_pte, .set_pte_pde = cik_sdma_vm_set_pte_pde, - .pad_ib = cik_sdma_vm_pad_ib, }; static void cik_sdma_set_vm_pte_funcs(struct amdgpu_device *adev) { + unsigned i; + if (adev->vm_manager.vm_pte_funcs == NULL) { adev->vm_manager.vm_pte_funcs = &cik_sdma_vm_pte_funcs; - adev->vm_manager.vm_pte_funcs_ring = &adev->sdma.instance[0].ring; - adev->vm_manager.vm_pte_funcs_ring->is_pte_ring = true; + for (i = 0; i < adev->sdma.num_instances; i++) + adev->vm_manager.vm_pte_rings[i] = + &adev->sdma.instance[i].ring; + + adev->vm_manager.vm_pte_num_rings = adev->sdma.num_instances; } } diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c index 093599aba64b..e3ff809a0cae 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v10_0.c @@ -2670,7 +2670,6 @@ static void dce_v10_0_crtc_destroy(struct drm_crtc *crtc) struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); drm_crtc_cleanup(crtc); - destroy_workqueue(amdgpu_crtc->pflip_queue); kfree(amdgpu_crtc); } @@ -2890,7 +2889,6 @@ static int dce_v10_0_crtc_init(struct amdgpu_device *adev, int index) drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256); amdgpu_crtc->crtc_id = index; - amdgpu_crtc->pflip_queue = create_singlethread_workqueue("amdgpu-pageflip-queue"); adev->mode_info.crtcs[index] = amdgpu_crtc; amdgpu_crtc->max_cursor_width = 128; @@ -3366,7 +3364,7 @@ static int dce_v10_0_pageflip_irq(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->ddev->event_lock, flags); drm_vblank_put(adev->ddev, amdgpu_crtc->crtc_id); - queue_work(amdgpu_crtc->pflip_queue, &works->unpin_work); + schedule_work(&works->unpin_work); return 0; } @@ -3624,16 +3622,8 @@ dce_v10_0_ext_dpms(struct drm_encoder *encoder, int mode) } -static bool dce_v10_0_ext_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static const struct drm_encoder_helper_funcs dce_v10_0_ext_helper_funcs = { .dpms = dce_v10_0_ext_dpms, - .mode_fixup = dce_v10_0_ext_mode_fixup, .prepare = dce_v10_0_ext_prepare, .mode_set = dce_v10_0_ext_mode_set, .commit = dce_v10_0_ext_commit, diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c index 8e67249d4367..6b6c9b6879ae 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v11_0.c @@ -2661,7 +2661,6 @@ static void dce_v11_0_crtc_destroy(struct drm_crtc *crtc) struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); drm_crtc_cleanup(crtc); - destroy_workqueue(amdgpu_crtc->pflip_queue); kfree(amdgpu_crtc); } @@ -2881,7 +2880,6 @@ static int dce_v11_0_crtc_init(struct amdgpu_device *adev, int index) drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256); amdgpu_crtc->crtc_id = index; - amdgpu_crtc->pflip_queue = create_singlethread_workqueue("amdgpu-pageflip-queue"); adev->mode_info.crtcs[index] = amdgpu_crtc; amdgpu_crtc->max_cursor_width = 128; @@ -3361,7 +3359,7 @@ static int dce_v11_0_pageflip_irq(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->ddev->event_lock, flags); drm_vblank_put(adev->ddev, amdgpu_crtc->crtc_id); - queue_work(amdgpu_crtc->pflip_queue, &works->unpin_work); + schedule_work(&works->unpin_work); return 0; } @@ -3619,16 +3617,8 @@ dce_v11_0_ext_dpms(struct drm_encoder *encoder, int mode) } -static bool dce_v11_0_ext_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static const struct drm_encoder_helper_funcs dce_v11_0_ext_helper_funcs = { .dpms = dce_v11_0_ext_dpms, - .mode_fixup = dce_v11_0_ext_mode_fixup, .prepare = dce_v11_0_ext_prepare, .mode_set = dce_v11_0_ext_mode_set, .commit = dce_v11_0_ext_commit, diff --git a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c index d0e128c24813..56bea36a6b18 100644 --- a/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/dce_v8_0.c @@ -2582,7 +2582,6 @@ static void dce_v8_0_crtc_destroy(struct drm_crtc *crtc) struct amdgpu_crtc *amdgpu_crtc = to_amdgpu_crtc(crtc); drm_crtc_cleanup(crtc); - destroy_workqueue(amdgpu_crtc->pflip_queue); kfree(amdgpu_crtc); } @@ -2809,7 +2808,6 @@ static int dce_v8_0_crtc_init(struct amdgpu_device *adev, int index) drm_mode_crtc_set_gamma_size(&amdgpu_crtc->base, 256); amdgpu_crtc->crtc_id = index; - amdgpu_crtc->pflip_queue = create_singlethread_workqueue("amdgpu-pageflip-queue"); adev->mode_info.crtcs[index] = amdgpu_crtc; amdgpu_crtc->max_cursor_width = CIK_CURSOR_WIDTH; @@ -3375,7 +3373,7 @@ static int dce_v8_0_pageflip_irq(struct amdgpu_device *adev, spin_unlock_irqrestore(&adev->ddev->event_lock, flags); drm_vblank_put(adev->ddev, amdgpu_crtc->crtc_id); - queue_work(amdgpu_crtc->pflip_queue, &works->unpin_work); + schedule_work(&works->unpin_work); return 0; } @@ -3554,16 +3552,8 @@ dce_v8_0_ext_dpms(struct drm_encoder *encoder, int mode) } -static bool dce_v8_0_ext_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static const struct drm_encoder_helper_funcs dce_v8_0_ext_helper_funcs = { .dpms = dce_v8_0_ext_dpms, - .mode_fixup = dce_v8_0_ext_mode_fixup, .prepare = dce_v8_0_ext_prepare, .mode_set = dce_v8_0_ext_mode_set, .commit = dce_v8_0_ext_commit, diff --git a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c index e35340afd3db..b336c918d6a7 100644 --- a/drivers/gpu/drm/amd/amdgpu/fiji_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/fiji_smc.c @@ -272,6 +272,12 @@ static int fiji_smu_upload_firmware_image(struct amdgpu_device *adev) if (!adev->pm.fw) return -EINVAL; + /* Skip SMC ucode loading on SR-IOV capable boards. + * vbios does this for us in asic_init in that case. + */ + if (adev->virtualization.supports_sr_iov) + return 0; + hdr = (const struct smc_firmware_header_v1_0 *)adev->pm.fw->data; amdgpu_ucode_print_smc_hdr(&hdr->header); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c index 6c76139de1c9..250bcbce7fdc 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v7_0.c @@ -31,8 +31,6 @@ #include "amdgpu_ucode.h" #include "clearstate_ci.h" -#include "uvd/uvd_4_2_d.h" - #include "dce/dce_8_0_d.h" #include "dce/dce_8_0_sh_mask.h" @@ -1006,9 +1004,15 @@ out: */ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) { - const u32 num_tile_mode_states = 32; - const u32 num_secondary_tile_mode_states = 16; - u32 reg_offset, gb_tile_moden, split_equal_to_row_size; + const u32 num_tile_mode_states = + ARRAY_SIZE(adev->gfx.config.tile_mode_array); + const u32 num_secondary_tile_mode_states = + ARRAY_SIZE(adev->gfx.config.macrotile_mode_array); + u32 reg_offset, split_equal_to_row_size; + uint32_t *tile, *macrotile; + + tile = adev->gfx.config.tile_mode_array; + macrotile = adev->gfx.config.macrotile_mode_array; switch (adev->gfx.config.mem_row_size_in_kb) { case 1: @@ -1023,832 +1027,531 @@ static void gfx_v7_0_tiling_mode_table_init(struct amdgpu_device *adev) break; } + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + tile[reg_offset] = 0; + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + macrotile[reg_offset] = 0; + switch (adev->asic_type) { case CHIP_BONAIRE: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P4_16x16)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P4_16x16)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (TILE_SPLIT(split_equal_to_row_size)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (TILE_SPLIT(split_equal_to_row_size)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (TILE_SPLIT(split_equal_to_row_size)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (TILE_SPLIT(split_equal_to_row_size)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; case CHIP_HAWAII: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P4_16x16) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | - NUM_BANKS(ADDR_SURF_4_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (ARRAY_MODE(ARRAY_PRT_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P16_32x32_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P4_16x16) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_1) | + NUM_BANKS(ADDR_SURF_4_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; case CHIP_KABINI: case CHIP_KAVERI: case CHIP_MULLINS: default: - for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 1: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 2: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 3: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 4: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 5: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); - break; - case 6: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | - TILE_SPLIT(split_equal_to_row_size)); - break; - case 7: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - - case 8: - gb_tile_moden = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | - PIPE_CONFIG(ADDR_SURF_P2)); - break; - case 9: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); - break; - case 10: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 11: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 12: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 13: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); - break; - case 14: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 15: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 16: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 17: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 18: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 19: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); - break; - case 20: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 21: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 22: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 23: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - case 24: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 25: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 26: - gb_tile_moden = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); - break; - case 27: - gb_tile_moden = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); - break; - case 28: - gb_tile_moden = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); - break; - case 29: - gb_tile_moden = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | - PIPE_CONFIG(ADDR_SURF_P2) | - MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | - SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); - break; - case 30: - gb_tile_moden = (TILE_SPLIT(split_equal_to_row_size)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.tile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_TILE_MODE0 + reg_offset, gb_tile_moden); - } - for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) { - switch (reg_offset) { - case 0: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 1: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 2: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 3: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 4: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 5: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 6: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - case 8: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 9: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 10: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 11: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 12: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 13: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | - NUM_BANKS(ADDR_SURF_16_BANK)); - break; - case 14: - gb_tile_moden = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | - BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | - MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | - NUM_BANKS(ADDR_SURF_8_BANK)); - break; - default: - gb_tile_moden = 0; - break; - } - adev->gfx.config.macrotile_mode_array[reg_offset] = gb_tile_moden; - WREG32(mmGB_MACROTILE_MODE0 + reg_offset, gb_tile_moden); - } + tile[0] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_64B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[1] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_128B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[2] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_256B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[3] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + TILE_SPLIT(ADDR_SURF_TILE_SPLIT_512B) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[4] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[5] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING)); + tile[6] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DEPTH_MICRO_TILING) | + TILE_SPLIT(split_equal_to_row_size)); + tile[7] = (TILE_SPLIT(split_equal_to_row_size)); + tile[8] = (ARRAY_MODE(ARRAY_LINEAR_ALIGNED) | + PIPE_CONFIG(ADDR_SURF_P2)); + tile[9] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING)); + tile[10] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[11] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_DISPLAY_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[12] = (TILE_SPLIT(split_equal_to_row_size)); + tile[13] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING)); + tile[14] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[15] = (ARRAY_MODE(ARRAY_3D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[16] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[17] = (TILE_SPLIT(split_equal_to_row_size)); + tile[18] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[19] = (ARRAY_MODE(ARRAY_1D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING)); + tile[20] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[21] = (ARRAY_MODE(ARRAY_3D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[22] = (ARRAY_MODE(ARRAY_PRT_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[23] = (TILE_SPLIT(split_equal_to_row_size)); + tile[24] = (ARRAY_MODE(ARRAY_2D_TILED_THICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THIN_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[25] = (ARRAY_MODE(ARRAY_2D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[26] = (ARRAY_MODE(ARRAY_3D_TILED_XTHICK) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_THICK_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_1)); + tile[27] = (ARRAY_MODE(ARRAY_1D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING)); + tile[28] = (ARRAY_MODE(ARRAY_2D_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_2)); + tile[29] = (ARRAY_MODE(ARRAY_PRT_TILED_THIN1) | + PIPE_CONFIG(ADDR_SURF_P2) | + MICRO_TILE_MODE_NEW(ADDR_SURF_ROTATED_MICRO_TILING) | + SAMPLE_SPLIT(ADDR_SURF_SAMPLE_SPLIT_8)); + tile[30] = (TILE_SPLIT(split_equal_to_row_size)); + + macrotile[0] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[1] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[2] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[3] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[4] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[5] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[6] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + macrotile[8] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_8) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[9] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_4) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[10] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_4) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[11] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_2) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[12] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_2) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[13] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_4) | + NUM_BANKS(ADDR_SURF_16_BANK)); + macrotile[14] = (BANK_WIDTH(ADDR_SURF_BANK_WIDTH_1) | + BANK_HEIGHT(ADDR_SURF_BANK_HEIGHT_1) | + MACRO_TILE_ASPECT(ADDR_SURF_MACRO_ASPECT_2) | + NUM_BANKS(ADDR_SURF_8_BANK)); + + for (reg_offset = 0; reg_offset < num_tile_mode_states; reg_offset++) + WREG32(mmGB_TILE_MODE0 + reg_offset, tile[reg_offset]); + for (reg_offset = 0; reg_offset < num_secondary_tile_mode_states; reg_offset++) + if (reg_offset != 7) + WREG32(mmGB_MACROTILE_MODE0 + reg_offset, macrotile[reg_offset]); break; } } @@ -1893,45 +1596,31 @@ void gfx_v7_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) */ static u32 gfx_v7_0_create_bitmask(u32 bit_width) { - u32 i, mask = 0; - - for (i = 0; i < bit_width; i++) { - mask <<= 1; - mask |= 1; - } - return mask; + return (u32)((1ULL << bit_width) - 1); } /** - * gfx_v7_0_get_rb_disabled - computes the mask of disabled RBs + * gfx_v7_0_get_rb_active_bitmap - computes the mask of enabled RBs * * @adev: amdgpu_device pointer - * @max_rb_num: max RBs (render backends) for the asic - * @se_num: number of SEs (shader engines) for the asic - * @sh_per_se: number of SH blocks per SE for the asic * - * Calculates the bitmask of disabled RBs (CIK). - * Returns the disabled RB bitmask. + * Calculates the bitmask of enabled RBs (CIK). + * Returns the enabled RB bitmask. */ -static u32 gfx_v7_0_get_rb_disabled(struct amdgpu_device *adev, - u32 max_rb_num_per_se, - u32 sh_per_se) +static u32 gfx_v7_0_get_rb_active_bitmap(struct amdgpu_device *adev) { u32 data, mask; data = RREG32(mmCC_RB_BACKEND_DISABLE); - if (data & 1) - data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; - else - data = 0; - data |= RREG32(mmGC_USER_RB_BACKEND_DISABLE); + data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; data >>= GC_USER_RB_BACKEND_DISABLE__BACKEND_DISABLE__SHIFT; - mask = gfx_v7_0_create_bitmask(max_rb_num_per_se / sh_per_se); + mask = gfx_v7_0_create_bitmask(adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se); - return data & mask; + return (~data) & mask; } /** @@ -1940,73 +1629,36 @@ static u32 gfx_v7_0_get_rb_disabled(struct amdgpu_device *adev, * @adev: amdgpu_device pointer * @se_num: number of SEs (shader engines) for the asic * @sh_per_se: number of SH blocks per SE for the asic - * @max_rb_num: max RBs (render backends) for the asic * * Configures per-SE/SH RB registers (CIK). */ -static void gfx_v7_0_setup_rb(struct amdgpu_device *adev, - u32 se_num, u32 sh_per_se, - u32 max_rb_num_per_se) +static void gfx_v7_0_setup_rb(struct amdgpu_device *adev) { int i, j; - u32 data, mask; - u32 disabled_rbs = 0; - u32 enabled_rbs = 0; + u32 data, tmp, num_rbs = 0; + u32 active_rbs = 0; mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - for (j = 0; j < sh_per_se; j++) { + for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { + for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { gfx_v7_0_select_se_sh(adev, i, j); - data = gfx_v7_0_get_rb_disabled(adev, max_rb_num_per_se, sh_per_se); + data = gfx_v7_0_get_rb_active_bitmap(adev); if (adev->asic_type == CHIP_HAWAII) - disabled_rbs |= data << ((i * sh_per_se + j) * HAWAII_RB_BITMAP_WIDTH_PER_SH); + active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * + HAWAII_RB_BITMAP_WIDTH_PER_SH); else - disabled_rbs |= data << ((i * sh_per_se + j) * CIK_RB_BITMAP_WIDTH_PER_SH); + active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * + CIK_RB_BITMAP_WIDTH_PER_SH); } } gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); - mask = 1; - for (i = 0; i < max_rb_num_per_se * se_num; i++) { - if (!(disabled_rbs & mask)) - enabled_rbs |= mask; - mask <<= 1; - } - - adev->gfx.config.backend_enable_mask = enabled_rbs; - - mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - gfx_v7_0_select_se_sh(adev, i, 0xffffffff); - data = 0; - for (j = 0; j < sh_per_se; j++) { - switch (enabled_rbs & 3) { - case 0: - if (j == 0) - data |= (RASTER_CONFIG_RB_MAP_3 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - else - data |= (RASTER_CONFIG_RB_MAP_0 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - break; - case 1: - data |= (RASTER_CONFIG_RB_MAP_0 << (i * sh_per_se + j) * 2); - break; - case 2: - data |= (RASTER_CONFIG_RB_MAP_3 << (i * sh_per_se + j) * 2); - break; - case 3: - default: - data |= (RASTER_CONFIG_RB_MAP_2 << (i * sh_per_se + j) * 2); - break; - } - enabled_rbs >>= 2; - } - WREG32(mmPA_SC_RASTER_CONFIG, data); - } - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); - mutex_unlock(&adev->grbm_idx_mutex); + adev->gfx.config.backend_enable_mask = active_rbs; + tmp = active_rbs; + while (tmp >>= 1) + num_rbs++; + adev->gfx.config.num_rbs = num_rbs; } /** @@ -2059,192 +1711,23 @@ static void gmc_v7_0_init_compute_vmid(struct amdgpu_device *adev) */ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) { - u32 gb_addr_config; - u32 mc_shared_chmap, mc_arb_ramcfg; - u32 dimm00_addr_map, dimm01_addr_map, dimm10_addr_map, dimm11_addr_map; - u32 sh_mem_cfg; - u32 tmp; + u32 tmp, sh_mem_cfg; int i; - switch (adev->asic_type) { - case CHIP_BONAIRE: - adev->gfx.config.max_shader_engines = 2; - adev->gfx.config.max_tile_pipes = 4; - adev->gfx.config.max_cu_per_sh = 7; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 2; - adev->gfx.config.max_texture_channel_caches = 4; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 32; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_HAWAII: - adev->gfx.config.max_shader_engines = 4; - adev->gfx.config.max_tile_pipes = 16; - adev->gfx.config.max_cu_per_sh = 11; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 4; - adev->gfx.config.max_texture_channel_caches = 16; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 32; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = HAWAII_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_KAVERI: - adev->gfx.config.max_shader_engines = 1; - adev->gfx.config.max_tile_pipes = 4; - if ((adev->pdev->device == 0x1304) || - (adev->pdev->device == 0x1305) || - (adev->pdev->device == 0x130C) || - (adev->pdev->device == 0x130F) || - (adev->pdev->device == 0x1310) || - (adev->pdev->device == 0x1311) || - (adev->pdev->device == 0x131C)) { - adev->gfx.config.max_cu_per_sh = 8; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1309) || - (adev->pdev->device == 0x130A) || - (adev->pdev->device == 0x130D) || - (adev->pdev->device == 0x1313) || - (adev->pdev->device == 0x131D)) { - adev->gfx.config.max_cu_per_sh = 6; - adev->gfx.config.max_backends_per_se = 2; - } else if ((adev->pdev->device == 0x1306) || - (adev->pdev->device == 0x1307) || - (adev->pdev->device == 0x130B) || - (adev->pdev->device == 0x130E) || - (adev->pdev->device == 0x1315) || - (adev->pdev->device == 0x131B)) { - adev->gfx.config.max_cu_per_sh = 4; - adev->gfx.config.max_backends_per_se = 1; - } else { - adev->gfx.config.max_cu_per_sh = 3; - adev->gfx.config.max_backends_per_se = 1; - } - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_texture_channel_caches = 4; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 16; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - case CHIP_KABINI: - case CHIP_MULLINS: - default: - adev->gfx.config.max_shader_engines = 1; - adev->gfx.config.max_tile_pipes = 2; - adev->gfx.config.max_cu_per_sh = 2; - adev->gfx.config.max_sh_per_se = 1; - adev->gfx.config.max_backends_per_se = 1; - adev->gfx.config.max_texture_channel_caches = 2; - adev->gfx.config.max_gprs = 256; - adev->gfx.config.max_gs_threads = 16; - adev->gfx.config.max_hw_contexts = 8; - - adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; - adev->gfx.config.sc_prim_fifo_size_backend = 0x100; - adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; - adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; - gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; - break; - } - WREG32(mmGRBM_CNTL, (0xff << GRBM_CNTL__READ_TIMEOUT__SHIFT)); - mc_shared_chmap = RREG32(mmMC_SHARED_CHMAP); - adev->gfx.config.mc_arb_ramcfg = RREG32(mmMC_ARB_RAMCFG); - mc_arb_ramcfg = adev->gfx.config.mc_arb_ramcfg; - - adev->gfx.config.num_tile_pipes = adev->gfx.config.max_tile_pipes; - adev->gfx.config.mem_max_burst_length_bytes = 256; - if (adev->flags & AMD_IS_APU) { - /* Get memory bank mapping mode. */ - tmp = RREG32(mmMC_FUS_DRAM0_BANK_ADDR_MAPPING); - dimm00_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM0ADDRMAP); - dimm01_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM1ADDRMAP); - - tmp = RREG32(mmMC_FUS_DRAM1_BANK_ADDR_MAPPING); - dimm10_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM0ADDRMAP); - dimm11_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM1ADDRMAP); - - /* Validate settings in case only one DIMM installed. */ - if ((dimm00_addr_map == 0) || (dimm00_addr_map == 3) || (dimm00_addr_map == 4) || (dimm00_addr_map > 12)) - dimm00_addr_map = 0; - if ((dimm01_addr_map == 0) || (dimm01_addr_map == 3) || (dimm01_addr_map == 4) || (dimm01_addr_map > 12)) - dimm01_addr_map = 0; - if ((dimm10_addr_map == 0) || (dimm10_addr_map == 3) || (dimm10_addr_map == 4) || (dimm10_addr_map > 12)) - dimm10_addr_map = 0; - if ((dimm11_addr_map == 0) || (dimm11_addr_map == 3) || (dimm11_addr_map == 4) || (dimm11_addr_map > 12)) - dimm11_addr_map = 0; - - /* If DIMM Addr map is 8GB, ROW size should be 2KB. Otherwise 1KB. */ - /* If ROW size(DIMM1) != ROW size(DMIMM0), ROW size should be larger one. */ - if ((dimm00_addr_map == 11) || (dimm01_addr_map == 11) || (dimm10_addr_map == 11) || (dimm11_addr_map == 11)) - adev->gfx.config.mem_row_size_in_kb = 2; - else - adev->gfx.config.mem_row_size_in_kb = 1; - } else { - tmp = (mc_arb_ramcfg & MC_ARB_RAMCFG__NOOFCOLS_MASK) >> MC_ARB_RAMCFG__NOOFCOLS__SHIFT; - adev->gfx.config.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024; - if (adev->gfx.config.mem_row_size_in_kb > 4) - adev->gfx.config.mem_row_size_in_kb = 4; - } - /* XXX use MC settings? */ - adev->gfx.config.shader_engine_tile_size = 32; - adev->gfx.config.num_gpus = 1; - adev->gfx.config.multi_gpu_tile_size = 64; - - /* fix up row size */ - gb_addr_config &= ~GB_ADDR_CONFIG__ROW_SIZE_MASK; - switch (adev->gfx.config.mem_row_size_in_kb) { - case 1: - default: - gb_addr_config |= (0 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - case 2: - gb_addr_config |= (1 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - case 4: - gb_addr_config |= (2 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); - break; - } - adev->gfx.config.gb_addr_config = gb_addr_config; - - WREG32(mmGB_ADDR_CONFIG, gb_addr_config); - WREG32(mmHDP_ADDR_CONFIG, gb_addr_config); - WREG32(mmDMIF_ADDR_CALC, gb_addr_config); - WREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET, gb_addr_config & 0x70); - WREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET, gb_addr_config & 0x70); - WREG32(mmUVD_UDEC_ADDR_CONFIG, gb_addr_config); - WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, gb_addr_config); - WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, gb_addr_config); + WREG32(mmGB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmHDP_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmDMIF_ADDR_CALC, adev->gfx.config.gb_addr_config); gfx_v7_0_tiling_mode_table_init(adev); - gfx_v7_0_setup_rb(adev, adev->gfx.config.max_shader_engines, - adev->gfx.config.max_sh_per_se, - adev->gfx.config.max_backends_per_se); + gfx_v7_0_setup_rb(adev); /* set HW defaults for 3D engine */ WREG32(mmCP_MEQ_THRESHOLDS, - (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | - (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); + (0x30 << CP_MEQ_THRESHOLDS__MEQ1_START__SHIFT) | + (0x60 << CP_MEQ_THRESHOLDS__MEQ2_START__SHIFT)); mutex_lock(&adev->grbm_idx_mutex); /* @@ -2255,7 +1738,7 @@ static void gfx_v7_0_gpu_init(struct amdgpu_device *adev) /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ - sh_mem_cfg = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, + sh_mem_cfg = REG_SET_FIELD(0, SH_MEM_CONFIG, ALIGNMENT_MODE, SH_MEM_ALIGNMENT_MODE_UNALIGNED); mutex_lock(&adev->srbm_mutex); @@ -2379,7 +1862,7 @@ static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring) return r; } WREG32(scratch, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); amdgpu_gfx_scratch_free(adev, scratch); @@ -2388,7 +1871,7 @@ static int gfx_v7_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1)); amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -2516,36 +1999,6 @@ static void gfx_v7_0_ring_emit_fence_compute(struct amdgpu_ring *ring, amdgpu_ring_write(ring, upper_32_bits(seq)); } -/** - * gfx_v7_0_ring_emit_semaphore - emit a semaphore on the CP ring - * - * @ring: amdgpu ring buffer object - * @semaphore: amdgpu semaphore object - * @emit_wait: Is this a sempahore wait? - * - * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP - * from running ahead of semaphore waits. - */ -static bool gfx_v7_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; - - amdgpu_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 1)); - amdgpu_ring_write(ring, addr & 0xffffffff); - amdgpu_ring_write(ring, (upper_32_bits(addr) & 0xffff) | sel); - - if (emit_wait && (ring->type == AMDGPU_RING_TYPE_GFX)) { - /* Prevent the PFP from running ahead of the semaphore wait */ - amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); - amdgpu_ring_write(ring, 0x0); - } - - return true; -} - /* * IB stuff */ @@ -2661,7 +2114,7 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) } WREG32(scratch, 0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err1; @@ -2671,9 +2124,8 @@ static int gfx_v7_0_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[2] = 0xDEADBEEF; ib.length_dw = 3; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err2; @@ -2842,7 +2294,7 @@ static int gfx_v7_0_cp_gfx_start(struct amdgpu_device *adev) gfx_v7_0_cp_gfx_enable(adev, true); - r = amdgpu_ring_lock(ring, gfx_v7_0_get_csb_size(adev) + 8); + r = amdgpu_ring_alloc(ring, gfx_v7_0_get_csb_size(adev) + 8); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r); return r; @@ -2911,7 +2363,7 @@ static int gfx_v7_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, 0x0000000e); /* VGT_VERTEX_REUSE_BLOCK_CNTL */ amdgpu_ring_write(ring, 0x00000010); /* VGT_OUT_DEALLOC_CNTL */ - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); return 0; } @@ -2989,21 +2441,14 @@ static int gfx_v7_0_cp_gfx_resume(struct amdgpu_device *adev) static u32 gfx_v7_0_ring_get_rptr_gfx(struct amdgpu_ring *ring) { - u32 rptr; - - rptr = ring->adev->wb.wb[ring->rptr_offs]; - - return rptr; + return ring->adev->wb.wb[ring->rptr_offs]; } static u32 gfx_v7_0_ring_get_wptr_gfx(struct amdgpu_ring *ring) { struct amdgpu_device *adev = ring->adev; - u32 wptr; - wptr = RREG32(mmCP_RB0_WPTR); - - return wptr; + return RREG32(mmCP_RB0_WPTR); } static void gfx_v7_0_ring_set_wptr_gfx(struct amdgpu_ring *ring) @@ -3016,21 +2461,13 @@ static void gfx_v7_0_ring_set_wptr_gfx(struct amdgpu_ring *ring) static u32 gfx_v7_0_ring_get_rptr_compute(struct amdgpu_ring *ring) { - u32 rptr; - - rptr = ring->adev->wb.wb[ring->rptr_offs]; - - return rptr; + return ring->adev->wb.wb[ring->rptr_offs]; } static u32 gfx_v7_0_ring_get_wptr_compute(struct amdgpu_ring *ring) { - u32 wptr; - /* XXX check if swapping is necessary on BE */ - wptr = ring->adev->wb.wb[ring->wptr_offs]; - - return wptr; + return ring->adev->wb.wb[ring->wptr_offs]; } static void gfx_v7_0_ring_set_wptr_compute(struct amdgpu_ring *ring) @@ -3126,21 +2563,6 @@ static int gfx_v7_0_cp_compute_load_microcode(struct amdgpu_device *adev) } /** - * gfx_v7_0_cp_compute_start - start the compute queues - * - * @adev: amdgpu_device pointer - * - * Enable the compute queues. - * Returns 0 for success, error for failure. - */ -static int gfx_v7_0_cp_compute_start(struct amdgpu_device *adev) -{ - gfx_v7_0_cp_compute_enable(adev, true); - - return 0; -} - -/** * gfx_v7_0_cp_compute_fini - stop the compute queues * * @adev: amdgpu_device pointer @@ -3330,9 +2752,7 @@ static int gfx_v7_0_cp_compute_resume(struct amdgpu_device *adev) u32 *buf; struct bonaire_mqd *mqd; - r = gfx_v7_0_cp_compute_start(adev); - if (r) - return r; + gfx_v7_0_cp_compute_enable(adev, true); /* fix up chicken bits */ tmp = RREG32(mmCP_CPF_DEBUG); @@ -4395,28 +3815,20 @@ static void gfx_v7_0_enable_gfx_cgpg(struct amdgpu_device *adev, } } -static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev, - u32 se, u32 sh) +static u32 gfx_v7_0_get_cu_active_bitmap(struct amdgpu_device *adev) { - u32 mask = 0, tmp, tmp1; - int i; - - gfx_v7_0_select_se_sh(adev, se, sh); - tmp = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); - tmp1 = RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + u32 data, mask; - tmp &= 0xffff0000; + data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); + data |= RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - tmp |= tmp1; - tmp >>= 16; + data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; - for (i = 0; i < adev->gfx.config.max_cu_per_sh; i ++) { - mask <<= 1; - mask |= 1; - } + mask = gfx_v7_0_create_bitmask(adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se); - return (~tmp) & mask; + return (~data) & mask; } static void gfx_v7_0_init_ao_cu_mask(struct amdgpu_device *adev) @@ -4754,6 +4166,172 @@ static int gfx_v7_0_late_init(void *handle) return 0; } +static void gfx_v7_0_gpu_early_init(struct amdgpu_device *adev) +{ + u32 gb_addr_config; + u32 mc_shared_chmap, mc_arb_ramcfg; + u32 dimm00_addr_map, dimm01_addr_map, dimm10_addr_map, dimm11_addr_map; + u32 tmp; + + switch (adev->asic_type) { + case CHIP_BONAIRE: + adev->gfx.config.max_shader_engines = 2; + adev->gfx.config.max_tile_pipes = 4; + adev->gfx.config.max_cu_per_sh = 7; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 2; + adev->gfx.config.max_texture_channel_caches = 4; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 32; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_HAWAII: + adev->gfx.config.max_shader_engines = 4; + adev->gfx.config.max_tile_pipes = 16; + adev->gfx.config.max_cu_per_sh = 11; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 4; + adev->gfx.config.max_texture_channel_caches = 16; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 32; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = HAWAII_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_KAVERI: + adev->gfx.config.max_shader_engines = 1; + adev->gfx.config.max_tile_pipes = 4; + if ((adev->pdev->device == 0x1304) || + (adev->pdev->device == 0x1305) || + (adev->pdev->device == 0x130C) || + (adev->pdev->device == 0x130F) || + (adev->pdev->device == 0x1310) || + (adev->pdev->device == 0x1311) || + (adev->pdev->device == 0x131C)) { + adev->gfx.config.max_cu_per_sh = 8; + adev->gfx.config.max_backends_per_se = 2; + } else if ((adev->pdev->device == 0x1309) || + (adev->pdev->device == 0x130A) || + (adev->pdev->device == 0x130D) || + (adev->pdev->device == 0x1313) || + (adev->pdev->device == 0x131D)) { + adev->gfx.config.max_cu_per_sh = 6; + adev->gfx.config.max_backends_per_se = 2; + } else if ((adev->pdev->device == 0x1306) || + (adev->pdev->device == 0x1307) || + (adev->pdev->device == 0x130B) || + (adev->pdev->device == 0x130E) || + (adev->pdev->device == 0x1315) || + (adev->pdev->device == 0x131B)) { + adev->gfx.config.max_cu_per_sh = 4; + adev->gfx.config.max_backends_per_se = 1; + } else { + adev->gfx.config.max_cu_per_sh = 3; + adev->gfx.config.max_backends_per_se = 1; + } + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_texture_channel_caches = 4; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 16; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + case CHIP_KABINI: + case CHIP_MULLINS: + default: + adev->gfx.config.max_shader_engines = 1; + adev->gfx.config.max_tile_pipes = 2; + adev->gfx.config.max_cu_per_sh = 2; + adev->gfx.config.max_sh_per_se = 1; + adev->gfx.config.max_backends_per_se = 1; + adev->gfx.config.max_texture_channel_caches = 2; + adev->gfx.config.max_gprs = 256; + adev->gfx.config.max_gs_threads = 16; + adev->gfx.config.max_hw_contexts = 8; + + adev->gfx.config.sc_prim_fifo_size_frontend = 0x20; + adev->gfx.config.sc_prim_fifo_size_backend = 0x100; + adev->gfx.config.sc_hiz_tile_fifo_size = 0x30; + adev->gfx.config.sc_earlyz_tile_fifo_size = 0x130; + gb_addr_config = BONAIRE_GB_ADDR_CONFIG_GOLDEN; + break; + } + + mc_shared_chmap = RREG32(mmMC_SHARED_CHMAP); + adev->gfx.config.mc_arb_ramcfg = RREG32(mmMC_ARB_RAMCFG); + mc_arb_ramcfg = adev->gfx.config.mc_arb_ramcfg; + + adev->gfx.config.num_tile_pipes = adev->gfx.config.max_tile_pipes; + adev->gfx.config.mem_max_burst_length_bytes = 256; + if (adev->flags & AMD_IS_APU) { + /* Get memory bank mapping mode. */ + tmp = RREG32(mmMC_FUS_DRAM0_BANK_ADDR_MAPPING); + dimm00_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM0ADDRMAP); + dimm01_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM0_BANK_ADDR_MAPPING, DIMM1ADDRMAP); + + tmp = RREG32(mmMC_FUS_DRAM1_BANK_ADDR_MAPPING); + dimm10_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM0ADDRMAP); + dimm11_addr_map = REG_GET_FIELD(tmp, MC_FUS_DRAM1_BANK_ADDR_MAPPING, DIMM1ADDRMAP); + + /* Validate settings in case only one DIMM installed. */ + if ((dimm00_addr_map == 0) || (dimm00_addr_map == 3) || (dimm00_addr_map == 4) || (dimm00_addr_map > 12)) + dimm00_addr_map = 0; + if ((dimm01_addr_map == 0) || (dimm01_addr_map == 3) || (dimm01_addr_map == 4) || (dimm01_addr_map > 12)) + dimm01_addr_map = 0; + if ((dimm10_addr_map == 0) || (dimm10_addr_map == 3) || (dimm10_addr_map == 4) || (dimm10_addr_map > 12)) + dimm10_addr_map = 0; + if ((dimm11_addr_map == 0) || (dimm11_addr_map == 3) || (dimm11_addr_map == 4) || (dimm11_addr_map > 12)) + dimm11_addr_map = 0; + + /* If DIMM Addr map is 8GB, ROW size should be 2KB. Otherwise 1KB. */ + /* If ROW size(DIMM1) != ROW size(DMIMM0), ROW size should be larger one. */ + if ((dimm00_addr_map == 11) || (dimm01_addr_map == 11) || (dimm10_addr_map == 11) || (dimm11_addr_map == 11)) + adev->gfx.config.mem_row_size_in_kb = 2; + else + adev->gfx.config.mem_row_size_in_kb = 1; + } else { + tmp = (mc_arb_ramcfg & MC_ARB_RAMCFG__NOOFCOLS_MASK) >> MC_ARB_RAMCFG__NOOFCOLS__SHIFT; + adev->gfx.config.mem_row_size_in_kb = (4 * (1 << (8 + tmp))) / 1024; + if (adev->gfx.config.mem_row_size_in_kb > 4) + adev->gfx.config.mem_row_size_in_kb = 4; + } + /* XXX use MC settings? */ + adev->gfx.config.shader_engine_tile_size = 32; + adev->gfx.config.num_gpus = 1; + adev->gfx.config.multi_gpu_tile_size = 64; + + /* fix up row size */ + gb_addr_config &= ~GB_ADDR_CONFIG__ROW_SIZE_MASK; + switch (adev->gfx.config.mem_row_size_in_kb) { + case 1: + default: + gb_addr_config |= (0 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + case 2: + gb_addr_config |= (1 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + case 4: + gb_addr_config |= (2 << GB_ADDR_CONFIG__ROW_SIZE__SHIFT); + break; + } + adev->gfx.config.gb_addr_config = gb_addr_config; +} + static int gfx_v7_0_sw_init(void *handle) { struct amdgpu_ring *ring; @@ -4857,6 +4435,10 @@ static int gfx_v7_0_sw_init(void *handle) if (r) return r; + adev->gfx.ce_ram_size = 0x8000; + + gfx_v7_0_gpu_early_init(adev); + return r; } @@ -4897,8 +4479,6 @@ static int gfx_v7_0_hw_init(void *handle) if (r) return r; - adev->gfx.ce_ram_size = 0x8000; - return r; } @@ -5015,16 +4595,6 @@ static void gfx_v7_0_print_status(void *handle) RREG32(mmHDP_ADDR_CONFIG)); dev_info(adev->dev, " DMIF_ADDR_CALC=0x%08X\n", RREG32(mmDMIF_ADDR_CALC)); - dev_info(adev->dev, " SDMA0_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET)); - dev_info(adev->dev, " SDMA1_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET)); - dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); dev_info(adev->dev, " CP_MEQ_THRESHOLDS=0x%08X\n", RREG32(mmCP_MEQ_THRESHOLDS)); @@ -5567,13 +5137,13 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_gfx = { .parse_cs = NULL, .emit_ib = gfx_v7_0_ring_emit_ib_gfx, .emit_fence = gfx_v7_0_ring_emit_fence_gfx, - .emit_semaphore = gfx_v7_0_ring_emit_semaphore, .emit_vm_flush = gfx_v7_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v7_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v7_0_ring_emit_hdp_flush, .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { @@ -5583,13 +5153,13 @@ static const struct amdgpu_ring_funcs gfx_v7_0_ring_funcs_compute = { .parse_cs = NULL, .emit_ib = gfx_v7_0_ring_emit_ib_compute, .emit_fence = gfx_v7_0_ring_emit_fence_compute, - .emit_semaphore = gfx_v7_0_ring_emit_semaphore, .emit_vm_flush = gfx_v7_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v7_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v7_0_ring_emit_hdp_flush, .test_ring = gfx_v7_0_ring_test_ring, .test_ib = gfx_v7_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void gfx_v7_0_set_ring_funcs(struct amdgpu_device *adev) @@ -5659,7 +5229,7 @@ static void gfx_v7_0_set_gds_init(struct amdgpu_device *adev) int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, - struct amdgpu_cu_info *cu_info) + struct amdgpu_cu_info *cu_info) { int i, j, k, counter, active_cu_number = 0; u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; @@ -5673,10 +5243,11 @@ int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, mask = 1; ao_bitmap = 0; counter = 0; - bitmap = gfx_v7_0_get_cu_active_bitmap(adev, i, j); + gfx_v7_0_select_se_sh(adev, i, j); + bitmap = gfx_v7_0_get_cu_active_bitmap(adev); cu_info->bitmap[i][j] = bitmap; - for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) { + for (k = 0; k < 16; k ++) { if (bitmap & mask) { if (counter < 2) ao_bitmap |= mask; @@ -5688,9 +5259,11 @@ int gfx_v7_0_get_cu_info(struct amdgpu_device *adev, ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); } } + gfx_v7_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; cu_info->ao_cu_mask = ao_cu_mask; - mutex_unlock(&adev->grbm_idx_mutex); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 8f8ec37ecd88..10c865087d0a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -43,9 +43,6 @@ #include "gca/gfx_8_0_sh_mask.h" #include "gca/gfx_8_0_enum.h" -#include "uvd/uvd_5_0_d.h" -#include "uvd/uvd_5_0_sh_mask.h" - #include "dce/dce_10_0_d.h" #include "dce/dce_10_0_sh_mask.h" @@ -652,7 +649,7 @@ static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring) return r; } WREG32(scratch, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); @@ -662,7 +659,7 @@ static int gfx_v8_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, PACKET3(PACKET3_SET_UCONFIG_REG, 1)); amdgpu_ring_write(ring, (scratch - PACKET3_SET_UCONFIG_REG_START)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(scratch); @@ -699,7 +696,7 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring) } WREG32(scratch, 0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err1; @@ -709,9 +706,8 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[2] = 0xDEADBEEF; ib.length_dw = 3; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err2; @@ -1171,7 +1167,7 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev) /* allocate an indirect buffer to put the commands in */ memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, total_size, &ib); + r = amdgpu_ib_get(adev, NULL, total_size, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); return r; @@ -1266,9 +1262,8 @@ static int gfx_v8_0_do_edc_gpr_workarounds(struct amdgpu_device *adev) ib.ptr[ib.length_dw++] = EVENT_TYPE(7) | EVENT_INDEX(4); /* shedule the ib on the ring */ - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) { DRM_ERROR("amdgpu: ib submit failed (%d).\n", r); goto fail; @@ -2574,11 +2569,6 @@ static void gfx_v8_0_tiling_mode_table_init(struct amdgpu_device *adev) } } -static u32 gfx_v8_0_create_bitmask(u32 bit_width) -{ - return (u32)((1ULL << bit_width) - 1); -} - void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) { u32 data = REG_SET_FIELD(0, GRBM_GFX_INDEX, INSTANCE_BROADCAST_WRITES, 1); @@ -2599,89 +2589,50 @@ void gfx_v8_0_select_se_sh(struct amdgpu_device *adev, u32 se_num, u32 sh_num) WREG32(mmGRBM_GFX_INDEX, data); } -static u32 gfx_v8_0_get_rb_disabled(struct amdgpu_device *adev, - u32 max_rb_num_per_se, - u32 sh_per_se) +static u32 gfx_v8_0_create_bitmask(u32 bit_width) +{ + return (u32)((1ULL << bit_width) - 1); +} + +static u32 gfx_v8_0_get_rb_active_bitmap(struct amdgpu_device *adev) { u32 data, mask; data = RREG32(mmCC_RB_BACKEND_DISABLE); - data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; - data |= RREG32(mmGC_USER_RB_BACKEND_DISABLE); + data &= CC_RB_BACKEND_DISABLE__BACKEND_DISABLE_MASK; data >>= GC_USER_RB_BACKEND_DISABLE__BACKEND_DISABLE__SHIFT; - mask = gfx_v8_0_create_bitmask(max_rb_num_per_se / sh_per_se); + mask = gfx_v8_0_create_bitmask(adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se); - return data & mask; + return (~data) & mask; } -static void gfx_v8_0_setup_rb(struct amdgpu_device *adev, - u32 se_num, u32 sh_per_se, - u32 max_rb_num_per_se) +static void gfx_v8_0_setup_rb(struct amdgpu_device *adev) { int i, j; - u32 data, mask; - u32 disabled_rbs = 0; - u32 enabled_rbs = 0; + u32 data, tmp, num_rbs = 0; + u32 active_rbs = 0; mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - for (j = 0; j < sh_per_se; j++) { + for (i = 0; i < adev->gfx.config.max_shader_engines; i++) { + for (j = 0; j < adev->gfx.config.max_sh_per_se; j++) { gfx_v8_0_select_se_sh(adev, i, j); - data = gfx_v8_0_get_rb_disabled(adev, - max_rb_num_per_se, sh_per_se); - disabled_rbs |= data << ((i * sh_per_se + j) * - RB_BITMAP_WIDTH_PER_SH); + data = gfx_v8_0_get_rb_active_bitmap(adev); + active_rbs |= data << ((i * adev->gfx.config.max_sh_per_se + j) * + RB_BITMAP_WIDTH_PER_SH); } } gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); mutex_unlock(&adev->grbm_idx_mutex); - mask = 1; - for (i = 0; i < max_rb_num_per_se * se_num; i++) { - if (!(disabled_rbs & mask)) - enabled_rbs |= mask; - mask <<= 1; - } - - adev->gfx.config.backend_enable_mask = enabled_rbs; - - mutex_lock(&adev->grbm_idx_mutex); - for (i = 0; i < se_num; i++) { - gfx_v8_0_select_se_sh(adev, i, 0xffffffff); - data = RREG32(mmPA_SC_RASTER_CONFIG); - for (j = 0; j < sh_per_se; j++) { - switch (enabled_rbs & 3) { - case 0: - if (j == 0) - data |= (RASTER_CONFIG_RB_MAP_3 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - else - data |= (RASTER_CONFIG_RB_MAP_0 << - PA_SC_RASTER_CONFIG__PKR_MAP__SHIFT); - break; - case 1: - data |= (RASTER_CONFIG_RB_MAP_0 << - (i * sh_per_se + j) * 2); - break; - case 2: - data |= (RASTER_CONFIG_RB_MAP_3 << - (i * sh_per_se + j) * 2); - break; - case 3: - default: - data |= (RASTER_CONFIG_RB_MAP_2 << - (i * sh_per_se + j) * 2); - break; - } - enabled_rbs >>= 2; - } - WREG32(mmPA_SC_RASTER_CONFIG, data); - } - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); - mutex_unlock(&adev->grbm_idx_mutex); + adev->gfx.config.backend_enable_mask = active_rbs; + tmp = active_rbs; + while (tmp >>= 1) + num_rbs++; + adev->gfx.config.num_rbs = num_rbs; } /** @@ -2741,19 +2692,10 @@ static void gfx_v8_0_gpu_init(struct amdgpu_device *adev) WREG32(mmGB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); WREG32(mmHDP_ADDR_CONFIG, adev->gfx.config.gb_addr_config); WREG32(mmDMIF_ADDR_CALC, adev->gfx.config.gb_addr_config); - WREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET, - adev->gfx.config.gb_addr_config & 0x70); - WREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET, - adev->gfx.config.gb_addr_config & 0x70); - WREG32(mmUVD_UDEC_ADDR_CONFIG, adev->gfx.config.gb_addr_config); - WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); - WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, adev->gfx.config.gb_addr_config); gfx_v8_0_tiling_mode_table_init(adev); - gfx_v8_0_setup_rb(adev, adev->gfx.config.max_shader_engines, - adev->gfx.config.max_sh_per_se, - adev->gfx.config.max_backends_per_se); + gfx_v8_0_setup_rb(adev); /* XXX SH_MEM regs */ /* where to put LDS, scratch, GPUVM in FSA64 space */ @@ -3062,7 +3004,7 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev) gfx_v8_0_cp_gfx_enable(adev, true); - r = amdgpu_ring_lock(ring, gfx_v8_0_get_csb_size(adev) + 4); + r = amdgpu_ring_alloc(ring, gfx_v8_0_get_csb_size(adev) + 4); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring (%d).\n", r); return r; @@ -3126,7 +3068,7 @@ static int gfx_v8_0_cp_gfx_start(struct amdgpu_device *adev) amdgpu_ring_write(ring, 0x8000); amdgpu_ring_write(ring, 0x8000); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); return 0; } @@ -3226,13 +3168,6 @@ static void gfx_v8_0_cp_compute_enable(struct amdgpu_device *adev, bool enable) udelay(50); } -static int gfx_v8_0_cp_compute_start(struct amdgpu_device *adev) -{ - gfx_v8_0_cp_compute_enable(adev, true); - - return 0; -} - static int gfx_v8_0_cp_compute_load_microcode(struct amdgpu_device *adev) { const struct gfx_firmware_header_v1_0 *mec_hdr; @@ -3802,9 +3737,7 @@ static int gfx_v8_0_cp_compute_resume(struct amdgpu_device *adev) WREG32(mmCP_PQ_STATUS, tmp); } - r = gfx_v8_0_cp_compute_start(adev); - if (r) - return r; + gfx_v8_0_cp_compute_enable(adev, true); for (i = 0; i < adev->gfx.num_compute_rings; i++) { struct amdgpu_ring *ring = &adev->gfx.compute_ring[i]; @@ -4016,16 +3949,6 @@ static void gfx_v8_0_print_status(void *handle) RREG32(mmHDP_ADDR_CONFIG)); dev_info(adev->dev, " DMIF_ADDR_CALC=0x%08X\n", RREG32(mmDMIF_ADDR_CALC)); - dev_info(adev->dev, " SDMA0_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA0_REGISTER_OFFSET)); - dev_info(adev->dev, " SDMA1_TILING_CONFIG=0x%08X\n", - RREG32(mmSDMA0_TILING_CONFIG + SDMA1_REGISTER_OFFSET)); - dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); - dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", - RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); dev_info(adev->dev, " CP_MEQ_THRESHOLDS=0x%08X\n", RREG32(mmCP_MEQ_THRESHOLDS)); @@ -4762,49 +4685,11 @@ static void gfx_v8_0_ring_emit_fence_gfx(struct amdgpu_ring *ring, u64 addr, } -/** - * gfx_v8_0_ring_emit_semaphore - emit a semaphore on the CP ring - * - * @ring: amdgpu ring buffer object - * @semaphore: amdgpu semaphore object - * @emit_wait: Is this a sempahore wait? - * - * Emits a semaphore signal/wait packet to the CP ring and prevents the PFP - * from running ahead of semaphore waits. - */ -static bool gfx_v8_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - unsigned sel = emit_wait ? PACKET3_SEM_SEL_WAIT : PACKET3_SEM_SEL_SIGNAL; - - if (ring->adev->asic_type == CHIP_TOPAZ || - ring->adev->asic_type == CHIP_TONGA || - ring->adev->asic_type == CHIP_FIJI) - /* we got a hw semaphore bug in VI TONGA, return false to switch back to sw fence wait */ - return false; - else { - amdgpu_ring_write(ring, PACKET3(PACKET3_MEM_SEMAPHORE, 2)); - amdgpu_ring_write(ring, lower_32_bits(addr)); - amdgpu_ring_write(ring, upper_32_bits(addr)); - amdgpu_ring_write(ring, sel); - } - - if (emit_wait && (ring->type == AMDGPU_RING_TYPE_GFX)) { - /* Prevent the PFP from running ahead of the semaphore wait */ - amdgpu_ring_write(ring, PACKET3(PACKET3_PFP_SYNC_ME, 0)); - amdgpu_ring_write(ring, 0x0); - } - - return true; -} - static void gfx_v8_0_ring_emit_vm_flush(struct amdgpu_ring *ring, unsigned vm_id, uint64_t pd_addr) { int usepfp = (ring->type == AMDGPU_RING_TYPE_GFX); - uint32_t seq = ring->fence_drv.sync_seq[ring->idx]; + uint32_t seq = ring->fence_drv.sync_seq; uint64_t addr = ring->fence_drv.gpu_addr; amdgpu_ring_write(ring, PACKET3(PACKET3_WAIT_REG_MEM, 5)); @@ -5145,13 +5030,13 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_gfx = { .parse_cs = NULL, .emit_ib = gfx_v8_0_ring_emit_ib_gfx, .emit_fence = gfx_v8_0_ring_emit_fence_gfx, - .emit_semaphore = gfx_v8_0_ring_emit_semaphore, .emit_vm_flush = gfx_v8_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v8_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v8_0_ring_emit_hdp_flush, .test_ring = gfx_v8_0_ring_test_ring, .test_ib = gfx_v8_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { @@ -5161,13 +5046,13 @@ static const struct amdgpu_ring_funcs gfx_v8_0_ring_funcs_compute = { .parse_cs = NULL, .emit_ib = gfx_v8_0_ring_emit_ib_compute, .emit_fence = gfx_v8_0_ring_emit_fence_compute, - .emit_semaphore = gfx_v8_0_ring_emit_semaphore, .emit_vm_flush = gfx_v8_0_ring_emit_vm_flush, .emit_gds_switch = gfx_v8_0_ring_emit_gds_switch, .emit_hdp_flush = gfx_v8_0_ring_emit_hdp_flush, .test_ring = gfx_v8_0_ring_test_ring, .test_ib = gfx_v8_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void gfx_v8_0_set_ring_funcs(struct amdgpu_device *adev) @@ -5236,32 +5121,24 @@ static void gfx_v8_0_set_gds_init(struct amdgpu_device *adev) } } -static u32 gfx_v8_0_get_cu_active_bitmap(struct amdgpu_device *adev, - u32 se, u32 sh) +static u32 gfx_v8_0_get_cu_active_bitmap(struct amdgpu_device *adev) { - u32 mask = 0, tmp, tmp1; - int i; - - gfx_v8_0_select_se_sh(adev, se, sh); - tmp = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); - tmp1 = RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + u32 data, mask; - tmp &= 0xffff0000; + data = RREG32(mmCC_GC_SHADER_ARRAY_CONFIG); + data |= RREG32(mmGC_USER_SHADER_ARRAY_CONFIG); - tmp |= tmp1; - tmp >>= 16; + data &= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS_MASK; + data >>= CC_GC_SHADER_ARRAY_CONFIG__INACTIVE_CUS__SHIFT; - for (i = 0; i < adev->gfx.config.max_cu_per_sh; i ++) { - mask <<= 1; - mask |= 1; - } + mask = gfx_v8_0_create_bitmask(adev->gfx.config.max_backends_per_se / + adev->gfx.config.max_sh_per_se); - return (~tmp) & mask; + return (~data) & mask; } int gfx_v8_0_get_cu_info(struct amdgpu_device *adev, - struct amdgpu_cu_info *cu_info) + struct amdgpu_cu_info *cu_info) { int i, j, k, counter, active_cu_number = 0; u32 mask, bitmap, ao_bitmap, ao_cu_mask = 0; @@ -5275,10 +5152,11 @@ int gfx_v8_0_get_cu_info(struct amdgpu_device *adev, mask = 1; ao_bitmap = 0; counter = 0; - bitmap = gfx_v8_0_get_cu_active_bitmap(adev, i, j); + gfx_v8_0_select_se_sh(adev, i, j); + bitmap = gfx_v8_0_get_cu_active_bitmap(adev); cu_info->bitmap[i][j] = bitmap; - for (k = 0; k < adev->gfx.config.max_cu_per_sh; k ++) { + for (k = 0; k < 16; k ++) { if (bitmap & mask) { if (counter < 2) ao_bitmap |= mask; @@ -5290,9 +5168,11 @@ int gfx_v8_0_get_cu_info(struct amdgpu_device *adev, ao_cu_mask |= (ao_bitmap << (i * 16 + j * 8)); } } + gfx_v8_0_select_se_sh(adev, 0xffffffff, 0xffffffff); + mutex_unlock(&adev->grbm_idx_mutex); cu_info->number = active_cu_number; cu_info->ao_cu_mask = ao_cu_mask; - mutex_unlock(&adev->grbm_idx_mutex); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c index 8aa2991ab379..68ee66b38e5c 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v7_0.c @@ -694,7 +694,8 @@ static int gmc_v7_0_vm_init(struct amdgpu_device *adev) * amdgpu graphics/compute will use VMIDs 1-7 * amdkfd will use VMIDs 8-15 */ - adev->vm_manager.nvm = AMDGPU_NUM_OF_VMIDS; + adev->vm_manager.num_ids = AMDGPU_NUM_OF_VMIDS; + amdgpu_vm_manager_init(adev); /* base offset of vram pages */ if (adev->flags & AMD_IS_APU) { @@ -926,10 +927,6 @@ static int gmc_v7_0_sw_init(void *handle) int dma_bits; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_gem_init(adev); - if (r) - return r; - r = amdgpu_irq_add_id(adev, 146, &adev->mc.vm_fault); if (r) return r; @@ -1010,7 +1007,7 @@ static int gmc_v7_0_sw_fini(void *handle) adev->vm_manager.enabled = false; } gmc_v7_0_gart_fini(adev); - amdgpu_gem_fini(adev); + amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c index 3efd45546241..757803ae7c4a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v8_0.c @@ -252,6 +252,12 @@ static int gmc_v8_0_mc_load_microcode(struct amdgpu_device *adev) if (!adev->mc.fw) return -EINVAL; + /* Skip MC ucode loading on SR-IOV capable boards. + * vbios does this for us in asic_init in that case. + */ + if (adev->virtualization.supports_sr_iov) + return 0; + hdr = (const struct mc_firmware_header_v1_0 *)adev->mc.fw->data; amdgpu_ucode_print_mc_hdr(&hdr->header); @@ -774,7 +780,8 @@ static int gmc_v8_0_vm_init(struct amdgpu_device *adev) * amdgpu graphics/compute will use VMIDs 1-7 * amdkfd will use VMIDs 8-15 */ - adev->vm_manager.nvm = AMDGPU_NUM_OF_VMIDS; + adev->vm_manager.num_ids = AMDGPU_NUM_OF_VMIDS; + amdgpu_vm_manager_init(adev); /* base offset of vram pages */ if (adev->flags & AMD_IS_APU) { @@ -880,10 +887,6 @@ static int gmc_v8_0_sw_init(void *handle) int dma_bits; struct amdgpu_device *adev = (struct amdgpu_device *)handle; - r = amdgpu_gem_init(adev); - if (r) - return r; - r = amdgpu_irq_add_id(adev, 146, &adev->mc.vm_fault); if (r) return r; @@ -964,7 +967,7 @@ static int gmc_v8_0_sw_fini(void *handle) adev->vm_manager.enabled = false; } gmc_v8_0_gart_fini(adev); - amdgpu_gem_fini(adev); + amdgpu_gem_force_release(adev); amdgpu_bo_fini(adev); return 0; diff --git a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c index 090486c18249..52ee08193295 100644 --- a/drivers/gpu/drm/amd/amdgpu/iceland_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/iceland_smc.c @@ -279,6 +279,12 @@ static int iceland_smu_upload_firmware_image(struct amdgpu_device *adev) if (!adev->pm.fw) return -EINVAL; + /* Skip SMC ucode loading on SR-IOV capable boards. + * vbios does this for us in asic_init in that case. + */ + if (adev->virtualization.supports_sr_iov) + return 0; + hdr = (const struct smc_firmware_header_v1_0 *)adev->pm.fw->data; amdgpu_ucode_print_smc_hdr(&hdr->header); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index 2cf50180cc51..29ec986dd6fc 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -335,31 +335,6 @@ static void sdma_v2_4_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se } /** - * sdma_v2_4_ring_emit_semaphore - emit a semaphore on the dma ring - * - * @ring: amdgpu_ring structure holding ring information - * @semaphore: amdgpu semaphore object - * @emit_wait: wait or signal semaphore - * - * Add a DMA semaphore packet to the ring wait on or signal - * other rings (VI). - */ -static bool sdma_v2_4_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - u64 addr = semaphore->gpu_addr; - u32 sig = emit_wait ? 0 : 1; - - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SEM) | - SDMA_PKT_SEMAPHORE_HEADER_SIGNAL(sig)); - amdgpu_ring_write(ring, lower_32_bits(addr) & 0xfffffff8); - amdgpu_ring_write(ring, upper_32_bits(addr)); - - return true; -} - -/** * sdma_v2_4_gfx_stop - stop the gfx async dma engines * * @adev: amdgpu_device pointer @@ -459,6 +434,9 @@ static int sdma_v2_4_gfx_resume(struct amdgpu_device *adev) vi_srbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); + WREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i], + adev->gfx.config.gb_addr_config & 0x70); + WREG32(mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); /* Set ring buffer size in dwords */ @@ -636,7 +614,7 @@ static int sdma_v2_4_ring_test_ring(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ring_lock(ring, 5); + r = amdgpu_ring_alloc(ring, 5); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); amdgpu_wb_free(adev, index); @@ -649,7 +627,7 @@ static int sdma_v2_4_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, upper_32_bits(gpu_addr)); amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = le32_to_cpu(adev->wb.wb[index]); @@ -699,7 +677,7 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err0; @@ -716,9 +694,8 @@ static int sdma_v2_4_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[7] = SDMA_PKT_HEADER_OP(SDMA_OP_NOP); ib.length_dw = 8; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err1; @@ -797,7 +774,7 @@ static void sdma_v2_4_vm_copy_pte(struct amdgpu_ib *ib, * Update PTEs by writing them manually using sDMA (CIK). */ static void sdma_v2_4_vm_write_pte(struct amdgpu_ib *ib, - uint64_t pe, + const dma_addr_t *pages_addr, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags) { @@ -816,14 +793,7 @@ static void sdma_v2_4_vm_write_pte(struct amdgpu_ib *ib, ib->ptr[ib->length_dw++] = upper_32_bits(pe); ib->ptr[ib->length_dw++] = ndw; for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & AMDGPU_PTE_SYSTEM) { - value = amdgpu_vm_map_gart(ib->ring->adev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & AMDGPU_PTE_VALID) { - value = addr; - } else { - value = 0; - } + value = amdgpu_vm_map_gart(pages_addr, addr); addr += incr; value |= flags; ib->ptr[ib->length_dw++] = value; @@ -881,14 +851,14 @@ static void sdma_v2_4_vm_set_pte_pde(struct amdgpu_ib *ib, } /** - * sdma_v2_4_vm_pad_ib - pad the IB to the required number of dw + * sdma_v2_4_ring_pad_ib - pad the IB to the required number of dw * * @ib: indirect buffer to fill with padding * */ -static void sdma_v2_4_vm_pad_ib(struct amdgpu_ib *ib) +static void sdma_v2_4_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib) { - struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ib->ring); + struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ring); u32 pad_count; int i; @@ -1111,6 +1081,8 @@ static void sdma_v2_4_print_status(void *handle) i, RREG32(mmSDMA0_GFX_RB_BASE + sdma_offsets[i])); dev_info(adev->dev, " SDMA%d_GFX_RB_BASE_HI=0x%08X\n", i, RREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i])); + dev_info(adev->dev, " SDMA%d_TILING_CONFIG=0x%08X\n", + i, RREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i])); mutex_lock(&adev->srbm_mutex); for (j = 0; j < 16; j++) { vi_srbm_select(adev, 0, 0, 0, j); @@ -1302,12 +1274,12 @@ static const struct amdgpu_ring_funcs sdma_v2_4_ring_funcs = { .parse_cs = NULL, .emit_ib = sdma_v2_4_ring_emit_ib, .emit_fence = sdma_v2_4_ring_emit_fence, - .emit_semaphore = sdma_v2_4_ring_emit_semaphore, .emit_vm_flush = sdma_v2_4_ring_emit_vm_flush, .emit_hdp_flush = sdma_v2_4_ring_emit_hdp_flush, .test_ring = sdma_v2_4_ring_test_ring, .test_ib = sdma_v2_4_ring_test_ib, .insert_nop = sdma_v2_4_ring_insert_nop, + .pad_ib = sdma_v2_4_ring_pad_ib, }; static void sdma_v2_4_set_ring_funcs(struct amdgpu_device *adev) @@ -1405,14 +1377,18 @@ static const struct amdgpu_vm_pte_funcs sdma_v2_4_vm_pte_funcs = { .copy_pte = sdma_v2_4_vm_copy_pte, .write_pte = sdma_v2_4_vm_write_pte, .set_pte_pde = sdma_v2_4_vm_set_pte_pde, - .pad_ib = sdma_v2_4_vm_pad_ib, }; static void sdma_v2_4_set_vm_pte_funcs(struct amdgpu_device *adev) { + unsigned i; + if (adev->vm_manager.vm_pte_funcs == NULL) { adev->vm_manager.vm_pte_funcs = &sdma_v2_4_vm_pte_funcs; - adev->vm_manager.vm_pte_funcs_ring = &adev->sdma.instance[0].ring; - adev->vm_manager.vm_pte_funcs_ring->is_pte_ring = true; + for (i = 0; i < adev->sdma.num_instances; i++) + adev->vm_manager.vm_pte_rings[i] = + &adev->sdma.instance[i].ring; + + adev->vm_manager.vm_pte_num_rings = adev->sdma.num_instances; } } diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c index ad54c46751b0..6f064d7076e6 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v3_0.c @@ -444,32 +444,6 @@ static void sdma_v3_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 se amdgpu_ring_write(ring, SDMA_PKT_TRAP_INT_CONTEXT_INT_CONTEXT(0)); } - -/** - * sdma_v3_0_ring_emit_semaphore - emit a semaphore on the dma ring - * - * @ring: amdgpu_ring structure holding ring information - * @semaphore: amdgpu semaphore object - * @emit_wait: wait or signal semaphore - * - * Add a DMA semaphore packet to the ring wait on or signal - * other rings (VI). - */ -static bool sdma_v3_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - u64 addr = semaphore->gpu_addr; - u32 sig = emit_wait ? 0 : 1; - - amdgpu_ring_write(ring, SDMA_PKT_HEADER_OP(SDMA_OP_SEM) | - SDMA_PKT_SEMAPHORE_HEADER_SIGNAL(sig)); - amdgpu_ring_write(ring, lower_32_bits(addr) & 0xfffffff8); - amdgpu_ring_write(ring, upper_32_bits(addr)); - - return true; -} - /** * sdma_v3_0_gfx_stop - stop the gfx async dma engines * @@ -596,6 +570,9 @@ static int sdma_v3_0_gfx_resume(struct amdgpu_device *adev) vi_srbm_select(adev, 0, 0, 0, 0); mutex_unlock(&adev->srbm_mutex); + WREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i], + adev->gfx.config.gb_addr_config & 0x70); + WREG32(mmSDMA0_SEM_WAIT_FAIL_TIMER_CNTL + sdma_offsets[i], 0); /* Set ring buffer size in dwords */ @@ -788,7 +765,7 @@ static int sdma_v3_0_ring_test_ring(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); - r = amdgpu_ring_lock(ring, 5); + r = amdgpu_ring_alloc(ring, 5); if (r) { DRM_ERROR("amdgpu: dma failed to lock ring %d (%d).\n", ring->idx, r); amdgpu_wb_free(adev, index); @@ -801,7 +778,7 @@ static int sdma_v3_0_ring_test_ring(struct amdgpu_ring *ring) amdgpu_ring_write(ring, upper_32_bits(gpu_addr)); amdgpu_ring_write(ring, SDMA_PKT_WRITE_UNTILED_DW_3_COUNT(1)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = le32_to_cpu(adev->wb.wb[index]); @@ -851,7 +828,7 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring) tmp = 0xCAFEDEAD; adev->wb.wb[index] = cpu_to_le32(tmp); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(ring, NULL, 256, &ib); + r = amdgpu_ib_get(adev, NULL, 256, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%d).\n", r); goto err0; @@ -868,9 +845,8 @@ static int sdma_v3_0_ring_test_ib(struct amdgpu_ring *ring) ib.ptr[7] = SDMA_PKT_NOP_HEADER_OP(SDMA_OP_NOP); ib.length_dw = 8; - r = amdgpu_sched_ib_submit_kernel_helper(adev, ring, &ib, 1, NULL, - AMDGPU_FENCE_OWNER_UNDEFINED, - &f); + r = amdgpu_ib_schedule(ring, 1, &ib, AMDGPU_FENCE_OWNER_UNDEFINED, + NULL, &f); if (r) goto err1; @@ -948,7 +924,7 @@ static void sdma_v3_0_vm_copy_pte(struct amdgpu_ib *ib, * Update PTEs by writing them manually using sDMA (CIK). */ static void sdma_v3_0_vm_write_pte(struct amdgpu_ib *ib, - uint64_t pe, + const dma_addr_t *pages_addr, uint64_t pe, uint64_t addr, unsigned count, uint32_t incr, uint32_t flags) { @@ -967,14 +943,7 @@ static void sdma_v3_0_vm_write_pte(struct amdgpu_ib *ib, ib->ptr[ib->length_dw++] = upper_32_bits(pe); ib->ptr[ib->length_dw++] = ndw; for (; ndw > 0; ndw -= 2, --count, pe += 8) { - if (flags & AMDGPU_PTE_SYSTEM) { - value = amdgpu_vm_map_gart(ib->ring->adev, addr); - value &= 0xFFFFFFFFFFFFF000ULL; - } else if (flags & AMDGPU_PTE_VALID) { - value = addr; - } else { - value = 0; - } + value = amdgpu_vm_map_gart(pages_addr, addr); addr += incr; value |= flags; ib->ptr[ib->length_dw++] = value; @@ -1032,14 +1001,14 @@ static void sdma_v3_0_vm_set_pte_pde(struct amdgpu_ib *ib, } /** - * sdma_v3_0_vm_pad_ib - pad the IB to the required number of dw + * sdma_v3_0_ring_pad_ib - pad the IB to the required number of dw * * @ib: indirect buffer to fill with padding * */ -static void sdma_v3_0_vm_pad_ib(struct amdgpu_ib *ib) +static void sdma_v3_0_ring_pad_ib(struct amdgpu_ring *ring, struct amdgpu_ib *ib) { - struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ib->ring); + struct amdgpu_sdma_instance *sdma = amdgpu_get_sdma_instance(ring); u32 pad_count; int i; @@ -1275,6 +1244,8 @@ static void sdma_v3_0_print_status(void *handle) i, RREG32(mmSDMA0_GFX_RB_BASE_HI + sdma_offsets[i])); dev_info(adev->dev, " SDMA%d_GFX_DOORBELL=0x%08X\n", i, RREG32(mmSDMA0_GFX_DOORBELL + sdma_offsets[i])); + dev_info(adev->dev, " SDMA%d_TILING_CONFIG=0x%08X\n", + i, RREG32(mmSDMA0_TILING_CONFIG + sdma_offsets[i])); mutex_lock(&adev->srbm_mutex); for (j = 0; j < 16; j++) { vi_srbm_select(adev, 0, 0, 0, j); @@ -1570,12 +1541,12 @@ static const struct amdgpu_ring_funcs sdma_v3_0_ring_funcs = { .parse_cs = NULL, .emit_ib = sdma_v3_0_ring_emit_ib, .emit_fence = sdma_v3_0_ring_emit_fence, - .emit_semaphore = sdma_v3_0_ring_emit_semaphore, .emit_vm_flush = sdma_v3_0_ring_emit_vm_flush, .emit_hdp_flush = sdma_v3_0_ring_emit_hdp_flush, .test_ring = sdma_v3_0_ring_test_ring, .test_ib = sdma_v3_0_ring_test_ib, .insert_nop = sdma_v3_0_ring_insert_nop, + .pad_ib = sdma_v3_0_ring_pad_ib, }; static void sdma_v3_0_set_ring_funcs(struct amdgpu_device *adev) @@ -1673,14 +1644,18 @@ static const struct amdgpu_vm_pte_funcs sdma_v3_0_vm_pte_funcs = { .copy_pte = sdma_v3_0_vm_copy_pte, .write_pte = sdma_v3_0_vm_write_pte, .set_pte_pde = sdma_v3_0_vm_set_pte_pde, - .pad_ib = sdma_v3_0_vm_pad_ib, }; static void sdma_v3_0_set_vm_pte_funcs(struct amdgpu_device *adev) { + unsigned i; + if (adev->vm_manager.vm_pte_funcs == NULL) { adev->vm_manager.vm_pte_funcs = &sdma_v3_0_vm_pte_funcs; - adev->vm_manager.vm_pte_funcs_ring = &adev->sdma.instance[0].ring; - adev->vm_manager.vm_pte_funcs_ring->is_pte_ring = true; + for (i = 0; i < adev->sdma.num_instances; i++) + adev->vm_manager.vm_pte_rings[i] = + &adev->sdma.instance[i].ring; + + adev->vm_manager.vm_pte_num_rings = adev->sdma.num_instances; } } diff --git a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c index 361c49a82323..083893dd68c0 100644 --- a/drivers/gpu/drm/amd/amdgpu/tonga_smc.c +++ b/drivers/gpu/drm/amd/amdgpu/tonga_smc.c @@ -272,6 +272,12 @@ static int tonga_smu_upload_firmware_image(struct amdgpu_device *adev) if (!adev->pm.fw) return -EINVAL; + /* Skip SMC ucode loading on SR-IOV capable boards. + * vbios does this for us in asic_init in that case. + */ + if (adev->virtualization.supports_sr_iov) + return 0; + hdr = (const struct smc_firmware_header_v1_0 *)adev->pm.fw->data; amdgpu_ucode_print_smc_hdr(&hdr->header); diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c index 5e9f73af83a8..70ed73fa5156 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v4_2.c @@ -164,7 +164,7 @@ static int uvd_v4_2_hw_init(void *handle) goto done; } - r = amdgpu_ring_lock(ring, 10); + r = amdgpu_ring_alloc(ring, 10); if (r) { DRM_ERROR("amdgpu: ring failed to lock UVD ring (%d).\n", r); goto done; @@ -189,7 +189,7 @@ static int uvd_v4_2_hw_init(void *handle) amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CNTL, 0)); amdgpu_ring_write(ring, 3); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); done: /* lower clocks again */ @@ -439,33 +439,6 @@ static void uvd_v4_2_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** - * uvd_v4_2_ring_emit_semaphore - emit semaphore command - * - * @ring: amdgpu_ring pointer - * @semaphore: semaphore to emit commands for - * @emit_wait: true if we should emit a wait command - * - * Emit a semaphore command (either wait or signal) to the UVD ring. - */ -static bool uvd_v4_2_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_LOW, 0)); - amdgpu_ring_write(ring, (addr >> 3) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_HIGH, 0)); - amdgpu_ring_write(ring, (addr >> 23) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CMD, 0)); - amdgpu_ring_write(ring, 0x80 | (emit_wait ? 1 : 0)); - - return true; -} - -/** * uvd_v4_2_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -480,7 +453,7 @@ static int uvd_v4_2_ring_test_ring(struct amdgpu_ring *ring) int r; WREG32(mmUVD_CONTEXT_ID, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); @@ -488,7 +461,7 @@ static int uvd_v4_2_ring_test_ring(struct amdgpu_ring *ring) } amdgpu_ring_write(ring, PACKET0(mmUVD_CONTEXT_ID, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(mmUVD_CONTEXT_ID); if (tmp == 0xDEADBEEF) @@ -549,7 +522,7 @@ static int uvd_v4_2_ring_test_ib(struct amdgpu_ring *ring) goto error; } - r = amdgpu_uvd_get_destroy_msg(ring, 1, &fence); + r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); if (r) { DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); goto error; @@ -603,6 +576,10 @@ static void uvd_v4_2_mc_resume(struct amdgpu_device *adev) addr = (adev->uvd.gpu_addr >> 32) & 0xFF; WREG32(mmUVD_LMI_EXT40_ADDR, addr | (0x9 << 16) | (0x1 << 31)); + WREG32(mmUVD_UDEC_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + uvd_v4_2_init_cg(adev); } @@ -804,6 +781,13 @@ static void uvd_v4_2_print_status(void *handle) RREG32(mmUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL)); dev_info(adev->dev, " UVD_CONTEXT_ID=0x%08X\n", RREG32(mmUVD_CONTEXT_ID)); + dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); + } static int uvd_v4_2_set_interrupt_state(struct amdgpu_device *adev, @@ -882,10 +866,10 @@ static const struct amdgpu_ring_funcs uvd_v4_2_ring_funcs = { .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v4_2_ring_emit_ib, .emit_fence = uvd_v4_2_ring_emit_fence, - .emit_semaphore = uvd_v4_2_ring_emit_semaphore, .test_ring = uvd_v4_2_ring_test_ring, .test_ib = uvd_v4_2_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void uvd_v4_2_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c index 38864f562981..578ffb62fdb2 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v5_0.c @@ -160,7 +160,7 @@ static int uvd_v5_0_hw_init(void *handle) goto done; } - r = amdgpu_ring_lock(ring, 10); + r = amdgpu_ring_alloc(ring, 10); if (r) { DRM_ERROR("amdgpu: ring failed to lock UVD ring (%d).\n", r); goto done; @@ -185,7 +185,7 @@ static int uvd_v5_0_hw_init(void *handle) amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CNTL, 0)); amdgpu_ring_write(ring, 3); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); done: /* lower clocks again */ @@ -279,6 +279,10 @@ static void uvd_v5_0_mc_resume(struct amdgpu_device *adev) size = AMDGPU_UVD_HEAP_SIZE; WREG32(mmUVD_VCPU_CACHE_OFFSET2, offset >> 3); WREG32(mmUVD_VCPU_CACHE_SIZE2, size); + + WREG32(mmUVD_UDEC_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, adev->gfx.config.gb_addr_config); } /** @@ -483,33 +487,6 @@ static void uvd_v5_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** - * uvd_v5_0_ring_emit_semaphore - emit semaphore command - * - * @ring: amdgpu_ring pointer - * @semaphore: semaphore to emit commands for - * @emit_wait: true if we should emit a wait command - * - * Emit a semaphore command (either wait or signal) to the UVD ring. - */ -static bool uvd_v5_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_LOW, 0)); - amdgpu_ring_write(ring, (addr >> 3) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_HIGH, 0)); - amdgpu_ring_write(ring, (addr >> 23) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CMD, 0)); - amdgpu_ring_write(ring, 0x80 | (emit_wait ? 1 : 0)); - - return true; -} - -/** * uvd_v5_0_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -524,7 +501,7 @@ static int uvd_v5_0_ring_test_ring(struct amdgpu_ring *ring) int r; WREG32(mmUVD_CONTEXT_ID, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); @@ -532,7 +509,7 @@ static int uvd_v5_0_ring_test_ring(struct amdgpu_ring *ring) } amdgpu_ring_write(ring, PACKET0(mmUVD_CONTEXT_ID, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(mmUVD_CONTEXT_ID); if (tmp == 0xDEADBEEF) @@ -595,7 +572,7 @@ static int uvd_v5_0_ring_test_ib(struct amdgpu_ring *ring) goto error; } - r = amdgpu_uvd_get_destroy_msg(ring, 1, &fence); + r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); if (r) { DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); goto error; @@ -751,6 +728,12 @@ static void uvd_v5_0_print_status(void *handle) RREG32(mmUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL)); dev_info(adev->dev, " UVD_CONTEXT_ID=0x%08X\n", RREG32(mmUVD_CONTEXT_ID)); + dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); } static int uvd_v5_0_set_interrupt_state(struct amdgpu_device *adev, @@ -821,10 +804,10 @@ static const struct amdgpu_ring_funcs uvd_v5_0_ring_funcs = { .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v5_0_ring_emit_ib, .emit_fence = uvd_v5_0_ring_emit_fence, - .emit_semaphore = uvd_v5_0_ring_emit_semaphore, .test_ring = uvd_v5_0_ring_test_ring, .test_ib = uvd_v5_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void uvd_v5_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c index 3d5913926436..d4da1f04378c 100644 --- a/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c +++ b/drivers/gpu/drm/amd/amdgpu/uvd_v6_0.c @@ -157,7 +157,7 @@ static int uvd_v6_0_hw_init(void *handle) goto done; } - r = amdgpu_ring_lock(ring, 10); + r = amdgpu_ring_alloc(ring, 10); if (r) { DRM_ERROR("amdgpu: ring failed to lock UVD ring (%d).\n", r); goto done; @@ -182,7 +182,7 @@ static int uvd_v6_0_hw_init(void *handle) amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CNTL, 0)); amdgpu_ring_write(ring, 3); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); done: if (!r) @@ -277,6 +277,10 @@ static void uvd_v6_0_mc_resume(struct amdgpu_device *adev) size = AMDGPU_UVD_HEAP_SIZE; WREG32(mmUVD_VCPU_CACHE_OFFSET2, offset >> 3); WREG32(mmUVD_VCPU_CACHE_SIZE2, size); + + WREG32(mmUVD_UDEC_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DB_ADDR_CONFIG, adev->gfx.config.gb_addr_config); + WREG32(mmUVD_UDEC_DBW_ADDR_CONFIG, adev->gfx.config.gb_addr_config); } static void cz_set_uvd_clock_gating_branches(struct amdgpu_device *adev, @@ -722,33 +726,6 @@ static void uvd_v6_0_ring_emit_fence(struct amdgpu_ring *ring, u64 addr, u64 seq } /** - * uvd_v6_0_ring_emit_semaphore - emit semaphore command - * - * @ring: amdgpu_ring pointer - * @semaphore: semaphore to emit commands for - * @emit_wait: true if we should emit a wait command - * - * Emit a semaphore command (either wait or signal) to the UVD ring. - */ -static bool uvd_v6_0_ring_emit_semaphore(struct amdgpu_ring *ring, - struct amdgpu_semaphore *semaphore, - bool emit_wait) -{ - uint64_t addr = semaphore->gpu_addr; - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_LOW, 0)); - amdgpu_ring_write(ring, (addr >> 3) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_ADDR_HIGH, 0)); - amdgpu_ring_write(ring, (addr >> 23) & 0x000FFFFF); - - amdgpu_ring_write(ring, PACKET0(mmUVD_SEMA_CMD, 0)); - amdgpu_ring_write(ring, 0x80 | (emit_wait ? 1 : 0)); - - return true; -} - -/** * uvd_v6_0_ring_test_ring - register write test * * @ring: amdgpu_ring pointer @@ -763,7 +740,7 @@ static int uvd_v6_0_ring_test_ring(struct amdgpu_ring *ring) int r; WREG32(mmUVD_CONTEXT_ID, 0xCAFEDEAD); - r = amdgpu_ring_lock(ring, 3); + r = amdgpu_ring_alloc(ring, 3); if (r) { DRM_ERROR("amdgpu: cp failed to lock ring %d (%d).\n", ring->idx, r); @@ -771,7 +748,7 @@ static int uvd_v6_0_ring_test_ring(struct amdgpu_ring *ring) } amdgpu_ring_write(ring, PACKET0(mmUVD_CONTEXT_ID, 0)); amdgpu_ring_write(ring, 0xDEADBEEF); - amdgpu_ring_unlock_commit(ring); + amdgpu_ring_commit(ring); for (i = 0; i < adev->usec_timeout; i++) { tmp = RREG32(mmUVD_CONTEXT_ID); if (tmp == 0xDEADBEEF) @@ -827,7 +804,7 @@ static int uvd_v6_0_ring_test_ib(struct amdgpu_ring *ring) goto error; } - r = amdgpu_uvd_get_destroy_msg(ring, 1, &fence); + r = amdgpu_uvd_get_destroy_msg(ring, 1, true, &fence); if (r) { DRM_ERROR("amdgpu: failed to get destroy ib (%d).\n", r); goto error; @@ -974,6 +951,12 @@ static void uvd_v6_0_print_status(void *handle) RREG32(mmUVD_SEMA_SIGNAL_INCOMPLETE_TIMEOUT_CNTL)); dev_info(adev->dev, " UVD_CONTEXT_ID=0x%08X\n", RREG32(mmUVD_CONTEXT_ID)); + dev_info(adev->dev, " UVD_UDEC_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DB_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DB_ADDR_CONFIG)); + dev_info(adev->dev, " UVD_UDEC_DBW_ADDR_CONFIG=0x%08X\n", + RREG32(mmUVD_UDEC_DBW_ADDR_CONFIG)); } static int uvd_v6_0_set_interrupt_state(struct amdgpu_device *adev, @@ -1062,10 +1045,10 @@ static const struct amdgpu_ring_funcs uvd_v6_0_ring_funcs = { .parse_cs = amdgpu_uvd_ring_parse_cs, .emit_ib = uvd_v6_0_ring_emit_ib, .emit_fence = uvd_v6_0_ring_emit_fence, - .emit_semaphore = uvd_v6_0_ring_emit_semaphore, .test_ring = uvd_v6_0_ring_test_ring, .test_ib = uvd_v6_0_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void uvd_v6_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c index 52ac7a8f1e58..9c804f436974 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v2_0.c @@ -639,10 +639,10 @@ static const struct amdgpu_ring_funcs vce_v2_0_ring_funcs = { .parse_cs = amdgpu_vce_ring_parse_cs, .emit_ib = amdgpu_vce_ring_emit_ib, .emit_fence = amdgpu_vce_ring_emit_fence, - .emit_semaphore = amdgpu_vce_ring_emit_semaphore, .test_ring = amdgpu_vce_ring_test_ring, .test_ib = amdgpu_vce_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void vce_v2_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c index e99af81e4aec..8f8d479061f8 100644 --- a/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c +++ b/drivers/gpu/drm/amd/amdgpu/vce_v3_0.c @@ -759,10 +759,10 @@ static const struct amdgpu_ring_funcs vce_v3_0_ring_funcs = { .parse_cs = amdgpu_vce_ring_parse_cs, .emit_ib = amdgpu_vce_ring_emit_ib, .emit_fence = amdgpu_vce_ring_emit_fence, - .emit_semaphore = amdgpu_vce_ring_emit_semaphore, .test_ring = amdgpu_vce_ring_test_ring, .test_ib = amdgpu_vce_ring_test_ib, .insert_nop = amdgpu_ring_insert_nop, + .pad_ib = amdgpu_ring_generic_pad_ib, }; static void vce_v3_0_set_ring_funcs(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/amdgpu/vi.c b/drivers/gpu/drm/amd/amdgpu/vi.c index 89f5a1ff6f43..125003517544 100644 --- a/drivers/gpu/drm/amd/amdgpu/vi.c +++ b/drivers/gpu/drm/amd/amdgpu/vi.c @@ -74,6 +74,9 @@ #include "uvd_v6_0.h" #include "vce_v3_0.h" #include "amdgpu_powerplay.h" +#if defined(CONFIG_DRM_AMD_ACP) +#include "amdgpu_acp.h" +#endif /* * Indirect registers accessor @@ -571,374 +574,12 @@ static int vi_read_register(struct amdgpu_device *adev, u32 se_num, return -EINVAL; } -static void vi_print_gpu_status_regs(struct amdgpu_device *adev) -{ - dev_info(adev->dev, " GRBM_STATUS=0x%08X\n", - RREG32(mmGRBM_STATUS)); - dev_info(adev->dev, " GRBM_STATUS2=0x%08X\n", - RREG32(mmGRBM_STATUS2)); - dev_info(adev->dev, " GRBM_STATUS_SE0=0x%08X\n", - RREG32(mmGRBM_STATUS_SE0)); - dev_info(adev->dev, " GRBM_STATUS_SE1=0x%08X\n", - RREG32(mmGRBM_STATUS_SE1)); - dev_info(adev->dev, " GRBM_STATUS_SE2=0x%08X\n", - RREG32(mmGRBM_STATUS_SE2)); - dev_info(adev->dev, " GRBM_STATUS_SE3=0x%08X\n", - RREG32(mmGRBM_STATUS_SE3)); - dev_info(adev->dev, " SRBM_STATUS=0x%08X\n", - RREG32(mmSRBM_STATUS)); - dev_info(adev->dev, " SRBM_STATUS2=0x%08X\n", - RREG32(mmSRBM_STATUS2)); - dev_info(adev->dev, " SDMA0_STATUS_REG = 0x%08X\n", - RREG32(mmSDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET)); - if (adev->sdma.num_instances > 1) { - dev_info(adev->dev, " SDMA1_STATUS_REG = 0x%08X\n", - RREG32(mmSDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET)); - } - dev_info(adev->dev, " CP_STAT = 0x%08x\n", RREG32(mmCP_STAT)); - dev_info(adev->dev, " CP_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT1)); - dev_info(adev->dev, " CP_STALLED_STAT2 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT2)); - dev_info(adev->dev, " CP_STALLED_STAT3 = 0x%08x\n", - RREG32(mmCP_STALLED_STAT3)); - dev_info(adev->dev, " CP_CPF_BUSY_STAT = 0x%08x\n", - RREG32(mmCP_CPF_BUSY_STAT)); - dev_info(adev->dev, " CP_CPF_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_CPF_STALLED_STAT1)); - dev_info(adev->dev, " CP_CPF_STATUS = 0x%08x\n", RREG32(mmCP_CPF_STATUS)); - dev_info(adev->dev, " CP_CPC_BUSY_STAT = 0x%08x\n", RREG32(mmCP_CPC_BUSY_STAT)); - dev_info(adev->dev, " CP_CPC_STALLED_STAT1 = 0x%08x\n", - RREG32(mmCP_CPC_STALLED_STAT1)); - dev_info(adev->dev, " CP_CPC_STATUS = 0x%08x\n", RREG32(mmCP_CPC_STATUS)); -} - -/** - * vi_gpu_check_soft_reset - check which blocks are busy - * - * @adev: amdgpu_device pointer - * - * Check which blocks are busy and return the relevant reset - * mask to be used by vi_gpu_soft_reset(). - * Returns a mask of the blocks to be reset. - */ -u32 vi_gpu_check_soft_reset(struct amdgpu_device *adev) -{ - u32 reset_mask = 0; - u32 tmp; - - /* GRBM_STATUS */ - tmp = RREG32(mmGRBM_STATUS); - if (tmp & (GRBM_STATUS__PA_BUSY_MASK | GRBM_STATUS__SC_BUSY_MASK | - GRBM_STATUS__BCI_BUSY_MASK | GRBM_STATUS__SX_BUSY_MASK | - GRBM_STATUS__TA_BUSY_MASK | GRBM_STATUS__VGT_BUSY_MASK | - GRBM_STATUS__DB_BUSY_MASK | GRBM_STATUS__CB_BUSY_MASK | - GRBM_STATUS__GDS_BUSY_MASK | GRBM_STATUS__SPI_BUSY_MASK | - GRBM_STATUS__IA_BUSY_MASK | GRBM_STATUS__IA_BUSY_NO_DMA_MASK)) - reset_mask |= AMDGPU_RESET_GFX; - - if (tmp & (GRBM_STATUS__CP_BUSY_MASK | GRBM_STATUS__CP_COHERENCY_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_CP; - - /* GRBM_STATUS2 */ - tmp = RREG32(mmGRBM_STATUS2); - if (tmp & GRBM_STATUS2__RLC_BUSY_MASK) - reset_mask |= AMDGPU_RESET_RLC; - - if (tmp & (GRBM_STATUS2__CPF_BUSY_MASK | - GRBM_STATUS2__CPC_BUSY_MASK | - GRBM_STATUS2__CPG_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_CP; - - /* SRBM_STATUS2 */ - tmp = RREG32(mmSRBM_STATUS2); - if (tmp & SRBM_STATUS2__SDMA_BUSY_MASK) - reset_mask |= AMDGPU_RESET_DMA; - - if (tmp & SRBM_STATUS2__SDMA1_BUSY_MASK) - reset_mask |= AMDGPU_RESET_DMA1; - - /* SRBM_STATUS */ - tmp = RREG32(mmSRBM_STATUS); - - if (tmp & SRBM_STATUS__IH_BUSY_MASK) - reset_mask |= AMDGPU_RESET_IH; - - if (tmp & SRBM_STATUS__SEM_BUSY_MASK) - reset_mask |= AMDGPU_RESET_SEM; - - if (tmp & SRBM_STATUS__GRBM_RQ_PENDING_MASK) - reset_mask |= AMDGPU_RESET_GRBM; - - if (adev->asic_type != CHIP_TOPAZ) { - if (tmp & (SRBM_STATUS__UVD_RQ_PENDING_MASK | - SRBM_STATUS__UVD_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_UVD; - } - - if (tmp & SRBM_STATUS__VMC_BUSY_MASK) - reset_mask |= AMDGPU_RESET_VMC; - - if (tmp & (SRBM_STATUS__MCB_BUSY_MASK | SRBM_STATUS__MCB_NON_DISPLAY_BUSY_MASK | - SRBM_STATUS__MCC_BUSY_MASK | SRBM_STATUS__MCD_BUSY_MASK)) - reset_mask |= AMDGPU_RESET_MC; - - /* SDMA0_STATUS_REG */ - tmp = RREG32(mmSDMA0_STATUS_REG + SDMA0_REGISTER_OFFSET); - if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK)) - reset_mask |= AMDGPU_RESET_DMA; - - /* SDMA1_STATUS_REG */ - if (adev->sdma.num_instances > 1) { - tmp = RREG32(mmSDMA0_STATUS_REG + SDMA1_REGISTER_OFFSET); - if (!(tmp & SDMA0_STATUS_REG__IDLE_MASK)) - reset_mask |= AMDGPU_RESET_DMA1; - } -#if 0 - /* VCE_STATUS */ - if (adev->asic_type != CHIP_TOPAZ) { - tmp = RREG32(mmVCE_STATUS); - if (tmp & VCE_STATUS__VCPU_REPORT_RB0_BUSY_MASK) - reset_mask |= AMDGPU_RESET_VCE; - if (tmp & VCE_STATUS__VCPU_REPORT_RB1_BUSY_MASK) - reset_mask |= AMDGPU_RESET_VCE1; - - } - - if (adev->asic_type != CHIP_TOPAZ) { - if (amdgpu_display_is_display_hung(adev)) - reset_mask |= AMDGPU_RESET_DISPLAY; - } -#endif - - /* Skip MC reset as it's mostly likely not hung, just busy */ - if (reset_mask & AMDGPU_RESET_MC) { - DRM_DEBUG("MC busy: 0x%08X, clearing.\n", reset_mask); - reset_mask &= ~AMDGPU_RESET_MC; - } - - return reset_mask; -} - -/** - * vi_gpu_soft_reset - soft reset GPU - * - * @adev: amdgpu_device pointer - * @reset_mask: mask of which blocks to reset - * - * Soft reset the blocks specified in @reset_mask. - */ -static void vi_gpu_soft_reset(struct amdgpu_device *adev, u32 reset_mask) -{ - struct amdgpu_mode_mc_save save; - u32 grbm_soft_reset = 0, srbm_soft_reset = 0; - u32 tmp; - - if (reset_mask == 0) - return; - - dev_info(adev->dev, "GPU softreset: 0x%08X\n", reset_mask); - - vi_print_gpu_status_regs(adev); - dev_info(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_ADDR 0x%08X\n", - RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_ADDR)); - dev_info(adev->dev, " VM_CONTEXT1_PROTECTION_FAULT_STATUS 0x%08X\n", - RREG32(mmVM_CONTEXT1_PROTECTION_FAULT_STATUS)); - - /* disable CG/PG */ - - /* stop the rlc */ - //XXX - //gfx_v8_0_rlc_stop(adev); - - /* Disable GFX parsing/prefetching */ - tmp = RREG32(mmCP_ME_CNTL); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, 1); - WREG32(mmCP_ME_CNTL, tmp); - - /* Disable MEC parsing/prefetching */ - tmp = RREG32(mmCP_MEC_CNTL); - tmp = REG_SET_FIELD(tmp, CP_MEC_CNTL, MEC_ME1_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_MEC_CNTL, MEC_ME2_HALT, 1); - WREG32(mmCP_MEC_CNTL, tmp); - - if (reset_mask & AMDGPU_RESET_DMA) { - /* sdma0 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); - tmp = REG_SET_FIELD(tmp, SDMA0_F32_CNTL, HALT, 1); - WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); - } - if (reset_mask & AMDGPU_RESET_DMA1) { - /* sdma1 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); - tmp = REG_SET_FIELD(tmp, SDMA0_F32_CNTL, HALT, 1); - WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); - } - - gmc_v8_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { - dev_warn(adev->dev, "Wait for MC idle timedout !\n"); - } - - if (reset_mask & (AMDGPU_RESET_GFX | AMDGPU_RESET_COMPUTE | AMDGPU_RESET_CP)) { - grbm_soft_reset = - REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP, 1); - grbm_soft_reset = - REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_GFX, 1); - } - - if (reset_mask & AMDGPU_RESET_CP) { - grbm_soft_reset = - REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_CP, 1); - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_GRBM, 1); - } - - if (reset_mask & AMDGPU_RESET_DMA) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_SDMA, 1); - - if (reset_mask & AMDGPU_RESET_DMA1) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_SDMA1, 1); - - if (reset_mask & AMDGPU_RESET_DISPLAY) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_DC, 1); - - if (reset_mask & AMDGPU_RESET_RLC) - grbm_soft_reset = - REG_SET_FIELD(grbm_soft_reset, GRBM_SOFT_RESET, SOFT_RESET_RLC, 1); - - if (reset_mask & AMDGPU_RESET_SEM) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_SEM, 1); - - if (reset_mask & AMDGPU_RESET_IH) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_IH, 1); - - if (reset_mask & AMDGPU_RESET_GRBM) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_GRBM, 1); - - if (reset_mask & AMDGPU_RESET_VMC) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VMC, 1); - - if (reset_mask & AMDGPU_RESET_UVD) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_UVD, 1); - - if (reset_mask & AMDGPU_RESET_VCE) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE0, 1); - - if (reset_mask & AMDGPU_RESET_VCE) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_VCE1, 1); - - if (!(adev->flags & AMD_IS_APU)) { - if (reset_mask & AMDGPU_RESET_MC) - srbm_soft_reset = - REG_SET_FIELD(srbm_soft_reset, SRBM_SOFT_RESET, SOFT_RESET_MC, 1); - } - - if (grbm_soft_reset) { - tmp = RREG32(mmGRBM_SOFT_RESET); - tmp |= grbm_soft_reset; - dev_info(adev->dev, "GRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(mmGRBM_SOFT_RESET, tmp); - tmp = RREG32(mmGRBM_SOFT_RESET); - - udelay(50); - - tmp &= ~grbm_soft_reset; - WREG32(mmGRBM_SOFT_RESET, tmp); - tmp = RREG32(mmGRBM_SOFT_RESET); - } - - if (srbm_soft_reset) { - tmp = RREG32(mmSRBM_SOFT_RESET); - tmp |= srbm_soft_reset; - dev_info(adev->dev, "SRBM_SOFT_RESET=0x%08X\n", tmp); - WREG32(mmSRBM_SOFT_RESET, tmp); - tmp = RREG32(mmSRBM_SOFT_RESET); - - udelay(50); - - tmp &= ~srbm_soft_reset; - WREG32(mmSRBM_SOFT_RESET, tmp); - tmp = RREG32(mmSRBM_SOFT_RESET); - } - - /* Wait a little for things to settle down */ - udelay(50); - - gmc_v8_0_mc_resume(adev, &save); - udelay(50); - - vi_print_gpu_status_regs(adev); -} - static void vi_gpu_pci_config_reset(struct amdgpu_device *adev) { - struct amdgpu_mode_mc_save save; - u32 tmp, i; + u32 i; dev_info(adev->dev, "GPU pci config reset\n"); - /* disable dpm? */ - - /* disable cg/pg */ - - /* Disable GFX parsing/prefetching */ - tmp = RREG32(mmCP_ME_CNTL); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, ME_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, PFP_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_ME_CNTL, CE_HALT, 1); - WREG32(mmCP_ME_CNTL, tmp); - - /* Disable MEC parsing/prefetching */ - tmp = RREG32(mmCP_MEC_CNTL); - tmp = REG_SET_FIELD(tmp, CP_MEC_CNTL, MEC_ME1_HALT, 1); - tmp = REG_SET_FIELD(tmp, CP_MEC_CNTL, MEC_ME2_HALT, 1); - WREG32(mmCP_MEC_CNTL, tmp); - - /* Disable GFX parsing/prefetching */ - WREG32(mmCP_ME_CNTL, CP_ME_CNTL__ME_HALT_MASK | - CP_ME_CNTL__PFP_HALT_MASK | CP_ME_CNTL__CE_HALT_MASK); - - /* Disable MEC parsing/prefetching */ - WREG32(mmCP_MEC_CNTL, - CP_MEC_CNTL__MEC_ME1_HALT_MASK | CP_MEC_CNTL__MEC_ME2_HALT_MASK); - - /* sdma0 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET); - tmp = REG_SET_FIELD(tmp, SDMA0_F32_CNTL, HALT, 1); - WREG32(mmSDMA0_F32_CNTL + SDMA0_REGISTER_OFFSET, tmp); - - /* sdma1 */ - tmp = RREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET); - tmp = REG_SET_FIELD(tmp, SDMA0_F32_CNTL, HALT, 1); - WREG32(mmSDMA0_F32_CNTL + SDMA1_REGISTER_OFFSET, tmp); - - /* XXX other engines? */ - - /* halt the rlc, disable cp internal ints */ - //XXX - //gfx_v8_0_rlc_stop(adev); - - udelay(50); - - /* disable mem access */ - gmc_v8_0_mc_stop(adev, &save); - if (amdgpu_asic_wait_for_mc_idle(adev)) { - dev_warn(adev->dev, "Wait for MC idle timed out !\n"); - } - /* disable BM */ pci_clear_master(adev->pdev); /* reset */ @@ -978,26 +619,11 @@ static void vi_set_bios_scratch_engine_hung(struct amdgpu_device *adev, bool hun */ static int vi_asic_reset(struct amdgpu_device *adev) { - u32 reset_mask; - - reset_mask = vi_gpu_check_soft_reset(adev); - - if (reset_mask) - vi_set_bios_scratch_engine_hung(adev, true); - - /* try soft reset */ - vi_gpu_soft_reset(adev, reset_mask); - - reset_mask = vi_gpu_check_soft_reset(adev); + vi_set_bios_scratch_engine_hung(adev, true); - /* try pci config reset */ - if (reset_mask && amdgpu_hard_reset) - vi_gpu_pci_config_reset(adev); + vi_gpu_pci_config_reset(adev); - reset_mask = vi_gpu_check_soft_reset(adev); - - if (!reset_mask) - vi_set_bios_scratch_engine_hung(adev, false); + vi_set_bios_scratch_engine_hung(adev, false); return 0; } @@ -1347,6 +973,15 @@ static const struct amdgpu_ip_block_version cz_ip_blocks[] = .rev = 0, .funcs = &vce_v3_0_ip_funcs, }, +#if defined(CONFIG_DRM_AMD_ACP) + { + .type = AMD_IP_BLOCK_TYPE_ACP, + .major = 2, + .minor = 2, + .rev = 0, + .funcs = &acp_ip_funcs, + }, +#endif }; int vi_set_ip_blocks(struct amdgpu_device *adev) diff --git a/drivers/gpu/drm/amd/include/amd_shared.h b/drivers/gpu/drm/amd/include/amd_shared.h index 1195d06f55bc..15ff8b2c26e7 100644 --- a/drivers/gpu/drm/amd/include/amd_shared.h +++ b/drivers/gpu/drm/amd/include/amd_shared.h @@ -73,6 +73,7 @@ enum amd_ip_block_type { AMD_IP_BLOCK_TYPE_SDMA, AMD_IP_BLOCK_TYPE_UVD, AMD_IP_BLOCK_TYPE_VCE, + AMD_IP_BLOCK_TYPE_ACP, }; enum amd_clockgating_state { diff --git a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_d.h b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_d.h index dc52ea0df4b4..d3ccf5a86de0 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_d.h +++ b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_d.h @@ -1379,6 +1379,7 @@ #define mmDC_GPIO_PAD_STRENGTH_1 0x1978 #define mmDC_GPIO_PAD_STRENGTH_2 0x1979 #define mmPHY_AUX_CNTL 0x197f +#define mmDC_GPIO_I2CPAD_MASK 0x1974 #define mmDC_GPIO_I2CPAD_A 0x1975 #define mmDC_GPIO_I2CPAD_EN 0x1976 #define mmDC_GPIO_I2CPAD_Y 0x1977 diff --git a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_enum.h b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_enum.h new file mode 100644 index 000000000000..6bea30ef3df5 --- /dev/null +++ b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_enum.h @@ -0,0 +1,1117 @@ +/* + * DCE_8_0 Register documentation + * + * Copyright (C) 2016 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef DCE_8_0_ENUM_H +#define DCE_8_0_ENUM_H + +typedef enum SurfaceEndian { + ENDIAN_NONE = 0x0, + ENDIAN_8IN16 = 0x1, + ENDIAN_8IN32 = 0x2, + ENDIAN_8IN64 = 0x3, +} SurfaceEndian; +typedef enum ArrayMode { + ARRAY_LINEAR_GENERAL = 0x0, + ARRAY_LINEAR_ALIGNED = 0x1, + ARRAY_1D_TILED_THIN1 = 0x2, + ARRAY_1D_TILED_THICK = 0x3, + ARRAY_2D_TILED_THIN1 = 0x4, + ARRAY_PRT_TILED_THIN1 = 0x5, + ARRAY_PRT_2D_TILED_THIN1 = 0x6, + ARRAY_2D_TILED_THICK = 0x7, + ARRAY_2D_TILED_XTHICK = 0x8, + ARRAY_PRT_TILED_THICK = 0x9, + ARRAY_PRT_2D_TILED_THICK = 0xa, + ARRAY_PRT_3D_TILED_THIN1 = 0xb, + ARRAY_3D_TILED_THIN1 = 0xc, + ARRAY_3D_TILED_THICK = 0xd, + ARRAY_3D_TILED_XTHICK = 0xe, + ARRAY_PRT_3D_TILED_THICK = 0xf, +} ArrayMode; +typedef enum PipeTiling { + CONFIG_1_PIPE = 0x0, + CONFIG_2_PIPE = 0x1, + CONFIG_4_PIPE = 0x2, + CONFIG_8_PIPE = 0x3, +} PipeTiling; +typedef enum BankTiling { + CONFIG_4_BANK = 0x0, + CONFIG_8_BANK = 0x1, +} BankTiling; +typedef enum GroupInterleave { + CONFIG_256B_GROUP = 0x0, + CONFIG_512B_GROUP = 0x1, +} GroupInterleave; +typedef enum RowTiling { + CONFIG_1KB_ROW = 0x0, + CONFIG_2KB_ROW = 0x1, + CONFIG_4KB_ROW = 0x2, + CONFIG_8KB_ROW = 0x3, + CONFIG_1KB_ROW_OPT = 0x4, + CONFIG_2KB_ROW_OPT = 0x5, + CONFIG_4KB_ROW_OPT = 0x6, + CONFIG_8KB_ROW_OPT = 0x7, +} RowTiling; +typedef enum BankSwapBytes { + CONFIG_128B_SWAPS = 0x0, + CONFIG_256B_SWAPS = 0x1, + CONFIG_512B_SWAPS = 0x2, + CONFIG_1KB_SWAPS = 0x3, +} BankSwapBytes; +typedef enum SampleSplitBytes { + CONFIG_1KB_SPLIT = 0x0, + CONFIG_2KB_SPLIT = 0x1, + CONFIG_4KB_SPLIT = 0x2, + CONFIG_8KB_SPLIT = 0x3, +} SampleSplitBytes; +typedef enum NumPipes { + ADDR_CONFIG_1_PIPE = 0x0, + ADDR_CONFIG_2_PIPE = 0x1, + ADDR_CONFIG_4_PIPE = 0x2, + ADDR_CONFIG_8_PIPE = 0x3, +} NumPipes; +typedef enum PipeInterleaveSize { + ADDR_CONFIG_PIPE_INTERLEAVE_256B = 0x0, + ADDR_CONFIG_PIPE_INTERLEAVE_512B = 0x1, +} PipeInterleaveSize; +typedef enum BankInterleaveSize { + ADDR_CONFIG_BANK_INTERLEAVE_1 = 0x0, + ADDR_CONFIG_BANK_INTERLEAVE_2 = 0x1, + ADDR_CONFIG_BANK_INTERLEAVE_4 = 0x2, + ADDR_CONFIG_BANK_INTERLEAVE_8 = 0x3, +} BankInterleaveSize; +typedef enum NumShaderEngines { + ADDR_CONFIG_1_SHADER_ENGINE = 0x0, + ADDR_CONFIG_2_SHADER_ENGINE = 0x1, +} NumShaderEngines; +typedef enum ShaderEngineTileSize { + ADDR_CONFIG_SE_TILE_16 = 0x0, + ADDR_CONFIG_SE_TILE_32 = 0x1, +} ShaderEngineTileSize; +typedef enum NumGPUs { + ADDR_CONFIG_1_GPU = 0x0, + ADDR_CONFIG_2_GPU = 0x1, + ADDR_CONFIG_4_GPU = 0x2, +} NumGPUs; +typedef enum MultiGPUTileSize { + ADDR_CONFIG_GPU_TILE_16 = 0x0, + ADDR_CONFIG_GPU_TILE_32 = 0x1, + ADDR_CONFIG_GPU_TILE_64 = 0x2, + ADDR_CONFIG_GPU_TILE_128 = 0x3, +} MultiGPUTileSize; +typedef enum RowSize { + ADDR_CONFIG_1KB_ROW = 0x0, + ADDR_CONFIG_2KB_ROW = 0x1, + ADDR_CONFIG_4KB_ROW = 0x2, +} RowSize; +typedef enum NumLowerPipes { + ADDR_CONFIG_1_LOWER_PIPES = 0x0, + ADDR_CONFIG_2_LOWER_PIPES = 0x1, +} NumLowerPipes; +typedef enum DebugBlockId { + DBG_CLIENT_BLKID_RESERVED = 0x0, + DBG_CLIENT_BLKID_dbg = 0x1, + DBG_CLIENT_BLKID_uvdu_0 = 0x2, + DBG_CLIENT_BLKID_uvdu_1 = 0x3, + DBG_CLIENT_BLKID_uvdu_2 = 0x4, + DBG_CLIENT_BLKID_uvdu_3 = 0x5, + DBG_CLIENT_BLKID_uvdu_4 = 0x6, + DBG_CLIENT_BLKID_uvdu_5 = 0x7, + DBG_CLIENT_BLKID_uvdu_6 = 0x8, + DBG_CLIENT_BLKID_uvdm_0 = 0x9, + DBG_CLIENT_BLKID_uvdm_1 = 0xa, + DBG_CLIENT_BLKID_uvdm_2 = 0xb, + DBG_CLIENT_BLKID_uvdm_3 = 0xc, + DBG_CLIENT_BLKID_vcea_0 = 0xd, + DBG_CLIENT_BLKID_vcea_1 = 0xe, + DBG_CLIENT_BLKID_vcea_2 = 0xf, + DBG_CLIENT_BLKID_vcea_3 = 0x10, + DBG_CLIENT_BLKID_vcea_4 = 0x11, + DBG_CLIENT_BLKID_vcea_5 = 0x12, + DBG_CLIENT_BLKID_vcea_6 = 0x13, + DBG_CLIENT_BLKID_vceb_0 = 0x14, + DBG_CLIENT_BLKID_vceb_1 = 0x15, + DBG_CLIENT_BLKID_vceb_2 = 0x16, + DBG_CLIENT_BLKID_dco = 0x17, + DBG_CLIENT_BLKID_xdma = 0x18, + DBG_CLIENT_BLKID_smu_0 = 0x19, + DBG_CLIENT_BLKID_smu_1 = 0x1a, + DBG_CLIENT_BLKID_smu_2 = 0x1b, + DBG_CLIENT_BLKID_gck = 0x1c, + DBG_CLIENT_BLKID_tmonw0 = 0x1d, + DBG_CLIENT_BLKID_tmonw1 = 0x1e, + DBG_CLIENT_BLKID_grbm = 0x1f, + DBG_CLIENT_BLKID_rlc = 0x20, + DBG_CLIENT_BLKID_ds0 = 0x21, + DBG_CLIENT_BLKID_cpg_0 = 0x22, + DBG_CLIENT_BLKID_cpg_1 = 0x23, + DBG_CLIENT_BLKID_cpc_0 = 0x24, + DBG_CLIENT_BLKID_cpc_1 = 0x25, + DBG_CLIENT_BLKID_cpf = 0x26, + DBG_CLIENT_BLKID_scf0 = 0x27, + DBG_CLIENT_BLKID_scf1 = 0x28, + DBG_CLIENT_BLKID_scf2 = 0x29, + DBG_CLIENT_BLKID_scf3 = 0x2a, + DBG_CLIENT_BLKID_pc0 = 0x2b, + DBG_CLIENT_BLKID_pc1 = 0x2c, + DBG_CLIENT_BLKID_pc2 = 0x2d, + DBG_CLIENT_BLKID_pc3 = 0x2e, + DBG_CLIENT_BLKID_vgt0 = 0x2f, + DBG_CLIENT_BLKID_vgt1 = 0x30, + DBG_CLIENT_BLKID_vgt2 = 0x31, + DBG_CLIENT_BLKID_vgt3 = 0x32, + DBG_CLIENT_BLKID_sx00 = 0x33, + DBG_CLIENT_BLKID_sx10 = 0x34, + DBG_CLIENT_BLKID_sx20 = 0x35, + DBG_CLIENT_BLKID_sx30 = 0x36, + DBG_CLIENT_BLKID_cb001 = 0x37, + DBG_CLIENT_BLKID_cb200 = 0x38, + DBG_CLIENT_BLKID_cb201 = 0x39, + DBG_CLIENT_BLKID_cbr0 = 0x3a, + DBG_CLIENT_BLKID_cb000 = 0x3b, + DBG_CLIENT_BLKID_cb101 = 0x3c, + DBG_CLIENT_BLKID_cb300 = 0x3d, + DBG_CLIENT_BLKID_cb301 = 0x3e, + DBG_CLIENT_BLKID_cbr1 = 0x3f, + DBG_CLIENT_BLKID_cb100 = 0x40, + DBG_CLIENT_BLKID_ia0 = 0x41, + DBG_CLIENT_BLKID_ia1 = 0x42, + DBG_CLIENT_BLKID_bci0 = 0x43, + DBG_CLIENT_BLKID_bci1 = 0x44, + DBG_CLIENT_BLKID_bci2 = 0x45, + DBG_CLIENT_BLKID_bci3 = 0x46, + DBG_CLIENT_BLKID_pa0 = 0x47, + DBG_CLIENT_BLKID_pa1 = 0x48, + DBG_CLIENT_BLKID_spim0 = 0x49, + DBG_CLIENT_BLKID_spim1 = 0x4a, + DBG_CLIENT_BLKID_spim2 = 0x4b, + DBG_CLIENT_BLKID_spim3 = 0x4c, + DBG_CLIENT_BLKID_sdma = 0x4d, + DBG_CLIENT_BLKID_ih = 0x4e, + DBG_CLIENT_BLKID_sem = 0x4f, + DBG_CLIENT_BLKID_srbm = 0x50, + DBG_CLIENT_BLKID_hdp = 0x51, + DBG_CLIENT_BLKID_acp_0 = 0x52, + DBG_CLIENT_BLKID_acp_1 = 0x53, + DBG_CLIENT_BLKID_sam = 0x54, + DBG_CLIENT_BLKID_mcc0 = 0x55, + DBG_CLIENT_BLKID_mcc1 = 0x56, + DBG_CLIENT_BLKID_mcc2 = 0x57, + DBG_CLIENT_BLKID_mcc3 = 0x58, + DBG_CLIENT_BLKID_mcd0 = 0x59, + DBG_CLIENT_BLKID_mcd1 = 0x5a, + DBG_CLIENT_BLKID_mcd2 = 0x5b, + DBG_CLIENT_BLKID_mcd3 = 0x5c, + DBG_CLIENT_BLKID_mcb = 0x5d, + DBG_CLIENT_BLKID_vmc = 0x5e, + DBG_CLIENT_BLKID_gmcon = 0x5f, + DBG_CLIENT_BLKID_gdc_0 = 0x60, + DBG_CLIENT_BLKID_gdc_1 = 0x61, + DBG_CLIENT_BLKID_gdc_2 = 0x62, + DBG_CLIENT_BLKID_gdc_3 = 0x63, + DBG_CLIENT_BLKID_gdc_4 = 0x64, + DBG_CLIENT_BLKID_gdc_5 = 0x65, + DBG_CLIENT_BLKID_gdc_6 = 0x66, + DBG_CLIENT_BLKID_gdc_7 = 0x67, + DBG_CLIENT_BLKID_gdc_8 = 0x68, + DBG_CLIENT_BLKID_gdc_9 = 0x69, + DBG_CLIENT_BLKID_gdc_10 = 0x6a, + DBG_CLIENT_BLKID_gdc_11 = 0x6b, + DBG_CLIENT_BLKID_gdc_12 = 0x6c, + DBG_CLIENT_BLKID_gdc_13 = 0x6d, + DBG_CLIENT_BLKID_gdc_14 = 0x6e, + DBG_CLIENT_BLKID_gdc_15 = 0x6f, + DBG_CLIENT_BLKID_gdc_16 = 0x70, + DBG_CLIENT_BLKID_gdc_17 = 0x71, + DBG_CLIENT_BLKID_gdc_18 = 0x72, + DBG_CLIENT_BLKID_gdc_19 = 0x73, + DBG_CLIENT_BLKID_gdc_20 = 0x74, + DBG_CLIENT_BLKID_gdc_21 = 0x75, + DBG_CLIENT_BLKID_gdc_22 = 0x76, + DBG_CLIENT_BLKID_wd = 0x77, + DBG_CLIENT_BLKID_sdma_0 = 0x78, + DBG_CLIENT_BLKID_sdma_1 = 0x79, +} DebugBlockId; +typedef enum DebugBlockId_OLD { + DBG_BLOCK_ID_RESERVED = 0x0, + DBG_BLOCK_ID_DBG = 0x1, + DBG_BLOCK_ID_VMC = 0x2, + DBG_BLOCK_ID_PDMA = 0x3, + DBG_BLOCK_ID_CG = 0x4, + DBG_BLOCK_ID_SRBM = 0x5, + DBG_BLOCK_ID_GRBM = 0x6, + DBG_BLOCK_ID_RLC = 0x7, + DBG_BLOCK_ID_CSC = 0x8, + DBG_BLOCK_ID_SEM = 0x9, + DBG_BLOCK_ID_IH = 0xa, + DBG_BLOCK_ID_SC = 0xb, + DBG_BLOCK_ID_SQ = 0xc, + DBG_BLOCK_ID_AVP = 0xd, + DBG_BLOCK_ID_GMCON = 0xe, + DBG_BLOCK_ID_SMU = 0xf, + DBG_BLOCK_ID_DMA0 = 0x10, + DBG_BLOCK_ID_DMA1 = 0x11, + DBG_BLOCK_ID_SPIM = 0x12, + DBG_BLOCK_ID_GDS = 0x13, + DBG_BLOCK_ID_SPIS = 0x14, + DBG_BLOCK_ID_UNUSED0 = 0x15, + DBG_BLOCK_ID_PA0 = 0x16, + DBG_BLOCK_ID_PA1 = 0x17, + DBG_BLOCK_ID_CP0 = 0x18, + DBG_BLOCK_ID_CP1 = 0x19, + DBG_BLOCK_ID_CP2 = 0x1a, + DBG_BLOCK_ID_UNUSED1 = 0x1b, + DBG_BLOCK_ID_UVDU = 0x1c, + DBG_BLOCK_ID_UVDM = 0x1d, + DBG_BLOCK_ID_VCE = 0x1e, + DBG_BLOCK_ID_UNUSED2 = 0x1f, + DBG_BLOCK_ID_VGT0 = 0x20, + DBG_BLOCK_ID_VGT1 = 0x21, + DBG_BLOCK_ID_IA = 0x22, + DBG_BLOCK_ID_UNUSED3 = 0x23, + DBG_BLOCK_ID_SCT0 = 0x24, + DBG_BLOCK_ID_SCT1 = 0x25, + DBG_BLOCK_ID_SPM0 = 0x26, + DBG_BLOCK_ID_SPM1 = 0x27, + DBG_BLOCK_ID_TCAA = 0x28, + DBG_BLOCK_ID_TCAB = 0x29, + DBG_BLOCK_ID_TCCA = 0x2a, + DBG_BLOCK_ID_TCCB = 0x2b, + DBG_BLOCK_ID_MCC0 = 0x2c, + DBG_BLOCK_ID_MCC1 = 0x2d, + DBG_BLOCK_ID_MCC2 = 0x2e, + DBG_BLOCK_ID_MCC3 = 0x2f, + DBG_BLOCK_ID_SX0 = 0x30, + DBG_BLOCK_ID_SX1 = 0x31, + DBG_BLOCK_ID_SX2 = 0x32, + DBG_BLOCK_ID_SX3 = 0x33, + DBG_BLOCK_ID_UNUSED4 = 0x34, + DBG_BLOCK_ID_UNUSED5 = 0x35, + DBG_BLOCK_ID_UNUSED6 = 0x36, + DBG_BLOCK_ID_UNUSED7 = 0x37, + DBG_BLOCK_ID_PC0 = 0x38, + DBG_BLOCK_ID_PC1 = 0x39, + DBG_BLOCK_ID_UNUSED8 = 0x3a, + DBG_BLOCK_ID_UNUSED9 = 0x3b, + DBG_BLOCK_ID_UNUSED10 = 0x3c, + DBG_BLOCK_ID_UNUSED11 = 0x3d, + DBG_BLOCK_ID_MCB = 0x3e, + DBG_BLOCK_ID_UNUSED12 = 0x3f, + DBG_BLOCK_ID_SCB0 = 0x40, + DBG_BLOCK_ID_SCB1 = 0x41, + DBG_BLOCK_ID_UNUSED13 = 0x42, + DBG_BLOCK_ID_UNUSED14 = 0x43, + DBG_BLOCK_ID_SCF0 = 0x44, + DBG_BLOCK_ID_SCF1 = 0x45, + DBG_BLOCK_ID_UNUSED15 = 0x46, + DBG_BLOCK_ID_UNUSED16 = 0x47, + DBG_BLOCK_ID_BCI0 = 0x48, + DBG_BLOCK_ID_BCI1 = 0x49, + DBG_BLOCK_ID_BCI2 = 0x4a, + DBG_BLOCK_ID_BCI3 = 0x4b, + DBG_BLOCK_ID_UNUSED17 = 0x4c, + DBG_BLOCK_ID_UNUSED18 = 0x4d, + DBG_BLOCK_ID_UNUSED19 = 0x4e, + DBG_BLOCK_ID_UNUSED20 = 0x4f, + DBG_BLOCK_ID_CB00 = 0x50, + DBG_BLOCK_ID_CB01 = 0x51, + DBG_BLOCK_ID_CB02 = 0x52, + DBG_BLOCK_ID_CB03 = 0x53, + DBG_BLOCK_ID_CB04 = 0x54, + DBG_BLOCK_ID_UNUSED21 = 0x55, + DBG_BLOCK_ID_UNUSED22 = 0x56, + DBG_BLOCK_ID_UNUSED23 = 0x57, + DBG_BLOCK_ID_CB10 = 0x58, + DBG_BLOCK_ID_CB11 = 0x59, + DBG_BLOCK_ID_CB12 = 0x5a, + DBG_BLOCK_ID_CB13 = 0x5b, + DBG_BLOCK_ID_CB14 = 0x5c, + DBG_BLOCK_ID_UNUSED24 = 0x5d, + DBG_BLOCK_ID_UNUSED25 = 0x5e, + DBG_BLOCK_ID_UNUSED26 = 0x5f, + DBG_BLOCK_ID_TCP0 = 0x60, + DBG_BLOCK_ID_TCP1 = 0x61, + DBG_BLOCK_ID_TCP2 = 0x62, + DBG_BLOCK_ID_TCP3 = 0x63, + DBG_BLOCK_ID_TCP4 = 0x64, + DBG_BLOCK_ID_TCP5 = 0x65, + DBG_BLOCK_ID_TCP6 = 0x66, + DBG_BLOCK_ID_TCP7 = 0x67, + DBG_BLOCK_ID_TCP8 = 0x68, + DBG_BLOCK_ID_TCP9 = 0x69, + DBG_BLOCK_ID_TCP10 = 0x6a, + DBG_BLOCK_ID_TCP11 = 0x6b, + DBG_BLOCK_ID_TCP12 = 0x6c, + DBG_BLOCK_ID_TCP13 = 0x6d, + DBG_BLOCK_ID_TCP14 = 0x6e, + DBG_BLOCK_ID_TCP15 = 0x6f, + DBG_BLOCK_ID_TCP16 = 0x70, + DBG_BLOCK_ID_TCP17 = 0x71, + DBG_BLOCK_ID_TCP18 = 0x72, + DBG_BLOCK_ID_TCP19 = 0x73, + DBG_BLOCK_ID_TCP20 = 0x74, + DBG_BLOCK_ID_TCP21 = 0x75, + DBG_BLOCK_ID_TCP22 = 0x76, + DBG_BLOCK_ID_TCP23 = 0x77, + DBG_BLOCK_ID_TCP_RESERVED0 = 0x78, + DBG_BLOCK_ID_TCP_RESERVED1 = 0x79, + DBG_BLOCK_ID_TCP_RESERVED2 = 0x7a, + DBG_BLOCK_ID_TCP_RESERVED3 = 0x7b, + DBG_BLOCK_ID_TCP_RESERVED4 = 0x7c, + DBG_BLOCK_ID_TCP_RESERVED5 = 0x7d, + DBG_BLOCK_ID_TCP_RESERVED6 = 0x7e, + DBG_BLOCK_ID_TCP_RESERVED7 = 0x7f, + DBG_BLOCK_ID_DB00 = 0x80, + DBG_BLOCK_ID_DB01 = 0x81, + DBG_BLOCK_ID_DB02 = 0x82, + DBG_BLOCK_ID_DB03 = 0x83, + DBG_BLOCK_ID_DB04 = 0x84, + DBG_BLOCK_ID_UNUSED27 = 0x85, + DBG_BLOCK_ID_UNUSED28 = 0x86, + DBG_BLOCK_ID_UNUSED29 = 0x87, + DBG_BLOCK_ID_DB10 = 0x88, + DBG_BLOCK_ID_DB11 = 0x89, + DBG_BLOCK_ID_DB12 = 0x8a, + DBG_BLOCK_ID_DB13 = 0x8b, + DBG_BLOCK_ID_DB14 = 0x8c, + DBG_BLOCK_ID_UNUSED30 = 0x8d, + DBG_BLOCK_ID_UNUSED31 = 0x8e, + DBG_BLOCK_ID_UNUSED32 = 0x8f, + DBG_BLOCK_ID_TCC0 = 0x90, + DBG_BLOCK_ID_TCC1 = 0x91, + DBG_BLOCK_ID_TCC2 = 0x92, + DBG_BLOCK_ID_TCC3 = 0x93, + DBG_BLOCK_ID_TCC4 = 0x94, + DBG_BLOCK_ID_TCC5 = 0x95, + DBG_BLOCK_ID_TCC6 = 0x96, + DBG_BLOCK_ID_TCC7 = 0x97, + DBG_BLOCK_ID_SPS00 = 0x98, + DBG_BLOCK_ID_SPS01 = 0x99, + DBG_BLOCK_ID_SPS02 = 0x9a, + DBG_BLOCK_ID_SPS10 = 0x9b, + DBG_BLOCK_ID_SPS11 = 0x9c, + DBG_BLOCK_ID_SPS12 = 0x9d, + DBG_BLOCK_ID_UNUSED33 = 0x9e, + DBG_BLOCK_ID_UNUSED34 = 0x9f, + DBG_BLOCK_ID_TA00 = 0xa0, + DBG_BLOCK_ID_TA01 = 0xa1, + DBG_BLOCK_ID_TA02 = 0xa2, + DBG_BLOCK_ID_TA03 = 0xa3, + DBG_BLOCK_ID_TA04 = 0xa4, + DBG_BLOCK_ID_TA05 = 0xa5, + DBG_BLOCK_ID_TA06 = 0xa6, + DBG_BLOCK_ID_TA07 = 0xa7, + DBG_BLOCK_ID_TA08 = 0xa8, + DBG_BLOCK_ID_TA09 = 0xa9, + DBG_BLOCK_ID_TA0A = 0xaa, + DBG_BLOCK_ID_TA0B = 0xab, + DBG_BLOCK_ID_UNUSED35 = 0xac, + DBG_BLOCK_ID_UNUSED36 = 0xad, + DBG_BLOCK_ID_UNUSED37 = 0xae, + DBG_BLOCK_ID_UNUSED38 = 0xaf, + DBG_BLOCK_ID_TA10 = 0xb0, + DBG_BLOCK_ID_TA11 = 0xb1, + DBG_BLOCK_ID_TA12 = 0xb2, + DBG_BLOCK_ID_TA13 = 0xb3, + DBG_BLOCK_ID_TA14 = 0xb4, + DBG_BLOCK_ID_TA15 = 0xb5, + DBG_BLOCK_ID_TA16 = 0xb6, + DBG_BLOCK_ID_TA17 = 0xb7, + DBG_BLOCK_ID_TA18 = 0xb8, + DBG_BLOCK_ID_TA19 = 0xb9, + DBG_BLOCK_ID_TA1A = 0xba, + DBG_BLOCK_ID_TA1B = 0xbb, + DBG_BLOCK_ID_UNUSED39 = 0xbc, + DBG_BLOCK_ID_UNUSED40 = 0xbd, + DBG_BLOCK_ID_UNUSED41 = 0xbe, + DBG_BLOCK_ID_UNUSED42 = 0xbf, + DBG_BLOCK_ID_TD00 = 0xc0, + DBG_BLOCK_ID_TD01 = 0xc1, + DBG_BLOCK_ID_TD02 = 0xc2, + DBG_BLOCK_ID_TD03 = 0xc3, + DBG_BLOCK_ID_TD04 = 0xc4, + DBG_BLOCK_ID_TD05 = 0xc5, + DBG_BLOCK_ID_TD06 = 0xc6, + DBG_BLOCK_ID_TD07 = 0xc7, + DBG_BLOCK_ID_TD08 = 0xc8, + DBG_BLOCK_ID_TD09 = 0xc9, + DBG_BLOCK_ID_TD0A = 0xca, + DBG_BLOCK_ID_TD0B = 0xcb, + DBG_BLOCK_ID_UNUSED43 = 0xcc, + DBG_BLOCK_ID_UNUSED44 = 0xcd, + DBG_BLOCK_ID_UNUSED45 = 0xce, + DBG_BLOCK_ID_UNUSED46 = 0xcf, + DBG_BLOCK_ID_TD10 = 0xd0, + DBG_BLOCK_ID_TD11 = 0xd1, + DBG_BLOCK_ID_TD12 = 0xd2, + DBG_BLOCK_ID_TD13 = 0xd3, + DBG_BLOCK_ID_TD14 = 0xd4, + DBG_BLOCK_ID_TD15 = 0xd5, + DBG_BLOCK_ID_TD16 = 0xd6, + DBG_BLOCK_ID_TD17 = 0xd7, + DBG_BLOCK_ID_TD18 = 0xd8, + DBG_BLOCK_ID_TD19 = 0xd9, + DBG_BLOCK_ID_TD1A = 0xda, + DBG_BLOCK_ID_TD1B = 0xdb, + DBG_BLOCK_ID_UNUSED47 = 0xdc, + DBG_BLOCK_ID_UNUSED48 = 0xdd, + DBG_BLOCK_ID_UNUSED49 = 0xde, + DBG_BLOCK_ID_UNUSED50 = 0xdf, + DBG_BLOCK_ID_MCD0 = 0xe0, + DBG_BLOCK_ID_MCD1 = 0xe1, + DBG_BLOCK_ID_MCD2 = 0xe2, + DBG_BLOCK_ID_MCD3 = 0xe3, + DBG_BLOCK_ID_MCD4 = 0xe4, + DBG_BLOCK_ID_MCD5 = 0xe5, + DBG_BLOCK_ID_UNUSED51 = 0xe6, + DBG_BLOCK_ID_UNUSED52 = 0xe7, +} DebugBlockId_OLD; +typedef enum DebugBlockId_BY2 { + DBG_BLOCK_ID_RESERVED_BY2 = 0x0, + DBG_BLOCK_ID_VMC_BY2 = 0x1, + DBG_BLOCK_ID_CG_BY2 = 0x2, + DBG_BLOCK_ID_GRBM_BY2 = 0x3, + DBG_BLOCK_ID_CSC_BY2 = 0x4, + DBG_BLOCK_ID_IH_BY2 = 0x5, + DBG_BLOCK_ID_SQ_BY2 = 0x6, + DBG_BLOCK_ID_GMCON_BY2 = 0x7, + DBG_BLOCK_ID_DMA0_BY2 = 0x8, + DBG_BLOCK_ID_SPIM_BY2 = 0x9, + DBG_BLOCK_ID_SPIS_BY2 = 0xa, + DBG_BLOCK_ID_PA0_BY2 = 0xb, + DBG_BLOCK_ID_CP0_BY2 = 0xc, + DBG_BLOCK_ID_CP2_BY2 = 0xd, + DBG_BLOCK_ID_UVDU_BY2 = 0xe, + DBG_BLOCK_ID_VCE_BY2 = 0xf, + DBG_BLOCK_ID_VGT0_BY2 = 0x10, + DBG_BLOCK_ID_IA_BY2 = 0x11, + DBG_BLOCK_ID_SCT0_BY2 = 0x12, + DBG_BLOCK_ID_SPM0_BY2 = 0x13, + DBG_BLOCK_ID_TCAA_BY2 = 0x14, + DBG_BLOCK_ID_TCCA_BY2 = 0x15, + DBG_BLOCK_ID_MCC0_BY2 = 0x16, + DBG_BLOCK_ID_MCC2_BY2 = 0x17, + DBG_BLOCK_ID_SX0_BY2 = 0x18, + DBG_BLOCK_ID_SX2_BY2 = 0x19, + DBG_BLOCK_ID_UNUSED4_BY2 = 0x1a, + DBG_BLOCK_ID_UNUSED6_BY2 = 0x1b, + DBG_BLOCK_ID_PC0_BY2 = 0x1c, + DBG_BLOCK_ID_UNUSED8_BY2 = 0x1d, + DBG_BLOCK_ID_UNUSED10_BY2 = 0x1e, + DBG_BLOCK_ID_MCB_BY2 = 0x1f, + DBG_BLOCK_ID_SCB0_BY2 = 0x20, + DBG_BLOCK_ID_UNUSED13_BY2 = 0x21, + DBG_BLOCK_ID_SCF0_BY2 = 0x22, + DBG_BLOCK_ID_UNUSED15_BY2 = 0x23, + DBG_BLOCK_ID_BCI0_BY2 = 0x24, + DBG_BLOCK_ID_BCI2_BY2 = 0x25, + DBG_BLOCK_ID_UNUSED17_BY2 = 0x26, + DBG_BLOCK_ID_UNUSED19_BY2 = 0x27, + DBG_BLOCK_ID_CB00_BY2 = 0x28, + DBG_BLOCK_ID_CB02_BY2 = 0x29, + DBG_BLOCK_ID_CB04_BY2 = 0x2a, + DBG_BLOCK_ID_UNUSED22_BY2 = 0x2b, + DBG_BLOCK_ID_CB10_BY2 = 0x2c, + DBG_BLOCK_ID_CB12_BY2 = 0x2d, + DBG_BLOCK_ID_CB14_BY2 = 0x2e, + DBG_BLOCK_ID_UNUSED25_BY2 = 0x2f, + DBG_BLOCK_ID_TCP0_BY2 = 0x30, + DBG_BLOCK_ID_TCP2_BY2 = 0x31, + DBG_BLOCK_ID_TCP4_BY2 = 0x32, + DBG_BLOCK_ID_TCP6_BY2 = 0x33, + DBG_BLOCK_ID_TCP8_BY2 = 0x34, + DBG_BLOCK_ID_TCP10_BY2 = 0x35, + DBG_BLOCK_ID_TCP12_BY2 = 0x36, + DBG_BLOCK_ID_TCP14_BY2 = 0x37, + DBG_BLOCK_ID_TCP16_BY2 = 0x38, + DBG_BLOCK_ID_TCP18_BY2 = 0x39, + DBG_BLOCK_ID_TCP20_BY2 = 0x3a, + DBG_BLOCK_ID_TCP22_BY2 = 0x3b, + DBG_BLOCK_ID_TCP_RESERVED0_BY2 = 0x3c, + DBG_BLOCK_ID_TCP_RESERVED2_BY2 = 0x3d, + DBG_BLOCK_ID_TCP_RESERVED4_BY2 = 0x3e, + DBG_BLOCK_ID_TCP_RESERVED6_BY2 = 0x3f, + DBG_BLOCK_ID_DB00_BY2 = 0x40, + DBG_BLOCK_ID_DB02_BY2 = 0x41, + DBG_BLOCK_ID_DB04_BY2 = 0x42, + DBG_BLOCK_ID_UNUSED28_BY2 = 0x43, + DBG_BLOCK_ID_DB10_BY2 = 0x44, + DBG_BLOCK_ID_DB12_BY2 = 0x45, + DBG_BLOCK_ID_DB14_BY2 = 0x46, + DBG_BLOCK_ID_UNUSED31_BY2 = 0x47, + DBG_BLOCK_ID_TCC0_BY2 = 0x48, + DBG_BLOCK_ID_TCC2_BY2 = 0x49, + DBG_BLOCK_ID_TCC4_BY2 = 0x4a, + DBG_BLOCK_ID_TCC6_BY2 = 0x4b, + DBG_BLOCK_ID_SPS00_BY2 = 0x4c, + DBG_BLOCK_ID_SPS02_BY2 = 0x4d, + DBG_BLOCK_ID_SPS11_BY2 = 0x4e, + DBG_BLOCK_ID_UNUSED33_BY2 = 0x4f, + DBG_BLOCK_ID_TA00_BY2 = 0x50, + DBG_BLOCK_ID_TA02_BY2 = 0x51, + DBG_BLOCK_ID_TA04_BY2 = 0x52, + DBG_BLOCK_ID_TA06_BY2 = 0x53, + DBG_BLOCK_ID_TA08_BY2 = 0x54, + DBG_BLOCK_ID_TA0A_BY2 = 0x55, + DBG_BLOCK_ID_UNUSED35_BY2 = 0x56, + DBG_BLOCK_ID_UNUSED37_BY2 = 0x57, + DBG_BLOCK_ID_TA10_BY2 = 0x58, + DBG_BLOCK_ID_TA12_BY2 = 0x59, + DBG_BLOCK_ID_TA14_BY2 = 0x5a, + DBG_BLOCK_ID_TA16_BY2 = 0x5b, + DBG_BLOCK_ID_TA18_BY2 = 0x5c, + DBG_BLOCK_ID_TA1A_BY2 = 0x5d, + DBG_BLOCK_ID_UNUSED39_BY2 = 0x5e, + DBG_BLOCK_ID_UNUSED41_BY2 = 0x5f, + DBG_BLOCK_ID_TD00_BY2 = 0x60, + DBG_BLOCK_ID_TD02_BY2 = 0x61, + DBG_BLOCK_ID_TD04_BY2 = 0x62, + DBG_BLOCK_ID_TD06_BY2 = 0x63, + DBG_BLOCK_ID_TD08_BY2 = 0x64, + DBG_BLOCK_ID_TD0A_BY2 = 0x65, + DBG_BLOCK_ID_UNUSED43_BY2 = 0x66, + DBG_BLOCK_ID_UNUSED45_BY2 = 0x67, + DBG_BLOCK_ID_TD10_BY2 = 0x68, + DBG_BLOCK_ID_TD12_BY2 = 0x69, + DBG_BLOCK_ID_TD14_BY2 = 0x6a, + DBG_BLOCK_ID_TD16_BY2 = 0x6b, + DBG_BLOCK_ID_TD18_BY2 = 0x6c, + DBG_BLOCK_ID_TD1A_BY2 = 0x6d, + DBG_BLOCK_ID_UNUSED47_BY2 = 0x6e, + DBG_BLOCK_ID_UNUSED49_BY2 = 0x6f, + DBG_BLOCK_ID_MCD0_BY2 = 0x70, + DBG_BLOCK_ID_MCD2_BY2 = 0x71, + DBG_BLOCK_ID_MCD4_BY2 = 0x72, + DBG_BLOCK_ID_UNUSED51_BY2 = 0x73, +} DebugBlockId_BY2; +typedef enum DebugBlockId_BY4 { + DBG_BLOCK_ID_RESERVED_BY4 = 0x0, + DBG_BLOCK_ID_CG_BY4 = 0x1, + DBG_BLOCK_ID_CSC_BY4 = 0x2, + DBG_BLOCK_ID_SQ_BY4 = 0x3, + DBG_BLOCK_ID_DMA0_BY4 = 0x4, + DBG_BLOCK_ID_SPIS_BY4 = 0x5, + DBG_BLOCK_ID_CP0_BY4 = 0x6, + DBG_BLOCK_ID_UVDU_BY4 = 0x7, + DBG_BLOCK_ID_VGT0_BY4 = 0x8, + DBG_BLOCK_ID_SCT0_BY4 = 0x9, + DBG_BLOCK_ID_TCAA_BY4 = 0xa, + DBG_BLOCK_ID_MCC0_BY4 = 0xb, + DBG_BLOCK_ID_SX0_BY4 = 0xc, + DBG_BLOCK_ID_UNUSED4_BY4 = 0xd, + DBG_BLOCK_ID_PC0_BY4 = 0xe, + DBG_BLOCK_ID_UNUSED10_BY4 = 0xf, + DBG_BLOCK_ID_SCB0_BY4 = 0x10, + DBG_BLOCK_ID_SCF0_BY4 = 0x11, + DBG_BLOCK_ID_BCI0_BY4 = 0x12, + DBG_BLOCK_ID_UNUSED17_BY4 = 0x13, + DBG_BLOCK_ID_CB00_BY4 = 0x14, + DBG_BLOCK_ID_CB04_BY4 = 0x15, + DBG_BLOCK_ID_CB10_BY4 = 0x16, + DBG_BLOCK_ID_CB14_BY4 = 0x17, + DBG_BLOCK_ID_TCP0_BY4 = 0x18, + DBG_BLOCK_ID_TCP4_BY4 = 0x19, + DBG_BLOCK_ID_TCP8_BY4 = 0x1a, + DBG_BLOCK_ID_TCP12_BY4 = 0x1b, + DBG_BLOCK_ID_TCP16_BY4 = 0x1c, + DBG_BLOCK_ID_TCP20_BY4 = 0x1d, + DBG_BLOCK_ID_TCP_RESERVED0_BY4 = 0x1e, + DBG_BLOCK_ID_TCP_RESERVED4_BY4 = 0x1f, + DBG_BLOCK_ID_DB_BY4 = 0x20, + DBG_BLOCK_ID_DB04_BY4 = 0x21, + DBG_BLOCK_ID_DB10_BY4 = 0x22, + DBG_BLOCK_ID_DB14_BY4 = 0x23, + DBG_BLOCK_ID_TCC0_BY4 = 0x24, + DBG_BLOCK_ID_TCC4_BY4 = 0x25, + DBG_BLOCK_ID_SPS00_BY4 = 0x26, + DBG_BLOCK_ID_SPS11_BY4 = 0x27, + DBG_BLOCK_ID_TA00_BY4 = 0x28, + DBG_BLOCK_ID_TA04_BY4 = 0x29, + DBG_BLOCK_ID_TA08_BY4 = 0x2a, + DBG_BLOCK_ID_UNUSED35_BY4 = 0x2b, + DBG_BLOCK_ID_TA10_BY4 = 0x2c, + DBG_BLOCK_ID_TA14_BY4 = 0x2d, + DBG_BLOCK_ID_TA18_BY4 = 0x2e, + DBG_BLOCK_ID_UNUSED39_BY4 = 0x2f, + DBG_BLOCK_ID_TD00_BY4 = 0x30, + DBG_BLOCK_ID_TD04_BY4 = 0x31, + DBG_BLOCK_ID_TD08_BY4 = 0x32, + DBG_BLOCK_ID_UNUSED43_BY4 = 0x33, + DBG_BLOCK_ID_TD10_BY4 = 0x34, + DBG_BLOCK_ID_TD14_BY4 = 0x35, + DBG_BLOCK_ID_TD18_BY4 = 0x36, + DBG_BLOCK_ID_UNUSED47_BY4 = 0x37, + DBG_BLOCK_ID_MCD0_BY4 = 0x38, + DBG_BLOCK_ID_MCD4_BY4 = 0x39, +} DebugBlockId_BY4; +typedef enum DebugBlockId_BY8 { + DBG_BLOCK_ID_RESERVED_BY8 = 0x0, + DBG_BLOCK_ID_CSC_BY8 = 0x1, + DBG_BLOCK_ID_DMA0_BY8 = 0x2, + DBG_BLOCK_ID_CP0_BY8 = 0x3, + DBG_BLOCK_ID_VGT0_BY8 = 0x4, + DBG_BLOCK_ID_TCAA_BY8 = 0x5, + DBG_BLOCK_ID_SX0_BY8 = 0x6, + DBG_BLOCK_ID_PC0_BY8 = 0x7, + DBG_BLOCK_ID_SCB0_BY8 = 0x8, + DBG_BLOCK_ID_BCI0_BY8 = 0x9, + DBG_BLOCK_ID_CB00_BY8 = 0xa, + DBG_BLOCK_ID_CB10_BY8 = 0xb, + DBG_BLOCK_ID_TCP0_BY8 = 0xc, + DBG_BLOCK_ID_TCP8_BY8 = 0xd, + DBG_BLOCK_ID_TCP16_BY8 = 0xe, + DBG_BLOCK_ID_TCP_RESERVED0_BY8 = 0xf, + DBG_BLOCK_ID_DB00_BY8 = 0x10, + DBG_BLOCK_ID_DB10_BY8 = 0x11, + DBG_BLOCK_ID_TCC0_BY8 = 0x12, + DBG_BLOCK_ID_SPS00_BY8 = 0x13, + DBG_BLOCK_ID_TA00_BY8 = 0x14, + DBG_BLOCK_ID_TA08_BY8 = 0x15, + DBG_BLOCK_ID_TA10_BY8 = 0x16, + DBG_BLOCK_ID_TA18_BY8 = 0x17, + DBG_BLOCK_ID_TD00_BY8 = 0x18, + DBG_BLOCK_ID_TD08_BY8 = 0x19, + DBG_BLOCK_ID_TD10_BY8 = 0x1a, + DBG_BLOCK_ID_TD18_BY8 = 0x1b, + DBG_BLOCK_ID_MCD0_BY8 = 0x1c, +} DebugBlockId_BY8; +typedef enum DebugBlockId_BY16 { + DBG_BLOCK_ID_RESERVED_BY16 = 0x0, + DBG_BLOCK_ID_DMA0_BY16 = 0x1, + DBG_BLOCK_ID_VGT0_BY16 = 0x2, + DBG_BLOCK_ID_SX0_BY16 = 0x3, + DBG_BLOCK_ID_SCB0_BY16 = 0x4, + DBG_BLOCK_ID_CB00_BY16 = 0x5, + DBG_BLOCK_ID_TCP0_BY16 = 0x6, + DBG_BLOCK_ID_TCP16_BY16 = 0x7, + DBG_BLOCK_ID_DB00_BY16 = 0x8, + DBG_BLOCK_ID_TCC0_BY16 = 0x9, + DBG_BLOCK_ID_TA00_BY16 = 0xa, + DBG_BLOCK_ID_TA10_BY16 = 0xb, + DBG_BLOCK_ID_TD00_BY16 = 0xc, + DBG_BLOCK_ID_TD10_BY16 = 0xd, + DBG_BLOCK_ID_MCD0_BY16 = 0xe, +} DebugBlockId_BY16; +typedef enum CompareRef { + REF_NEVER = 0x0, + REF_LESS = 0x1, + REF_EQUAL = 0x2, + REF_LEQUAL = 0x3, + REF_GREATER = 0x4, + REF_NOTEQUAL = 0x5, + REF_GEQUAL = 0x6, + REF_ALWAYS = 0x7, +} CompareRef; +typedef enum ReadSize { + READ_256_BITS = 0x0, + READ_512_BITS = 0x1, +} ReadSize; +typedef enum DepthFormat { + DEPTH_INVALID = 0x0, + DEPTH_16 = 0x1, + DEPTH_X8_24 = 0x2, + DEPTH_8_24 = 0x3, + DEPTH_X8_24_FLOAT = 0x4, + DEPTH_8_24_FLOAT = 0x5, + DEPTH_32_FLOAT = 0x6, + DEPTH_X24_8_32_FLOAT = 0x7, +} DepthFormat; +typedef enum ZFormat { + Z_INVALID = 0x0, + Z_16 = 0x1, + Z_24 = 0x2, + Z_32_FLOAT = 0x3, +} ZFormat; +typedef enum StencilFormat { + STENCIL_INVALID = 0x0, + STENCIL_8 = 0x1, +} StencilFormat; +typedef enum CmaskMode { + CMASK_CLEAR_NONE = 0x0, + CMASK_CLEAR_ONE = 0x1, + CMASK_CLEAR_ALL = 0x2, + CMASK_ANY_EXPANDED = 0x3, + CMASK_ALPHA0_FRAG1 = 0x4, + CMASK_ALPHA0_FRAG2 = 0x5, + CMASK_ALPHA0_FRAG4 = 0x6, + CMASK_ALPHA0_FRAGS = 0x7, + CMASK_ALPHA1_FRAG1 = 0x8, + CMASK_ALPHA1_FRAG2 = 0x9, + CMASK_ALPHA1_FRAG4 = 0xa, + CMASK_ALPHA1_FRAGS = 0xb, + CMASK_ALPHAX_FRAG1 = 0xc, + CMASK_ALPHAX_FRAG2 = 0xd, + CMASK_ALPHAX_FRAG4 = 0xe, + CMASK_ALPHAX_FRAGS = 0xf, +} CmaskMode; +typedef enum QuadExportFormat { + EXPORT_UNUSED = 0x0, + EXPORT_32_R = 0x1, + EXPORT_32_GR = 0x2, + EXPORT_32_AR = 0x3, + EXPORT_FP16_ABGR = 0x4, + EXPORT_UNSIGNED16_ABGR = 0x5, + EXPORT_SIGNED16_ABGR = 0x6, + EXPORT_32_ABGR = 0x7, +} QuadExportFormat; +typedef enum QuadExportFormatOld { + EXPORT_4P_32BPC_ABGR = 0x0, + EXPORT_4P_16BPC_ABGR = 0x1, + EXPORT_4P_32BPC_GR = 0x2, + EXPORT_4P_32BPC_AR = 0x3, + EXPORT_2P_32BPC_ABGR = 0x4, + EXPORT_8P_32BPC_R = 0x5, +} QuadExportFormatOld; +typedef enum ColorFormat { + COLOR_INVALID = 0x0, + COLOR_8 = 0x1, + COLOR_16 = 0x2, + COLOR_8_8 = 0x3, + COLOR_32 = 0x4, + COLOR_16_16 = 0x5, + COLOR_10_11_11 = 0x6, + COLOR_11_11_10 = 0x7, + COLOR_10_10_10_2 = 0x8, + COLOR_2_10_10_10 = 0x9, + COLOR_8_8_8_8 = 0xa, + COLOR_32_32 = 0xb, + COLOR_16_16_16_16 = 0xc, + COLOR_RESERVED_13 = 0xd, + COLOR_32_32_32_32 = 0xe, + COLOR_RESERVED_15 = 0xf, + COLOR_5_6_5 = 0x10, + COLOR_1_5_5_5 = 0x11, + COLOR_5_5_5_1 = 0x12, + COLOR_4_4_4_4 = 0x13, + COLOR_8_24 = 0x14, + COLOR_24_8 = 0x15, + COLOR_X24_8_32_FLOAT = 0x16, + COLOR_RESERVED_23 = 0x17, +} ColorFormat; +typedef enum SurfaceFormat { + FMT_INVALID = 0x0, + FMT_8 = 0x1, + FMT_16 = 0x2, + FMT_8_8 = 0x3, + FMT_32 = 0x4, + FMT_16_16 = 0x5, + FMT_10_11_11 = 0x6, + FMT_11_11_10 = 0x7, + FMT_10_10_10_2 = 0x8, + FMT_2_10_10_10 = 0x9, + FMT_8_8_8_8 = 0xa, + FMT_32_32 = 0xb, + FMT_16_16_16_16 = 0xc, + FMT_32_32_32 = 0xd, + FMT_32_32_32_32 = 0xe, + FMT_RESERVED_4 = 0xf, + FMT_5_6_5 = 0x10, + FMT_1_5_5_5 = 0x11, + FMT_5_5_5_1 = 0x12, + FMT_4_4_4_4 = 0x13, + FMT_8_24 = 0x14, + FMT_24_8 = 0x15, + FMT_X24_8_32_FLOAT = 0x16, + FMT_RESERVED_33 = 0x17, + FMT_11_11_10_FLOAT = 0x18, + FMT_16_FLOAT = 0x19, + FMT_32_FLOAT = 0x1a, + FMT_16_16_FLOAT = 0x1b, + FMT_8_24_FLOAT = 0x1c, + FMT_24_8_FLOAT = 0x1d, + FMT_32_32_FLOAT = 0x1e, + FMT_10_11_11_FLOAT = 0x1f, + FMT_16_16_16_16_FLOAT = 0x20, + FMT_3_3_2 = 0x21, + FMT_6_5_5 = 0x22, + FMT_32_32_32_32_FLOAT = 0x23, + FMT_RESERVED_36 = 0x24, + FMT_1 = 0x25, + FMT_1_REVERSED = 0x26, + FMT_GB_GR = 0x27, + FMT_BG_RG = 0x28, + FMT_32_AS_8 = 0x29, + FMT_32_AS_8_8 = 0x2a, + FMT_5_9_9_9_SHAREDEXP = 0x2b, + FMT_8_8_8 = 0x2c, + FMT_16_16_16 = 0x2d, + FMT_16_16_16_FLOAT = 0x2e, + FMT_4_4 = 0x2f, + FMT_32_32_32_FLOAT = 0x30, + FMT_BC1 = 0x31, + FMT_BC2 = 0x32, + FMT_BC3 = 0x33, + FMT_BC4 = 0x34, + FMT_BC5 = 0x35, + FMT_BC6 = 0x36, + FMT_BC7 = 0x37, + FMT_32_AS_32_32_32_32 = 0x38, + FMT_APC3 = 0x39, + FMT_APC4 = 0x3a, + FMT_APC5 = 0x3b, + FMT_APC6 = 0x3c, + FMT_APC7 = 0x3d, + FMT_CTX1 = 0x3e, + FMT_RESERVED_63 = 0x3f, +} SurfaceFormat; +typedef enum BUF_DATA_FORMAT { + BUF_DATA_FORMAT_INVALID = 0x0, + BUF_DATA_FORMAT_8 = 0x1, + BUF_DATA_FORMAT_16 = 0x2, + BUF_DATA_FORMAT_8_8 = 0x3, + BUF_DATA_FORMAT_32 = 0x4, + BUF_DATA_FORMAT_16_16 = 0x5, + BUF_DATA_FORMAT_10_11_11 = 0x6, + BUF_DATA_FORMAT_11_11_10 = 0x7, + BUF_DATA_FORMAT_10_10_10_2 = 0x8, + BUF_DATA_FORMAT_2_10_10_10 = 0x9, + BUF_DATA_FORMAT_8_8_8_8 = 0xa, + BUF_DATA_FORMAT_32_32 = 0xb, + BUF_DATA_FORMAT_16_16_16_16 = 0xc, + BUF_DATA_FORMAT_32_32_32 = 0xd, + BUF_DATA_FORMAT_32_32_32_32 = 0xe, + BUF_DATA_FORMAT_RESERVED_15 = 0xf, +} BUF_DATA_FORMAT; +typedef enum IMG_DATA_FORMAT { + IMG_DATA_FORMAT_INVALID = 0x0, + IMG_DATA_FORMAT_8 = 0x1, + IMG_DATA_FORMAT_16 = 0x2, + IMG_DATA_FORMAT_8_8 = 0x3, + IMG_DATA_FORMAT_32 = 0x4, + IMG_DATA_FORMAT_16_16 = 0x5, + IMG_DATA_FORMAT_10_11_11 = 0x6, + IMG_DATA_FORMAT_11_11_10 = 0x7, + IMG_DATA_FORMAT_10_10_10_2 = 0x8, + IMG_DATA_FORMAT_2_10_10_10 = 0x9, + IMG_DATA_FORMAT_8_8_8_8 = 0xa, + IMG_DATA_FORMAT_32_32 = 0xb, + IMG_DATA_FORMAT_16_16_16_16 = 0xc, + IMG_DATA_FORMAT_32_32_32 = 0xd, + IMG_DATA_FORMAT_32_32_32_32 = 0xe, + IMG_DATA_FORMAT_RESERVED_15 = 0xf, + IMG_DATA_FORMAT_5_6_5 = 0x10, + IMG_DATA_FORMAT_1_5_5_5 = 0x11, + IMG_DATA_FORMAT_5_5_5_1 = 0x12, + IMG_DATA_FORMAT_4_4_4_4 = 0x13, + IMG_DATA_FORMAT_8_24 = 0x14, + IMG_DATA_FORMAT_24_8 = 0x15, + IMG_DATA_FORMAT_X24_8_32 = 0x16, + IMG_DATA_FORMAT_RESERVED_23 = 0x17, + IMG_DATA_FORMAT_RESERVED_24 = 0x18, + IMG_DATA_FORMAT_RESERVED_25 = 0x19, + IMG_DATA_FORMAT_RESERVED_26 = 0x1a, + IMG_DATA_FORMAT_RESERVED_27 = 0x1b, + IMG_DATA_FORMAT_RESERVED_28 = 0x1c, + IMG_DATA_FORMAT_RESERVED_29 = 0x1d, + IMG_DATA_FORMAT_RESERVED_30 = 0x1e, + IMG_DATA_FORMAT_RESERVED_31 = 0x1f, + IMG_DATA_FORMAT_GB_GR = 0x20, + IMG_DATA_FORMAT_BG_RG = 0x21, + IMG_DATA_FORMAT_5_9_9_9 = 0x22, + IMG_DATA_FORMAT_BC1 = 0x23, + IMG_DATA_FORMAT_BC2 = 0x24, + IMG_DATA_FORMAT_BC3 = 0x25, + IMG_DATA_FORMAT_BC4 = 0x26, + IMG_DATA_FORMAT_BC5 = 0x27, + IMG_DATA_FORMAT_BC6 = 0x28, + IMG_DATA_FORMAT_BC7 = 0x29, + IMG_DATA_FORMAT_RESERVED_42 = 0x2a, + IMG_DATA_FORMAT_RESERVED_43 = 0x2b, + IMG_DATA_FORMAT_FMASK8_S2_F1 = 0x2c, + IMG_DATA_FORMAT_FMASK8_S4_F1 = 0x2d, + IMG_DATA_FORMAT_FMASK8_S8_F1 = 0x2e, + IMG_DATA_FORMAT_FMASK8_S2_F2 = 0x2f, + IMG_DATA_FORMAT_FMASK8_S4_F2 = 0x30, + IMG_DATA_FORMAT_FMASK8_S4_F4 = 0x31, + IMG_DATA_FORMAT_FMASK16_S16_F1 = 0x32, + IMG_DATA_FORMAT_FMASK16_S8_F2 = 0x33, + IMG_DATA_FORMAT_FMASK32_S16_F2 = 0x34, + IMG_DATA_FORMAT_FMASK32_S8_F4 = 0x35, + IMG_DATA_FORMAT_FMASK32_S8_F8 = 0x36, + IMG_DATA_FORMAT_FMASK64_S16_F4 = 0x37, + IMG_DATA_FORMAT_FMASK64_S16_F8 = 0x38, + IMG_DATA_FORMAT_4_4 = 0x39, + IMG_DATA_FORMAT_6_5_5 = 0x3a, + IMG_DATA_FORMAT_1 = 0x3b, + IMG_DATA_FORMAT_1_REVERSED = 0x3c, + IMG_DATA_FORMAT_32_AS_8 = 0x3d, + IMG_DATA_FORMAT_32_AS_8_8 = 0x3e, + IMG_DATA_FORMAT_32_AS_32_32_32_32 = 0x3f, +} IMG_DATA_FORMAT; +typedef enum BUF_NUM_FORMAT { + BUF_NUM_FORMAT_UNORM = 0x0, + BUF_NUM_FORMAT_SNORM = 0x1, + BUF_NUM_FORMAT_USCALED = 0x2, + BUF_NUM_FORMAT_SSCALED = 0x3, + BUF_NUM_FORMAT_UINT = 0x4, + BUF_NUM_FORMAT_SINT = 0x5, + BUF_NUM_FORMAT_SNORM_OGL = 0x6, + BUF_NUM_FORMAT_FLOAT = 0x7, +} BUF_NUM_FORMAT; +typedef enum IMG_NUM_FORMAT { + IMG_NUM_FORMAT_UNORM = 0x0, + IMG_NUM_FORMAT_SNORM = 0x1, + IMG_NUM_FORMAT_USCALED = 0x2, + IMG_NUM_FORMAT_SSCALED = 0x3, + IMG_NUM_FORMAT_UINT = 0x4, + IMG_NUM_FORMAT_SINT = 0x5, + IMG_NUM_FORMAT_SNORM_OGL = 0x6, + IMG_NUM_FORMAT_FLOAT = 0x7, + IMG_NUM_FORMAT_RESERVED_8 = 0x8, + IMG_NUM_FORMAT_SRGB = 0x9, + IMG_NUM_FORMAT_UBNORM = 0xa, + IMG_NUM_FORMAT_UBNORM_OGL = 0xb, + IMG_NUM_FORMAT_UBINT = 0xc, + IMG_NUM_FORMAT_UBSCALED = 0xd, + IMG_NUM_FORMAT_RESERVED_14 = 0xe, + IMG_NUM_FORMAT_RESERVED_15 = 0xf, +} IMG_NUM_FORMAT; +typedef enum TileType { + ARRAY_COLOR_TILE = 0x0, + ARRAY_DEPTH_TILE = 0x1, +} TileType; +typedef enum NonDispTilingOrder { + ADDR_SURF_MICRO_TILING_DISPLAY = 0x0, + ADDR_SURF_MICRO_TILING_NON_DISPLAY = 0x1, +} NonDispTilingOrder; +typedef enum MicroTileMode { + ADDR_SURF_DISPLAY_MICRO_TILING = 0x0, + ADDR_SURF_THIN_MICRO_TILING = 0x1, + ADDR_SURF_DEPTH_MICRO_TILING = 0x2, + ADDR_SURF_ROTATED_MICRO_TILING = 0x3, + ADDR_SURF_THICK_MICRO_TILING = 0x4, +} MicroTileMode; +typedef enum TileSplit { + ADDR_SURF_TILE_SPLIT_64B = 0x0, + ADDR_SURF_TILE_SPLIT_128B = 0x1, + ADDR_SURF_TILE_SPLIT_256B = 0x2, + ADDR_SURF_TILE_SPLIT_512B = 0x3, + ADDR_SURF_TILE_SPLIT_1KB = 0x4, + ADDR_SURF_TILE_SPLIT_2KB = 0x5, + ADDR_SURF_TILE_SPLIT_4KB = 0x6, +} TileSplit; +typedef enum SampleSplit { + ADDR_SURF_SAMPLE_SPLIT_1 = 0x0, + ADDR_SURF_SAMPLE_SPLIT_2 = 0x1, + ADDR_SURF_SAMPLE_SPLIT_4 = 0x2, + ADDR_SURF_SAMPLE_SPLIT_8 = 0x3, +} SampleSplit; +typedef enum PipeConfig { + ADDR_SURF_P2 = 0x0, + ADDR_SURF_P2_RESERVED0 = 0x1, + ADDR_SURF_P2_RESERVED1 = 0x2, + ADDR_SURF_P2_RESERVED2 = 0x3, + ADDR_SURF_P4_8x16 = 0x4, + ADDR_SURF_P4_16x16 = 0x5, + ADDR_SURF_P4_16x32 = 0x6, + ADDR_SURF_P4_32x32 = 0x7, + ADDR_SURF_P8_16x16_8x16 = 0x8, + ADDR_SURF_P8_16x32_8x16 = 0x9, + ADDR_SURF_P8_32x32_8x16 = 0xa, + ADDR_SURF_P8_16x32_16x16 = 0xb, + ADDR_SURF_P8_32x32_16x16 = 0xc, + ADDR_SURF_P8_32x32_16x32 = 0xd, + ADDR_SURF_P8_32x64_32x32 = 0xe, +} PipeConfig; +typedef enum NumBanks { + ADDR_SURF_2_BANK = 0x0, + ADDR_SURF_4_BANK = 0x1, + ADDR_SURF_8_BANK = 0x2, + ADDR_SURF_16_BANK = 0x3, +} NumBanks; +typedef enum BankWidth { + ADDR_SURF_BANK_WIDTH_1 = 0x0, + ADDR_SURF_BANK_WIDTH_2 = 0x1, + ADDR_SURF_BANK_WIDTH_4 = 0x2, + ADDR_SURF_BANK_WIDTH_8 = 0x3, +} BankWidth; +typedef enum BankHeight { + ADDR_SURF_BANK_HEIGHT_1 = 0x0, + ADDR_SURF_BANK_HEIGHT_2 = 0x1, + ADDR_SURF_BANK_HEIGHT_4 = 0x2, + ADDR_SURF_BANK_HEIGHT_8 = 0x3, +} BankHeight; +typedef enum BankWidthHeight { + ADDR_SURF_BANK_WH_1 = 0x0, + ADDR_SURF_BANK_WH_2 = 0x1, + ADDR_SURF_BANK_WH_4 = 0x2, + ADDR_SURF_BANK_WH_8 = 0x3, +} BankWidthHeight; +typedef enum MacroTileAspect { + ADDR_SURF_MACRO_ASPECT_1 = 0x0, + ADDR_SURF_MACRO_ASPECT_2 = 0x1, + ADDR_SURF_MACRO_ASPECT_4 = 0x2, + ADDR_SURF_MACRO_ASPECT_8 = 0x3, +} MacroTileAspect; +typedef enum TCC_CACHE_POLICIES { + TCC_CACHE_POLICY_LRU = 0x0, + TCC_CACHE_POLICY_STREAM = 0x1, + TCC_CACHE_POLICY_BYPASS = 0x2, +} TCC_CACHE_POLICIES; +typedef enum PERFMON_COUNTER_MODE { + PERFMON_COUNTER_MODE_ACCUM = 0x0, + PERFMON_COUNTER_MODE_ACTIVE_CYCLES = 0x1, + PERFMON_COUNTER_MODE_MAX = 0x2, + PERFMON_COUNTER_MODE_DIRTY = 0x3, + PERFMON_COUNTER_MODE_SAMPLE = 0x4, + PERFMON_COUNTER_MODE_CYCLES_SINCE_FIRST_EVENT = 0x5, + PERFMON_COUNTER_MODE_CYCLES_SINCE_LAST_EVENT = 0x6, + PERFMON_COUNTER_MODE_CYCLES_GE_HI = 0x7, + PERFMON_COUNTER_MODE_CYCLES_EQ_HI = 0x8, + PERFMON_COUNTER_MODE_INACTIVE_CYCLES = 0x9, + PERFMON_COUNTER_MODE_RESERVED = 0xf, +} PERFMON_COUNTER_MODE; +typedef enum PERFMON_SPM_MODE { + PERFMON_SPM_MODE_OFF = 0x0, + PERFMON_SPM_MODE_16BIT_CLAMP = 0x1, + PERFMON_SPM_MODE_16BIT_NO_CLAMP = 0x2, + PERFMON_SPM_MODE_32BIT_CLAMP = 0x3, + PERFMON_SPM_MODE_32BIT_NO_CLAMP = 0x4, + PERFMON_SPM_MODE_RESERVED_5 = 0x5, + PERFMON_SPM_MODE_RESERVED_6 = 0x6, + PERFMON_SPM_MODE_RESERVED_7 = 0x7, + PERFMON_SPM_MODE_TEST_MODE_0 = 0x8, + PERFMON_SPM_MODE_TEST_MODE_1 = 0x9, + PERFMON_SPM_MODE_TEST_MODE_2 = 0xa, +} PERFMON_SPM_MODE; +typedef enum SurfaceTiling { + ARRAY_LINEAR = 0x0, + ARRAY_TILED = 0x1, +} SurfaceTiling; +typedef enum SurfaceArray { + ARRAY_1D = 0x0, + ARRAY_2D = 0x1, + ARRAY_3D = 0x2, + ARRAY_3D_SLICE = 0x3, +} SurfaceArray; +typedef enum ColorArray { + ARRAY_2D_ALT_COLOR = 0x0, + ARRAY_2D_COLOR = 0x1, + ARRAY_3D_SLICE_COLOR = 0x3, +} ColorArray; +typedef enum DepthArray { + ARRAY_2D_ALT_DEPTH = 0x0, + ARRAY_2D_DEPTH = 0x1, +} DepthArray; + +#endif /* DCE_8_0_ENUM_H */ diff --git a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_sh_mask.h index 8a2930734477..c331c9fe7b81 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/dce/dce_8_0_sh_mask.h @@ -4130,6 +4130,18 @@ #define PHY_AUX_CNTL__AUX_PAD_WAKE__SHIFT 0xe #define PHY_AUX_CNTL__AUX_PAD_RXSEL_MASK 0x10000 #define PHY_AUX_CNTL__AUX_PAD_RXSEL__SHIFT 0x10 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_MASK_MASK 0x1 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_MASK__SHIFT 0x0 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_PD_DIS_MASK 0x2 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_PD_DIS__SHIFT 0x1 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_RECV_MASK 0x4 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SCL_RECV__SHIFT 0x2 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_MASK_MASK 0x10 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_MASK__SHIFT 0x4 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_PD_DIS_MASK 0x20 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_PD_DIS__SHIFT 0x5 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_RECV_MASK 0x40 +#define DC_GPIO_I2CPAD_MASK__DC_GPIO_SDA_RECV__SHIFT 0x6 #define DC_GPIO_I2CPAD_A__DC_GPIO_SCL_A_MASK 0x1 #define DC_GPIO_I2CPAD_A__DC_GPIO_SCL_A__SHIFT 0x0 #define DC_GPIO_I2CPAD_A__DC_GPIO_SDA_A_MASK 0x2 diff --git a/drivers/gpu/drm/amd/include/ivsrcid/ivsrcid_vislands30.h b/drivers/gpu/drm/amd/include/ivsrcid/ivsrcid_vislands30.h new file mode 100644 index 000000000000..d21c6b14662f --- /dev/null +++ b/drivers/gpu/drm/amd/include/ivsrcid/ivsrcid_vislands30.h @@ -0,0 +1,102 @@ +/* + * Volcanic Islands IV SRC Register documentation + * + * Copyright (C) 2015 Advanced Micro Devices, Inc. + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included + * in all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS + * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN + * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN + * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. + */ + +#ifndef _IVSRCID_VISLANDS30_H_ +#define _IVSRCID_VISLANDS30_H_ + + +// IV Source IDs + +#define VISLANDS30_IV_SRCID_D1_V_UPDATE_INT 7 // 0x07 +#define VISLANDS30_IV_EXTID_D1_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D1_GRPH_PFLIP 8 // 0x08 +#define VISLANDS30_IV_EXTID_D1_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_D2_V_UPDATE_INT 9 // 0x09 +#define VISLANDS30_IV_EXTID_D2_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D2_GRPH_PFLIP 10 // 0x0a +#define VISLANDS30_IV_EXTID_D2_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_D3_V_UPDATE_INT 11 // 0x0b +#define VISLANDS30_IV_EXTID_D3_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D3_GRPH_PFLIP 12 // 0x0c +#define VISLANDS30_IV_EXTID_D3_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_D4_V_UPDATE_INT 13 // 0x0d +#define VISLANDS30_IV_EXTID_D4_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D4_GRPH_PFLIP 14 // 0x0e +#define VISLANDS30_IV_EXTID_D4_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_D5_V_UPDATE_INT 15 // 0x0f +#define VISLANDS30_IV_EXTID_D5_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D5_GRPH_PFLIP 16 // 0x10 +#define VISLANDS30_IV_EXTID_D5_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_D6_V_UPDATE_INT 17 // 0x11 +#define VISLANDS30_IV_EXTID_D6_V_UPDATE_INT 0 + +#define VISLANDS30_IV_SRCID_D6_GRPH_PFLIP 18 // 0x12 +#define VISLANDS30_IV_EXTID_D6_GRPH_PFLIP 0 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_A 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_A 0 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_B 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_B 1 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_C 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_C 2 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_D 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_D 3 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_E 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_E 4 + +#define VISLANDS30_IV_SRCID_HOTPLUG_DETECT_F 42 // 0x2a +#define VISLANDS30_IV_EXTID_HOTPLUG_DETECT_F 5 + +#define VISLANDS30_IV_SRCID_HPD_RX_A 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_A 6 + +#define VISLANDS30_IV_SRCID_HPD_RX_B 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_B 7 + +#define VISLANDS30_IV_SRCID_HPD_RX_C 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_C 8 + +#define VISLANDS30_IV_SRCID_HPD_RX_D 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_D 9 + +#define VISLANDS30_IV_SRCID_HPD_RX_E 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_E 10 + +#define VISLANDS30_IV_SRCID_HPD_RX_F 42 // 0x2a +#define VISLANDS30_IV_EXTID_HPD_RX_F 11 + +#endif // _IVSRCID_VISLANDS30_H_ diff --git a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c index aa67244a77ae..2ee4190f8c89 100644 --- a/drivers/gpu/drm/amd/powerplay/amd_powerplay.c +++ b/drivers/gpu/drm/amd/powerplay/amd_powerplay.c @@ -29,6 +29,7 @@ #include "pp_instance.h" #include "power_state.h" #include "eventmanager.h" +#include "pp_debug.h" #define PP_CHECK(handle) \ do { \ @@ -433,7 +434,10 @@ enum amd_pm_state_type pp_dpm_get_current_power_state(void *handle) case PP_StateUILabel_Performance: return POWER_STATE_TYPE_PERFORMANCE; default: - return POWER_STATE_TYPE_DEFAULT; + if (state->classification.flags & PP_StateClassificationFlag_Boot) + return POWER_STATE_TYPE_INTERNAL_BOOT; + else + return POWER_STATE_TYPE_DEFAULT; } } @@ -535,6 +539,112 @@ static int pp_dpm_get_temperature(void *handle) return hwmgr->hwmgr_func->get_temperature(hwmgr); } +static int pp_dpm_get_pp_num_states(void *handle, + struct pp_states_info *data) +{ + struct pp_hwmgr *hwmgr; + int i; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (hwmgr == NULL || hwmgr->ps == NULL) + return -EINVAL; + + data->nums = hwmgr->num_ps; + + for (i = 0; i < hwmgr->num_ps; i++) { + struct pp_power_state *state = (struct pp_power_state *) + ((unsigned long)hwmgr->ps + i * hwmgr->ps_size); + switch (state->classification.ui_label) { + case PP_StateUILabel_Battery: + data->states[i] = POWER_STATE_TYPE_BATTERY; + break; + case PP_StateUILabel_Balanced: + data->states[i] = POWER_STATE_TYPE_BALANCED; + break; + case PP_StateUILabel_Performance: + data->states[i] = POWER_STATE_TYPE_PERFORMANCE; + break; + default: + if (state->classification.flags & PP_StateClassificationFlag_Boot) + data->states[i] = POWER_STATE_TYPE_INTERNAL_BOOT; + else + data->states[i] = POWER_STATE_TYPE_DEFAULT; + } + } + + return 0; +} + +static int pp_dpm_get_pp_table(void *handle, char **table) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (hwmgr == NULL || hwmgr->hwmgr_func == NULL || + hwmgr->hwmgr_func->get_pp_table == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->get_pp_table(hwmgr, table); +} + +static int pp_dpm_set_pp_table(void *handle, const char *buf, size_t size) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (hwmgr == NULL || hwmgr->hwmgr_func == NULL || + hwmgr->hwmgr_func->set_pp_table == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->set_pp_table(hwmgr, buf, size); +} + +static int pp_dpm_force_clock_level(void *handle, + enum pp_clock_type type, int level) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (hwmgr == NULL || hwmgr->hwmgr_func == NULL || + hwmgr->hwmgr_func->force_clock_level == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->force_clock_level(hwmgr, type, level); +} + +static int pp_dpm_print_clock_levels(void *handle, + enum pp_clock_type type, char *buf) +{ + struct pp_hwmgr *hwmgr; + + if (!handle) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (hwmgr == NULL || hwmgr->hwmgr_func == NULL || + hwmgr->hwmgr_func->print_clock_levels == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->print_clock_levels(hwmgr, type, buf); +} + const struct amd_powerplay_funcs pp_dpm_funcs = { .get_temperature = pp_dpm_get_temperature, .load_firmware = pp_dpm_load_fw, @@ -552,6 +662,11 @@ const struct amd_powerplay_funcs pp_dpm_funcs = { .get_fan_control_mode = pp_dpm_get_fan_control_mode, .set_fan_speed_percent = pp_dpm_set_fan_speed_percent, .get_fan_speed_percent = pp_dpm_get_fan_speed_percent, + .get_pp_num_states = pp_dpm_get_pp_num_states, + .get_pp_table = pp_dpm_get_pp_table, + .set_pp_table = pp_dpm_set_pp_table, + .force_clock_level = pp_dpm_force_clock_level, + .print_clock_levels = pp_dpm_print_clock_levels, }; static int amd_pp_instance_init(struct amd_pp_init *pp_init, @@ -635,10 +750,10 @@ int amd_powerplay_fini(void *handle) /* export this function to DAL */ -int amd_powerplay_display_configuration_change(void *handle, const void *input) +int amd_powerplay_display_configuration_change(void *handle, + const struct amd_pp_display_configuration *display_config) { struct pp_hwmgr *hwmgr; - const struct amd_pp_display_configuration *display_config = input; PP_CHECK((struct pp_instance *)handle); @@ -650,7 +765,7 @@ int amd_powerplay_display_configuration_change(void *handle, const void *input) } int amd_powerplay_get_display_power_level(void *handle, - struct amd_pp_dal_clock_info *output) + struct amd_pp_simple_clock_info *output) { struct pp_hwmgr *hwmgr; @@ -663,3 +778,86 @@ int amd_powerplay_get_display_power_level(void *handle, return phm_get_dal_power_level(hwmgr, output); } + +int amd_powerplay_get_current_clocks(void *handle, + struct amd_pp_clock_info *clocks) +{ + struct pp_hwmgr *hwmgr; + struct amd_pp_simple_clock_info simple_clocks; + struct pp_clock_info hw_clocks; + + PP_CHECK((struct pp_instance *)handle); + + if (clocks == NULL) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + phm_get_dal_power_level(hwmgr, &simple_clocks); + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_PowerContainment)) { + if (0 != phm_get_clock_info(hwmgr, &hwmgr->current_ps->hardware, &hw_clocks, PHM_PerformanceLevelDesignation_PowerContainment)) + PP_ASSERT_WITH_CODE(0, "Error in PHM_GetPowerContainmentClockInfo", return -1); + } else { + if (0 != phm_get_clock_info(hwmgr, &hwmgr->current_ps->hardware, &hw_clocks, PHM_PerformanceLevelDesignation_Activity)) + PP_ASSERT_WITH_CODE(0, "Error in PHM_GetClockInfo", return -1); + } + + clocks->min_engine_clock = hw_clocks.min_eng_clk; + clocks->max_engine_clock = hw_clocks.max_eng_clk; + clocks->min_memory_clock = hw_clocks.min_mem_clk; + clocks->max_memory_clock = hw_clocks.max_mem_clk; + clocks->min_bus_bandwidth = hw_clocks.min_bus_bandwidth; + clocks->max_bus_bandwidth = hw_clocks.max_bus_bandwidth; + + clocks->max_engine_clock_in_sr = hw_clocks.max_eng_clk; + clocks->min_engine_clock_in_sr = hw_clocks.min_eng_clk; + + clocks->max_clocks_state = simple_clocks.level; + + if (0 == phm_get_current_shallow_sleep_clocks(hwmgr, &hwmgr->current_ps->hardware, &hw_clocks)) { + clocks->max_engine_clock_in_sr = hw_clocks.max_eng_clk; + clocks->min_engine_clock_in_sr = hw_clocks.min_eng_clk; + } + + return 0; + +} + +int amd_powerplay_get_clock_by_type(void *handle, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks) +{ + int result = -1; + + struct pp_hwmgr *hwmgr; + + PP_CHECK((struct pp_instance *)handle); + + if (clocks == NULL) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + result = phm_get_clock_by_type(hwmgr, type, clocks); + + return result; +} + +int amd_powerplay_get_display_mode_validation_clocks(void *handle, + struct amd_pp_simple_clock_info *clocks) +{ + int result = -1; + struct pp_hwmgr *hwmgr; + + PP_CHECK((struct pp_instance *)handle); + + if (clocks == NULL) + return -EINVAL; + + hwmgr = ((struct pp_instance *)handle)->hwmgr; + + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_DynamicPatchPowerState)) + result = phm_get_max_high_clocks(hwmgr, clocks); + + return result; +} + diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c index 0874ab42ee95..ef1daf1251c7 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/cz_hwmgr.c @@ -715,7 +715,6 @@ static int cz_tf_update_sclk_limit(struct pp_hwmgr *hwmgr, unsigned long clock = 0; unsigned long level; unsigned long stable_pstate_sclk; - struct PP_Clocks clocks; unsigned long percentage; cz_hwmgr->sclk_dpm.soft_min_clk = table->entries[0].clk; @@ -726,8 +725,9 @@ static int cz_tf_update_sclk_limit(struct pp_hwmgr *hwmgr, else cz_hwmgr->sclk_dpm.soft_max_clk = table->entries[table->count - 1].clk; - /*PECI_GetMinClockSettings(pHwMgr->pPECI, &clocks);*/ - clock = clocks.engineClock; + clock = hwmgr->display_config.min_core_set_clock; + if (clock == 0) + printk(KERN_ERR "[ powerplay ] min_core_set_clock not set\n"); if (cz_hwmgr->sclk_dpm.hard_min_clk != clock) { cz_hwmgr->sclk_dpm.hard_min_clk = clock; @@ -883,9 +883,9 @@ static int cz_tf_update_low_mem_pstate(struct pp_hwmgr *hwmgr, if (pnew_state->action == FORCE_HIGH) cz_nbdpm_pstate_enable_disable(hwmgr, false, disable_switch); - else if(pnew_state->action == CANCEL_FORCE_HIGH) - cz_nbdpm_pstate_enable_disable(hwmgr, false, disable_switch); - else + else if (pnew_state->action == CANCEL_FORCE_HIGH) + cz_nbdpm_pstate_enable_disable(hwmgr, true, disable_switch); + else cz_nbdpm_pstate_enable_disable(hwmgr, enable_low_mem_state, disable_switch); } return 0; @@ -1110,9 +1110,10 @@ static int cz_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, cast_const_PhwCzPowerState(&pcurrent_ps->hardware); struct cz_hwmgr *cz_hwmgr = (struct cz_hwmgr *)(hwmgr->backend); - struct PP_Clocks clocks; + struct PP_Clocks clocks = {0, 0, 0, 0}; bool force_high; - unsigned long num_of_active_displays = 4; + uint32_t num_of_active_displays = 0; + struct cgs_display_info info = {0}; cz_ps->evclk = hwmgr->vce_arbiter.evclk; cz_ps->ecclk = hwmgr->vce_arbiter.ecclk; @@ -1124,12 +1125,15 @@ static int cz_apply_state_adjust_rules(struct pp_hwmgr *hwmgr, cz_hwmgr->battery_state = (PP_StateUILabel_Battery == prequest_ps->classification.ui_label); - /* to do PECI_GetMinClockSettings(pHwMgr->pPECI, &clocks); */ - /* PECI_GetNumberOfActiveDisplays(pHwMgr->pPECI, &numOfActiveDisplays); */ + clocks.memoryClock = hwmgr->display_config.min_mem_set_clock != 0 ? + hwmgr->display_config.min_mem_set_clock : + cz_hwmgr->sys_info.nbp_memory_clock[1]; + + cgs_get_active_displays_info(hwmgr->device, &info); + num_of_active_displays = info.display_count; + if (phm_cap_enabled(hwmgr->platform_descriptor.platformCaps, PHM_PlatformCaps_StablePState)) clocks.memoryClock = hwmgr->dyn_state.max_clock_voltage_on_ac.mclk; - else - clocks.memoryClock = 0; if (clocks.memoryClock < hwmgr->gfx_arbiter.mclk) clocks.memoryClock = hwmgr->gfx_arbiter.mclk; @@ -1199,6 +1203,7 @@ static int cz_hwmgr_backend_init(struct pp_hwmgr *hwmgr) printk(KERN_ERR "[ powerplay ] Fail to construct set_power_state\n"); return result; } + hwmgr->platform_descriptor.hardwareActivityPerformanceLevels = CZ_MAX_HARDWARE_POWERLEVELS; result = phm_construct_table(hwmgr, &cz_phm_enable_clock_power_gatings_master, &(hwmgr->enable_clock_power_gatings)); if (result != 0) { @@ -1630,10 +1635,10 @@ static void cz_hw_print_display_cfg( & PWRMGT_SEPARATION_TIME_MASK) << PWRMGT_SEPARATION_TIME_SHIFT; - data|= (hw_data->cc6_settings.cpu_cc6_disable ? 0x1 : 0x0) + data |= (hw_data->cc6_settings.cpu_cc6_disable ? 0x1 : 0x0) << PWRMGT_DISABLE_CPU_CSTATES_SHIFT; - data|= (hw_data->cc6_settings.cpu_pstate_disable ? 0x1 : 0x0) + data |= (hw_data->cc6_settings.cpu_pstate_disable ? 0x1 : 0x0) << PWRMGT_DISABLE_CPU_PSTATES_SHIFT; PP_DBG_LOG("SetDisplaySizePowerParams data: 0x%X\n", @@ -1648,9 +1653,9 @@ static void cz_hw_print_display_cfg( } - static int cz_store_cc6_data(struct pp_hwmgr *hwmgr, uint32_t separation_time, +static int cz_store_cc6_data(struct pp_hwmgr *hwmgr, uint32_t separation_time, bool cc6_disable, bool pstate_disable, bool pstate_switch_disable) - { +{ struct cz_hwmgr *hw_data = (struct cz_hwmgr *)(hwmgr->backend); if (separation_time != @@ -1678,20 +1683,19 @@ static void cz_hw_print_display_cfg( return 0; } - static int cz_get_dal_power_level(struct pp_hwmgr *hwmgr, - struct amd_pp_dal_clock_info*info) +static int cz_get_dal_power_level(struct pp_hwmgr *hwmgr, + struct amd_pp_simple_clock_info *info) { uint32_t i; - const struct phm_clock_voltage_dependency_table * table = + const struct phm_clock_voltage_dependency_table *table = hwmgr->dyn_state.vddc_dep_on_dal_pwrl; - const struct phm_clock_and_voltage_limits* limits = + const struct phm_clock_and_voltage_limits *limits = &hwmgr->dyn_state.max_clock_voltage_on_ac; info->engine_max_clock = limits->sclk; info->memory_max_clock = limits->mclk; for (i = table->count - 1; i > 0; i--) { - if (limits->vddc >= table->entries[i].v) { info->level = table->entries[i].clk; return 0; @@ -1700,6 +1704,158 @@ static void cz_hw_print_display_cfg( return -EINVAL; } +static int cz_force_clock_level(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, int level) +{ + if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) + return -EINVAL; + + switch (type) { + case PP_SCLK: + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetSclkSoftMin, + (1 << level)); + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SetSclkSoftMax, + (1 << level)); + break; + default: + break; + } + + return 0; +} + +static int cz_print_clock_levels(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, char *buf) +{ + struct phm_clock_voltage_dependency_table *sclk_table = + hwmgr->dyn_state.vddc_dependency_on_sclk; + int i, now, size = 0; + + switch (type) { + case PP_SCLK: + now = PHM_GET_FIELD(cgs_read_ind_register(hwmgr->device, + CGS_IND_REG__SMC, + ixTARGET_AND_CURRENT_PROFILE_INDEX), + TARGET_AND_CURRENT_PROFILE_INDEX, + CURR_SCLK_INDEX); + + for (i = 0; i < sclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, sclk_table->entries[i].clk / 100, + (i == now) ? "*" : ""); + break; + default: + break; + } + return size; +} + +static int cz_get_performance_level(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, + PHM_PerformanceLevelDesignation designation, uint32_t index, + PHM_PerformanceLevel *level) +{ + const struct cz_power_state *ps; + struct cz_hwmgr *data; + uint32_t level_index; + uint32_t i; + + if (level == NULL || hwmgr == NULL || state == NULL) + return -EINVAL; + + data = (struct cz_hwmgr *)(hwmgr->backend); + ps = cast_const_PhwCzPowerState(state); + + level_index = index > ps->level - 1 ? ps->level - 1 : index; + + level->coreClock = ps->levels[level_index].engineClock; + + if (designation == PHM_PerformanceLevelDesignation_PowerContainment) { + for (i = 1; i < ps->level; i++) { + if (ps->levels[i].engineClock > data->dce_slow_sclk_threshold) { + level->coreClock = ps->levels[i].engineClock; + break; + } + } + } + + if (level_index == 0) + level->memory_clock = data->sys_info.nbp_memory_clock[CZ_NUM_NBPMEMORYCLOCK - 1]; + else + level->memory_clock = data->sys_info.nbp_memory_clock[0]; + + level->vddc = (cz_convert_8Bit_index_to_voltage(hwmgr, ps->levels[level_index].vddcIndex) + 2) / 4; + level->nonLocalMemoryFreq = 0; + level->nonLocalMemoryWidth = 0; + + return 0; +} + +static int cz_get_current_shallow_sleep_clocks(struct pp_hwmgr *hwmgr, + const struct pp_hw_power_state *state, struct pp_clock_info *clock_info) +{ + const struct cz_power_state *ps = cast_const_PhwCzPowerState(state); + + clock_info->min_eng_clk = ps->levels[0].engineClock / (1 << (ps->levels[0].ssDividerIndex)); + clock_info->max_eng_clk = ps->levels[ps->level - 1].engineClock / (1 << (ps->levels[ps->level - 1].ssDividerIndex)); + + return 0; +} + +static int cz_get_clock_by_type(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, + struct amd_pp_clocks *clocks) +{ + struct cz_hwmgr *data = (struct cz_hwmgr *)(hwmgr->backend); + int i; + struct phm_clock_voltage_dependency_table *table; + + clocks->count = cz_get_max_sclk_level(hwmgr); + switch (type) { + case amd_pp_disp_clock: + for (i = 0; i < clocks->count; i++) + clocks->clock[i] = data->sys_info.display_clock[i]; + break; + case amd_pp_sys_clock: + table = hwmgr->dyn_state.vddc_dependency_on_sclk; + for (i = 0; i < clocks->count; i++) + clocks->clock[i] = table->entries[i].clk; + break; + case amd_pp_mem_clock: + clocks->count = CZ_NUM_NBPMEMORYCLOCK; + for (i = 0; i < clocks->count; i++) + clocks->clock[i] = data->sys_info.nbp_memory_clock[clocks->count - 1 - i]; + break; + default: + return -1; + } + + return 0; +} + +static int cz_get_max_high_clocks(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks) +{ + struct phm_clock_voltage_dependency_table *table = + hwmgr->dyn_state.vddc_dependency_on_sclk; + unsigned long level; + const struct phm_clock_and_voltage_limits *limits = + &hwmgr->dyn_state.max_clock_voltage_on_ac; + + if ((NULL == table) || (table->count <= 0) || (clocks == NULL)) + return -EINVAL; + + level = cz_get_max_sclk_level(hwmgr) - 1; + + if (level < table->count) + clocks->engine_max_clock = table->entries[level].clk; + else + clocks->engine_max_clock = table->entries[table->count - 1].clk; + + clocks->memory_max_clock = limits->mclk; + + return 0; +} + static const struct pp_hwmgr_func cz_hwmgr_funcs = { .backend_init = cz_hwmgr_backend_init, .backend_fini = cz_hwmgr_backend_fini, @@ -1718,7 +1874,13 @@ static const struct pp_hwmgr_func cz_hwmgr_funcs = { .print_current_perforce_level = cz_print_current_perforce_level, .set_cpu_power_state = cz_set_cpu_power_state, .store_cc6_data = cz_store_cc6_data, - .get_dal_power_level= cz_get_dal_power_level, + .force_clock_level = cz_force_clock_level, + .print_clock_levels = cz_print_clock_levels, + .get_dal_power_level = cz_get_dal_power_level, + .get_performance_level = cz_get_performance_level, + .get_current_shallow_sleep_clocks = cz_get_current_shallow_sleep_clocks, + .get_clock_by_type = cz_get_clock_by_type, + .get_max_high_clocks = cz_get_max_high_clocks, }; int cz_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c index 28031a7eddba..5cca2ecc6bea 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/fiji_hwmgr.c @@ -5073,6 +5073,125 @@ static int fiji_get_fan_control_mode(struct pp_hwmgr *hwmgr) CG_FDO_CTRL2, FDO_PWM_MODE); } +static int fiji_get_pp_table(struct pp_hwmgr *hwmgr, char **table) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + *table = (char *)&data->smc_state_table; + + return sizeof(struct SMU73_Discrete_DpmTable); +} + +static int fiji_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + void *table = (void *)&data->smc_state_table; + + memcpy(table, buf, size); + + return 0; +} + +static int fiji_force_clock_level(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, int level) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + + if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) + return -EINVAL; + + switch (type) { + case PP_SCLK: + if (!data->sclk_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + (1 << level)); + break; + case PP_MCLK: + if (!data->mclk_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + (1 << level)); + break; + case PP_PCIE: + if (!data->pcie_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_PCIeDPM_ForceLevel, + (1 << level)); + break; + default: + break; + } + + return 0; +} + +static int fiji_print_clock_levels(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, char *buf) +{ + struct fiji_hwmgr *data = (struct fiji_hwmgr *)(hwmgr->backend); + struct fiji_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table); + struct fiji_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table); + struct fiji_single_dpm_table *pcie_table = &(data->dpm_table.pcie_speed_table); + int i, now, size = 0; + uint32_t clock, pcie_speed; + + switch (type) { + case PP_SCLK: + smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_API_GetSclkFrequency); + clock = cgs_read_register(hwmgr->device, mmSMC_MSG_ARG_0); + + for (i = 0; i < sclk_table->count; i++) { + if (clock > sclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < sclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, sclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_MCLK: + smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_API_GetMclkFrequency); + clock = cgs_read_register(hwmgr->device, mmSMC_MSG_ARG_0); + + for (i = 0; i < mclk_table->count; i++) { + if (clock > mclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < mclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, mclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_PCIE: + pcie_speed = fiji_get_current_pcie_speed(hwmgr); + for (i = 0; i < pcie_table->count; i++) { + if (pcie_speed != pcie_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < pcie_table->count; i++) + size += sprintf(buf + size, "%d: %s %s\n", i, + (pcie_table->dpm_levels[i].value == 0) ? "2.5GB, x1" : + (pcie_table->dpm_levels[i].value == 1) ? "5.0GB, x16" : + (pcie_table->dpm_levels[i].value == 2) ? "8.0GB, x16" : "", + (i == now) ? "*" : ""); + break; + default: + break; + } + return size; +} + static const struct pp_hwmgr_func fiji_hwmgr_funcs = { .backend_init = &fiji_hwmgr_backend_init, .backend_fini = &tonga_hwmgr_backend_fini, @@ -5108,6 +5227,10 @@ static const struct pp_hwmgr_func fiji_hwmgr_funcs = { .register_internal_thermal_interrupt = fiji_register_internal_thermal_interrupt, .set_fan_control_mode = fiji_set_fan_control_mode, .get_fan_control_mode = fiji_get_fan_control_mode, + .get_pp_table = fiji_get_pp_table, + .set_pp_table = fiji_set_pp_table, + .force_clock_level = fiji_force_clock_level, + .print_clock_levels = fiji_print_clock_levels, }; int fiji_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c index 0f2d5e4bc241..be31bed2538a 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/hardwaremanager.c @@ -26,7 +26,7 @@ #include "power_state.h" #include "pp_acpi.h" #include "amd_acpi.h" -#include "amd_powerplay.h" +#include "pp_debug.h" #define PHM_FUNC_CHECK(hw) \ do { \ @@ -313,13 +313,12 @@ int phm_store_dal_configuration_data(struct pp_hwmgr *hwmgr, } int phm_get_dal_power_level(struct pp_hwmgr *hwmgr, - struct amd_pp_dal_clock_info *info) + struct amd_pp_simple_clock_info *info) { PHM_FUNC_CHECK(hwmgr); if (info == NULL || hwmgr->hwmgr_func->get_dal_power_level == NULL) return -EINVAL; - return hwmgr->hwmgr_func->get_dal_power_level(hwmgr, info); } @@ -332,3 +331,91 @@ int phm_set_cpu_power_state(struct pp_hwmgr *hwmgr) return 0; } + + +int phm_get_performance_level(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, + PHM_PerformanceLevelDesignation designation, uint32_t index, + PHM_PerformanceLevel *level) +{ + PHM_FUNC_CHECK(hwmgr); + if (hwmgr->hwmgr_func->get_performance_level == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->get_performance_level(hwmgr, state, designation, index, level); + + +} + + +/** +* Gets Clock Info. +* +* @param pHwMgr the address of the powerplay hardware manager. +* @param pPowerState the address of the Power State structure. +* @param pClockInfo the address of PP_ClockInfo structure where the result will be returned. +* @exception PP_Result_Failed if any of the paramters is NULL, otherwise the return value from the back-end. +*/ +int phm_get_clock_info(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, struct pp_clock_info *pclock_info, + PHM_PerformanceLevelDesignation designation) +{ + int result; + PHM_PerformanceLevel performance_level; + + PHM_FUNC_CHECK(hwmgr); + + PP_ASSERT_WITH_CODE((NULL != state), "Invalid Input!", return -EINVAL); + PP_ASSERT_WITH_CODE((NULL != pclock_info), "Invalid Input!", return -EINVAL); + + result = phm_get_performance_level(hwmgr, state, PHM_PerformanceLevelDesignation_Activity, 0, &performance_level); + + PP_ASSERT_WITH_CODE((0 == result), "Failed to retrieve minimum clocks.", return result); + + + pclock_info->min_mem_clk = performance_level.memory_clock; + pclock_info->min_eng_clk = performance_level.coreClock; + pclock_info->min_bus_bandwidth = performance_level.nonLocalMemoryFreq * performance_level.nonLocalMemoryWidth; + + + result = phm_get_performance_level(hwmgr, state, designation, + (hwmgr->platform_descriptor.hardwareActivityPerformanceLevels - 1), &performance_level); + + PP_ASSERT_WITH_CODE((0 == result), "Failed to retrieve maximum clocks.", return result); + + pclock_info->max_mem_clk = performance_level.memory_clock; + pclock_info->max_eng_clk = performance_level.coreClock; + pclock_info->max_bus_bandwidth = performance_level.nonLocalMemoryFreq * performance_level.nonLocalMemoryWidth; + + return 0; +} + +int phm_get_current_shallow_sleep_clocks(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, struct pp_clock_info *clock_info) +{ + PHM_FUNC_CHECK(hwmgr); + + if (hwmgr->hwmgr_func->get_current_shallow_sleep_clocks == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->get_current_shallow_sleep_clocks(hwmgr, state, clock_info); + +} + +int phm_get_clock_by_type(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks) +{ + PHM_FUNC_CHECK(hwmgr); + + if (hwmgr->hwmgr_func->get_clock_by_type == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->get_clock_by_type(hwmgr, type, clocks); + +} + +int phm_get_max_high_clocks(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks) +{ + PHM_FUNC_CHECK(hwmgr); + + if (hwmgr->hwmgr_func->get_max_high_clocks == NULL) + return -EINVAL; + + return hwmgr->hwmgr_func->get_max_high_clocks(hwmgr, clocks); +} diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h b/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h index b7429a527828..b10df328d58c 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/ppevvmath.h @@ -293,7 +293,7 @@ fInt GetScaledFraction(int X, int factor) } if (factor == 1) - return (ConvertToFraction(X)); + return ConvertToFraction(X); fValue = fDivide(ConvertToFraction(X * uPow(-1, bNEGATED)), ConvertToFraction(factor)); @@ -371,7 +371,7 @@ fInt fDivide (fInt X, fInt Y) fZERO = ConvertToFraction(0); if (Equal(Y, fZERO)) - return fZERO; + return fZERO; longlongX = (int64_t)X.full; longlongY = (int64_t)Y.full; diff --git a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c index 44a925006479..bc83fa35ec46 100644 --- a/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c +++ b/drivers/gpu/drm/amd/powerplay/hwmgr/tonga_hwmgr.c @@ -6018,6 +6018,125 @@ static int tonga_get_fan_control_mode(struct pp_hwmgr *hwmgr) CG_FDO_CTRL2, FDO_PWM_MODE); } +static int tonga_get_pp_table(struct pp_hwmgr *hwmgr, char **table) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + + *table = (char *)&data->smc_state_table; + + return sizeof(struct SMU72_Discrete_DpmTable); +} + +static int tonga_set_pp_table(struct pp_hwmgr *hwmgr, const char *buf, size_t size) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + + void *table = (void *)&data->smc_state_table; + + memcpy(table, buf, size); + + return 0; +} + +static int tonga_force_clock_level(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, int level) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + + if (hwmgr->dpm_level != AMD_DPM_FORCED_LEVEL_MANUAL) + return -EINVAL; + + switch (type) { + case PP_SCLK: + if (!data->sclk_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_SCLKDPM_SetEnabledMask, + (1 << level)); + break; + case PP_MCLK: + if (!data->mclk_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_MCLKDPM_SetEnabledMask, + (1 << level)); + break; + case PP_PCIE: + if (!data->pcie_dpm_key_disabled) + smum_send_msg_to_smc_with_parameter(hwmgr->smumgr, + PPSMC_MSG_PCIeDPM_ForceLevel, + (1 << level)); + break; + default: + break; + } + + return 0; +} + +static int tonga_print_clock_levels(struct pp_hwmgr *hwmgr, + enum pp_clock_type type, char *buf) +{ + struct tonga_hwmgr *data = (struct tonga_hwmgr *)(hwmgr->backend); + struct tonga_single_dpm_table *sclk_table = &(data->dpm_table.sclk_table); + struct tonga_single_dpm_table *mclk_table = &(data->dpm_table.mclk_table); + struct tonga_single_dpm_table *pcie_table = &(data->dpm_table.pcie_speed_table); + int i, now, size = 0; + uint32_t clock, pcie_speed; + + switch (type) { + case PP_SCLK: + smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_API_GetSclkFrequency); + clock = cgs_read_register(hwmgr->device, mmSMC_MSG_ARG_0); + + for (i = 0; i < sclk_table->count; i++) { + if (clock > sclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < sclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, sclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_MCLK: + smum_send_msg_to_smc(hwmgr->smumgr, PPSMC_MSG_API_GetMclkFrequency); + clock = cgs_read_register(hwmgr->device, mmSMC_MSG_ARG_0); + + for (i = 0; i < mclk_table->count; i++) { + if (clock > mclk_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < mclk_table->count; i++) + size += sprintf(buf + size, "%d: %uMhz %s\n", + i, mclk_table->dpm_levels[i].value / 100, + (i == now) ? "*" : ""); + break; + case PP_PCIE: + pcie_speed = tonga_get_current_pcie_speed(hwmgr); + for (i = 0; i < pcie_table->count; i++) { + if (pcie_speed != pcie_table->dpm_levels[i].value) + continue; + break; + } + now = i; + + for (i = 0; i < pcie_table->count; i++) + size += sprintf(buf + size, "%d: %s %s\n", i, + (pcie_table->dpm_levels[i].value == 0) ? "2.5GB, x8" : + (pcie_table->dpm_levels[i].value == 1) ? "5.0GB, x16" : + (pcie_table->dpm_levels[i].value == 2) ? "8.0GB, x16" : "", + (i == now) ? "*" : ""); + break; + default: + break; + } + return size; +} + static const struct pp_hwmgr_func tonga_hwmgr_funcs = { .backend_init = &tonga_hwmgr_backend_init, .backend_fini = &tonga_hwmgr_backend_fini, @@ -6055,6 +6174,10 @@ static const struct pp_hwmgr_func tonga_hwmgr_funcs = { .check_states_equal = tonga_check_states_equal, .set_fan_control_mode = tonga_set_fan_control_mode, .get_fan_control_mode = tonga_get_fan_control_mode, + .get_pp_table = tonga_get_pp_table, + .set_pp_table = tonga_set_pp_table, + .force_clock_level = tonga_force_clock_level, + .print_clock_levels = tonga_print_clock_levels, }; int tonga_hwmgr_init(struct pp_hwmgr *hwmgr) diff --git a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h index e61a3e67852e..7255f7ddf93a 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h +++ b/drivers/gpu/drm/amd/powerplay/inc/amd_powerplay.h @@ -29,6 +29,7 @@ #include "amd_shared.h" #include "cgs_common.h" + enum amd_pp_event { AMD_PP_EVENT_INITIALIZE = 0, AMD_PP_EVENT_UNINITIALIZE, @@ -123,6 +124,7 @@ enum amd_dpm_forced_level { AMD_DPM_FORCED_LEVEL_AUTO = 0, AMD_DPM_FORCED_LEVEL_LOW = 1, AMD_DPM_FORCED_LEVEL_HIGH = 2, + AMD_DPM_FORCED_LEVEL_MANUAL = 3, }; struct amd_pp_init { @@ -212,12 +214,55 @@ struct amd_pp_display_configuration { uint32_t dce_tolerable_mclk_in_active_latency; }; -struct amd_pp_dal_clock_info { +struct amd_pp_simple_clock_info { uint32_t engine_max_clock; uint32_t memory_max_clock; uint32_t level; }; +enum PP_DAL_POWERLEVEL { + PP_DAL_POWERLEVEL_INVALID = 0, + PP_DAL_POWERLEVEL_ULTRALOW, + PP_DAL_POWERLEVEL_LOW, + PP_DAL_POWERLEVEL_NOMINAL, + PP_DAL_POWERLEVEL_PERFORMANCE, + + PP_DAL_POWERLEVEL_0 = PP_DAL_POWERLEVEL_ULTRALOW, + PP_DAL_POWERLEVEL_1 = PP_DAL_POWERLEVEL_LOW, + PP_DAL_POWERLEVEL_2 = PP_DAL_POWERLEVEL_NOMINAL, + PP_DAL_POWERLEVEL_3 = PP_DAL_POWERLEVEL_PERFORMANCE, + PP_DAL_POWERLEVEL_4 = PP_DAL_POWERLEVEL_3+1, + PP_DAL_POWERLEVEL_5 = PP_DAL_POWERLEVEL_4+1, + PP_DAL_POWERLEVEL_6 = PP_DAL_POWERLEVEL_5+1, + PP_DAL_POWERLEVEL_7 = PP_DAL_POWERLEVEL_6+1, +}; + +struct amd_pp_clock_info { + uint32_t min_engine_clock; + uint32_t max_engine_clock; + uint32_t min_memory_clock; + uint32_t max_memory_clock; + uint32_t min_bus_bandwidth; + uint32_t max_bus_bandwidth; + uint32_t max_engine_clock_in_sr; + uint32_t min_engine_clock_in_sr; + enum PP_DAL_POWERLEVEL max_clocks_state; +}; + +enum amd_pp_clock_type { + amd_pp_disp_clock = 1, + amd_pp_sys_clock, + amd_pp_mem_clock +}; + +#define MAX_NUM_CLOCKS 16 + +struct amd_pp_clocks { + uint32_t count; + uint32_t clock[MAX_NUM_CLOCKS]; +}; + + enum { PP_GROUP_UNKNOWN = 0, PP_GROUP_GFX = 1, @@ -225,6 +270,17 @@ enum { PP_GROUP_MAX }; +enum pp_clock_type { + PP_SCLK, + PP_MCLK, + PP_PCIE, +}; + +struct pp_states_info { + uint32_t nums; + uint32_t states[16]; +}; + #define PP_GROUP_MASK 0xF0000000 #define PP_GROUP_SHIFT 28 @@ -278,6 +334,11 @@ struct amd_powerplay_funcs { int (*get_fan_control_mode)(void *handle); int (*set_fan_speed_percent)(void *handle, uint32_t percent); int (*get_fan_speed_percent)(void *handle, uint32_t *speed); + int (*get_pp_num_states)(void *handle, struct pp_states_info *data); + int (*get_pp_table)(void *handle, char **table); + int (*set_pp_table)(void *handle, const char *buf, size_t size); + int (*force_clock_level)(void *handle, enum pp_clock_type type, int level); + int (*print_clock_levels)(void *handle, enum pp_clock_type type, char *buf); }; struct amd_powerplay { @@ -288,12 +349,23 @@ struct amd_powerplay { int amd_powerplay_init(struct amd_pp_init *pp_init, struct amd_powerplay *amd_pp); + int amd_powerplay_fini(void *handle); -int amd_powerplay_display_configuration_change(void *handle, const void *input); +int amd_powerplay_display_configuration_change(void *handle, + const struct amd_pp_display_configuration *input); int amd_powerplay_get_display_power_level(void *handle, - struct amd_pp_dal_clock_info *output); + struct amd_pp_simple_clock_info *output); + +int amd_powerplay_get_current_clocks(void *handle, + struct amd_pp_clock_info *output); + +int amd_powerplay_get_clock_by_type(void *handle, + enum amd_pp_clock_type type, + struct amd_pp_clocks *clocks); +int amd_powerplay_get_display_mode_validation_clocks(void *handle, + struct amd_pp_simple_clock_info *output); #endif /* _AMD_POWERPLAY_H_ */ diff --git a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h index 91795efe1336..040d3f7cbf49 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hardwaremanager.h @@ -31,6 +31,7 @@ struct pp_power_state; enum amd_dpm_forced_level; struct PP_TemperatureRange; + struct phm_fan_speed_info { uint32_t min_percent; uint32_t max_percent; @@ -290,6 +291,15 @@ struct PP_Clocks { uint32_t engineClockInSR; }; +struct pp_clock_info { + uint32_t min_mem_clk; + uint32_t max_mem_clk; + uint32_t min_eng_clk; + uint32_t max_eng_clk; + uint32_t min_bus_bandwidth; + uint32_t max_bus_bandwidth; +}; + struct phm_platform_descriptor { uint32_t platformCaps[PHM_MAX_NUM_CAPS_ULONG_ENTRIES]; uint32_t vbiosInterruptId; @@ -323,24 +333,6 @@ struct phm_clocks { uint32_t clock[MAX_NUM_CLOCKS]; }; -enum PP_DAL_POWERLEVEL { - PP_DAL_POWERLEVEL_INVALID = 0, - PP_DAL_POWERLEVEL_ULTRALOW, - PP_DAL_POWERLEVEL_LOW, - PP_DAL_POWERLEVEL_NOMINAL, - PP_DAL_POWERLEVEL_PERFORMANCE, - - PP_DAL_POWERLEVEL_0 = PP_DAL_POWERLEVEL_ULTRALOW, - PP_DAL_POWERLEVEL_1 = PP_DAL_POWERLEVEL_LOW, - PP_DAL_POWERLEVEL_2 = PP_DAL_POWERLEVEL_NOMINAL, - PP_DAL_POWERLEVEL_3 = PP_DAL_POWERLEVEL_PERFORMANCE, - PP_DAL_POWERLEVEL_4 = PP_DAL_POWERLEVEL_3+1, - PP_DAL_POWERLEVEL_5 = PP_DAL_POWERLEVEL_4+1, - PP_DAL_POWERLEVEL_6 = PP_DAL_POWERLEVEL_5+1, - PP_DAL_POWERLEVEL_7 = PP_DAL_POWERLEVEL_6+1, -}; - - extern int phm_enable_clock_power_gatings(struct pp_hwmgr *hwmgr); extern int phm_powergate_uvd(struct pp_hwmgr *hwmgr, bool gate); extern int phm_powergate_vce(struct pp_hwmgr *hwmgr, bool gate); @@ -375,11 +367,25 @@ extern int phm_store_dal_configuration_data(struct pp_hwmgr *hwmgr, const struct amd_pp_display_configuration *display_config); extern int phm_get_dal_power_level(struct pp_hwmgr *hwmgr, - struct amd_pp_dal_clock_info*info); + struct amd_pp_simple_clock_info *info); extern int phm_set_cpu_power_state(struct pp_hwmgr *hwmgr); extern int phm_power_down_asic(struct pp_hwmgr *hwmgr); +extern int phm_get_performance_level(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, + PHM_PerformanceLevelDesignation designation, uint32_t index, + PHM_PerformanceLevel *level); + +extern int phm_get_clock_info(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, + struct pp_clock_info *pclock_info, + PHM_PerformanceLevelDesignation designation); + +extern int phm_get_current_shallow_sleep_clocks(struct pp_hwmgr *hwmgr, const struct pp_hw_power_state *state, struct pp_clock_info *clock_info); + +extern int phm_get_clock_by_type(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks); + +extern int phm_get_max_high_clocks(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks); + #endif /* _HARDWARE_MANAGER_H_ */ diff --git a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h index aeaa3dbba525..928f5a740cba 100644 --- a/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h +++ b/drivers/gpu/drm/amd/powerplay/inc/hwmgr.h @@ -325,8 +325,18 @@ struct pp_hwmgr_func { bool cc6_disable, bool pstate_disable, bool pstate_switch_disable); int (*get_dal_power_level)(struct pp_hwmgr *hwmgr, - struct amd_pp_dal_clock_info *info); + struct amd_pp_simple_clock_info *info); + int (*get_performance_level)(struct pp_hwmgr *, const struct pp_hw_power_state *, + PHM_PerformanceLevelDesignation, uint32_t, PHM_PerformanceLevel *); + int (*get_current_shallow_sleep_clocks)(struct pp_hwmgr *hwmgr, + const struct pp_hw_power_state *state, struct pp_clock_info *clock_info); + int (*get_clock_by_type)(struct pp_hwmgr *hwmgr, enum amd_pp_clock_type type, struct amd_pp_clocks *clocks); + int (*get_max_high_clocks)(struct pp_hwmgr *hwmgr, struct amd_pp_simple_clock_info *clocks); int (*power_off_asic)(struct pp_hwmgr *hwmgr); + int (*get_pp_table)(struct pp_hwmgr *hwmgr, char **table); + int (*set_pp_table)(struct pp_hwmgr *hwmgr, const char *buf, size_t size); + int (*force_clock_level)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, int level); + int (*print_clock_levels)(struct pp_hwmgr *hwmgr, enum pp_clock_type type, char *buf); }; struct pp_table_func { diff --git a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c index 8b2becd1aa07..a5ff9458d359 100644 --- a/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c +++ b/drivers/gpu/drm/amd/scheduler/gpu_scheduler.c @@ -229,6 +229,14 @@ static void amd_sched_entity_wakeup(struct fence *f, struct fence_cb *cb) amd_sched_wakeup(entity->sched); } +static void amd_sched_entity_clear_dep(struct fence *f, struct fence_cb *cb) +{ + struct amd_sched_entity *entity = + container_of(cb, struct amd_sched_entity, cb); + entity->dependency = NULL; + fence_put(f); +} + static bool amd_sched_entity_add_dependency_cb(struct amd_sched_entity *entity) { struct amd_gpu_scheduler *sched = entity->sched; @@ -251,7 +259,7 @@ static bool amd_sched_entity_add_dependency_cb(struct amd_sched_entity *entity) } /* Wait for fence to be scheduled */ - entity->cb.func = amd_sched_entity_wakeup; + entity->cb.func = amd_sched_entity_clear_dep; list_add_tail(&entity->cb.node, &s_fence->scheduled_cb); return true; } diff --git a/drivers/gpu/drm/arm/Kconfig b/drivers/gpu/drm/arm/Kconfig new file mode 100644 index 000000000000..eaed454e043c --- /dev/null +++ b/drivers/gpu/drm/arm/Kconfig @@ -0,0 +1,27 @@ +config DRM_ARM + bool + help + Choose this option to select drivers for ARM's devices + +config DRM_HDLCD + tristate "ARM HDLCD" + depends on DRM && OF && (ARM || ARM64) + depends on COMMON_CLK + select DRM_ARM + select DRM_KMS_HELPER + select DRM_KMS_FB_HELPER + select DRM_KMS_CMA_HELPER + help + Choose this option if you have an ARM High Definition Colour LCD + controller. + + If M is selected the module will be called hdlcd. + +config DRM_HDLCD_SHOW_UNDERRUN + bool "Show underrun conditions" + depends on DRM_HDLCD + default n + help + Enable this option to show in red colour the pixels that the + HDLCD device did not fetch from framebuffer due to underrun + conditions. diff --git a/drivers/gpu/drm/arm/Makefile b/drivers/gpu/drm/arm/Makefile new file mode 100644 index 000000000000..89dcb7bab93a --- /dev/null +++ b/drivers/gpu/drm/arm/Makefile @@ -0,0 +1,2 @@ +hdlcd-y := hdlcd_drv.o hdlcd_crtc.o +obj-$(CONFIG_DRM_HDLCD) += hdlcd.o diff --git a/drivers/gpu/drm/arm/hdlcd_crtc.c b/drivers/gpu/drm/arm/hdlcd_crtc.c new file mode 100644 index 000000000000..fef1b04c2aab --- /dev/null +++ b/drivers/gpu/drm/arm/hdlcd_crtc.c @@ -0,0 +1,327 @@ +/* + * Copyright (C) 2013-2015 ARM Limited + * Author: Liviu Dudau <[email protected]> + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file COPYING in the main directory of this archive + * for more details. + * + * Implementation of a CRTC class for the HDLCD driver. + */ + +#include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drm_fb_helper.h> +#include <drm/drm_fb_cma_helper.h> +#include <drm/drm_gem_cma_helper.h> +#include <drm/drm_of.h> +#include <drm/drm_plane_helper.h> +#include <linux/clk.h> +#include <linux/of_graph.h> +#include <linux/platform_data/simplefb.h> +#include <video/videomode.h> + +#include "hdlcd_drv.h" +#include "hdlcd_regs.h" + +/* + * The HDLCD controller is a dumb RGB streamer that gets connected to + * a single HDMI transmitter or in the case of the ARM Models it gets + * emulated by the software that does the actual rendering. + * + */ + +static const struct drm_crtc_funcs hdlcd_crtc_funcs = { + .destroy = drm_crtc_cleanup, + .set_config = drm_atomic_helper_set_config, + .page_flip = drm_atomic_helper_page_flip, + .reset = drm_atomic_helper_crtc_reset, + .atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_crtc_destroy_state, +}; + +static struct simplefb_format supported_formats[] = SIMPLEFB_FORMATS; + +/* + * Setup the HDLCD registers for decoding the pixels out of the framebuffer + */ +static int hdlcd_set_pxl_fmt(struct drm_crtc *crtc) +{ + unsigned int btpp; + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + uint32_t pixel_format; + struct simplefb_format *format = NULL; + int i; + + pixel_format = crtc->primary->state->fb->pixel_format; + + for (i = 0; i < ARRAY_SIZE(supported_formats); i++) { + if (supported_formats[i].fourcc == pixel_format) + format = &supported_formats[i]; + } + + if (WARN_ON(!format)) + return 0; + + /* HDLCD uses 'bytes per pixel', zero means 1 byte */ + btpp = (format->bits_per_pixel + 7) / 8; + hdlcd_write(hdlcd, HDLCD_REG_PIXEL_FORMAT, (btpp - 1) << 3); + + /* + * The format of the HDLCD_REG_<color>_SELECT register is: + * - bits[23:16] - default value for that color component + * - bits[11:8] - number of bits to extract for each color component + * - bits[4:0] - index of the lowest bit to extract + * + * The default color value is used when bits[11:8] are zero, when the + * pixel is outside the visible frame area or when there is a + * buffer underrun. + */ + hdlcd_write(hdlcd, HDLCD_REG_RED_SELECT, format->red.offset | +#ifdef CONFIG_DRM_HDLCD_SHOW_UNDERRUN + 0x00ff0000 | /* show underruns in red */ +#endif + ((format->red.length & 0xf) << 8)); + hdlcd_write(hdlcd, HDLCD_REG_GREEN_SELECT, format->green.offset | + ((format->green.length & 0xf) << 8)); + hdlcd_write(hdlcd, HDLCD_REG_BLUE_SELECT, format->blue.offset | + ((format->blue.length & 0xf) << 8)); + + return 0; +} + +static void hdlcd_crtc_mode_set_nofb(struct drm_crtc *crtc) +{ + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + struct drm_display_mode *m = &crtc->state->adjusted_mode; + struct videomode vm; + unsigned int polarities, line_length, err; + + vm.vfront_porch = m->crtc_vsync_start - m->crtc_vdisplay; + vm.vback_porch = m->crtc_vtotal - m->crtc_vsync_end; + vm.vsync_len = m->crtc_vsync_end - m->crtc_vsync_start; + vm.hfront_porch = m->crtc_hsync_start - m->crtc_hdisplay; + vm.hback_porch = m->crtc_htotal - m->crtc_hsync_end; + vm.hsync_len = m->crtc_hsync_end - m->crtc_hsync_start; + + polarities = HDLCD_POLARITY_DATAEN | HDLCD_POLARITY_DATA; + + if (m->flags & DRM_MODE_FLAG_PHSYNC) + polarities |= HDLCD_POLARITY_HSYNC; + if (m->flags & DRM_MODE_FLAG_PVSYNC) + polarities |= HDLCD_POLARITY_VSYNC; + + line_length = crtc->primary->state->fb->pitches[0]; + + /* Allow max number of outstanding requests and largest burst size */ + hdlcd_write(hdlcd, HDLCD_REG_BUS_OPTIONS, + HDLCD_BUS_MAX_OUTSTAND | HDLCD_BUS_BURST_16); + + hdlcd_write(hdlcd, HDLCD_REG_FB_LINE_LENGTH, line_length); + hdlcd_write(hdlcd, HDLCD_REG_FB_LINE_PITCH, line_length); + hdlcd_write(hdlcd, HDLCD_REG_FB_LINE_COUNT, m->crtc_vdisplay - 1); + hdlcd_write(hdlcd, HDLCD_REG_V_DATA, m->crtc_vdisplay - 1); + hdlcd_write(hdlcd, HDLCD_REG_V_BACK_PORCH, vm.vback_porch - 1); + hdlcd_write(hdlcd, HDLCD_REG_V_FRONT_PORCH, vm.vfront_porch - 1); + hdlcd_write(hdlcd, HDLCD_REG_V_SYNC, vm.vsync_len - 1); + hdlcd_write(hdlcd, HDLCD_REG_H_BACK_PORCH, vm.hback_porch - 1); + hdlcd_write(hdlcd, HDLCD_REG_H_FRONT_PORCH, vm.hfront_porch - 1); + hdlcd_write(hdlcd, HDLCD_REG_H_SYNC, vm.hsync_len - 1); + hdlcd_write(hdlcd, HDLCD_REG_H_DATA, m->crtc_hdisplay - 1); + hdlcd_write(hdlcd, HDLCD_REG_POLARITIES, polarities); + + err = hdlcd_set_pxl_fmt(crtc); + if (err) + return; + + clk_set_rate(hdlcd->clk, m->crtc_clock * 1000); +} + +static void hdlcd_crtc_enable(struct drm_crtc *crtc) +{ + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + + clk_prepare_enable(hdlcd->clk); + hdlcd_write(hdlcd, HDLCD_REG_COMMAND, 1); + drm_crtc_vblank_on(crtc); +} + +static void hdlcd_crtc_disable(struct drm_crtc *crtc) +{ + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + + if (!crtc->primary->fb) + return; + + clk_disable_unprepare(hdlcd->clk); + hdlcd_write(hdlcd, HDLCD_REG_COMMAND, 0); + drm_crtc_vblank_off(crtc); +} + +static int hdlcd_crtc_atomic_check(struct drm_crtc *crtc, + struct drm_crtc_state *state) +{ + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + struct drm_display_mode *mode = &state->adjusted_mode; + long rate, clk_rate = mode->clock * 1000; + + rate = clk_round_rate(hdlcd->clk, clk_rate); + if (rate != clk_rate) { + /* clock required by mode not supported by hardware */ + return -EINVAL; + } + + return 0; +} + +static void hdlcd_crtc_atomic_begin(struct drm_crtc *crtc, + struct drm_crtc_state *state) +{ + struct hdlcd_drm_private *hdlcd = crtc_to_hdlcd_priv(crtc); + unsigned long flags; + + if (crtc->state->event) { + struct drm_pending_vblank_event *event = crtc->state->event; + + crtc->state->event = NULL; + event->pipe = drm_crtc_index(crtc); + + WARN_ON(drm_crtc_vblank_get(crtc) != 0); + + spin_lock_irqsave(&crtc->dev->event_lock, flags); + list_add_tail(&event->base.link, &hdlcd->event_list); + spin_unlock_irqrestore(&crtc->dev->event_lock, flags); + } +} + +static void hdlcd_crtc_atomic_flush(struct drm_crtc *crtc, + struct drm_crtc_state *state) +{ +} + +static bool hdlcd_crtc_mode_fixup(struct drm_crtc *crtc, + const struct drm_display_mode *mode, + struct drm_display_mode *adjusted_mode) +{ + return true; +} + +static const struct drm_crtc_helper_funcs hdlcd_crtc_helper_funcs = { + .mode_fixup = hdlcd_crtc_mode_fixup, + .mode_set = drm_helper_crtc_mode_set, + .mode_set_base = drm_helper_crtc_mode_set_base, + .mode_set_nofb = hdlcd_crtc_mode_set_nofb, + .enable = hdlcd_crtc_enable, + .disable = hdlcd_crtc_disable, + .prepare = hdlcd_crtc_disable, + .commit = hdlcd_crtc_enable, + .atomic_check = hdlcd_crtc_atomic_check, + .atomic_begin = hdlcd_crtc_atomic_begin, + .atomic_flush = hdlcd_crtc_atomic_flush, +}; + +static int hdlcd_plane_atomic_check(struct drm_plane *plane, + struct drm_plane_state *state) +{ + return 0; +} + +static void hdlcd_plane_atomic_update(struct drm_plane *plane, + struct drm_plane_state *state) +{ + struct hdlcd_drm_private *hdlcd; + struct drm_gem_cma_object *gem; + dma_addr_t scanout_start; + + if (!plane->state->crtc || !plane->state->fb) + return; + + hdlcd = crtc_to_hdlcd_priv(plane->state->crtc); + gem = drm_fb_cma_get_gem_obj(plane->state->fb, 0); + scanout_start = gem->paddr; + hdlcd_write(hdlcd, HDLCD_REG_FB_BASE, scanout_start); +} + +static const struct drm_plane_helper_funcs hdlcd_plane_helper_funcs = { + .prepare_fb = NULL, + .cleanup_fb = NULL, + .atomic_check = hdlcd_plane_atomic_check, + .atomic_update = hdlcd_plane_atomic_update, +}; + +static void hdlcd_plane_destroy(struct drm_plane *plane) +{ + drm_plane_helper_disable(plane); + drm_plane_cleanup(plane); +} + +static const struct drm_plane_funcs hdlcd_plane_funcs = { + .update_plane = drm_atomic_helper_update_plane, + .disable_plane = drm_atomic_helper_disable_plane, + .destroy = hdlcd_plane_destroy, + .reset = drm_atomic_helper_plane_reset, + .atomic_duplicate_state = drm_atomic_helper_plane_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_plane_destroy_state, +}; + +static struct drm_plane *hdlcd_plane_init(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + struct drm_plane *plane = NULL; + u32 formats[ARRAY_SIZE(supported_formats)], i; + int ret; + + plane = devm_kzalloc(drm->dev, sizeof(*plane), GFP_KERNEL); + if (!plane) + return ERR_PTR(-ENOMEM); + + for (i = 0; i < ARRAY_SIZE(supported_formats); i++) + formats[i] = supported_formats[i].fourcc; + + ret = drm_universal_plane_init(drm, plane, 0xff, &hdlcd_plane_funcs, + formats, ARRAY_SIZE(formats), + DRM_PLANE_TYPE_PRIMARY, NULL); + if (ret) { + devm_kfree(drm->dev, plane); + return ERR_PTR(ret); + } + + drm_plane_helper_add(plane, &hdlcd_plane_helper_funcs); + hdlcd->plane = plane; + + return plane; +} + +void hdlcd_crtc_suspend(struct drm_crtc *crtc) +{ + hdlcd_crtc_disable(crtc); +} + +void hdlcd_crtc_resume(struct drm_crtc *crtc) +{ + hdlcd_crtc_enable(crtc); +} + +int hdlcd_setup_crtc(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + struct drm_plane *primary; + int ret; + + primary = hdlcd_plane_init(drm); + if (IS_ERR(primary)) + return PTR_ERR(primary); + + ret = drm_crtc_init_with_planes(drm, &hdlcd->crtc, primary, NULL, + &hdlcd_crtc_funcs, NULL); + if (ret) { + hdlcd_plane_destroy(primary); + devm_kfree(drm->dev, primary); + return ret; + } + + drm_crtc_helper_add(&hdlcd->crtc, &hdlcd_crtc_helper_funcs); + return 0; +} diff --git a/drivers/gpu/drm/arm/hdlcd_drv.c b/drivers/gpu/drm/arm/hdlcd_drv.c new file mode 100644 index 000000000000..56b829f97699 --- /dev/null +++ b/drivers/gpu/drm/arm/hdlcd_drv.c @@ -0,0 +1,550 @@ +/* + * Copyright (C) 2013-2015 ARM Limited + * Author: Liviu Dudau <[email protected]> + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file COPYING in the main directory of this archive + * for more details. + * + * ARM HDLCD Driver + */ + +#include <linux/module.h> +#include <linux/spinlock.h> +#include <linux/clk.h> +#include <linux/component.h> +#include <linux/list.h> +#include <linux/of_graph.h> +#include <linux/of_reserved_mem.h> +#include <linux/pm_runtime.h> + +#include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drm_fb_helper.h> +#include <drm/drm_fb_cma_helper.h> +#include <drm/drm_gem_cma_helper.h> +#include <drm/drm_of.h> + +#include "hdlcd_drv.h" +#include "hdlcd_regs.h" + +static int hdlcd_load(struct drm_device *drm, unsigned long flags) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + struct platform_device *pdev = to_platform_device(drm->dev); + struct resource *res; + u32 version; + int ret; + + hdlcd->clk = devm_clk_get(drm->dev, "pxlclk"); + if (IS_ERR(hdlcd->clk)) + return PTR_ERR(hdlcd->clk); + +#ifdef CONFIG_DEBUG_FS + atomic_set(&hdlcd->buffer_underrun_count, 0); + atomic_set(&hdlcd->bus_error_count, 0); + atomic_set(&hdlcd->vsync_count, 0); + atomic_set(&hdlcd->dma_end_count, 0); +#endif + + INIT_LIST_HEAD(&hdlcd->event_list); + + res = platform_get_resource(pdev, IORESOURCE_MEM, 0); + hdlcd->mmio = devm_ioremap_resource(drm->dev, res); + if (IS_ERR(hdlcd->mmio)) { + DRM_ERROR("failed to map control registers area\n"); + ret = PTR_ERR(hdlcd->mmio); + hdlcd->mmio = NULL; + goto fail; + } + + version = hdlcd_read(hdlcd, HDLCD_REG_VERSION); + if ((version & HDLCD_PRODUCT_MASK) != HDLCD_PRODUCT_ID) { + DRM_ERROR("unknown product id: 0x%x\n", version); + ret = -EINVAL; + goto fail; + } + DRM_INFO("found ARM HDLCD version r%dp%d\n", + (version & HDLCD_VERSION_MAJOR_MASK) >> 8, + version & HDLCD_VERSION_MINOR_MASK); + + /* Get the optional framebuffer memory resource */ + ret = of_reserved_mem_device_init(drm->dev); + if (ret && ret != -ENODEV) + goto fail; + + ret = dma_set_mask_and_coherent(drm->dev, DMA_BIT_MASK(32)); + if (ret) + goto setup_fail; + + ret = hdlcd_setup_crtc(drm); + if (ret < 0) { + DRM_ERROR("failed to create crtc\n"); + goto setup_fail; + } + + pm_runtime_enable(drm->dev); + + pm_runtime_get_sync(drm->dev); + ret = drm_irq_install(drm, platform_get_irq(pdev, 0)); + pm_runtime_put_sync(drm->dev); + if (ret < 0) { + DRM_ERROR("failed to install IRQ handler\n"); + goto irq_fail; + } + + return 0; + +irq_fail: + drm_crtc_cleanup(&hdlcd->crtc); +setup_fail: + of_reserved_mem_device_release(drm->dev); +fail: + devm_clk_put(drm->dev, hdlcd->clk); + + return ret; +} + +static void hdlcd_fb_output_poll_changed(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + + if (hdlcd->fbdev) + drm_fbdev_cma_hotplug_event(hdlcd->fbdev); +} + +static int hdlcd_atomic_commit(struct drm_device *dev, + struct drm_atomic_state *state, bool async) +{ + return drm_atomic_helper_commit(dev, state, false); +} + +static const struct drm_mode_config_funcs hdlcd_mode_config_funcs = { + .fb_create = drm_fb_cma_create, + .output_poll_changed = hdlcd_fb_output_poll_changed, + .atomic_check = drm_atomic_helper_check, + .atomic_commit = hdlcd_atomic_commit, +}; + +static void hdlcd_setup_mode_config(struct drm_device *drm) +{ + drm_mode_config_init(drm); + drm->mode_config.min_width = 0; + drm->mode_config.min_height = 0; + drm->mode_config.max_width = HDLCD_MAX_XRES; + drm->mode_config.max_height = HDLCD_MAX_YRES; + drm->mode_config.funcs = &hdlcd_mode_config_funcs; +} + +static void hdlcd_lastclose(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + + drm_fbdev_cma_restore_mode(hdlcd->fbdev); +} + +static irqreturn_t hdlcd_irq(int irq, void *arg) +{ + struct drm_device *drm = arg; + struct hdlcd_drm_private *hdlcd = drm->dev_private; + unsigned long irq_status; + + irq_status = hdlcd_read(hdlcd, HDLCD_REG_INT_STATUS); + +#ifdef CONFIG_DEBUG_FS + if (irq_status & HDLCD_INTERRUPT_UNDERRUN) + atomic_inc(&hdlcd->buffer_underrun_count); + + if (irq_status & HDLCD_INTERRUPT_DMA_END) + atomic_inc(&hdlcd->dma_end_count); + + if (irq_status & HDLCD_INTERRUPT_BUS_ERROR) + atomic_inc(&hdlcd->bus_error_count); + + if (irq_status & HDLCD_INTERRUPT_VSYNC) + atomic_inc(&hdlcd->vsync_count); + +#endif + if (irq_status & HDLCD_INTERRUPT_VSYNC) { + bool events_sent = false; + unsigned long flags; + struct drm_pending_vblank_event *e, *t; + + drm_crtc_handle_vblank(&hdlcd->crtc); + + spin_lock_irqsave(&drm->event_lock, flags); + list_for_each_entry_safe(e, t, &hdlcd->event_list, base.link) { + list_del(&e->base.link); + drm_crtc_send_vblank_event(&hdlcd->crtc, e); + events_sent = true; + } + if (events_sent) + drm_crtc_vblank_put(&hdlcd->crtc); + spin_unlock_irqrestore(&drm->event_lock, flags); + } + + /* acknowledge interrupt(s) */ + hdlcd_write(hdlcd, HDLCD_REG_INT_CLEAR, irq_status); + + return IRQ_HANDLED; +} + +static void hdlcd_irq_preinstall(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + /* Ensure interrupts are disabled */ + hdlcd_write(hdlcd, HDLCD_REG_INT_MASK, 0); + hdlcd_write(hdlcd, HDLCD_REG_INT_CLEAR, ~0); +} + +static int hdlcd_irq_postinstall(struct drm_device *drm) +{ +#ifdef CONFIG_DEBUG_FS + struct hdlcd_drm_private *hdlcd = drm->dev_private; + unsigned long irq_mask = hdlcd_read(hdlcd, HDLCD_REG_INT_MASK); + + /* enable debug interrupts */ + irq_mask |= HDLCD_DEBUG_INT_MASK; + + hdlcd_write(hdlcd, HDLCD_REG_INT_MASK, irq_mask); +#endif + return 0; +} + +static void hdlcd_irq_uninstall(struct drm_device *drm) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + /* disable all the interrupts that we might have enabled */ + unsigned long irq_mask = hdlcd_read(hdlcd, HDLCD_REG_INT_MASK); + +#ifdef CONFIG_DEBUG_FS + /* disable debug interrupts */ + irq_mask &= ~HDLCD_DEBUG_INT_MASK; +#endif + + /* disable vsync interrupts */ + irq_mask &= ~HDLCD_INTERRUPT_VSYNC; + + hdlcd_write(hdlcd, HDLCD_REG_INT_MASK, irq_mask); +} + +static int hdlcd_enable_vblank(struct drm_device *drm, unsigned int crtc) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + unsigned int mask = hdlcd_read(hdlcd, HDLCD_REG_INT_MASK); + + hdlcd_write(hdlcd, HDLCD_REG_INT_MASK, mask | HDLCD_INTERRUPT_VSYNC); + + return 0; +} + +static void hdlcd_disable_vblank(struct drm_device *drm, unsigned int crtc) +{ + struct hdlcd_drm_private *hdlcd = drm->dev_private; + unsigned int mask = hdlcd_read(hdlcd, HDLCD_REG_INT_MASK); + + hdlcd_write(hdlcd, HDLCD_REG_INT_MASK, mask & ~HDLCD_INTERRUPT_VSYNC); +} + +#ifdef CONFIG_DEBUG_FS +static int hdlcd_show_underrun_count(struct seq_file *m, void *arg) +{ + struct drm_info_node *node = (struct drm_info_node *)m->private; + struct drm_device *drm = node->minor->dev; + struct hdlcd_drm_private *hdlcd = drm->dev_private; + + seq_printf(m, "underrun : %d\n", atomic_read(&hdlcd->buffer_underrun_count)); + seq_printf(m, "dma_end : %d\n", atomic_read(&hdlcd->dma_end_count)); + seq_printf(m, "bus_error: %d\n", atomic_read(&hdlcd->bus_error_count)); + seq_printf(m, "vsync : %d\n", atomic_read(&hdlcd->vsync_count)); + return 0; +} + +static int hdlcd_show_pxlclock(struct seq_file *m, void *arg) +{ + struct drm_info_node *node = (struct drm_info_node *)m->private; + struct drm_device *drm = node->minor->dev; + struct hdlcd_drm_private *hdlcd = drm->dev_private; + unsigned long clkrate = clk_get_rate(hdlcd->clk); + unsigned long mode_clock = hdlcd->crtc.mode.crtc_clock * 1000; + + seq_printf(m, "hw : %lu\n", clkrate); + seq_printf(m, "mode: %lu\n", mode_clock); + return 0; +} + +static struct drm_info_list hdlcd_debugfs_list[] = { + { "interrupt_count", hdlcd_show_underrun_count, 0 }, + { "clocks", hdlcd_show_pxlclock, 0 }, +}; + +static int hdlcd_debugfs_init(struct drm_minor *minor) +{ + return drm_debugfs_create_files(hdlcd_debugfs_list, + ARRAY_SIZE(hdlcd_debugfs_list), minor->debugfs_root, minor); +} + +static void hdlcd_debugfs_cleanup(struct drm_minor *minor) +{ + drm_debugfs_remove_files(hdlcd_debugfs_list, + ARRAY_SIZE(hdlcd_debugfs_list), minor); +} +#endif + +static const struct file_operations fops = { + .owner = THIS_MODULE, + .open = drm_open, + .release = drm_release, + .unlocked_ioctl = drm_ioctl, +#ifdef CONFIG_COMPAT + .compat_ioctl = drm_compat_ioctl, +#endif + .poll = drm_poll, + .read = drm_read, + .llseek = noop_llseek, + .mmap = drm_gem_cma_mmap, +}; + +static struct drm_driver hdlcd_driver = { + .driver_features = DRIVER_HAVE_IRQ | DRIVER_GEM | + DRIVER_MODESET | DRIVER_PRIME | + DRIVER_ATOMIC, + .lastclose = hdlcd_lastclose, + .irq_handler = hdlcd_irq, + .irq_preinstall = hdlcd_irq_preinstall, + .irq_postinstall = hdlcd_irq_postinstall, + .irq_uninstall = hdlcd_irq_uninstall, + .get_vblank_counter = drm_vblank_no_hw_counter, + .enable_vblank = hdlcd_enable_vblank, + .disable_vblank = hdlcd_disable_vblank, + .gem_free_object = drm_gem_cma_free_object, + .gem_vm_ops = &drm_gem_cma_vm_ops, + .dumb_create = drm_gem_cma_dumb_create, + .dumb_map_offset = drm_gem_cma_dumb_map_offset, + .dumb_destroy = drm_gem_dumb_destroy, + .prime_handle_to_fd = drm_gem_prime_handle_to_fd, + .prime_fd_to_handle = drm_gem_prime_fd_to_handle, + .gem_prime_export = drm_gem_prime_export, + .gem_prime_import = drm_gem_prime_import, + .gem_prime_get_sg_table = drm_gem_cma_prime_get_sg_table, + .gem_prime_import_sg_table = drm_gem_cma_prime_import_sg_table, + .gem_prime_vmap = drm_gem_cma_prime_vmap, + .gem_prime_vunmap = drm_gem_cma_prime_vunmap, + .gem_prime_mmap = drm_gem_cma_prime_mmap, +#ifdef CONFIG_DEBUG_FS + .debugfs_init = hdlcd_debugfs_init, + .debugfs_cleanup = hdlcd_debugfs_cleanup, +#endif + .fops = &fops, + .name = "hdlcd", + .desc = "ARM HDLCD Controller DRM", + .date = "20151021", + .major = 1, + .minor = 0, +}; + +static int hdlcd_drm_bind(struct device *dev) +{ + struct drm_device *drm; + struct hdlcd_drm_private *hdlcd; + int ret; + + hdlcd = devm_kzalloc(dev, sizeof(*hdlcd), GFP_KERNEL); + if (!hdlcd) + return -ENOMEM; + + drm = drm_dev_alloc(&hdlcd_driver, dev); + if (!drm) + return -ENOMEM; + + drm->dev_private = hdlcd; + hdlcd_setup_mode_config(drm); + ret = hdlcd_load(drm, 0); + if (ret) + goto err_free; + + ret = drm_dev_register(drm, 0); + if (ret) + goto err_unload; + + dev_set_drvdata(dev, drm); + + ret = component_bind_all(dev, drm); + if (ret) { + DRM_ERROR("Failed to bind all components\n"); + goto err_unregister; + } + + ret = drm_vblank_init(drm, drm->mode_config.num_crtc); + if (ret < 0) { + DRM_ERROR("failed to initialise vblank\n"); + goto err_vblank; + } + drm->vblank_disable_allowed = true; + + drm_mode_config_reset(drm); + drm_kms_helper_poll_init(drm); + + hdlcd->fbdev = drm_fbdev_cma_init(drm, 32, drm->mode_config.num_crtc, + drm->mode_config.num_connector); + + if (IS_ERR(hdlcd->fbdev)) { + ret = PTR_ERR(hdlcd->fbdev); + hdlcd->fbdev = NULL; + goto err_fbdev; + } + + return 0; + +err_fbdev: + drm_kms_helper_poll_fini(drm); + drm_mode_config_cleanup(drm); + drm_vblank_cleanup(drm); +err_vblank: + component_unbind_all(dev, drm); +err_unregister: + drm_dev_unregister(drm); +err_unload: + pm_runtime_get_sync(drm->dev); + drm_irq_uninstall(drm); + pm_runtime_put_sync(drm->dev); + pm_runtime_disable(drm->dev); + of_reserved_mem_device_release(drm->dev); + devm_clk_put(dev, hdlcd->clk); +err_free: + drm_dev_unref(drm); + + return ret; +} + +static void hdlcd_drm_unbind(struct device *dev) +{ + struct drm_device *drm = dev_get_drvdata(dev); + struct hdlcd_drm_private *hdlcd = drm->dev_private; + + if (hdlcd->fbdev) { + drm_fbdev_cma_fini(hdlcd->fbdev); + hdlcd->fbdev = NULL; + } + drm_kms_helper_poll_fini(drm); + component_unbind_all(dev, drm); + drm_vblank_cleanup(drm); + pm_runtime_get_sync(drm->dev); + drm_irq_uninstall(drm); + pm_runtime_put_sync(drm->dev); + pm_runtime_disable(drm->dev); + of_reserved_mem_device_release(drm->dev); + if (!IS_ERR(hdlcd->clk)) { + devm_clk_put(drm->dev, hdlcd->clk); + hdlcd->clk = NULL; + } + drm_mode_config_cleanup(drm); + drm_dev_unregister(drm); + drm_dev_unref(drm); + drm->dev_private = NULL; + dev_set_drvdata(dev, NULL); +} + +static const struct component_master_ops hdlcd_master_ops = { + .bind = hdlcd_drm_bind, + .unbind = hdlcd_drm_unbind, +}; + +static int compare_dev(struct device *dev, void *data) +{ + return dev->of_node == data; +} + +static int hdlcd_probe(struct platform_device *pdev) +{ + struct device_node *port, *ep; + struct component_match *match = NULL; + + if (!pdev->dev.of_node) + return -ENODEV; + + /* there is only one output port inside each device, find it */ + ep = of_graph_get_next_endpoint(pdev->dev.of_node, NULL); + if (!ep) + return -ENODEV; + + if (!of_device_is_available(ep)) { + of_node_put(ep); + return -ENODEV; + } + + /* add the remote encoder port as component */ + port = of_graph_get_remote_port_parent(ep); + of_node_put(ep); + if (!port || !of_device_is_available(port)) { + of_node_put(port); + return -EAGAIN; + } + + component_match_add(&pdev->dev, &match, compare_dev, port); + + return component_master_add_with_match(&pdev->dev, &hdlcd_master_ops, + match); +} + +static int hdlcd_remove(struct platform_device *pdev) +{ + component_master_del(&pdev->dev, &hdlcd_master_ops); + return 0; +} + +static const struct of_device_id hdlcd_of_match[] = { + { .compatible = "arm,hdlcd" }, + {}, +}; +MODULE_DEVICE_TABLE(of, hdlcd_of_match); + +static int __maybe_unused hdlcd_pm_suspend(struct device *dev) +{ + struct drm_device *drm = dev_get_drvdata(dev); + struct drm_crtc *crtc; + + if (pm_runtime_suspended(dev)) + return 0; + + drm_modeset_lock_all(drm); + list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) + hdlcd_crtc_suspend(crtc); + drm_modeset_unlock_all(drm); + return 0; +} + +static int __maybe_unused hdlcd_pm_resume(struct device *dev) +{ + struct drm_device *drm = dev_get_drvdata(dev); + struct drm_crtc *crtc; + + if (!pm_runtime_suspended(dev)) + return 0; + + drm_modeset_lock_all(drm); + list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) + hdlcd_crtc_resume(crtc); + drm_modeset_unlock_all(drm); + return 0; +} + +static SIMPLE_DEV_PM_OPS(hdlcd_pm_ops, hdlcd_pm_suspend, hdlcd_pm_resume); + +static struct platform_driver hdlcd_platform_driver = { + .probe = hdlcd_probe, + .remove = hdlcd_remove, + .driver = { + .name = "hdlcd", + .pm = &hdlcd_pm_ops, + .of_match_table = hdlcd_of_match, + }, +}; + +module_platform_driver(hdlcd_platform_driver); + +MODULE_AUTHOR("Liviu Dudau"); +MODULE_DESCRIPTION("ARM HDLCD DRM driver"); +MODULE_LICENSE("GPL v2"); diff --git a/drivers/gpu/drm/arm/hdlcd_drv.h b/drivers/gpu/drm/arm/hdlcd_drv.h new file mode 100644 index 000000000000..aa234784f053 --- /dev/null +++ b/drivers/gpu/drm/arm/hdlcd_drv.h @@ -0,0 +1,42 @@ +/* + * ARM HDLCD Controller register definition + */ + +#ifndef __HDLCD_DRV_H__ +#define __HDLCD_DRV_H__ + +struct hdlcd_drm_private { + void __iomem *mmio; + struct clk *clk; + struct drm_fbdev_cma *fbdev; + struct drm_framebuffer *fb; + struct list_head event_list; + struct drm_crtc crtc; + struct drm_plane *plane; +#ifdef CONFIG_DEBUG_FS + atomic_t buffer_underrun_count; + atomic_t bus_error_count; + atomic_t vsync_count; + atomic_t dma_end_count; +#endif +}; + +#define crtc_to_hdlcd_priv(x) container_of(x, struct hdlcd_drm_private, crtc) + +static inline void hdlcd_write(struct hdlcd_drm_private *hdlcd, + unsigned int reg, u32 value) +{ + writel(value, hdlcd->mmio + reg); +} + +static inline u32 hdlcd_read(struct hdlcd_drm_private *hdlcd, unsigned int reg) +{ + return readl(hdlcd->mmio + reg); +} + +int hdlcd_setup_crtc(struct drm_device *dev); +void hdlcd_set_scanout(struct hdlcd_drm_private *hdlcd); +void hdlcd_crtc_suspend(struct drm_crtc *crtc); +void hdlcd_crtc_resume(struct drm_crtc *crtc); + +#endif /* __HDLCD_DRV_H__ */ diff --git a/drivers/gpu/drm/arm/hdlcd_regs.h b/drivers/gpu/drm/arm/hdlcd_regs.h new file mode 100644 index 000000000000..66799ebef6d3 --- /dev/null +++ b/drivers/gpu/drm/arm/hdlcd_regs.h @@ -0,0 +1,87 @@ +/* + * Copyright (C) 2013,2014 ARM Limited + * + * This file is subject to the terms and conditions of the GNU General Public + * License. See the file COPYING in the main directory of this archive + * for more details. + * + * ARM HDLCD Controller register definition + */ + +#ifndef __HDLCD_REGS_H__ +#define __HDLCD_REGS_H__ + +/* register offsets */ +#define HDLCD_REG_VERSION 0x0000 /* ro */ +#define HDLCD_REG_INT_RAWSTAT 0x0010 /* rw */ +#define HDLCD_REG_INT_CLEAR 0x0014 /* wo */ +#define HDLCD_REG_INT_MASK 0x0018 /* rw */ +#define HDLCD_REG_INT_STATUS 0x001c /* ro */ +#define HDLCD_REG_FB_BASE 0x0100 /* rw */ +#define HDLCD_REG_FB_LINE_LENGTH 0x0104 /* rw */ +#define HDLCD_REG_FB_LINE_COUNT 0x0108 /* rw */ +#define HDLCD_REG_FB_LINE_PITCH 0x010c /* rw */ +#define HDLCD_REG_BUS_OPTIONS 0x0110 /* rw */ +#define HDLCD_REG_V_SYNC 0x0200 /* rw */ +#define HDLCD_REG_V_BACK_PORCH 0x0204 /* rw */ +#define HDLCD_REG_V_DATA 0x0208 /* rw */ +#define HDLCD_REG_V_FRONT_PORCH 0x020c /* rw */ +#define HDLCD_REG_H_SYNC 0x0210 /* rw */ +#define HDLCD_REG_H_BACK_PORCH 0x0214 /* rw */ +#define HDLCD_REG_H_DATA 0x0218 /* rw */ +#define HDLCD_REG_H_FRONT_PORCH 0x021c /* rw */ +#define HDLCD_REG_POLARITIES 0x0220 /* rw */ +#define HDLCD_REG_COMMAND 0x0230 /* rw */ +#define HDLCD_REG_PIXEL_FORMAT 0x0240 /* rw */ +#define HDLCD_REG_RED_SELECT 0x0244 /* rw */ +#define HDLCD_REG_GREEN_SELECT 0x0248 /* rw */ +#define HDLCD_REG_BLUE_SELECT 0x024c /* rw */ + +/* version */ +#define HDLCD_PRODUCT_ID 0x1CDC0000 +#define HDLCD_PRODUCT_MASK 0xFFFF0000 +#define HDLCD_VERSION_MAJOR_MASK 0x0000FF00 +#define HDLCD_VERSION_MINOR_MASK 0x000000FF + +/* interrupts */ +#define HDLCD_INTERRUPT_DMA_END (1 << 0) +#define HDLCD_INTERRUPT_BUS_ERROR (1 << 1) +#define HDLCD_INTERRUPT_VSYNC (1 << 2) +#define HDLCD_INTERRUPT_UNDERRUN (1 << 3) +#define HDLCD_DEBUG_INT_MASK (HDLCD_INTERRUPT_DMA_END | \ + HDLCD_INTERRUPT_BUS_ERROR | \ + HDLCD_INTERRUPT_UNDERRUN) + +/* polarities */ +#define HDLCD_POLARITY_VSYNC (1 << 0) +#define HDLCD_POLARITY_HSYNC (1 << 1) +#define HDLCD_POLARITY_DATAEN (1 << 2) +#define HDLCD_POLARITY_DATA (1 << 3) +#define HDLCD_POLARITY_PIXELCLK (1 << 4) + +/* commands */ +#define HDLCD_COMMAND_DISABLE (0 << 0) +#define HDLCD_COMMAND_ENABLE (1 << 0) + +/* pixel format */ +#define HDLCD_PIXEL_FMT_LITTLE_ENDIAN (0 << 31) +#define HDLCD_PIXEL_FMT_BIG_ENDIAN (1 << 31) +#define HDLCD_BYTES_PER_PIXEL_MASK (3 << 3) + +/* bus options */ +#define HDLCD_BUS_BURST_MASK 0x01f +#define HDLCD_BUS_MAX_OUTSTAND 0xf00 +#define HDLCD_BUS_BURST_NONE (0 << 0) +#define HDLCD_BUS_BURST_1 (1 << 0) +#define HDLCD_BUS_BURST_2 (1 << 1) +#define HDLCD_BUS_BURST_4 (1 << 2) +#define HDLCD_BUS_BURST_8 (1 << 3) +#define HDLCD_BUS_BURST_16 (1 << 4) + +/* Max resolution supported is 4096x4096, 32bpp */ +#define HDLCD_MAX_XRES 4096 +#define HDLCD_MAX_YRES 4096 + +#define NR_PALETTE 256 + +#endif /* __HDLCD_REGS_H__ */ diff --git a/drivers/gpu/drm/armada/armada_drv.c b/drivers/gpu/drm/armada/armada_drv.c index 3bd7e1cde99e..82043c204b76 100644 --- a/drivers/gpu/drm/armada/armada_drv.c +++ b/drivers/gpu/drm/armada/armada_drv.c @@ -188,9 +188,6 @@ static const struct file_operations armada_drm_fops = { static struct drm_driver armada_drm_driver = { .load = armada_drm_load, - .open = NULL, - .preclose = NULL, - .postclose = NULL, .lastclose = armada_drm_lastclose, .unload = armada_drm_unload, .set_busid = drm_platform_set_busid, diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c index 0123458cbd83..f221e2dc1b0d 100644 --- a/drivers/gpu/drm/ast/ast_mode.c +++ b/drivers/gpu/drm/ast/ast_mode.c @@ -710,13 +710,6 @@ static void ast_encoder_dpms(struct drm_encoder *encoder, int mode) } -static bool ast_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void ast_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -736,7 +729,6 @@ static void ast_encoder_commit(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs ast_enc_helper_funcs = { .dpms = ast_encoder_dpms, - .mode_fixup = ast_mode_fixup, .prepare = ast_encoder_prepare, .commit = ast_encoder_commit, .mode_set = ast_encoder_mode_set, diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c index 468a14f266a7..9863291a9a54 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_crtc.c @@ -280,24 +280,6 @@ static void atmel_hlcdc_crtc_destroy(struct drm_crtc *c) kfree(crtc); } -void atmel_hlcdc_crtc_cancel_page_flip(struct drm_crtc *c, - struct drm_file *file) -{ - struct atmel_hlcdc_crtc *crtc = drm_crtc_to_atmel_hlcdc_crtc(c); - struct drm_pending_vblank_event *event; - struct drm_device *dev = c->dev; - unsigned long flags; - - spin_lock_irqsave(&dev->event_lock, flags); - event = crtc->event; - if (event && event->base.file_priv == file) { - event->base.destroy(&event->base); - drm_vblank_put(dev, crtc->id); - crtc->event = NULL; - } - spin_unlock_irqrestore(&dev->event_lock, flags); -} - static void atmel_hlcdc_crtc_finish_page_flip(struct atmel_hlcdc_crtc *crtc) { struct drm_device *dev = crtc->base.dev; diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c index a45b32ba029e..3d8d16402d07 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.c @@ -619,15 +619,6 @@ static void atmel_hlcdc_dc_connector_unplug_all(struct drm_device *dev) mutex_unlock(&dev->mode_config.mutex); } -static void atmel_hlcdc_dc_preclose(struct drm_device *dev, - struct drm_file *file) -{ - struct drm_crtc *crtc; - - list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) - atmel_hlcdc_crtc_cancel_page_flip(crtc, file); -} - static void atmel_hlcdc_dc_lastclose(struct drm_device *dev) { struct atmel_hlcdc_dc *dc = dev->dev_private; @@ -698,7 +689,6 @@ static struct drm_driver atmel_hlcdc_dc_driver = { .driver_features = DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET | DRIVER_PRIME | DRIVER_ATOMIC, - .preclose = atmel_hlcdc_dc_preclose, .lastclose = atmel_hlcdc_dc_lastclose, .irq_handler = atmel_hlcdc_dc_irq_handler, .irq_preinstall = atmel_hlcdc_dc_irq_uninstall, diff --git a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.h b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.h index cf6b375bc38d..fed517f297da 100644 --- a/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.h +++ b/drivers/gpu/drm/atmel-hlcdc/atmel_hlcdc_dc.h @@ -152,9 +152,6 @@ int atmel_hlcdc_plane_prepare_disc_area(struct drm_crtc_state *c_state); void atmel_hlcdc_crtc_irq(struct drm_crtc *c); -void atmel_hlcdc_crtc_cancel_page_flip(struct drm_crtc *crtc, - struct drm_file *file); - void atmel_hlcdc_crtc_suspend(struct drm_crtc *crtc); void atmel_hlcdc_crtc_resume(struct drm_crtc *crtc); diff --git a/drivers/gpu/drm/bochs/bochs_kms.c b/drivers/gpu/drm/bochs/bochs_kms.c index 2849f1b95eec..317c27f2a50b 100644 --- a/drivers/gpu/drm/bochs/bochs_kms.c +++ b/drivers/gpu/drm/bochs/bochs_kms.c @@ -152,13 +152,6 @@ static void bochs_crtc_init(struct drm_device *dev) drm_crtc_helper_add(crtc, &bochs_helper_funcs); } -static bool bochs_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void bochs_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -179,7 +172,6 @@ static void bochs_encoder_commit(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs bochs_encoder_helper_funcs = { .dpms = bochs_encoder_dpms, - .mode_fixup = bochs_encoder_mode_fixup, .mode_set = bochs_encoder_mode_set, .prepare = bochs_encoder_prepare, .commit = bochs_encoder_commit, diff --git a/drivers/gpu/drm/bridge/dw-hdmi.c b/drivers/gpu/drm/bridge/dw-hdmi.c index b0aac4733020..9795b72472ba 100644 --- a/drivers/gpu/drm/bridge/dw-hdmi.c +++ b/drivers/gpu/drm/bridge/dw-hdmi.c @@ -1391,13 +1391,6 @@ static void dw_hdmi_bridge_mode_set(struct drm_bridge *bridge, mutex_unlock(&hdmi->mutex); } -static bool dw_hdmi_bridge_mode_fixup(struct drm_bridge *bridge, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void dw_hdmi_bridge_disable(struct drm_bridge *bridge) { struct dw_hdmi *hdmi = bridge->driver_private; @@ -1546,7 +1539,6 @@ static const struct drm_bridge_funcs dw_hdmi_bridge_funcs = { .pre_enable = dw_hdmi_bridge_nop, .post_disable = dw_hdmi_bridge_nop, .mode_set = dw_hdmi_bridge_mode_set, - .mode_fixup = dw_hdmi_bridge_mode_fixup, }; static irqreturn_t dw_hdmi_hardirq(int irq, void *dev_id) diff --git a/drivers/gpu/drm/cirrus/cirrus_mode.c b/drivers/gpu/drm/cirrus/cirrus_mode.c index 4a02854a6963..432ce9440e09 100644 --- a/drivers/gpu/drm/cirrus/cirrus_mode.c +++ b/drivers/gpu/drm/cirrus/cirrus_mode.c @@ -430,14 +430,6 @@ void cirrus_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, *blue = cirrus_crtc->lut_b[regno]; } - -static bool cirrus_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void cirrus_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -466,7 +458,6 @@ static void cirrus_encoder_destroy(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs cirrus_encoder_helper_funcs = { .dpms = cirrus_encoder_dpms, - .mode_fixup = cirrus_encoder_mode_fixup, .mode_set = cirrus_encoder_mode_set, .prepare = cirrus_encoder_prepare, .commit = cirrus_encoder_commit, diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c index 3f74193885f1..8fb469c4e4b8 100644 --- a/drivers/gpu/drm/drm_atomic.c +++ b/drivers/gpu/drm/drm_atomic.c @@ -1347,44 +1347,23 @@ static struct drm_pending_vblank_event *create_vblank_event( struct drm_device *dev, struct drm_file *file_priv, uint64_t user_data) { struct drm_pending_vblank_event *e = NULL; - unsigned long flags; - - spin_lock_irqsave(&dev->event_lock, flags); - if (file_priv->event_space < sizeof e->event) { - spin_unlock_irqrestore(&dev->event_lock, flags); - goto out; - } - file_priv->event_space -= sizeof e->event; - spin_unlock_irqrestore(&dev->event_lock, flags); + int ret; e = kzalloc(sizeof *e, GFP_KERNEL); - if (e == NULL) { - spin_lock_irqsave(&dev->event_lock, flags); - file_priv->event_space += sizeof e->event; - spin_unlock_irqrestore(&dev->event_lock, flags); - goto out; - } + if (!e) + return NULL; e->event.base.type = DRM_EVENT_FLIP_COMPLETE; - e->event.base.length = sizeof e->event; + e->event.base.length = sizeof(e->event); e->event.user_data = user_data; - e->base.event = &e->event.base; - e->base.file_priv = file_priv; - e->base.destroy = (void (*) (struct drm_pending_event *)) kfree; - -out: - return e; -} -static void destroy_vblank_event(struct drm_device *dev, - struct drm_file *file_priv, struct drm_pending_vblank_event *e) -{ - unsigned long flags; + ret = drm_event_reserve_init(dev, file_priv, &e->base, &e->event.base); + if (ret) { + kfree(e); + return NULL; + } - spin_lock_irqsave(&dev->event_lock, flags); - file_priv->event_space += sizeof e->event; - spin_unlock_irqrestore(&dev->event_lock, flags); - kfree(e); + return e; } static int atomic_set_prop(struct drm_atomic_state *state, @@ -1646,8 +1625,7 @@ out: if (!crtc_state->event) continue; - destroy_vblank_event(dev, file_priv, - crtc_state->event); + drm_event_cancel_free(dev, &crtc_state->event->base); } } diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c index 7c523060a076..2b430b05f35d 100644 --- a/drivers/gpu/drm/drm_atomic_helper.c +++ b/drivers/gpu/drm/drm_atomic_helper.c @@ -125,6 +125,47 @@ get_current_crtc_for_encoder(struct drm_device *dev, return NULL; } +static void +set_best_encoder(struct drm_atomic_state *state, + struct drm_connector_state *conn_state, + struct drm_encoder *encoder) +{ + struct drm_crtc_state *crtc_state; + struct drm_crtc *crtc; + + if (conn_state->best_encoder) { + /* Unset the encoder_mask in the old crtc state. */ + crtc = conn_state->connector->state->crtc; + + /* A NULL crtc is an error here because we should have + * duplicated a NULL best_encoder when crtc was NULL. + * As an exception restoring duplicated atomic state + * during resume is allowed, so don't warn when + * best_encoder is equal to encoder we intend to set. + */ + WARN_ON(!crtc && encoder != conn_state->best_encoder); + if (crtc) { + crtc_state = drm_atomic_get_existing_crtc_state(state, crtc); + + crtc_state->encoder_mask &= + ~(1 << drm_encoder_index(conn_state->best_encoder)); + } + } + + if (encoder) { + crtc = conn_state->crtc; + WARN_ON(!crtc); + if (crtc) { + crtc_state = drm_atomic_get_existing_crtc_state(state, crtc); + + crtc_state->encoder_mask |= + 1 << drm_encoder_index(encoder); + } + } + + conn_state->best_encoder = encoder; +} + static int steal_encoder(struct drm_atomic_state *state, struct drm_encoder *encoder, @@ -134,7 +175,6 @@ steal_encoder(struct drm_atomic_state *state, struct drm_crtc_state *crtc_state; struct drm_connector *connector; struct drm_connector_state *connector_state; - int ret; /* * We can only steal an encoder coming from a connector, which means we @@ -165,10 +205,10 @@ steal_encoder(struct drm_atomic_state *state, if (IS_ERR(connector_state)) return PTR_ERR(connector_state); - ret = drm_atomic_set_crtc_for_connector(connector_state, NULL); - if (ret) - return ret; - connector_state->best_encoder = NULL; + if (connector_state->best_encoder != encoder) + continue; + + set_best_encoder(state, connector_state, NULL); } return 0; @@ -216,7 +256,7 @@ update_connector_routing(struct drm_atomic_state *state, int conn_idx) connector->base.id, connector->name); - connector_state->best_encoder = NULL; + set_best_encoder(state, connector_state, NULL); return 0; } @@ -245,6 +285,8 @@ update_connector_routing(struct drm_atomic_state *state, int conn_idx) } if (new_encoder == connector_state->best_encoder) { + set_best_encoder(state, connector_state, new_encoder); + DRM_DEBUG_ATOMIC("[CONNECTOR:%d:%s] keeps [ENCODER:%d:%s], now on [CRTC:%d:%s]\n", connector->base.id, connector->name, @@ -279,7 +321,8 @@ update_connector_routing(struct drm_atomic_state *state, int conn_idx) if (WARN_ON(!connector_state->crtc)) return -EINVAL; - connector_state->best_encoder = new_encoder; + set_best_encoder(state, connector_state, new_encoder); + idx = drm_crtc_index(connector_state->crtc); crtc_state = state->crtc_states[idx]; @@ -617,7 +660,6 @@ disable_outputs(struct drm_device *dev, struct drm_atomic_state *old_state) for_each_connector_in_state(old_state, connector, old_conn_state, i) { const struct drm_encoder_helper_funcs *funcs; struct drm_encoder *encoder; - struct drm_crtc_state *old_crtc_state; /* Shut down everything that's in the changeset and currently * still on. So need to check the old, saved state. */ @@ -2549,8 +2591,10 @@ void drm_atomic_helper_plane_reset(struct drm_plane *plane) kfree(plane->state); plane->state = kzalloc(sizeof(*plane->state), GFP_KERNEL); - if (plane->state) + if (plane->state) { plane->state->plane = plane; + plane->state->rotation = BIT(DRM_ROTATE_0); + } } EXPORT_SYMBOL(drm_atomic_helper_plane_reset); diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c index d40bab29747e..65258acddb90 100644 --- a/drivers/gpu/drm/drm_crtc.c +++ b/drivers/gpu/drm/drm_crtc.c @@ -1161,6 +1161,29 @@ out_unlock: EXPORT_SYMBOL(drm_encoder_init); /** + * drm_encoder_index - find the index of a registered encoder + * @encoder: encoder to find index for + * + * Given a registered encoder, return the index of that encoder within a DRM + * device's list of encoders. + */ +unsigned int drm_encoder_index(struct drm_encoder *encoder) +{ + unsigned int index = 0; + struct drm_encoder *tmp; + + drm_for_each_encoder(tmp, encoder->dev) { + if (tmp == encoder) + return index; + + index++; + } + + BUG(); +} +EXPORT_SYMBOL(drm_encoder_index); + +/** * drm_encoder_cleanup - cleans up an initialised encoder * @encoder: encoder to cleanup * @@ -5265,7 +5288,6 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, struct drm_crtc *crtc; struct drm_framebuffer *fb = NULL; struct drm_pending_vblank_event *e = NULL; - unsigned long flags; int ret = -EINVAL; if (page_flip->flags & ~DRM_MODE_PAGE_FLIP_FLAGS || @@ -5316,41 +5338,26 @@ int drm_mode_page_flip_ioctl(struct drm_device *dev, } if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) { - ret = -ENOMEM; - spin_lock_irqsave(&dev->event_lock, flags); - if (file_priv->event_space < sizeof(e->event)) { - spin_unlock_irqrestore(&dev->event_lock, flags); - goto out; - } - file_priv->event_space -= sizeof(e->event); - spin_unlock_irqrestore(&dev->event_lock, flags); - - e = kzalloc(sizeof(*e), GFP_KERNEL); - if (e == NULL) { - spin_lock_irqsave(&dev->event_lock, flags); - file_priv->event_space += sizeof(e->event); - spin_unlock_irqrestore(&dev->event_lock, flags); + e = kzalloc(sizeof *e, GFP_KERNEL); + if (!e) { + ret = -ENOMEM; goto out; } - e->event.base.type = DRM_EVENT_FLIP_COMPLETE; e->event.base.length = sizeof(e->event); e->event.user_data = page_flip->user_data; - e->base.event = &e->event.base; - e->base.file_priv = file_priv; - e->base.destroy = - (void (*) (struct drm_pending_event *)) kfree; + ret = drm_event_reserve_init(dev, file_priv, &e->base, &e->event.base); + if (ret) { + kfree(e); + goto out; + } } crtc->primary->old_fb = crtc->primary->fb; ret = crtc->funcs->page_flip(crtc, fb, e, page_flip->flags); if (ret) { - if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) { - spin_lock_irqsave(&dev->event_lock, flags); - file_priv->event_space += sizeof(e->event); - spin_unlock_irqrestore(&dev->event_lock, flags); - kfree(e); - } + if (page_flip->flags & DRM_MODE_PAGE_FLIP_EVENT) + drm_event_cancel_free(dev, &e->base); /* Keep the old fb, don't unref it. */ crtc->primary->old_fb = NULL; } else { @@ -5731,6 +5738,48 @@ int drm_format_vert_chroma_subsampling(uint32_t format) EXPORT_SYMBOL(drm_format_vert_chroma_subsampling); /** + * drm_format_plane_width - width of the plane given the first plane + * @width: width of the first plane + * @format: pixel format + * @plane: plane index + * + * Returns: + * The width of @plane, given that the width of the first plane is @width. + */ +int drm_format_plane_width(int width, uint32_t format, int plane) +{ + if (plane >= drm_format_num_planes(format)) + return 0; + + if (plane == 0) + return width; + + return width / drm_format_horz_chroma_subsampling(format); +} +EXPORT_SYMBOL(drm_format_plane_width); + +/** + * drm_format_plane_height - height of the plane given the first plane + * @height: height of the first plane + * @format: pixel format + * @plane: plane index + * + * Returns: + * The height of @plane, given that the height of the first plane is @height. + */ +int drm_format_plane_height(int height, uint32_t format, int plane) +{ + if (plane >= drm_format_num_planes(format)) + return 0; + + if (plane == 0) + return height; + + return height / drm_format_vert_chroma_subsampling(format); +} +EXPORT_SYMBOL(drm_format_plane_height); + +/** * drm_rotation_simplify() - Try to simplify the rotation * @rotation: Rotation to be simplified * @supported_rotations: Supported rotations diff --git a/drivers/gpu/drm/drm_crtc_helper.c b/drivers/gpu/drm/drm_crtc_helper.c index a02a7f9a6a9d..7539eea4ccbc 100644 --- a/drivers/gpu/drm/drm_crtc_helper.c +++ b/drivers/gpu/drm/drm_crtc_helper.c @@ -73,9 +73,6 @@ * &drm_crtc_helper_funcs, struct &drm_encoder_helper_funcs and struct * &drm_connector_helper_funcs. */ -MODULE_AUTHOR("David Airlie, Jesse Barnes"); -MODULE_DESCRIPTION("DRM KMS helper"); -MODULE_LICENSE("GPL and additional rights"); /** * drm_helper_move_panel_connectors_to_head() - move panels to the front in the @@ -220,6 +217,15 @@ static void __drm_helper_disable_unused_functions(struct drm_device *dev) * disconnected connectors. Then it will disable all unused encoders and CRTCs * either by calling their disable callback if available or by calling their * dpms callback with DRM_MODE_DPMS_OFF. + * + * NOTE: + * + * This function is part of the legacy modeset helper library and will cause + * major confusion with atomic drivers. This is because atomic helpers guarantee + * to never call ->disable() hooks on a disabled function, or ->enable() hooks + * on an enabled functions. drm_helper_disable_unused_functions() on the other + * hand throws such guarantees into the wind and calls disable hooks + * unconditionally on unused functions. */ void drm_helper_disable_unused_functions(struct drm_device *dev) { @@ -328,16 +334,21 @@ bool drm_crtc_helper_set_mode(struct drm_crtc *crtc, } encoder_funcs = encoder->helper_private; - if (!(ret = encoder_funcs->mode_fixup(encoder, mode, - adjusted_mode))) { - DRM_DEBUG_KMS("Encoder fixup failed\n"); - goto done; + if (encoder_funcs->mode_fixup) { + if (!(ret = encoder_funcs->mode_fixup(encoder, mode, + adjusted_mode))) { + DRM_DEBUG_KMS("Encoder fixup failed\n"); + goto done; + } } } - if (!(ret = crtc_funcs->mode_fixup(crtc, mode, adjusted_mode))) { - DRM_DEBUG_KMS("CRTC fixup failed\n"); - goto done; + if (crtc_funcs->mode_fixup) { + if (!(ret = crtc_funcs->mode_fixup(crtc, mode, + adjusted_mode))) { + DRM_DEBUG_KMS("CRTC fixup failed\n"); + goto done; + } } DRM_DEBUG_KMS("[CRTC:%d:%s]\n", crtc->base.id, crtc->name); @@ -578,8 +589,6 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) if (set->crtc->primary->fb == NULL) { DRM_DEBUG_KMS("crtc has no fb, full mode set\n"); mode_changed = true; - } else if (set->fb == NULL) { - mode_changed = true; } else if (set->fb->pixel_format != set->crtc->primary->fb->pixel_format) { mode_changed = true; @@ -590,7 +599,7 @@ int drm_crtc_helper_set_config(struct drm_mode_set *set) if (set->x != set->crtc->x || set->y != set->crtc->y) fb_changed = true; - if (set->mode && !drm_mode_equal(set->mode, &set->crtc->mode)) { + if (!drm_mode_equal(set->mode, &set->crtc->mode)) { DRM_DEBUG_KMS("modes are different, full mode set\n"); drm_mode_debug_printmodeline(&set->crtc->mode); drm_mode_debug_printmodeline(set->mode); diff --git a/drivers/gpu/drm/drm_dp_aux_dev.c b/drivers/gpu/drm/drm_dp_aux_dev.c new file mode 100644 index 000000000000..f73b38b33a8e --- /dev/null +++ b/drivers/gpu/drm/drm_dp_aux_dev.c @@ -0,0 +1,368 @@ +/* + * Copyright © 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Rafael Antognolli <[email protected]> + * + */ + +#include <linux/device.h> +#include <linux/fs.h> +#include <linux/slab.h> +#include <linux/init.h> +#include <linux/kernel.h> +#include <linux/module.h> +#include <linux/uaccess.h> +#include <drm/drm_dp_helper.h> +#include <drm/drm_crtc.h> +#include <drm/drmP.h> + +struct drm_dp_aux_dev { + unsigned index; + struct drm_dp_aux *aux; + struct device *dev; + struct kref refcount; + atomic_t usecount; +}; + +#define DRM_AUX_MINORS 256 +#define AUX_MAX_OFFSET (1 << 20) +static DEFINE_IDR(aux_idr); +static DEFINE_MUTEX(aux_idr_mutex); +static struct class *drm_dp_aux_dev_class; +static int drm_dev_major = -1; + +static struct drm_dp_aux_dev *drm_dp_aux_dev_get_by_minor(unsigned index) +{ + struct drm_dp_aux_dev *aux_dev = NULL; + + mutex_lock(&aux_idr_mutex); + aux_dev = idr_find(&aux_idr, index); + if (!kref_get_unless_zero(&aux_dev->refcount)) + aux_dev = NULL; + mutex_unlock(&aux_idr_mutex); + + return aux_dev; +} + +static struct drm_dp_aux_dev *alloc_drm_dp_aux_dev(struct drm_dp_aux *aux) +{ + struct drm_dp_aux_dev *aux_dev; + int index; + + aux_dev = kzalloc(sizeof(*aux_dev), GFP_KERNEL); + if (!aux_dev) + return ERR_PTR(-ENOMEM); + aux_dev->aux = aux; + atomic_set(&aux_dev->usecount, 1); + kref_init(&aux_dev->refcount); + + mutex_lock(&aux_idr_mutex); + index = idr_alloc_cyclic(&aux_idr, aux_dev, 0, DRM_AUX_MINORS, + GFP_KERNEL); + mutex_unlock(&aux_idr_mutex); + if (index < 0) { + kfree(aux_dev); + return ERR_PTR(index); + } + aux_dev->index = index; + + return aux_dev; +} + +static void release_drm_dp_aux_dev(struct kref *ref) +{ + struct drm_dp_aux_dev *aux_dev = + container_of(ref, struct drm_dp_aux_dev, refcount); + + kfree(aux_dev); +} + +static ssize_t name_show(struct device *dev, + struct device_attribute *attr, char *buf) +{ + ssize_t res; + struct drm_dp_aux_dev *aux_dev = + drm_dp_aux_dev_get_by_minor(MINOR(dev->devt)); + + if (!aux_dev) + return -ENODEV; + + res = sprintf(buf, "%s\n", aux_dev->aux->name); + kref_put(&aux_dev->refcount, release_drm_dp_aux_dev); + + return res; +} +static DEVICE_ATTR_RO(name); + +static struct attribute *drm_dp_aux_attrs[] = { + &dev_attr_name.attr, + NULL, +}; +ATTRIBUTE_GROUPS(drm_dp_aux); + +static int auxdev_open(struct inode *inode, struct file *file) +{ + unsigned int minor = iminor(inode); + struct drm_dp_aux_dev *aux_dev; + + aux_dev = drm_dp_aux_dev_get_by_minor(minor); + if (!aux_dev) + return -ENODEV; + + file->private_data = aux_dev; + return 0; +} + +static loff_t auxdev_llseek(struct file *file, loff_t offset, int whence) +{ + return fixed_size_llseek(file, offset, whence, AUX_MAX_OFFSET); +} + +static ssize_t auxdev_read(struct file *file, char __user *buf, size_t count, + loff_t *offset) +{ + size_t bytes_pending, num_bytes_processed = 0; + struct drm_dp_aux_dev *aux_dev = file->private_data; + ssize_t res = 0; + + if (!atomic_inc_not_zero(&aux_dev->usecount)) + return -ENODEV; + + bytes_pending = min((loff_t)count, AUX_MAX_OFFSET - (*offset)); + + if (!access_ok(VERIFY_WRITE, buf, bytes_pending)) { + res = -EFAULT; + goto out; + } + + while (bytes_pending > 0) { + uint8_t localbuf[DP_AUX_MAX_PAYLOAD_BYTES]; + ssize_t todo = min_t(size_t, bytes_pending, sizeof(localbuf)); + + res = drm_dp_dpcd_read(aux_dev->aux, *offset, localbuf, todo); + if (res <= 0) { + res = num_bytes_processed ? num_bytes_processed : res; + goto out; + } + if (__copy_to_user(buf + num_bytes_processed, localbuf, res)) { + res = num_bytes_processed ? + num_bytes_processed : -EFAULT; + goto out; + } + bytes_pending -= res; + *offset += res; + num_bytes_processed += res; + res = num_bytes_processed; + } + +out: + atomic_dec(&aux_dev->usecount); + wake_up_atomic_t(&aux_dev->usecount); + return res; +} + +static ssize_t auxdev_write(struct file *file, const char __user *buf, + size_t count, loff_t *offset) +{ + size_t bytes_pending, num_bytes_processed = 0; + struct drm_dp_aux_dev *aux_dev = file->private_data; + ssize_t res = 0; + + if (!atomic_inc_not_zero(&aux_dev->usecount)) + return -ENODEV; + + bytes_pending = min((loff_t)count, AUX_MAX_OFFSET - *offset); + + if (!access_ok(VERIFY_READ, buf, bytes_pending)) { + res = -EFAULT; + goto out; + } + + while (bytes_pending > 0) { + uint8_t localbuf[DP_AUX_MAX_PAYLOAD_BYTES]; + ssize_t todo = min_t(size_t, bytes_pending, sizeof(localbuf)); + + if (__copy_from_user(localbuf, + buf + num_bytes_processed, todo)) { + res = num_bytes_processed ? + num_bytes_processed : -EFAULT; + goto out; + } + + res = drm_dp_dpcd_write(aux_dev->aux, *offset, localbuf, todo); + if (res <= 0) { + res = num_bytes_processed ? num_bytes_processed : res; + goto out; + } + bytes_pending -= res; + *offset += res; + num_bytes_processed += res; + res = num_bytes_processed; + } + +out: + atomic_dec(&aux_dev->usecount); + wake_up_atomic_t(&aux_dev->usecount); + return res; +} + +static int auxdev_release(struct inode *inode, struct file *file) +{ + struct drm_dp_aux_dev *aux_dev = file->private_data; + + kref_put(&aux_dev->refcount, release_drm_dp_aux_dev); + return 0; +} + +static const struct file_operations auxdev_fops = { + .owner = THIS_MODULE, + .llseek = auxdev_llseek, + .read = auxdev_read, + .write = auxdev_write, + .open = auxdev_open, + .release = auxdev_release, +}; + +#define to_auxdev(d) container_of(d, struct drm_dp_aux_dev, aux) + +static struct drm_dp_aux_dev *drm_dp_aux_dev_get_by_aux(struct drm_dp_aux *aux) +{ + struct drm_dp_aux_dev *iter, *aux_dev = NULL; + int id; + + /* don't increase kref count here because this function should only be + * used by drm_dp_aux_unregister_devnode. Thus, it will always have at + * least one reference - the one that drm_dp_aux_register_devnode + * created + */ + mutex_lock(&aux_idr_mutex); + idr_for_each_entry(&aux_idr, iter, id) { + if (iter->aux == aux) { + aux_dev = iter; + break; + } + } + mutex_unlock(&aux_idr_mutex); + return aux_dev; +} + +static int auxdev_wait_atomic_t(atomic_t *p) +{ + schedule(); + return 0; +} +/** + * drm_dp_aux_unregister_devnode() - unregister a devnode for this aux channel + * @aux: DisplayPort AUX channel + * + * Returns 0 on success or a negative error code on failure. + */ +void drm_dp_aux_unregister_devnode(struct drm_dp_aux *aux) +{ + struct drm_dp_aux_dev *aux_dev; + unsigned int minor; + + aux_dev = drm_dp_aux_dev_get_by_aux(aux); + if (!aux_dev) /* attach must have failed */ + return; + + mutex_lock(&aux_idr_mutex); + idr_remove(&aux_idr, aux_dev->index); + mutex_unlock(&aux_idr_mutex); + + atomic_dec(&aux_dev->usecount); + wait_on_atomic_t(&aux_dev->usecount, auxdev_wait_atomic_t, + TASK_UNINTERRUPTIBLE); + + minor = aux_dev->index; + if (aux_dev->dev) + device_destroy(drm_dp_aux_dev_class, + MKDEV(drm_dev_major, minor)); + + DRM_DEBUG("drm_dp_aux_dev: aux [%s] unregistering\n", aux->name); + kref_put(&aux_dev->refcount, release_drm_dp_aux_dev); +} +EXPORT_SYMBOL(drm_dp_aux_unregister_devnode); + +/** + * drm_dp_aux_register_devnode() - register a devnode for this aux channel + * @aux: DisplayPort AUX channel + * + * Returns 0 on success or a negative error code on failure. + */ +int drm_dp_aux_register_devnode(struct drm_dp_aux *aux) +{ + struct drm_dp_aux_dev *aux_dev; + int res; + + aux_dev = alloc_drm_dp_aux_dev(aux); + if (IS_ERR(aux_dev)) + return PTR_ERR(aux_dev); + + aux_dev->dev = device_create(drm_dp_aux_dev_class, aux->dev, + MKDEV(drm_dev_major, aux_dev->index), NULL, + "drm_dp_aux%d", aux_dev->index); + if (IS_ERR(aux_dev->dev)) { + res = PTR_ERR(aux_dev->dev); + aux_dev->dev = NULL; + goto error; + } + + DRM_DEBUG("drm_dp_aux_dev: aux [%s] registered as minor %d\n", + aux->name, aux_dev->index); + return 0; +error: + drm_dp_aux_unregister_devnode(aux); + return res; +} +EXPORT_SYMBOL(drm_dp_aux_register_devnode); + +int drm_dp_aux_dev_init(void) +{ + int res; + + drm_dp_aux_dev_class = class_create(THIS_MODULE, "drm_dp_aux_dev"); + if (IS_ERR(drm_dp_aux_dev_class)) { + res = PTR_ERR(drm_dp_aux_dev_class); + goto out; + } + drm_dp_aux_dev_class->dev_groups = drm_dp_aux_groups; + + res = register_chrdev(0, "aux", &auxdev_fops); + if (res < 0) + goto out; + drm_dev_major = res; + + return 0; +out: + class_destroy(drm_dp_aux_dev_class); + return res; +} +EXPORT_SYMBOL(drm_dp_aux_dev_init); + +void drm_dp_aux_dev_exit(void) +{ + unregister_chrdev(drm_dev_major, "aux"); + class_destroy(drm_dp_aux_dev_class); +} +EXPORT_SYMBOL(drm_dp_aux_dev_exit); diff --git a/drivers/gpu/drm/drm_dp_helper.c b/drivers/gpu/drm/drm_dp_helper.c index 9535c5b60387..7d58f594cffe 100644 --- a/drivers/gpu/drm/drm_dp_helper.c +++ b/drivers/gpu/drm/drm_dp_helper.c @@ -28,6 +28,7 @@ #include <linux/sched.h> #include <linux/i2c.h> #include <drm/drm_dp_helper.h> +#include <drm/drm_dp_aux_dev.h> #include <drm/drmP.h> /** @@ -754,6 +755,8 @@ static const struct i2c_algorithm drm_dp_i2c_algo = { */ int drm_dp_aux_register(struct drm_dp_aux *aux) { + int ret; + mutex_init(&aux->hw_mutex); aux->ddc.algo = &drm_dp_i2c_algo; @@ -768,7 +771,17 @@ int drm_dp_aux_register(struct drm_dp_aux *aux) strlcpy(aux->ddc.name, aux->name ? aux->name : dev_name(aux->dev), sizeof(aux->ddc.name)); - return i2c_add_adapter(&aux->ddc); + ret = drm_dp_aux_register_devnode(aux); + if (ret) + return ret; + + ret = i2c_add_adapter(&aux->ddc); + if (ret) { + drm_dp_aux_unregister_devnode(aux); + return ret; + } + + return 0; } EXPORT_SYMBOL(drm_dp_aux_register); @@ -778,6 +791,7 @@ EXPORT_SYMBOL(drm_dp_aux_register); */ void drm_dp_aux_unregister(struct drm_dp_aux *aux) { + drm_dp_aux_unregister_devnode(aux); i2c_del_adapter(&aux->ddc); } EXPORT_SYMBOL(drm_dp_aux_unregister); diff --git a/drivers/gpu/drm/drm_edid.c b/drivers/gpu/drm/drm_edid.c index 04cb4877fabd..fdb1eb014586 100644 --- a/drivers/gpu/drm/drm_edid.c +++ b/drivers/gpu/drm/drm_edid.c @@ -32,6 +32,7 @@ #include <linux/hdmi.h> #include <linux/i2c.h> #include <linux/module.h> +#include <linux/vga_switcheroo.h> #include <drm/drmP.h> #include <drm/drm_edid.h> #include <drm/drm_displayid.h> @@ -1395,6 +1396,31 @@ struct edid *drm_get_edid(struct drm_connector *connector, EXPORT_SYMBOL(drm_get_edid); /** + * drm_get_edid_switcheroo - get EDID data for a vga_switcheroo output + * @connector: connector we're probing + * @adapter: I2C adapter to use for DDC + * + * Wrapper around drm_get_edid() for laptops with dual GPUs using one set of + * outputs. The wrapper adds the requisite vga_switcheroo calls to temporarily + * switch DDC to the GPU which is retrieving EDID. + * + * Return: Pointer to valid EDID or %NULL if we couldn't find any. + */ +struct edid *drm_get_edid_switcheroo(struct drm_connector *connector, + struct i2c_adapter *adapter) +{ + struct pci_dev *pdev = connector->dev->pdev; + struct edid *edid; + + vga_switcheroo_lock_ddc(pdev); + edid = drm_get_edid(connector, adapter); + vga_switcheroo_unlock_ddc(pdev); + + return edid; +} +EXPORT_SYMBOL(drm_get_edid_switcheroo); + +/** * drm_edid_duplicate - duplicate an EDID and the extensions * @edid: EDID to duplicate * diff --git a/drivers/gpu/drm/drm_encoder_slave.c b/drivers/gpu/drm/drm_encoder_slave.c index e8629076de32..4484785cd9ac 100644 --- a/drivers/gpu/drm/drm_encoder_slave.c +++ b/drivers/gpu/drm/drm_encoder_slave.c @@ -140,6 +140,9 @@ bool drm_i2c_encoder_mode_fixup(struct drm_encoder *encoder, const struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) { + if (!get_slave_funcs(encoder)->mode_fixup) + return true; + return get_slave_funcs(encoder)->mode_fixup(encoder, mode, adjusted_mode); } EXPORT_SYMBOL(drm_i2c_encoder_mode_fixup); diff --git a/drivers/gpu/drm/drm_fb_cma_helper.c b/drivers/gpu/drm/drm_fb_cma_helper.c index c895b6fddbd8..bb88e3df9257 100644 --- a/drivers/gpu/drm/drm_fb_cma_helper.c +++ b/drivers/gpu/drm/drm_fb_cma_helper.c @@ -74,7 +74,8 @@ static struct drm_framebuffer_funcs drm_fb_cma_funcs = { }; static struct drm_fb_cma *drm_fb_cma_alloc(struct drm_device *dev, - const const struct drm_mode_fb_cmd2 *mode_cmd, struct drm_gem_cma_object **obj, + const struct drm_mode_fb_cmd2 *mode_cmd, + struct drm_gem_cma_object **obj, unsigned int num_planes) { struct drm_fb_cma *fb_cma; diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c index 1e103c4c6ee0..855108e6e1bd 100644 --- a/drivers/gpu/drm/drm_fb_helper.c +++ b/drivers/gpu/drm/drm_fb_helper.c @@ -104,21 +104,17 @@ int drm_fb_helper_single_add_all_connectors(struct drm_fb_helper *fb_helper) { struct drm_device *dev = fb_helper->dev; struct drm_connector *connector; - int i; + int i, ret; if (!drm_fbdev_emulation) return 0; mutex_lock(&dev->mode_config.mutex); drm_for_each_connector(connector, dev) { - struct drm_fb_helper_connector *fb_helper_connector; + ret = drm_fb_helper_add_one_connector(fb_helper, connector); - fb_helper_connector = kzalloc(sizeof(struct drm_fb_helper_connector), GFP_KERNEL); - if (!fb_helper_connector) + if (ret) goto fail; - - fb_helper_connector->connector = connector; - fb_helper->connector_info[fb_helper->connector_count++] = fb_helper_connector; } mutex_unlock(&dev->mode_config.mutex); return 0; @@ -130,7 +126,7 @@ fail: fb_helper->connector_count = 0; mutex_unlock(&dev->mode_config.mutex); - return -ENOMEM; + return ret; } EXPORT_SYMBOL(drm_fb_helper_single_add_all_connectors); @@ -1989,13 +1985,13 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper) width = dev->mode_config.max_width; height = dev->mode_config.max_height; - crtcs = kcalloc(dev->mode_config.num_connector, + crtcs = kcalloc(fb_helper->connector_count, sizeof(struct drm_fb_helper_crtc *), GFP_KERNEL); - modes = kcalloc(dev->mode_config.num_connector, + modes = kcalloc(fb_helper->connector_count, sizeof(struct drm_display_mode *), GFP_KERNEL); - offsets = kcalloc(dev->mode_config.num_connector, + offsets = kcalloc(fb_helper->connector_count, sizeof(struct drm_fb_offset), GFP_KERNEL); - enabled = kcalloc(dev->mode_config.num_connector, + enabled = kcalloc(fb_helper->connector_count, sizeof(bool), GFP_KERNEL); if (!crtcs || !modes || !enabled || !offsets) { DRM_ERROR("Memory allocation failed\n"); @@ -2009,9 +2005,9 @@ static void drm_setup_crtcs(struct drm_fb_helper *fb_helper) fb_helper->funcs->initial_config(fb_helper, crtcs, modes, offsets, enabled, width, height))) { - memset(modes, 0, dev->mode_config.num_connector*sizeof(modes[0])); - memset(crtcs, 0, dev->mode_config.num_connector*sizeof(crtcs[0])); - memset(offsets, 0, dev->mode_config.num_connector*sizeof(offsets[0])); + memset(modes, 0, fb_helper->connector_count*sizeof(modes[0])); + memset(crtcs, 0, fb_helper->connector_count*sizeof(crtcs[0])); + memset(offsets, 0, fb_helper->connector_count*sizeof(offsets[0])); if (!drm_target_cloned(fb_helper, modes, offsets, enabled, width, height) && @@ -2091,6 +2087,27 @@ out: * drm_fb_helper_fill_fix() are provided as helpers to setup simple default * values for the fbdev info structure. * + * HANG DEBUGGING: + * + * When you have fbcon support built-in or already loaded, this function will do + * a full modeset to setup the fbdev console. Due to locking misdesign in the + * VT/fbdev subsystem that entire modeset sequence has to be done while holding + * console_lock. Until console_unlock is called no dmesg lines will be sent out + * to consoles, not even serial console. This means when your driver crashes, + * you will see absolutely nothing else but a system stuck in this function, + * with no further output. Any kind of printk() you place within your own driver + * or in the drm core modeset code will also never show up. + * + * Standard debug practice is to run the fbcon setup without taking the + * console_lock as a hack, to be able to see backtraces and crashes on the + * serial line. This can be done by setting the fb.lockless_register_fb=1 kernel + * cmdline option. + * + * The other option is to just disable fbdev emulation since very likely the + * first modest from userspace will crash in the same way, and is even easier to + * debug. This can be done by setting the drm_kms_helper.fbdev_emulation=0 + * kernel cmdline option. + * * RETURNS: * Zero if everything went ok, nonzero otherwise. */ @@ -2175,9 +2192,9 @@ EXPORT_SYMBOL(drm_fb_helper_hotplug_event); * but the module doesn't depend on any fb console symbols. At least * attempt to load fbcon to avoid leaving the system without a usable console. */ -#if defined(CONFIG_FRAMEBUFFER_CONSOLE_MODULE) && !defined(CONFIG_EXPERT) -static int __init drm_fb_helper_modinit(void) +int __init drm_fb_helper_modinit(void) { +#if defined(CONFIG_FRAMEBUFFER_CONSOLE_MODULE) && !defined(CONFIG_EXPERT) const char *name = "fbcon"; struct module *fbcon; @@ -2187,8 +2204,7 @@ static int __init drm_fb_helper_modinit(void) if (!fbcon) request_module_nowait(name); +#endif return 0; } - -module_init(drm_fb_helper_modinit); -#endif +EXPORT_SYMBOL(drm_fb_helper_modinit); diff --git a/drivers/gpu/drm/drm_fops.c b/drivers/gpu/drm/drm_fops.c index 1ea8790e5090..aeef58ed359b 100644 --- a/drivers/gpu/drm/drm_fops.c +++ b/drivers/gpu/drm/drm_fops.c @@ -1,4 +1,4 @@ -/** +/* * \file drm_fops.c * File operations for DRM * @@ -44,6 +44,46 @@ /* from BKL pushdown */ DEFINE_MUTEX(drm_global_mutex); +/** + * DOC: file operations + * + * Drivers must define the file operations structure that forms the DRM + * userspace API entry point, even though most of those operations are + * implemented in the DRM core. The mandatory functions are drm_open(), + * drm_read(), drm_ioctl() and drm_compat_ioctl if CONFIG_COMPAT is enabled. + * Drivers which implement private ioctls that require 32/64 bit compatibility + * support must provided their onw .compat_ioctl() handler that processes + * private ioctls and calls drm_compat_ioctl() for core ioctls. + * + * In addition drm_read() and drm_poll() provide support for DRM events. DRM + * events are a generic and extensible means to send asynchronous events to + * userspace through the file descriptor. They are used to send vblank event and + * page flip completions by the KMS API. But drivers can also use it for their + * own needs, e.g. to signal completion of rendering. + * + * The memory mapping implementation will vary depending on how the driver + * manages memory. Legacy drivers will use the deprecated drm_legacy_mmap() + * function, modern drivers should use one of the provided memory-manager + * specific implementations. For GEM-based drivers this is drm_gem_mmap(). + * + * No other file operations are supported by the DRM userspace API. Overall the + * following is an example #file_operations structure: + * + * static const example_drm_fops = { + * .owner = THIS_MODULE, + * .open = drm_open, + * .release = drm_release, + * .unlocked_ioctl = drm_ioctl, + * #ifdef CONFIG_COMPAT + * .compat_ioctl = drm_compat_ioctl, + * #endif + * .poll = drm_poll, + * .read = drm_read, + * .llseek = no_llseek, + * .mmap = drm_gem_mmap, + * }; + */ + static int drm_open_helper(struct file *filp, struct drm_minor *minor); static int drm_setup(struct drm_device * dev) @@ -67,15 +107,17 @@ static int drm_setup(struct drm_device * dev) } /** - * Open file. + * drm_open - open method for DRM file + * @inode: device inode + * @filp: file pointer. * - * \param inode device inode - * \param filp file pointer. - * \return zero on success or a negative number on failure. + * This function must be used by drivers as their .open() #file_operations + * method. It looks up the correct DRM device and instantiates all the per-file + * resources for it. + * + * RETURNS: * - * Searches the DRM device with the same minor number, calls open_helper(), and - * increments the device open count. If the open count was previous at zero, - * i.e., it's the first that the device is open, then calls setup(). + * 0 on success or negative errno value on falure. */ int drm_open(struct inode *inode, struct file *filp) { @@ -112,7 +154,7 @@ err_undo: } EXPORT_SYMBOL(drm_open); -/** +/* * Check whether DRI will run on this CPU. * * \return non-zero if the DRI will run on this CPU, or zero otherwise. @@ -125,7 +167,7 @@ static int drm_cpu_valid(void) return 1; } -/** +/* * drm_new_set_master - Allocate a new master object and become master for the * associated master realm. * @@ -179,7 +221,7 @@ out_err: return ret; } -/** +/* * Called whenever a process opens /dev/drm. * * \param filp file pointer. @@ -222,6 +264,7 @@ static int drm_open_helper(struct file *filp, struct drm_minor *minor) INIT_LIST_HEAD(&priv->fbs); mutex_init(&priv->fbs_lock); INIT_LIST_HEAD(&priv->blobs); + INIT_LIST_HEAD(&priv->pending_event_list); INIT_LIST_HEAD(&priv->event_list); init_waitqueue_head(&priv->event_wait); priv->event_space = 4096; /* set aside 4k for event buffer */ @@ -311,18 +354,16 @@ static void drm_events_release(struct drm_file *file_priv) { struct drm_device *dev = file_priv->minor->dev; struct drm_pending_event *e, *et; - struct drm_pending_vblank_event *v, *vt; unsigned long flags; spin_lock_irqsave(&dev->event_lock, flags); - /* Remove pending flips */ - list_for_each_entry_safe(v, vt, &dev->vblank_event_list, base.link) - if (v->base.file_priv == file_priv) { - list_del(&v->base.link); - drm_vblank_put(dev, v->pipe); - v->base.destroy(&v->base); - } + /* Unlink pending events */ + list_for_each_entry_safe(e, et, &file_priv->pending_event_list, + pending_link) { + list_del(&e->pending_link); + e->file_priv = NULL; + } /* Remove unconsumed events */ list_for_each_entry_safe(e, et, &file_priv->event_list, link) { @@ -333,7 +374,7 @@ static void drm_events_release(struct drm_file *file_priv) spin_unlock_irqrestore(&dev->event_lock, flags); } -/** +/* * drm_legacy_dev_reinit * * Reinitializes a legacy/ums drm device in it's lastclose function. @@ -350,7 +391,7 @@ static void drm_legacy_dev_reinit(struct drm_device *dev) dev->if_version = 0; } -/** +/* * Take down the DRM device. * * \param dev DRM device structure. @@ -387,16 +428,17 @@ int drm_lastclose(struct drm_device * dev) } /** - * Release file. + * drm_release - release method for DRM file + * @inode: device inode + * @filp: file pointer. * - * \param inode device inode - * \param file_priv DRM file private. - * \return zero on success or a negative number on failure. + * This function must be used by drivers as their .release() #file_operations + * method. It frees any resources associated with the open file, and if this is + * the last open file for the DRM device also proceeds to call drm_lastclose(). * - * If the hardware lock is held then free it, and take it again for the kernel - * context since it's necessary to reclaim buffers. Unlink the file private - * data from its list and free it. Decreases the open count and if it reaches - * zero calls drm_lastclose(). + * RETURNS: + * + * Always succeeds and returns 0. */ int drm_release(struct inode *inode, struct file *filp) { @@ -451,7 +493,7 @@ int drm_release(struct inode *inode, struct file *filp) if (file_priv->is_master) { struct drm_master *master = file_priv->master; - /** + /* * Since the master is disappearing, so is the * possibility to lock. */ @@ -508,6 +550,32 @@ int drm_release(struct inode *inode, struct file *filp) } EXPORT_SYMBOL(drm_release); +/** + * drm_read - read method for DRM file + * @filp: file pointer + * @buffer: userspace destination pointer for the read + * @count: count in bytes to read + * @offset: offset to read + * + * This function must be used by drivers as their .read() #file_operations + * method iff they use DRM events for asynchronous signalling to userspace. + * Since events are used by the KMS API for vblank and page flip completion this + * means all modern display drivers must use it. + * + * @offset is ignore, DRM events are read like a pipe. Therefore drivers also + * must set the .llseek() #file_operation to no_llseek(). Polling support is + * provided by drm_poll(). + * + * This function will only ever read a full event. Therefore userspace must + * supply a big enough buffer to fit any event to ensure forward progress. Since + * the maximum event space is currently 4K it's recommended to just use that for + * safety. + * + * RETURNS: + * + * Number of bytes read (always aligned to full events, and can be 0) or a + * negative error code on failure. + */ ssize_t drm_read(struct file *filp, char __user *buffer, size_t count, loff_t *offset) { @@ -578,6 +646,22 @@ put_back_event: } EXPORT_SYMBOL(drm_read); +/** + * drm_poll - poll method for DRM file + * @filp: file pointer + * @wait: poll waiter table + * + * This function must be used by drivers as their .read() #file_operations + * method iff they use DRM events for asynchronous signalling to userspace. + * Since events are used by the KMS API for vblank and page flip completion this + * means all modern display drivers must use it. + * + * See also drm_read(). + * + * RETURNS: + * + * Mask of POLL flags indicating the current status of the file. + */ unsigned int drm_poll(struct file *filp, struct poll_table_struct *wait) { struct drm_file *file_priv = filp->private_data; @@ -591,3 +675,164 @@ unsigned int drm_poll(struct file *filp, struct poll_table_struct *wait) return mask; } EXPORT_SYMBOL(drm_poll); + +/** + * drm_event_reserve_init_locked - init a DRM event and reserve space for it + * @dev: DRM device + * @file_priv: DRM file private data + * @p: tracking structure for the pending event + * @e: actual event data to deliver to userspace + * + * This function prepares the passed in event for eventual delivery. If the event + * doesn't get delivered (because the IOCTL fails later on, before queuing up + * anything) then the even must be cancelled and freed using + * drm_event_cancel_free(). Successfully initialized events should be sent out + * using drm_send_event() or drm_send_event_locked() to signal completion of the + * asynchronous event to userspace. + * + * If callers embedded @p into a larger structure it must be allocated with + * kmalloc and @p must be the first member element. + * + * This is the locked version of drm_event_reserve_init() for callers which + * already hold dev->event_lock. + * + * RETURNS: + * + * 0 on success or a negative error code on failure. + */ +int drm_event_reserve_init_locked(struct drm_device *dev, + struct drm_file *file_priv, + struct drm_pending_event *p, + struct drm_event *e) +{ + if (file_priv->event_space < e->length) + return -ENOMEM; + + file_priv->event_space -= e->length; + + p->event = e; + list_add(&p->pending_link, &file_priv->pending_event_list); + p->file_priv = file_priv; + + /* we *could* pass this in as arg, but everyone uses kfree: */ + p->destroy = (void (*) (struct drm_pending_event *)) kfree; + + return 0; +} +EXPORT_SYMBOL(drm_event_reserve_init_locked); + +/** + * drm_event_reserve_init - init a DRM event and reserve space for it + * @dev: DRM device + * @file_priv: DRM file private data + * @p: tracking structure for the pending event + * @e: actual event data to deliver to userspace + * + * This function prepares the passed in event for eventual delivery. If the event + * doesn't get delivered (because the IOCTL fails later on, before queuing up + * anything) then the even must be cancelled and freed using + * drm_event_cancel_free(). Successfully initialized events should be sent out + * using drm_send_event() or drm_send_event_locked() to signal completion of the + * asynchronous event to userspace. + * + * If callers embedded @p into a larger structure it must be allocated with + * kmalloc and @p must be the first member element. + * + * Callers which already hold dev->event_lock should use + * drm_event_reserve_init() instead. + * + * RETURNS: + * + * 0 on success or a negative error code on failure. + */ +int drm_event_reserve_init(struct drm_device *dev, + struct drm_file *file_priv, + struct drm_pending_event *p, + struct drm_event *e) +{ + unsigned long flags; + int ret; + + spin_lock_irqsave(&dev->event_lock, flags); + ret = drm_event_reserve_init_locked(dev, file_priv, p, e); + spin_unlock_irqrestore(&dev->event_lock, flags); + + return ret; +} +EXPORT_SYMBOL(drm_event_reserve_init); + +/** + * drm_event_cancel_free - free a DRM event and release it's space + * @dev: DRM device + * @p: tracking structure for the pending event + * + * This function frees the event @p initialized with drm_event_reserve_init() + * and releases any allocated space. + */ +void drm_event_cancel_free(struct drm_device *dev, + struct drm_pending_event *p) +{ + unsigned long flags; + spin_lock_irqsave(&dev->event_lock, flags); + if (p->file_priv) { + p->file_priv->event_space += p->event->length; + list_del(&p->pending_link); + } + spin_unlock_irqrestore(&dev->event_lock, flags); + p->destroy(p); +} +EXPORT_SYMBOL(drm_event_cancel_free); + +/** + * drm_send_event_locked - send DRM event to file descriptor + * @dev: DRM device + * @e: DRM event to deliver + * + * This function sends the event @e, initialized with drm_event_reserve_init(), + * to its associated userspace DRM file. Callers must already hold + * dev->event_lock, see drm_send_event() for the unlocked version. + * + * Note that the core will take care of unlinking and disarming events when the + * corresponding DRM file is closed. Drivers need not worry about whether the + * DRM file for this event still exists and can call this function upon + * completion of the asynchronous work unconditionally. + */ +void drm_send_event_locked(struct drm_device *dev, struct drm_pending_event *e) +{ + assert_spin_locked(&dev->event_lock); + + if (!e->file_priv) { + e->destroy(e); + return; + } + + list_del(&e->pending_link); + list_add_tail(&e->link, + &e->file_priv->event_list); + wake_up_interruptible(&e->file_priv->event_wait); +} +EXPORT_SYMBOL(drm_send_event_locked); + +/** + * drm_send_event - send DRM event to file descriptor + * @dev: DRM device + * @e: DRM event to deliver + * + * This function sends the event @e, initialized with drm_event_reserve_init(), + * to its associated userspace DRM file. This function acquires dev->event_lock, + * see drm_send_event_locked() for callers which already hold this lock. + * + * Note that the core will take care of unlinking and disarming events when the + * corresponding DRM file is closed. Drivers need not worry about whether the + * DRM file for this event still exists and can call this function upon + * completion of the asynchronous work unconditionally. + */ +void drm_send_event(struct drm_device *dev, struct drm_pending_event *e) +{ + unsigned long irqflags; + + spin_lock_irqsave(&dev->event_lock, irqflags); + drm_send_event_locked(dev, e); + spin_unlock_irqrestore(&dev->event_lock, irqflags); +} +EXPORT_SYMBOL(drm_send_event); diff --git a/drivers/gpu/drm/drm_irq.c b/drivers/gpu/drm/drm_irq.c index d12a4efa651b..96d03ac38ef7 100644 --- a/drivers/gpu/drm/drm_irq.c +++ b/drivers/gpu/drm/drm_irq.c @@ -983,15 +983,12 @@ static void send_vblank_event(struct drm_device *dev, struct drm_pending_vblank_event *e, unsigned long seq, struct timeval *now) { - assert_spin_locked(&dev->event_lock); - e->event.sequence = seq; e->event.tv_sec = now->tv_sec; e->event.tv_usec = now->tv_usec; - list_add_tail(&e->base.link, - &e->base.file_priv->event_list); - wake_up_interruptible(&e->base.file_priv->event_wait); + drm_send_event_locked(dev, &e->base); + trace_drm_vblank_event_delivered(e->base.pid, e->pipe, e->event.sequence); } @@ -1601,9 +1598,6 @@ static int drm_queue_vblank_event(struct drm_device *dev, unsigned int pipe, e->event.base.type = DRM_EVENT_VBLANK; e->event.base.length = sizeof(e->event); e->event.user_data = vblwait->request.signal; - e->base.event = &e->event.base; - e->base.file_priv = file_priv; - e->base.destroy = (void (*) (struct drm_pending_event *)) kfree; spin_lock_irqsave(&dev->event_lock, flags); @@ -1619,12 +1613,12 @@ static int drm_queue_vblank_event(struct drm_device *dev, unsigned int pipe, goto err_unlock; } - if (file_priv->event_space < sizeof(e->event)) { - ret = -EBUSY; + ret = drm_event_reserve_init_locked(dev, file_priv, &e->base, + &e->event.base); + + if (ret) goto err_unlock; - } - file_priv->event_space -= sizeof(e->event); seq = drm_vblank_count_and_time(dev, pipe, &now); if ((vblwait->request.type & _DRM_VBLANK_NEXTONMISS) && diff --git a/drivers/gpu/drm/drm_kms_helper_common.c b/drivers/gpu/drm/drm_kms_helper_common.c new file mode 100644 index 000000000000..3187c4bb01cb --- /dev/null +++ b/drivers/gpu/drm/drm_kms_helper_common.c @@ -0,0 +1,60 @@ +/* + * Copyright © 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Rafael Antognolli <[email protected]> + * + */ + +#include <drm/drmP.h> +#include <drm/drm_fb_helper.h> +#include <drm/drm_dp_aux_dev.h> + +MODULE_AUTHOR("David Airlie, Jesse Barnes"); +MODULE_DESCRIPTION("DRM KMS helper"); +MODULE_LICENSE("GPL and additional rights"); + +static int __init drm_kms_helper_init(void) +{ + int ret; + + /* Call init functions from specific kms helpers here */ + ret = drm_fb_helper_modinit(); + if (ret < 0) + goto out; + + ret = drm_dp_aux_dev_init(); + if (ret < 0) + goto out; + +out: + return ret; +} + +static void __exit drm_kms_helper_exit(void) +{ + /* Call exit functions from specific kms helpers here */ + drm_dp_aux_dev_exit(); +} + +module_init(drm_kms_helper_init); +module_exit(drm_kms_helper_exit); diff --git a/drivers/gpu/drm/drm_modes.c b/drivers/gpu/drm/drm_modes.c index 20775c05235a..f7448a5e95a9 100644 --- a/drivers/gpu/drm/drm_modes.c +++ b/drivers/gpu/drm/drm_modes.c @@ -1371,8 +1371,7 @@ bool drm_mode_parse_command_line_for_connector(const char *mode_option, } done: if (i >= 0) { - printk(KERN_WARNING - "parse error at position %i in video mode '%s'\n", + pr_warn("[drm] parse error at position %i in video mode '%s'\n", i, name); mode->specified = false; return false; diff --git a/drivers/gpu/drm/drm_prime.c b/drivers/gpu/drm/drm_prime.c index 27aa7183b20b..df6cdc76a16e 100644 --- a/drivers/gpu/drm/drm_prime.c +++ b/drivers/gpu/drm/drm_prime.c @@ -329,7 +329,7 @@ static const struct dma_buf_ops drm_gem_prime_dmabuf_ops = { * drm_gem_prime_export - helper library implementation of the export callback * @dev: drm_device to export from * @obj: GEM object to export - * @flags: flags like DRM_CLOEXEC + * @flags: flags like DRM_CLOEXEC and DRM_RDWR * * This is the implementation of the gem_prime_export functions for GEM drivers * using the PRIME helpers. @@ -628,7 +628,6 @@ int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv) { struct drm_prime_handle *args = data; - uint32_t flags; if (!drm_core_check_feature(dev, DRIVER_PRIME)) return -EINVAL; @@ -637,14 +636,11 @@ int drm_prime_handle_to_fd_ioctl(struct drm_device *dev, void *data, return -ENOSYS; /* check flags are valid */ - if (args->flags & ~DRM_CLOEXEC) + if (args->flags & ~(DRM_CLOEXEC | DRM_RDWR)) return -EINVAL; - /* we only want to pass DRM_CLOEXEC which is == O_CLOEXEC */ - flags = args->flags & DRM_CLOEXEC; - return dev->driver->prime_handle_to_fd(dev, file_priv, - args->handle, flags, &args->fd); + args->handle, args->flags, &args->fd); } int drm_prime_fd_to_handle_ioctl(struct drm_device *dev, void *data, diff --git a/drivers/gpu/drm/exynos/exynos_dp_core.c b/drivers/gpu/drm/exynos/exynos_dp_core.c index 673164b331c8..9fd12c621270 100644 --- a/drivers/gpu/drm/exynos/exynos_dp_core.c +++ b/drivers/gpu/drm/exynos/exynos_dp_core.c @@ -1155,13 +1155,6 @@ static int exynos_dp_create_connector(struct drm_encoder *encoder) return 0; } -static bool exynos_dp_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void exynos_dp_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -1177,7 +1170,6 @@ static void exynos_dp_disable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs exynos_dp_encoder_helper_funcs = { - .mode_fixup = exynos_dp_mode_fixup, .mode_set = exynos_dp_mode_set, .enable = exynos_dp_enable, .disable = exynos_dp_disable, diff --git a/drivers/gpu/drm/exynos/exynos_drm_dpi.c b/drivers/gpu/drm/exynos/exynos_drm_dpi.c index 05350ae0785b..75e570f45259 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_dpi.c +++ b/drivers/gpu/drm/exynos/exynos_drm_dpi.c @@ -128,13 +128,6 @@ static int exynos_dpi_create_connector(struct drm_encoder *encoder) return 0; } -static bool exynos_dpi_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void exynos_dpi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -162,7 +155,6 @@ static void exynos_dpi_disable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs exynos_dpi_encoder_helper_funcs = { - .mode_fixup = exynos_dpi_mode_fixup, .mode_set = exynos_dpi_mode_set, .enable = exynos_dpi_enable, .disable = exynos_dpi_disable, diff --git a/drivers/gpu/drm/exynos/exynos_drm_drv.c b/drivers/gpu/drm/exynos/exynos_drm_drv.c index 68f0f36f6e7e..1e535f981240 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_drv.c +++ b/drivers/gpu/drm/exynos/exynos_drm_drv.c @@ -340,20 +340,6 @@ static void exynos_drm_preclose(struct drm_device *dev, static void exynos_drm_postclose(struct drm_device *dev, struct drm_file *file) { - struct drm_pending_event *e, *et; - unsigned long flags; - - if (!file->driver_priv) - return; - - spin_lock_irqsave(&dev->event_lock, flags); - /* Release all events handled by page flip handler but not freed. */ - list_for_each_entry_safe(e, et, &file->event_list, link) { - list_del(&e->link); - e->destroy(e); - } - spin_unlock_irqrestore(&dev->event_lock, flags); - kfree(file->driver_priv); file->driver_priv = NULL; } diff --git a/drivers/gpu/drm/exynos/exynos_drm_dsi.c b/drivers/gpu/drm/exynos/exynos_drm_dsi.c index e977a81af2e6..736115c580fc 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_dsi.c +++ b/drivers/gpu/drm/exynos/exynos_drm_dsi.c @@ -1597,13 +1597,6 @@ static int exynos_dsi_create_connector(struct drm_encoder *encoder) return 0; } -static bool exynos_dsi_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void exynos_dsi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -1623,7 +1616,6 @@ static void exynos_dsi_mode_set(struct drm_encoder *encoder, } static const struct drm_encoder_helper_funcs exynos_dsi_encoder_helper_funcs = { - .mode_fixup = exynos_dsi_mode_fixup, .mode_set = exynos_dsi_mode_set, .enable = exynos_dsi_enable, .disable = exynos_dsi_disable, diff --git a/drivers/gpu/drm/exynos/exynos_drm_g2d.c b/drivers/gpu/drm/exynos/exynos_drm_g2d.c index c17efdb238a6..99369816ff97 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_g2d.c +++ b/drivers/gpu/drm/exynos/exynos_drm_g2d.c @@ -880,7 +880,6 @@ static void g2d_finish_event(struct g2d_data *g2d, u32 cmdlist_no) struct g2d_runqueue_node *runqueue_node = g2d->runqueue_node; struct drm_exynos_pending_g2d_event *e; struct timeval now; - unsigned long flags; if (list_empty(&runqueue_node->event_list)) return; @@ -893,10 +892,7 @@ static void g2d_finish_event(struct g2d_data *g2d, u32 cmdlist_no) e->event.tv_usec = now.tv_usec; e->event.cmdlist_no = cmdlist_no; - spin_lock_irqsave(&drm_dev->event_lock, flags); - list_move_tail(&e->base.link, &e->base.file_priv->event_list); - wake_up_interruptible(&e->base.file_priv->event_wait); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); + drm_send_event(drm_dev, &e->base); } static irqreturn_t g2d_irq_handler(int irq, void *dev_id) @@ -1072,7 +1068,6 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data, struct drm_exynos_pending_g2d_event *e; struct g2d_cmdlist_node *node; struct g2d_cmdlist *cmdlist; - unsigned long flags; int size; int ret; @@ -1094,21 +1089,8 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data, node->event = NULL; if (req->event_type != G2D_EVENT_NOT) { - spin_lock_irqsave(&drm_dev->event_lock, flags); - if (file->event_space < sizeof(e->event)) { - spin_unlock_irqrestore(&drm_dev->event_lock, flags); - ret = -ENOMEM; - goto err; - } - file->event_space -= sizeof(e->event); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); - e = kzalloc(sizeof(*node->event), GFP_KERNEL); if (!e) { - spin_lock_irqsave(&drm_dev->event_lock, flags); - file->event_space += sizeof(e->event); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); - ret = -ENOMEM; goto err; } @@ -1116,9 +1098,12 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data, e->event.base.type = DRM_EXYNOS_G2D_EVENT; e->event.base.length = sizeof(e->event); e->event.user_data = req->user_data; - e->base.event = &e->event.base; - e->base.file_priv = file; - e->base.destroy = (void (*) (struct drm_pending_event *)) kfree; + + ret = drm_event_reserve_init(drm_dev, file, &e->base, &e->event.base); + if (ret) { + kfree(e); + goto err; + } node->event = e; } @@ -1219,12 +1204,8 @@ int exynos_g2d_set_cmdlist_ioctl(struct drm_device *drm_dev, void *data, err_unmap: g2d_unmap_cmdlist_gem(g2d, node, file); err_free_event: - if (node->event) { - spin_lock_irqsave(&drm_dev->event_lock, flags); - file->event_space += sizeof(e->event); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); - kfree(node->event); - } + if (node->event) + drm_event_cancel_free(drm_dev, &node->event->base); err: g2d_put_cmdlist(g2d, node); return ret; diff --git a/drivers/gpu/drm/exynos/exynos_drm_ipp.c b/drivers/gpu/drm/exynos/exynos_drm_ipp.c index 67d24236e745..3eab0d15f0b4 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_ipp.c +++ b/drivers/gpu/drm/exynos/exynos_drm_ipp.c @@ -618,27 +618,18 @@ static void ipp_clean_mem_nodes(struct drm_device *drm_dev, mutex_unlock(&c_node->mem_lock); } -static void ipp_free_event(struct drm_pending_event *event) -{ - kfree(event); -} - static int ipp_get_event(struct drm_device *drm_dev, struct drm_exynos_ipp_cmd_node *c_node, struct drm_exynos_ipp_queue_buf *qbuf) { struct drm_exynos_ipp_send_event *e; - unsigned long flags; + int ret; DRM_DEBUG_KMS("ops_id[%d]buf_id[%d]\n", qbuf->ops_id, qbuf->buf_id); e = kzalloc(sizeof(*e), GFP_KERNEL); - if (!e) { - spin_lock_irqsave(&drm_dev->event_lock, flags); - c_node->filp->event_space += sizeof(e->event); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); + if (!e) return -ENOMEM; - } /* make event */ e->event.base.type = DRM_EXYNOS_IPP_EVENT; @@ -646,9 +637,13 @@ static int ipp_get_event(struct drm_device *drm_dev, e->event.user_data = qbuf->user_data; e->event.prop_id = qbuf->prop_id; e->event.buf_id[EXYNOS_DRM_OPS_DST] = qbuf->buf_id; - e->base.event = &e->event.base; - e->base.file_priv = c_node->filp; - e->base.destroy = ipp_free_event; + + ret = drm_event_reserve_init(drm_dev, c_node->filp, &e->base, &e->event.base); + if (ret) { + kfree(e); + return ret; + } + mutex_lock(&c_node->event_lock); list_add_tail(&e->base.link, &c_node->event_list); mutex_unlock(&c_node->event_lock); @@ -1412,7 +1407,6 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv, struct drm_exynos_ipp_send_event *e; struct list_head *head; struct timeval now; - unsigned long flags; u32 tbuf_id[EXYNOS_DRM_OPS_MAX] = {0, }; int ret, i; @@ -1525,10 +1519,7 @@ static int ipp_send_event(struct exynos_drm_ippdrv *ippdrv, for_each_ipp_ops(i) e->event.buf_id[i] = tbuf_id[i]; - spin_lock_irqsave(&drm_dev->event_lock, flags); - list_move_tail(&e->base.link, &e->base.file_priv->event_list); - wake_up_interruptible(&e->base.file_priv->event_wait); - spin_unlock_irqrestore(&drm_dev->event_lock, flags); + drm_send_event(drm_dev, &e->base); mutex_unlock(&c_node->event_lock); DRM_DEBUG_KMS("done cmd[%d]prop_id[%d]buf_id[%d]\n", diff --git a/drivers/gpu/drm/exynos/exynos_drm_vidi.c b/drivers/gpu/drm/exynos/exynos_drm_vidi.c index 62ac4e5fa51d..65108cbd79d4 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_vidi.c +++ b/drivers/gpu/drm/exynos/exynos_drm_vidi.c @@ -410,13 +410,6 @@ static int vidi_create_connector(struct drm_encoder *encoder) return 0; } -static bool exynos_vidi_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void exynos_vidi_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -432,7 +425,6 @@ static void exynos_vidi_disable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs exynos_vidi_encoder_helper_funcs = { - .mode_fixup = exynos_vidi_mode_fixup, .mode_set = exynos_vidi_mode_set, .enable = exynos_vidi_enable, .disable = exynos_vidi_disable, diff --git a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c index fca97d3fc846..9648b7f9a31c 100644 --- a/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c +++ b/drivers/gpu/drm/fsl-dcu/fsl_dcu_drm_drv.c @@ -112,10 +112,6 @@ static int fsl_dcu_unload(struct drm_device *dev) return 0; } -static void fsl_dcu_drm_preclose(struct drm_device *dev, struct drm_file *file) -{ -} - static irqreturn_t fsl_dcu_drm_irq(int irq, void *arg) { struct drm_device *dev = arg; @@ -191,7 +187,6 @@ static struct drm_driver fsl_dcu_drm_driver = { | DRIVER_PRIME | DRIVER_ATOMIC, .load = fsl_dcu_load, .unload = fsl_dcu_unload, - .preclose = fsl_dcu_drm_preclose, .irq_handler = fsl_dcu_drm_irq, .get_vblank_counter = drm_vblank_no_hw_counter, .enable_vblank = fsl_dcu_drm_enable_vblank, diff --git a/drivers/gpu/drm/gma500/cdv_intel_crt.c b/drivers/gpu/drm/gma500/cdv_intel_crt.c index d0717a85c7ec..b837e7a92196 100644 --- a/drivers/gpu/drm/gma500/cdv_intel_crt.c +++ b/drivers/gpu/drm/gma500/cdv_intel_crt.c @@ -217,7 +217,6 @@ static int cdv_intel_crt_set_property(struct drm_connector *connector, static const struct drm_encoder_helper_funcs cdv_intel_crt_helper_funcs = { .dpms = cdv_intel_crt_dpms, - .mode_fixup = gma_encoder_mode_fixup, .prepare = gma_encoder_prepare, .commit = gma_encoder_commit, .mode_set = cdv_intel_crt_mode_set, diff --git a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c index ddf2d7700759..28f9d90988ff 100644 --- a/drivers/gpu/drm/gma500/cdv_intel_hdmi.c +++ b/drivers/gpu/drm/gma500/cdv_intel_hdmi.c @@ -255,7 +255,6 @@ static void cdv_hdmi_destroy(struct drm_connector *connector) static const struct drm_encoder_helper_funcs cdv_hdmi_helper_funcs = { .dpms = cdv_hdmi_dpms, - .mode_fixup = gma_encoder_mode_fixup, .prepare = gma_encoder_prepare, .mode_set = cdv_hdmi_mode_set, .commit = gma_encoder_commit, diff --git a/drivers/gpu/drm/gma500/framebuffer.c b/drivers/gpu/drm/gma500/framebuffer.c index cb95765050cc..033d894d030e 100644 --- a/drivers/gpu/drm/gma500/framebuffer.c +++ b/drivers/gpu/drm/gma500/framebuffer.c @@ -674,29 +674,17 @@ static const struct drm_mode_config_funcs psb_mode_funcs = { .output_poll_changed = psbfb_output_poll_changed, }; -static int psb_create_backlight_property(struct drm_device *dev) -{ - struct drm_psb_private *dev_priv = dev->dev_private; - struct drm_property *backlight; - - if (dev_priv->backlight_property) - return 0; - - backlight = drm_property_create_range(dev, 0, "backlight", 0, 100); - - dev_priv->backlight_property = backlight; - - return 0; -} - static void psb_setup_outputs(struct drm_device *dev) { struct drm_psb_private *dev_priv = dev->dev_private; struct drm_connector *connector; drm_mode_create_scaling_mode_property(dev); - psb_create_backlight_property(dev); + /* It is ok for this to fail - we just don't get backlight control */ + if (!dev_priv->backlight_property) + dev_priv->backlight_property = drm_property_create_range(dev, 0, + "backlight", 0, 100); dev_priv->ops->output_init(dev); list_for_each_entry(connector, &dev->mode_config.connector_list, diff --git a/drivers/gpu/drm/gma500/gma_display.c b/drivers/gpu/drm/gma500/gma_display.c index ff17af4cfc64..927082148d4d 100644 --- a/drivers/gpu/drm/gma500/gma_display.c +++ b/drivers/gpu/drm/gma500/gma_display.c @@ -478,13 +478,6 @@ int gma_crtc_cursor_move(struct drm_crtc *crtc, int x, int y) return 0; } -bool gma_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - bool gma_crtc_mode_fixup(struct drm_crtc *crtc, const struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) diff --git a/drivers/gpu/drm/gma500/gma_display.h b/drivers/gpu/drm/gma500/gma_display.h index ed569d8a6af3..78b9f986a6e5 100644 --- a/drivers/gpu/drm/gma500/gma_display.h +++ b/drivers/gpu/drm/gma500/gma_display.h @@ -90,9 +90,6 @@ extern void gma_crtc_restore(struct drm_crtc *crtc); extern void gma_encoder_prepare(struct drm_encoder *encoder); extern void gma_encoder_commit(struct drm_encoder *encoder); extern void gma_encoder_destroy(struct drm_encoder *encoder); -extern bool gma_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode); /* Common clock related functions */ extern const struct gma_limit_t *gma_limit(struct drm_crtc *crtc, int refclk); diff --git a/drivers/gpu/drm/gma500/intel_gmbus.c b/drivers/gpu/drm/gma500/intel_gmbus.c index 566d330aaeea..e7e22187c539 100644 --- a/drivers/gpu/drm/gma500/intel_gmbus.c +++ b/drivers/gpu/drm/gma500/intel_gmbus.c @@ -436,7 +436,7 @@ int gma_intel_setup_gmbus(struct drm_device *dev) return 0; err: - while (--i) { + while (i--) { struct intel_gmbus *bus = &dev_priv->gmbus[i]; i2c_del_adapter(&bus->adapter); } diff --git a/drivers/gpu/drm/gma500/mdfld_dsi_output.c b/drivers/gpu/drm/gma500/mdfld_dsi_output.c index d758f4cc6805..907cb51795c3 100644 --- a/drivers/gpu/drm/gma500/mdfld_dsi_output.c +++ b/drivers/gpu/drm/gma500/mdfld_dsi_output.c @@ -382,16 +382,6 @@ static int mdfld_dsi_connector_mode_valid(struct drm_connector *connector, return MODE_OK; } -static void mdfld_dsi_connector_dpms(struct drm_connector *connector, int mode) -{ - if (mode == connector->dpms) - return; - - /*first, execute dpms*/ - - drm_helper_connector_dpms(connector, mode); -} - static struct drm_encoder *mdfld_dsi_connector_best_encoder( struct drm_connector *connector) { @@ -404,7 +394,7 @@ static struct drm_encoder *mdfld_dsi_connector_best_encoder( /*DSI connector funcs*/ static const struct drm_connector_funcs mdfld_dsi_connector_funcs = { - .dpms = /*drm_helper_connector_dpms*/mdfld_dsi_connector_dpms, + .dpms = drm_helper_connector_dpms, .detect = mdfld_dsi_connector_detect, .fill_modes = drm_helper_probe_single_connector_modes, .set_property = mdfld_dsi_connector_set_property, diff --git a/drivers/gpu/drm/gma500/oaktrail_hdmi.c b/drivers/gpu/drm/gma500/oaktrail_hdmi.c index 2d18499d6060..8b2eb32ee988 100644 --- a/drivers/gpu/drm/gma500/oaktrail_hdmi.c +++ b/drivers/gpu/drm/gma500/oaktrail_hdmi.c @@ -601,7 +601,6 @@ static void oaktrail_hdmi_destroy(struct drm_connector *connector) static const struct drm_encoder_helper_funcs oaktrail_hdmi_helper_funcs = { .dpms = oaktrail_hdmi_dpms, - .mode_fixup = gma_encoder_mode_fixup, .prepare = gma_encoder_prepare, .mode_set = oaktrail_hdmi_mode_set, .commit = gma_encoder_commit, diff --git a/drivers/gpu/drm/gma500/psb_drv.c b/drivers/gpu/drm/gma500/psb_drv.c index 92e7e5795398..4e1c6850520e 100644 --- a/drivers/gpu/drm/gma500/psb_drv.c +++ b/drivers/gpu/drm/gma500/psb_drv.c @@ -442,14 +442,6 @@ static long psb_unlocked_ioctl(struct file *filp, unsigned int cmd, /* FIXME: do we need to wrap the other side of this */ } -/* - * When a client dies: - * - Check for and clean up flipped page state - */ -static void psb_driver_preclose(struct drm_device *dev, struct drm_file *priv) -{ -} - static int psb_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) { return drm_get_pci_dev(pdev, ent, &driver); @@ -495,7 +487,6 @@ static struct drm_driver driver = { .load = psb_driver_load, .unload = psb_driver_unload, .lastclose = psb_driver_lastclose, - .preclose = psb_driver_preclose, .set_busid = drm_pci_set_busid, .num_ioctls = ARRAY_SIZE(psb_ioctls), diff --git a/drivers/gpu/drm/i2c/ch7006_drv.c b/drivers/gpu/drm/i2c/ch7006_drv.c index 90db5f4dcce5..0594c45f7164 100644 --- a/drivers/gpu/drm/i2c/ch7006_drv.c +++ b/drivers/gpu/drm/i2c/ch7006_drv.c @@ -253,6 +253,8 @@ static int ch7006_encoder_create_resources(struct drm_encoder *encoder, drm_mode_create_tv_properties(dev, NUM_TV_NORMS, ch7006_tv_norm_names); priv->scale_property = drm_property_create_range(dev, 0, "scale", 0, 2); + if (!priv->scale_property) + return -ENOMEM; drm_object_attach_property(&connector->base, conf->tv_select_subconnector_property, priv->select_subconnector); diff --git a/drivers/gpu/drm/i2c/sil164_drv.c b/drivers/gpu/drm/i2c/sil164_drv.c index c400428f6c8c..db0b03fb0ff1 100644 --- a/drivers/gpu/drm/i2c/sil164_drv.c +++ b/drivers/gpu/drm/i2c/sil164_drv.c @@ -252,14 +252,6 @@ sil164_encoder_restore(struct drm_encoder *encoder) priv->saved_slave_state); } -static bool -sil164_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static int sil164_encoder_mode_valid(struct drm_encoder *encoder, struct drm_display_mode *mode) @@ -347,7 +339,6 @@ static const struct drm_encoder_slave_funcs sil164_encoder_funcs = { .dpms = sil164_encoder_dpms, .save = sil164_encoder_save, .restore = sil164_encoder_restore, - .mode_fixup = sil164_encoder_mode_fixup, .mode_valid = sil164_encoder_mode_valid, .mode_set = sil164_encoder_mode_set, .detect = sil164_encoder_detect, diff --git a/drivers/gpu/drm/i2c/tda998x_drv.c b/drivers/gpu/drm/i2c/tda998x_drv.c index 34e38749a817..b61282d89aa3 100644 --- a/drivers/gpu/drm/i2c/tda998x_drv.c +++ b/drivers/gpu/drm/i2c/tda998x_drv.c @@ -856,14 +856,6 @@ static void tda998x_encoder_dpms(struct drm_encoder *encoder, int mode) priv->dpms = mode; } -static bool -tda998x_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static int tda998x_connector_mode_valid(struct drm_connector *connector, struct drm_display_mode *mode) { @@ -1343,7 +1335,6 @@ static void tda998x_encoder_commit(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs tda998x_encoder_helper_funcs = { .dpms = tda998x_encoder_dpms, - .mode_fixup = tda998x_encoder_mode_fixup, .prepare = tda998x_encoder_prepare, .commit = tda998x_encoder_commit, .mode_set = tda998x_encoder_mode_set, diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c index 0fc38bb7276c..c5db23511184 100644 --- a/drivers/gpu/drm/i915/i915_debugfs.c +++ b/drivers/gpu/drm/i915/i915_debugfs.c @@ -1331,7 +1331,8 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) struct intel_engine_cs *ring; u64 acthd[I915_NUM_RINGS]; u32 seqno[I915_NUM_RINGS]; - int i; + u32 instdone[I915_NUM_INSTDONE_REG]; + int i, j; if (!i915.enable_hangcheck) { seq_printf(m, "Hangcheck disabled\n"); @@ -1345,6 +1346,8 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) acthd[i] = intel_ring_get_active_head(ring); } + i915_get_extra_instdone(dev, instdone); + intel_runtime_pm_put(dev_priv); if (delayed_work_pending(&dev_priv->gpu_error.hangcheck_work)) { @@ -1365,6 +1368,21 @@ static int i915_hangcheck_info(struct seq_file *m, void *unused) (long long)ring->hangcheck.max_acthd); seq_printf(m, "\tscore = %d\n", ring->hangcheck.score); seq_printf(m, "\taction = %d\n", ring->hangcheck.action); + + if (ring->id == RCS) { + seq_puts(m, "\tinstdone read ="); + + for (j = 0; j < I915_NUM_INSTDONE_REG; j++) + seq_printf(m, " 0x%08x", instdone[j]); + + seq_puts(m, "\n\tinstdone accu ="); + + for (j = 0; j < I915_NUM_INSTDONE_REG; j++) + seq_printf(m, " 0x%08x", + ring->hangcheck.instdone[j]); + + seq_puts(m, "\n"); + } } return 0; @@ -1942,11 +1960,8 @@ static int i915_context_status(struct seq_file *m, void *unused) seq_puts(m, "HW context "); describe_ctx(m, ctx); - for_each_ring(ring, dev_priv, i) { - if (ring->default_context == ctx) - seq_printf(m, "(default context %s) ", - ring->name); - } + if (ctx == dev_priv->kernel_context) + seq_printf(m, "(kernel context) "); if (i915.enable_execlists) { seq_putc(m, '\n'); @@ -1976,12 +1991,13 @@ static int i915_context_status(struct seq_file *m, void *unused) } static void i915_dump_lrc_obj(struct seq_file *m, - struct intel_engine_cs *ring, - struct drm_i915_gem_object *ctx_obj) + struct intel_context *ctx, + struct intel_engine_cs *ring) { struct page *page; uint32_t *reg_state; int j; + struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state; unsigned long ggtt_offset = 0; if (ctx_obj == NULL) { @@ -1991,7 +2007,7 @@ static void i915_dump_lrc_obj(struct seq_file *m, } seq_printf(m, "CONTEXT: %s %u\n", ring->name, - intel_execlists_ctx_id(ctx_obj)); + intel_execlists_ctx_id(ctx, ring)); if (!i915_gem_obj_ggtt_bound(ctx_obj)) seq_puts(m, "\tNot bound in GGTT\n"); @@ -2037,13 +2053,10 @@ static int i915_dump_lrc(struct seq_file *m, void *unused) if (ret) return ret; - list_for_each_entry(ctx, &dev_priv->context_list, link) { - for_each_ring(ring, dev_priv, i) { - if (ring->default_context != ctx) - i915_dump_lrc_obj(m, ring, - ctx->engine[i].state); - } - } + list_for_each_entry(ctx, &dev_priv->context_list, link) + if (ctx != dev_priv->kernel_context) + for_each_ring(ring, dev_priv, i) + i915_dump_lrc_obj(m, ctx, ring); mutex_unlock(&dev->struct_mutex); @@ -2092,13 +2105,13 @@ static int i915_execlists(struct seq_file *m, void *data) seq_printf(m, "\tStatus pointer: 0x%08X\n", status_pointer); read_pointer = ring->next_context_status_buffer; - write_pointer = status_pointer & 0x07; + write_pointer = GEN8_CSB_WRITE_PTR(status_pointer); if (read_pointer > write_pointer) - write_pointer += 6; + write_pointer += GEN8_CSB_ENTRIES; seq_printf(m, "\tRead pointer: 0x%08X, write pointer 0x%08X\n", read_pointer, write_pointer); - for (i = 0; i < 6; i++) { + for (i = 0; i < GEN8_CSB_ENTRIES; i++) { status = I915_READ(RING_CONTEXT_STATUS_BUF_LO(ring, i)); ctx_id = I915_READ(RING_CONTEXT_STATUS_BUF_HI(ring, i)); @@ -2115,11 +2128,8 @@ static int i915_execlists(struct seq_file *m, void *data) seq_printf(m, "\t%d requests in queue\n", count); if (head_req) { - struct drm_i915_gem_object *ctx_obj; - - ctx_obj = head_req->ctx->engine[ring_id].state; seq_printf(m, "\tHead request id: %u\n", - intel_execlists_ctx_id(ctx_obj)); + intel_execlists_ctx_id(head_req->ctx, ring)); seq_printf(m, "\tHead request tail: %u\n", head_req->tail); } diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c index d70d96fe553b..1c3d2544fec4 100644 --- a/drivers/gpu/drm/i915/i915_dma.c +++ b/drivers/gpu/drm/i915/i915_dma.c @@ -1261,8 +1261,6 @@ void i915_driver_preclose(struct drm_device *dev, struct drm_file *file) i915_gem_context_close(dev, file); i915_gem_release(dev, file); mutex_unlock(&dev->struct_mutex); - - intel_modeset_preclose(dev, file); } void i915_driver_postclose(struct drm_device *dev, struct drm_file *file) diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index f357058c74d9..44912ecebc1a 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -35,9 +35,12 @@ #include "i915_trace.h" #include "intel_drv.h" +#include <linux/apple-gmux.h> #include <linux/console.h> #include <linux/module.h> #include <linux/pm_runtime.h> +#include <linux/vgaarb.h> +#include <linux/vga_switcheroo.h> #include <drm/drm_crtc_helper.h> static struct drm_driver driver; @@ -969,6 +972,15 @@ static int i915_pci_probe(struct pci_dev *pdev, const struct pci_device_id *ent) if (PCI_FUNC(pdev->devfn)) return -ENODEV; + /* + * apple-gmux is needed on dual GPU MacBook Pro + * to probe the panel if we're the inactive GPU. + */ + if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) && + apple_gmux_present() && pdev != vga_default_device() && + !vga_switcheroo_handler_flags()) + return -EPROBE_DEFER; + return drm_get_pci_dev(pdev, ent, &driver); } @@ -1079,7 +1091,6 @@ static int bxt_resume_prepare(struct drm_i915_private *dev_priv) */ broxton_init_cdclk(dev); broxton_ddi_phy_init(dev); - intel_prepare_ddi(dev); return 0; } @@ -1338,8 +1349,8 @@ static int vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv, return 0; DRM_DEBUG_KMS("waiting for GT wells to go %s (%08x)\n", - wait_for_on ? "on" : "off", - I915_READ(VLV_GTLC_PW_STATUS)); + onoff(wait_for_on), + I915_READ(VLV_GTLC_PW_STATUS)); /* * RC6 transitioning can be delayed up to 2 msec (see @@ -1348,7 +1359,7 @@ static int vlv_wait_for_gt_wells(struct drm_i915_private *dev_priv, err = wait_for(COND, 3); if (err) DRM_ERROR("timeout waiting for GT wells to go %s\n", - wait_for_on ? "on" : "off"); + onoff(wait_for_on)); return err; #undef COND @@ -1359,7 +1370,7 @@ static void vlv_check_no_gt_access(struct drm_i915_private *dev_priv) if (!(I915_READ(VLV_GTLC_PW_STATUS) & VLV_GTLC_ALLOWWAKEERR)) return; - DRM_ERROR("GT register access while GT waking disabled\n"); + DRM_DEBUG_DRIVER("GT register access while GT waking disabled\n"); I915_WRITE(VLV_GTLC_PW_STATUS, VLV_GTLC_ALLOWWAKEERR); } @@ -1503,6 +1514,10 @@ static int intel_runtime_suspend(struct device *device) enable_rpm_wakeref_asserts(dev_priv); WARN_ON_ONCE(atomic_read(&dev_priv->pm.wakeref_count)); + + if (intel_uncore_arm_unclaimed_mmio_detection(dev_priv)) + DRM_ERROR("Unclaimed access detected prior to suspending\n"); + dev_priv->pm.suspended = true; /* @@ -1551,6 +1566,8 @@ static int intel_runtime_resume(struct device *device) intel_opregion_notify_adapter(dev, PCI_D0); dev_priv->pm.suspended = false; + if (intel_uncore_unclaimed_mmio(dev_priv)) + DRM_DEBUG_DRIVER("Unclaimed access during suspend, bios?\n"); intel_guc_resume(dev); diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index f0f75d7c0d94..afb0beee9975 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -34,6 +34,7 @@ #include <uapi/drm/drm_fourcc.h> #include <drm/drmP.h> +#include "i915_params.h" #include "i915_reg.h" #include "intel_bios.h" #include "intel_ringbuffer.h" @@ -58,7 +59,7 @@ #define DRIVER_NAME "i915" #define DRIVER_DESC "Intel Graphics" -#define DRIVER_DATE "20151218" +#define DRIVER_DATE "20160124" #undef WARN_ON /* Many gcc seem to no see through this and fall over :( */ @@ -69,11 +70,11 @@ BUILD_BUG_ON(__i915_warn_cond); \ WARN(__i915_warn_cond, "WARN_ON(" #x ")"); }) #else -#define WARN_ON(x) WARN((x), "WARN_ON(%s)", #x ) +#define WARN_ON(x) WARN((x), "%s", "WARN_ON(" __stringify(x) ")") #endif #undef WARN_ON_ONCE -#define WARN_ON_ONCE(x) WARN_ONCE((x), "WARN_ON_ONCE(%s)", #x ) +#define WARN_ON_ONCE(x) WARN_ONCE((x), "%s", "WARN_ON_ONCE(" __stringify(x) ")") #define MISSING_CASE(x) WARN(1, "Missing switch case (%lu) in %s\n", \ (long) (x), __func__); @@ -87,31 +88,25 @@ */ #define I915_STATE_WARN(condition, format...) ({ \ int __ret_warn_on = !!(condition); \ - if (unlikely(__ret_warn_on)) { \ - if (i915.verbose_state_checks) \ - WARN(1, format); \ - else \ + if (unlikely(__ret_warn_on)) \ + if (!WARN(i915.verbose_state_checks, format)) \ DRM_ERROR(format); \ - } \ unlikely(__ret_warn_on); \ }) -#define I915_STATE_WARN_ON(condition) ({ \ - int __ret_warn_on = !!(condition); \ - if (unlikely(__ret_warn_on)) { \ - if (i915.verbose_state_checks) \ - WARN(1, "WARN_ON(" #condition ")\n"); \ - else \ - DRM_ERROR("WARN_ON(" #condition ")\n"); \ - } \ - unlikely(__ret_warn_on); \ -}) +#define I915_STATE_WARN_ON(x) \ + I915_STATE_WARN((x), "%s", "WARN_ON(" __stringify(x) ")") static inline const char *yesno(bool v) { return v ? "yes" : "no"; } +static inline const char *onoff(bool v) +{ + return v ? "on" : "off"; +} + enum pipe { INVALID_PIPE = -1, PIPE_A = 0, @@ -339,7 +334,7 @@ struct drm_i915_file_private { unsigned boosts; } rps; - struct intel_engine_cs *bsd_ring; + unsigned int bsd_ring; }; enum intel_dpll_id { @@ -633,6 +628,7 @@ struct drm_i915_display_funcs { struct dpll *best_clock); int (*compute_pipe_wm)(struct intel_crtc *crtc, struct drm_atomic_state *state); + void (*program_watermarks)(struct intel_crtc_state *cstate); void (*update_wm)(struct drm_crtc *crtc); int (*modeset_calc_cdclk)(struct drm_atomic_state *state); void (*modeset_commit_cdclk)(struct drm_atomic_state *state); @@ -657,9 +653,6 @@ struct drm_i915_display_funcs { struct drm_i915_gem_object *obj, struct drm_i915_gem_request *req, uint32_t flags); - void (*update_primary_plane)(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y); void (*hpd_irq_setup)(struct drm_device *dev); /* clock updates for mode set */ /* cursor updates */ @@ -726,6 +719,8 @@ struct intel_uncore { i915_reg_t reg_post; u32 val_reset; } fw_domain[FW_DOMAIN_ID_COUNT]; + + int unclaimed_mmio_check; }; /* Iterate over initialised fw domains */ @@ -889,6 +884,9 @@ struct intel_context { struct drm_i915_gem_object *state; struct intel_ringbuffer *ringbuf; int pin_count; + struct i915_vma *lrc_vma; + u64 lrc_desc; + uint32_t *lrc_reg_state; } engine[I915_NUM_RINGS]; struct list_head link; @@ -1301,7 +1299,7 @@ struct i915_gem_mm { bool busy; /* the indicator for dispatch video commands on two BSD rings */ - int bsd_ring_dispatch_index; + unsigned int bsd_ring_dispatch_index; /** Bit 6 swizzling required for X tiling */ uint32_t bit_6_swizzle_x; @@ -1487,7 +1485,7 @@ struct intel_vbt_data { u8 seq_version; u32 size; u8 *data; - u8 *sequence[MIPI_SEQ_MAX]; + const u8 *sequence[MIPI_SEQ_MAX]; } dsi; int crt_ddc_pin; @@ -1784,7 +1782,7 @@ struct drm_i915_private { unsigned int fsb_freq, mem_freq, is_ddr3; unsigned int skl_boot_cdclk; - unsigned int cdclk_freq, max_cdclk_freq; + unsigned int cdclk_freq, max_cdclk_freq, atomic_cdclk_freq; unsigned int max_dotclk_freq; unsigned int hpll_freq; unsigned int czclk_freq; @@ -1829,8 +1827,13 @@ struct drm_i915_private { struct intel_pipe_crc pipe_crc[I915_MAX_PIPES]; #endif + /* dpll and cdclk state is protected by connection_mutex */ int num_shared_dpll; struct intel_shared_dpll shared_dplls[I915_NUM_PLLS]; + + unsigned int active_crtcs; + unsigned int min_pixclk[I915_MAX_PIPES]; + int dpio_phy_iosf_port[I915_NUM_PHYS_VLV]; struct i915_workarounds workarounds; @@ -1945,6 +1948,8 @@ struct drm_i915_private { void (*stop_ring)(struct intel_engine_cs *ring); } gt; + struct intel_context *kernel_context; + bool edp_low_vswing; /* perform PHY state sanity checks? */ @@ -2265,9 +2270,9 @@ struct drm_i915_gem_request { }; -int i915_gem_request_alloc(struct intel_engine_cs *ring, - struct intel_context *ctx, - struct drm_i915_gem_request **req_out); +struct drm_i915_gem_request * __must_check +i915_gem_request_alloc(struct intel_engine_cs *engine, + struct intel_context *ctx); void i915_gem_request_cancel(struct drm_i915_gem_request *req); void i915_gem_request_free(struct kref *req_ref); int i915_gem_request_add_to_client(struct drm_i915_gem_request *req, @@ -2576,6 +2581,11 @@ struct drm_i915_cmd_table { /* Early gen2 have a totally busted CS tlb and require pinned batches. */ #define HAS_BROKEN_CS_TLB(dev) (IS_I830(dev) || IS_845G(dev)) + +/* WaRsDisableCoarsePowerGating:skl,bxt */ +#define NEEDS_WaRsDisableCoarsePowerGating(dev) (IS_BXT_REVID(dev, 0, BXT_REVID_A1) || \ + ((IS_SKL_GT3(dev) || IS_SKL_GT4(dev)) && \ + IS_SKL_REVID(dev, 0, SKL_REVID_F0))) /* * dp aux and gmbus irq on gen4 seems to be able to generate legacy interrupts * even when in MSI mode. This results in spurious interrupt warnings if the @@ -2665,44 +2675,7 @@ extern int i915_max_ioctl; extern int i915_suspend_switcheroo(struct drm_device *dev, pm_message_t state); extern int i915_resume_switcheroo(struct drm_device *dev); -/* i915_params.c */ -struct i915_params { - int modeset; - int panel_ignore_lid; - int semaphores; - int lvds_channel_mode; - int panel_use_ssc; - int vbt_sdvo_panel_type; - int enable_rc6; - int enable_dc; - int enable_fbc; - int enable_ppgtt; - int enable_execlists; - int enable_psr; - unsigned int preliminary_hw_support; - int disable_power_well; - int enable_ips; - int invert_brightness; - int enable_cmd_parser; - /* leave bools at the end to not create holes */ - bool enable_hangcheck; - bool fastboot; - bool prefault_disable; - bool load_detect_test; - bool reset; - bool disable_display; - bool disable_vtd_wa; - bool enable_guc_submission; - int guc_log_level; - int use_mmio_flip; - int mmio_debug; - bool verbose_state_checks; - bool nuclear_pageflip; - int edp_vswing; -}; -extern struct i915_params i915 __read_mostly; - - /* i915_dma.c */ +/* i915_dma.c */ extern int i915_driver_load(struct drm_device *, unsigned long flags); extern int i915_driver_unload(struct drm_device *); extern int i915_driver_open(struct drm_device *dev, struct drm_file *file); @@ -2745,7 +2718,8 @@ extern void intel_uncore_sanitize(struct drm_device *dev); extern void intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake); extern void intel_uncore_init(struct drm_device *dev); -extern void intel_uncore_check_errors(struct drm_device *dev); +extern bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv); +extern bool intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv); extern void intel_uncore_fini(struct drm_device *dev); extern void intel_uncore_forcewake_reset(struct drm_device *dev, bool restore); const char *intel_uncore_forcewake_domain_to_str(const enum forcewake_domain_id id); diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index ddc21d4b388d..371bbb28c471 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1251,7 +1251,7 @@ int __i915_wait_request(struct drm_i915_gem_request *req, int state = interruptible ? TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE; DEFINE_WAIT(wait); unsigned long timeout_expire; - s64 before, now; + s64 before = 0; /* Only to silence a compiler warning. */ int ret; WARN(!intel_irqs_enabled(dev_priv), "IRQs disabled"); @@ -1271,14 +1271,17 @@ int __i915_wait_request(struct drm_i915_gem_request *req, return -ETIME; timeout_expire = jiffies + nsecs_to_jiffies_timeout(*timeout); + + /* + * Record current time in case interrupted by signal, or wedged. + */ + before = ktime_get_raw_ns(); } if (INTEL_INFO(dev_priv)->gen >= 6) gen6_rps_boost(dev_priv, rps, req->emitted_jiffies); - /* Record current time in case interrupted by signal, or wedged */ trace_i915_gem_request_wait_begin(req); - before = ktime_get_raw_ns(); /* Optimistic spin for the next jiffie before touching IRQs */ ret = __i915_spin_request(req, state); @@ -1343,11 +1346,10 @@ int __i915_wait_request(struct drm_i915_gem_request *req, finish_wait(&ring->irq_queue, &wait); out: - now = ktime_get_raw_ns(); trace_i915_gem_request_wait_end(req); if (timeout) { - s64 tres = *timeout - (now - before); + s64 tres = *timeout - (ktime_get_raw_ns() - before); *timeout = tres < 0 ? 0 : tres; @@ -2677,10 +2679,8 @@ void i915_gem_request_free(struct kref *req_ref) i915_gem_request_remove_from_client(req); if (ctx) { - if (i915.enable_execlists) { - if (ctx != req->ring->default_context) - intel_lr_context_unpin(req); - } + if (i915.enable_execlists && ctx != req->i915->kernel_context) + intel_lr_context_unpin(req); i915_gem_context_unreference(ctx); } @@ -2688,9 +2688,10 @@ void i915_gem_request_free(struct kref *req_ref) kmem_cache_free(req->i915->requests, req); } -int i915_gem_request_alloc(struct intel_engine_cs *ring, - struct intel_context *ctx, - struct drm_i915_gem_request **req_out) +static inline int +__i915_gem_request_alloc(struct intel_engine_cs *ring, + struct intel_context *ctx, + struct drm_i915_gem_request **req_out) { struct drm_i915_private *dev_priv = to_i915(ring->dev); struct drm_i915_gem_request *req; @@ -2753,6 +2754,31 @@ err: return ret; } +/** + * i915_gem_request_alloc - allocate a request structure + * + * @engine: engine that we wish to issue the request on. + * @ctx: context that the request will be associated with. + * This can be NULL if the request is not directly related to + * any specific user context, in which case this function will + * choose an appropriate context to use. + * + * Returns a pointer to the allocated request if successful, + * or an error code if not. + */ +struct drm_i915_gem_request * +i915_gem_request_alloc(struct intel_engine_cs *engine, + struct intel_context *ctx) +{ + struct drm_i915_gem_request *req; + int err; + + if (ctx == NULL) + ctx = to_i915(engine->dev)->kernel_context; + err = __i915_gem_request_alloc(engine, ctx, &req); + return err ? ERR_PTR(err) : req; +} + void i915_gem_request_cancel(struct drm_i915_gem_request *req) { intel_ring_reserved_space_cancel(req->ringbuf); @@ -3170,9 +3196,13 @@ __i915_gem_object_sync(struct drm_i915_gem_object *obj, return 0; if (*to_req == NULL) { - ret = i915_gem_request_alloc(to, to->default_context, to_req); - if (ret) - return ret; + struct drm_i915_gem_request *req; + + req = i915_gem_request_alloc(to, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); + + *to_req = req; } trace_i915_gem_ring_sync_to(*to_req, from, from_req); @@ -3372,9 +3402,9 @@ int i915_gpu_idle(struct drm_device *dev) if (!i915.enable_execlists) { struct drm_i915_gem_request *req; - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) - return ret; + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); ret = i915_switch_context(req); if (ret) { @@ -4328,10 +4358,20 @@ i915_gem_busy_ioctl(struct drm_device *dev, void *data, if (ret) goto unref; - BUILD_BUG_ON(I915_NUM_RINGS > 16); - args->busy = obj->active << 16; - if (obj->last_write_req) - args->busy |= obj->last_write_req->ring->id; + args->busy = 0; + if (obj->active) { + int i; + + for (i = 0; i < I915_NUM_RINGS; i++) { + struct drm_i915_gem_request *req; + + req = obj->last_read_req[i]; + if (req) + args->busy |= 1 << (16 + req->ring->exec_id); + } + if (obj->last_write_req) + args->busy |= obj->last_write_req->ring->exec_id; + } unref: drm_gem_object_unreference(&obj->base); @@ -4832,7 +4872,7 @@ i915_gem_init_hw(struct drm_device *dev) */ init_unused_rings(dev); - BUG_ON(!dev_priv->ring[RCS].default_context); + BUG_ON(!dev_priv->kernel_context); ret = i915_ppgtt_init_hw(dev); if (ret) { @@ -4869,10 +4909,9 @@ i915_gem_init_hw(struct drm_device *dev) for_each_ring(ring, dev_priv, i) { struct drm_i915_gem_request *req; - WARN_ON(!ring->default_context); - - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) { + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) { + ret = PTR_ERR(req); i915_gem_cleanup_ringbuffer(dev); goto out; } @@ -5112,6 +5151,8 @@ int i915_gem_open(struct drm_device *dev, struct drm_file *file) spin_lock_init(&file_priv->mm.lock); INIT_LIST_HEAD(&file_priv->mm.request_list); + file_priv->bsd_ring = -1; + ret = i915_gem_context_open(dev, file); if (ret) kfree(file_priv); diff --git a/drivers/gpu/drm/i915/i915_gem_context.c b/drivers/gpu/drm/i915/i915_gem_context.c index c25083c78ba7..6a4f64b03db6 100644 --- a/drivers/gpu/drm/i915/i915_gem_context.c +++ b/drivers/gpu/drm/i915/i915_gem_context.c @@ -347,22 +347,20 @@ void i915_gem_context_reset(struct drm_device *dev) i915_gem_context_unreference(lctx); ring->last_context = NULL; } - - /* Force the GPU state to be reinitialised on enabling */ - if (ring->default_context) - ring->default_context->legacy_hw_ctx.initialized = false; } + + /* Force the GPU state to be reinitialised on enabling */ + dev_priv->kernel_context->legacy_hw_ctx.initialized = false; } int i915_gem_context_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; struct intel_context *ctx; - int i; /* Init should only be called once per module load. Eventually the * restriction on the context_disabled check can be loosened. */ - if (WARN_ON(dev_priv->ring[RCS].default_context)) + if (WARN_ON(dev_priv->kernel_context)) return 0; if (intel_vgpu_active(dev) && HAS_LOGICAL_RING_CONTEXTS(dev)) { @@ -392,12 +390,7 @@ int i915_gem_context_init(struct drm_device *dev) return PTR_ERR(ctx); } - for (i = 0; i < I915_NUM_RINGS; i++) { - struct intel_engine_cs *ring = &dev_priv->ring[i]; - - /* NB: RCS will hold a ref for all rings */ - ring->default_context = ctx; - } + dev_priv->kernel_context = ctx; DRM_DEBUG_DRIVER("%s context support initialized\n", i915.enable_execlists ? "LR" : @@ -408,7 +401,7 @@ int i915_gem_context_init(struct drm_device *dev) void i915_gem_context_fini(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_context *dctx = dev_priv->ring[RCS].default_context; + struct intel_context *dctx = dev_priv->kernel_context; int i; if (dctx->legacy_hw_ctx.rcs_state) { @@ -435,17 +428,17 @@ void i915_gem_context_fini(struct drm_device *dev) i915_gem_object_ggtt_unpin(dctx->legacy_hw_ctx.rcs_state); } - for (i = 0; i < I915_NUM_RINGS; i++) { + for (i = I915_NUM_RINGS; --i >= 0;) { struct intel_engine_cs *ring = &dev_priv->ring[i]; - if (ring->last_context) + if (ring->last_context) { i915_gem_context_unreference(ring->last_context); - - ring->default_context = NULL; - ring->last_context = NULL; + ring->last_context = NULL; + } } i915_gem_context_unreference(dctx); + dev_priv->kernel_context = NULL; } int i915_gem_context_enable(struct drm_i915_gem_request *req) diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c index e9c2bfd85b52..1f3eef6fb345 100644 --- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c +++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c @@ -193,10 +193,26 @@ static void i915_gem_dmabuf_kunmap(struct dma_buf *dma_buf, unsigned long page_n static int i915_gem_dmabuf_mmap(struct dma_buf *dma_buf, struct vm_area_struct *vma) { - return -EINVAL; + struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf); + int ret; + + if (obj->base.size < vma->vm_end - vma->vm_start) + return -EINVAL; + + if (!obj->base.filp) + return -ENODEV; + + ret = obj->base.filp->f_op->mmap(obj->base.filp, vma); + if (ret) + return ret; + + fput(vma->vm_file); + vma->vm_file = get_file(obj->base.filp); + + return 0; } -static int i915_gem_begin_cpu_access(struct dma_buf *dma_buf, size_t start, size_t length, enum dma_data_direction direction) +static int i915_gem_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction) { struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf); struct drm_device *dev = obj->base.dev; @@ -212,6 +228,27 @@ static int i915_gem_begin_cpu_access(struct dma_buf *dma_buf, size_t start, size return ret; } +static void i915_gem_end_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction direction) +{ + struct drm_i915_gem_object *obj = dma_buf_to_obj(dma_buf); + struct drm_device *dev = obj->base.dev; + struct drm_i915_private *dev_priv = to_i915(dev); + bool was_interruptible; + int ret; + + mutex_lock(&dev->struct_mutex); + was_interruptible = dev_priv->mm.interruptible; + dev_priv->mm.interruptible = false; + + ret = i915_gem_object_set_to_gtt_domain(obj, false); + + dev_priv->mm.interruptible = was_interruptible; + mutex_unlock(&dev->struct_mutex); + + if (unlikely(ret)) + DRM_ERROR("unable to flush buffer following CPU access; rendering may be corrupt\n"); +} + static const struct dma_buf_ops i915_dmabuf_ops = { .map_dma_buf = i915_gem_map_dma_buf, .unmap_dma_buf = i915_gem_unmap_dma_buf, @@ -224,6 +261,7 @@ static const struct dma_buf_ops i915_dmabuf_ops = { .vmap = i915_gem_dmabuf_vmap, .vunmap = i915_gem_dmabuf_vunmap, .begin_cpu_access = i915_gem_begin_cpu_access, + .end_cpu_access = i915_gem_end_cpu_access, }; struct dma_buf *i915_gem_prime_export(struct drm_device *dev, diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c index dccb517361b3..2dc08ce1079a 100644 --- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c +++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c @@ -193,13 +193,10 @@ static struct i915_vma *eb_get_vma(struct eb_vmas *eb, unsigned long handle) return eb->lut[handle]; } else { struct hlist_head *head; - struct hlist_node *node; + struct i915_vma *vma; head = &eb->buckets[handle & eb->and]; - hlist_for_each(node, head) { - struct i915_vma *vma; - - vma = hlist_entry(node, struct i915_vma, exec_node); + hlist_for_each_entry(vma, head, exec_node) { if (vma->exec_handle == handle) return vma; } @@ -1309,6 +1306,9 @@ i915_gem_ringbuffer_submission(struct i915_execbuffer_params *params, exec_start = params->batch_obj_vm_offset + params->args_batch_start_offset; + if (exec_len == 0) + exec_len = params->batch_obj->base.size; + ret = ring->dispatch_execbuffer(params->request, exec_start, exec_len, params->dispatch_flags); @@ -1325,33 +1325,23 @@ i915_gem_ringbuffer_submission(struct i915_execbuffer_params *params, /** * Find one BSD ring to dispatch the corresponding BSD command. - * The Ring ID is returned. + * The ring index is returned. */ -static int gen8_dispatch_bsd_ring(struct drm_device *dev, - struct drm_file *file) +static unsigned int +gen8_dispatch_bsd_ring(struct drm_i915_private *dev_priv, struct drm_file *file) { - struct drm_i915_private *dev_priv = dev->dev_private; struct drm_i915_file_private *file_priv = file->driver_priv; - /* Check whether the file_priv is using one ring */ - if (file_priv->bsd_ring) - return file_priv->bsd_ring->id; - else { - /* If no, use the ping-pong mechanism to select one ring */ - int ring_id; - - mutex_lock(&dev->struct_mutex); - if (dev_priv->mm.bsd_ring_dispatch_index == 0) { - ring_id = VCS; - dev_priv->mm.bsd_ring_dispatch_index = 1; - } else { - ring_id = VCS2; - dev_priv->mm.bsd_ring_dispatch_index = 0; - } - file_priv->bsd_ring = &dev_priv->ring[ring_id]; - mutex_unlock(&dev->struct_mutex); - return ring_id; + /* Check whether the file_priv has already selected one ring. */ + if ((int)file_priv->bsd_ring < 0) { + /* If not, use the ping-pong mechanism to select one. */ + mutex_lock(&dev_priv->dev->struct_mutex); + file_priv->bsd_ring = dev_priv->mm.bsd_ring_dispatch_index; + dev_priv->mm.bsd_ring_dispatch_index ^= 1; + mutex_unlock(&dev_priv->dev->struct_mutex); } + + return file_priv->bsd_ring; } static struct drm_i915_gem_object * @@ -1374,6 +1364,63 @@ eb_get_batch(struct eb_vmas *eb) return vma->obj; } +#define I915_USER_RINGS (4) + +static const enum intel_ring_id user_ring_map[I915_USER_RINGS + 1] = { + [I915_EXEC_DEFAULT] = RCS, + [I915_EXEC_RENDER] = RCS, + [I915_EXEC_BLT] = BCS, + [I915_EXEC_BSD] = VCS, + [I915_EXEC_VEBOX] = VECS +}; + +static int +eb_select_ring(struct drm_i915_private *dev_priv, + struct drm_file *file, + struct drm_i915_gem_execbuffer2 *args, + struct intel_engine_cs **ring) +{ + unsigned int user_ring_id = args->flags & I915_EXEC_RING_MASK; + + if (user_ring_id > I915_USER_RINGS) { + DRM_DEBUG("execbuf with unknown ring: %u\n", user_ring_id); + return -EINVAL; + } + + if ((user_ring_id != I915_EXEC_BSD) && + ((args->flags & I915_EXEC_BSD_MASK) != 0)) { + DRM_DEBUG("execbuf with non bsd ring but with invalid " + "bsd dispatch flags: %d\n", (int)(args->flags)); + return -EINVAL; + } + + if (user_ring_id == I915_EXEC_BSD && HAS_BSD2(dev_priv)) { + unsigned int bsd_idx = args->flags & I915_EXEC_BSD_MASK; + + if (bsd_idx == I915_EXEC_BSD_DEFAULT) { + bsd_idx = gen8_dispatch_bsd_ring(dev_priv, file); + } else if (bsd_idx >= I915_EXEC_BSD_RING1 && + bsd_idx <= I915_EXEC_BSD_RING2) { + bsd_idx--; + } else { + DRM_DEBUG("execbuf with unknown bsd ring: %u\n", + bsd_idx); + return -EINVAL; + } + + *ring = &dev_priv->ring[_VCS(bsd_idx)]; + } else { + *ring = &dev_priv->ring[user_ring_map[user_ring_id]]; + } + + if (!intel_ring_initialized(*ring)) { + DRM_DEBUG("execbuf with invalid ring: %u\n", user_ring_id); + return -EINVAL; + } + + return 0; +} + static int i915_gem_do_execbuffer(struct drm_device *dev, void *data, struct drm_file *file, @@ -1381,6 +1428,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, struct drm_i915_gem_exec_object2 *exec) { struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_gem_request *req = NULL; struct eb_vmas *eb; struct drm_i915_gem_object *batch_obj; struct drm_i915_gem_exec_object2 shadow_exec_entry; @@ -1411,51 +1459,9 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, if (args->flags & I915_EXEC_IS_PINNED) dispatch_flags |= I915_DISPATCH_PINNED; - if ((args->flags & I915_EXEC_RING_MASK) > LAST_USER_RING) { - DRM_DEBUG("execbuf with unknown ring: %d\n", - (int)(args->flags & I915_EXEC_RING_MASK)); - return -EINVAL; - } - - if (((args->flags & I915_EXEC_RING_MASK) != I915_EXEC_BSD) && - ((args->flags & I915_EXEC_BSD_MASK) != 0)) { - DRM_DEBUG("execbuf with non bsd ring but with invalid " - "bsd dispatch flags: %d\n", (int)(args->flags)); - return -EINVAL; - } - - if ((args->flags & I915_EXEC_RING_MASK) == I915_EXEC_DEFAULT) - ring = &dev_priv->ring[RCS]; - else if ((args->flags & I915_EXEC_RING_MASK) == I915_EXEC_BSD) { - if (HAS_BSD2(dev)) { - int ring_id; - - switch (args->flags & I915_EXEC_BSD_MASK) { - case I915_EXEC_BSD_DEFAULT: - ring_id = gen8_dispatch_bsd_ring(dev, file); - ring = &dev_priv->ring[ring_id]; - break; - case I915_EXEC_BSD_RING1: - ring = &dev_priv->ring[VCS]; - break; - case I915_EXEC_BSD_RING2: - ring = &dev_priv->ring[VCS2]; - break; - default: - DRM_DEBUG("execbuf with unknown bsd ring: %d\n", - (int)(args->flags & I915_EXEC_BSD_MASK)); - return -EINVAL; - } - } else - ring = &dev_priv->ring[VCS]; - } else - ring = &dev_priv->ring[(args->flags & I915_EXEC_RING_MASK) - 1]; - - if (!intel_ring_initialized(ring)) { - DRM_DEBUG("execbuf with invalid ring: %d\n", - (int)(args->flags & I915_EXEC_RING_MASK)); - return -EINVAL; - } + ret = eb_select_ring(dev_priv, file, args, &ring); + if (ret) + return ret; if (args->buffer_count < 1) { DRM_DEBUG("execbuf with %d buffers\n", args->buffer_count); @@ -1602,11 +1608,13 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, params->batch_obj_vm_offset = i915_gem_obj_offset(batch_obj, vm); /* Allocate a request for this batch buffer nice and early. */ - ret = i915_gem_request_alloc(ring, ctx, ¶ms->request); - if (ret) + req = i915_gem_request_alloc(ring, ctx); + if (IS_ERR(req)) { + ret = PTR_ERR(req); goto err_batch_unpin; + } - ret = i915_gem_request_add_to_client(params->request, file); + ret = i915_gem_request_add_to_client(req, file); if (ret) goto err_batch_unpin; @@ -1622,6 +1630,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data, params->dispatch_flags = dispatch_flags; params->batch_obj = batch_obj; params->ctx = ctx; + params->request = req; ret = dev_priv->gt.execbuf_submit(params, args, &eb->vmas); @@ -1645,8 +1654,8 @@ err: * must be freed again. If it was submitted then it is being tracked * on the active request list and no clean up is required here. */ - if (ret && params->request) - i915_gem_request_cancel(params->request); + if (ret && req) + i915_gem_request_cancel(req); mutex_unlock(&dev->struct_mutex); diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c index 56f4f2e58d53..7377b6725c33 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.c +++ b/drivers/gpu/drm/i915/i915_gem_gtt.c @@ -96,9 +96,11 @@ static int i915_get_ggtt_vma_pages(struct i915_vma *vma); -const struct i915_ggtt_view i915_ggtt_view_normal; +const struct i915_ggtt_view i915_ggtt_view_normal = { + .type = I915_GGTT_VIEW_NORMAL, +}; const struct i915_ggtt_view i915_ggtt_view_rotated = { - .type = I915_GGTT_VIEW_ROTATED + .type = I915_GGTT_VIEW_ROTATED, }; static int sanitize_enable_ppgtt(struct drm_device *dev, int enable_ppgtt) @@ -3329,7 +3331,7 @@ i915_gem_obj_lookup_or_create_ggtt_vma(struct drm_i915_gem_object *obj, } static struct scatterlist * -rotate_pages(dma_addr_t *in, unsigned int offset, +rotate_pages(const dma_addr_t *in, unsigned int offset, unsigned int width, unsigned int height, struct sg_table *st, struct scatterlist *sg) { diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h index b448ad832dcf..e5737963ab79 100644 --- a/drivers/gpu/drm/i915/i915_gem_gtt.h +++ b/drivers/gpu/drm/i915/i915_gem_gtt.h @@ -44,7 +44,6 @@ typedef uint64_t gen8_ppgtt_pml4e_t; #define gtt_total_entries(gtt) ((gtt).base.total >> PAGE_SHIFT) - /* gen6-hsw has bit 11-4 for physical addr bit 39-32 */ #define GEN6_GTT_ADDR_ENCODE(addr) ((addr) | (((addr) >> 28) & 0xff0)) #define GEN6_PTE_ADDR_ENCODE(addr) GEN6_GTT_ADDR_ENCODE(addr) diff --git a/drivers/gpu/drm/i915/i915_gem_shrinker.c b/drivers/gpu/drm/i915/i915_gem_shrinker.c index f7df54a8ee2b..16da9c1422cc 100644 --- a/drivers/gpu/drm/i915/i915_gem_shrinker.c +++ b/drivers/gpu/drm/i915/i915_gem_shrinker.c @@ -47,6 +47,46 @@ static bool mutex_is_locked_by(struct mutex *mutex, struct task_struct *task) #endif } +static int num_vma_bound(struct drm_i915_gem_object *obj) +{ + struct i915_vma *vma; + int count = 0; + + list_for_each_entry(vma, &obj->vma_list, vma_link) { + if (drm_mm_node_allocated(&vma->node)) + count++; + if (vma->pin_count) + count++; + } + + return count; +} + +static bool swap_available(void) +{ + return get_nr_swap_pages() > 0; +} + +static bool can_release_pages(struct drm_i915_gem_object *obj) +{ + /* Only report true if by unbinding the object and putting its pages + * we can actually make forward progress towards freeing physical + * pages. + * + * If the pages are pinned for any other reason than being bound + * to the GPU, simply unbinding from the GPU is not going to succeed + * in releasing our pin count on the pages themselves. + */ + if (obj->pages_pin_count != num_vma_bound(obj)) + return false; + + /* We can only return physical pages to the system if we can either + * discard the contents (because the user has marked them as being + * purgeable) or if we can move their contents out to swap. + */ + return swap_available() || obj->madv == I915_MADV_DONTNEED; +} + /** * i915_gem_shrink - Shrink buffer object caches * @dev_priv: i915 device @@ -129,6 +169,9 @@ i915_gem_shrink(struct drm_i915_private *dev_priv, if ((flags & I915_SHRINK_ACTIVE) == 0 && obj->active) continue; + if (!can_release_pages(obj)) + continue; + drm_gem_object_reference(&obj->base); /* For the unbound phase, this should be a no-op! */ @@ -188,21 +231,6 @@ static bool i915_gem_shrinker_lock(struct drm_device *dev, bool *unlock) return true; } -static int num_vma_bound(struct drm_i915_gem_object *obj) -{ - struct i915_vma *vma; - int count = 0; - - list_for_each_entry(vma, &obj->vma_list, vma_link) { - if (drm_mm_node_allocated(&vma->node)) - count++; - if (vma->pin_count) - count++; - } - - return count; -} - static unsigned long i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc) { @@ -222,7 +250,7 @@ i915_gem_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc) count += obj->base.size >> PAGE_SHIFT; list_for_each_entry(obj, &dev_priv->mm.bound_list, global_list) { - if (!obj->active && obj->pages_pin_count == num_vma_bound(obj)) + if (!obj->active && can_release_pages(obj)) count += obj->base.size >> PAGE_SHIFT; } diff --git a/drivers/gpu/drm/i915/i915_gem_stolen.c b/drivers/gpu/drm/i915/i915_gem_stolen.c index 3476877fc0d6..c384dc9c8a63 100644 --- a/drivers/gpu/drm/i915/i915_gem_stolen.c +++ b/drivers/gpu/drm/i915/i915_gem_stolen.c @@ -569,6 +569,9 @@ _i915_gem_object_create_stolen(struct drm_device *dev, if (obj->pages == NULL) goto cleanup; + obj->get_page.sg = obj->pages->sgl; + obj->get_page.last = 0; + i915_gem_object_pin_pages(obj); obj->stolen = stolen; diff --git a/drivers/gpu/drm/i915/i915_gpu_error.c b/drivers/gpu/drm/i915/i915_gpu_error.c index 06ca4082735b..7eeb24427785 100644 --- a/drivers/gpu/drm/i915/i915_gpu_error.c +++ b/drivers/gpu/drm/i915/i915_gpu_error.c @@ -1050,7 +1050,7 @@ static void i915_gem_record_rings(struct drm_device *dev, if (request) rbuf = request->ctx->engine[ring->id].ringbuf; else - rbuf = ring->default_context->engine[ring->id].ringbuf; + rbuf = dev_priv->kernel_context->engine[ring->id].ringbuf; } else rbuf = ring->buffer; diff --git a/drivers/gpu/drm/i915/i915_guc_reg.h b/drivers/gpu/drm/i915/i915_guc_reg.h index 685c7991e24f..e4ba5822289b 100644 --- a/drivers/gpu/drm/i915/i915_guc_reg.h +++ b/drivers/gpu/drm/i915/i915_guc_reg.h @@ -40,6 +40,7 @@ #define GS_MIA_CORE_STATE (1 << GS_MIA_SHIFT) #define SOFT_SCRATCH(n) _MMIO(0xc180 + (n) * 4) +#define SOFT_SCRATCH_COUNT 16 #define UOS_RSA_SCRATCH(i) _MMIO(0xc200 + (i) * 4) #define UOS_RSA_SCRATCH_MAX_COUNT 64 diff --git a/drivers/gpu/drm/i915/i915_guc_submission.c b/drivers/gpu/drm/i915/i915_guc_submission.c index 05aa7e61cbe0..51ae5c1f806d 100644 --- a/drivers/gpu/drm/i915/i915_guc_submission.c +++ b/drivers/gpu/drm/i915/i915_guc_submission.c @@ -158,10 +158,8 @@ static int host2guc_sample_forcewake(struct intel_guc *guc, data[0] = HOST2GUC_ACTION_SAMPLE_FORCEWAKE; /* WaRsDisableCoarsePowerGating:skl,bxt */ - if (!intel_enable_rc6(dev_priv->dev) || - IS_BXT_REVID(dev, 0, BXT_REVID_A1) || - (IS_SKL_GT3(dev) && IS_SKL_REVID(dev, 0, SKL_REVID_E0)) || - (IS_SKL_GT4(dev) && IS_SKL_REVID(dev, 0, SKL_REVID_E0))) + if (!intel_enable_rc6(dev) || + NEEDS_WaRsDisableCoarsePowerGating(dev)) data[1] = 0; else /* bit 0 and 1 are for Render and Media domain separately */ @@ -246,6 +244,9 @@ static int guc_ring_doorbell(struct i915_guc_client *gc) db_exc.cookie = 1; } + /* Finally, update the cached copy of the GuC's WQ head */ + gc->wq_head = desc->head; + kunmap_atomic(base); return ret; } @@ -471,28 +472,30 @@ static void guc_fini_ctx_desc(struct intel_guc *guc, sizeof(desc) * client->ctx_index); } -/* Get valid workqueue item and return it back to offset */ -static int guc_get_workqueue_space(struct i915_guc_client *gc, u32 *offset) +int i915_guc_wq_check_space(struct i915_guc_client *gc) { struct guc_process_desc *desc; void *base; u32 size = sizeof(struct guc_wq_item); int ret = -ETIMEDOUT, timeout_counter = 200; + if (!gc) + return 0; + + /* Quickly return if wq space is available since last time we cache the + * head position. */ + if (CIRC_SPACE(gc->wq_tail, gc->wq_head, gc->wq_size) >= size) + return 0; + base = kmap_atomic(i915_gem_object_get_page(gc->client_obj, 0)); desc = base + gc->proc_desc_offset; while (timeout_counter-- > 0) { - if (CIRC_SPACE(gc->wq_tail, desc->head, gc->wq_size) >= size) { - *offset = gc->wq_tail; - - /* advance the tail for next workqueue item */ - gc->wq_tail += size; - gc->wq_tail &= gc->wq_size - 1; + gc->wq_head = desc->head; - /* this will break the loop */ - timeout_counter = 0; + if (CIRC_SPACE(gc->wq_tail, gc->wq_head, gc->wq_size) >= size) { ret = 0; + break; } if (timeout_counter) @@ -510,12 +513,16 @@ static int guc_add_workqueue_item(struct i915_guc_client *gc, enum intel_ring_id ring_id = rq->ring->id; struct guc_wq_item *wqi; void *base; - u32 tail, wq_len, wq_off = 0; - int ret; + u32 tail, wq_len, wq_off, space; + + space = CIRC_SPACE(gc->wq_tail, gc->wq_head, gc->wq_size); + if (WARN_ON(space < sizeof(struct guc_wq_item))) + return -ENOSPC; /* shouldn't happen */ - ret = guc_get_workqueue_space(gc, &wq_off); - if (ret) - return ret; + /* postincrement WQ tail for next time */ + wq_off = gc->wq_tail; + gc->wq_tail += sizeof(struct guc_wq_item); + gc->wq_tail &= gc->wq_size - 1; /* For now workqueue item is 4 DWs; workqueue buffer is 2 pages. So we * should not have the case where structure wqi is across page, neither @@ -832,6 +839,96 @@ static void guc_create_log(struct intel_guc *guc) guc->log_flags = (offset << GUC_LOG_BUF_ADDR_SHIFT) | flags; } +static void init_guc_policies(struct guc_policies *policies) +{ + struct guc_policy *policy; + u32 p, i; + + policies->dpc_promote_time = 500000; + policies->max_num_work_items = POLICY_MAX_NUM_WI; + + for (p = 0; p < GUC_CTX_PRIORITY_NUM; p++) { + for (i = 0; i < I915_NUM_RINGS; i++) { + policy = &policies->policy[p][i]; + + policy->execution_quantum = 1000000; + policy->preemption_time = 500000; + policy->fault_time = 250000; + policy->policy_flags = 0; + } + } + + policies->is_valid = 1; +} + +static void guc_create_ads(struct intel_guc *guc) +{ + struct drm_i915_private *dev_priv = guc_to_i915(guc); + struct drm_i915_gem_object *obj; + struct guc_ads *ads; + struct guc_policies *policies; + struct guc_mmio_reg_state *reg_state; + struct intel_engine_cs *ring; + struct page *page; + u32 size, i; + + /* The ads obj includes the struct itself and buffers passed to GuC */ + size = sizeof(struct guc_ads) + sizeof(struct guc_policies) + + sizeof(struct guc_mmio_reg_state) + + GUC_S3_SAVE_SPACE_PAGES * PAGE_SIZE; + + obj = guc->ads_obj; + if (!obj) { + obj = gem_allocate_guc_obj(dev_priv->dev, PAGE_ALIGN(size)); + if (!obj) + return; + + guc->ads_obj = obj; + } + + page = i915_gem_object_get_page(obj, 0); + ads = kmap(page); + + /* + * The GuC requires a "Golden Context" when it reinitialises + * engines after a reset. Here we use the Render ring default + * context, which must already exist and be pinned in the GGTT, + * so its address won't change after we've told the GuC where + * to find it. + */ + ring = &dev_priv->ring[RCS]; + ads->golden_context_lrca = ring->status_page.gfx_addr; + + for_each_ring(ring, dev_priv, i) + ads->eng_state_size[i] = intel_lr_context_size(ring); + + /* GuC scheduling policies */ + policies = (void *)ads + sizeof(struct guc_ads); + init_guc_policies(policies); + + ads->scheduler_policies = i915_gem_obj_ggtt_offset(obj) + + sizeof(struct guc_ads); + + /* MMIO reg state */ + reg_state = (void *)policies + sizeof(struct guc_policies); + + for (i = 0; i < I915_NUM_RINGS; i++) { + reg_state->mmio_white_list[i].mmio_start = + dev_priv->ring[i].mmio_base + GUC_MMIO_WHITE_LIST_START; + + /* Nothing to be saved or restored for now. */ + reg_state->mmio_white_list[i].count = 0; + } + + ads->reg_state_addr = ads->scheduler_policies + + sizeof(struct guc_policies); + + ads->reg_state_buffer = ads->reg_state_addr + + sizeof(struct guc_mmio_reg_state); + + kunmap(page); +} + /* * Set up the memory resources to be shared with the GuC. At this point, * we require just one object that can be mapped through the GGTT. @@ -858,6 +955,8 @@ int i915_guc_submission_init(struct drm_device *dev) guc_create_log(guc); + guc_create_ads(guc); + return 0; } @@ -865,7 +964,7 @@ int i915_guc_submission_enable(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; struct intel_guc *guc = &dev_priv->guc; - struct intel_context *ctx = dev_priv->ring[RCS].default_context; + struct intel_context *ctx = dev_priv->kernel_context; struct i915_guc_client *client; /* client for execbuf submission */ @@ -896,6 +995,9 @@ void i915_guc_submission_fini(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; struct intel_guc *guc = &dev_priv->guc; + gem_release_guc_obj(dev_priv->guc.ads_obj); + guc->ads_obj = NULL; + gem_release_guc_obj(dev_priv->guc.log_obj); guc->log_obj = NULL; @@ -919,7 +1021,7 @@ int intel_guc_suspend(struct drm_device *dev) if (!i915.enable_guc_submission) return 0; - ctx = dev_priv->ring[RCS].default_context; + ctx = dev_priv->kernel_context; data[0] = HOST2GUC_ACTION_ENTER_S_STATE; /* any value greater than GUC_POWER_D0 */ @@ -945,7 +1047,7 @@ int intel_guc_resume(struct drm_device *dev) if (!i915.enable_guc_submission) return 0; - ctx = dev_priv->ring[RCS].default_context; + ctx = dev_priv->kernel_context; data[0] = HOST2GUC_ACTION_EXIT_S_STATE; data[1] = GUC_POWER_D0; diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c index fa8afa7860ae..25a89373df63 100644 --- a/drivers/gpu/drm/i915/i915_irq.c +++ b/drivers/gpu/drm/i915/i915_irq.c @@ -2188,10 +2188,6 @@ static irqreturn_t ironlake_irq_handler(int irq, void *arg) /* IRQs are synced during runtime_suspend, we don't require a wakeref */ disable_rpm_wakeref_asserts(dev_priv); - /* We get interrupts on unclaimed registers, so check for this before we - * do any I915_{READ,WRITE}. */ - intel_uncore_check_errors(dev); - /* disable master interrupt before clearing iir */ de_ier = I915_READ(DEIER); I915_WRITE(DEIER, de_ier & ~DE_MASTER_IRQ_CONTROL); @@ -2268,43 +2264,20 @@ static void bxt_hpd_irq_handler(struct drm_device *dev, u32 hotplug_trigger, intel_hpd_irq_handler(dev, pin_mask, long_mask); } -static irqreturn_t gen8_irq_handler(int irq, void *arg) +static irqreturn_t +gen8_de_irq_handler(struct drm_i915_private *dev_priv, u32 master_ctl) { - struct drm_device *dev = arg; - struct drm_i915_private *dev_priv = dev->dev_private; - u32 master_ctl; + struct drm_device *dev = dev_priv->dev; irqreturn_t ret = IRQ_NONE; - uint32_t tmp = 0; + u32 iir; enum pipe pipe; - u32 aux_mask = GEN8_AUX_CHANNEL_A; - - if (!intel_irqs_enabled(dev_priv)) - return IRQ_NONE; - - /* IRQs are synced during runtime_suspend, we don't require a wakeref */ - disable_rpm_wakeref_asserts(dev_priv); - - if (INTEL_INFO(dev_priv)->gen >= 9) - aux_mask |= GEN9_AUX_CHANNEL_B | GEN9_AUX_CHANNEL_C | - GEN9_AUX_CHANNEL_D; - - master_ctl = I915_READ_FW(GEN8_MASTER_IRQ); - master_ctl &= ~GEN8_MASTER_IRQ_CONTROL; - if (!master_ctl) - goto out; - - I915_WRITE_FW(GEN8_MASTER_IRQ, 0); - - /* Find, clear, then process each source of interrupt */ - - ret = gen8_gt_irq_handler(dev_priv, master_ctl); if (master_ctl & GEN8_DE_MISC_IRQ) { - tmp = I915_READ(GEN8_DE_MISC_IIR); - if (tmp) { - I915_WRITE(GEN8_DE_MISC_IIR, tmp); + iir = I915_READ(GEN8_DE_MISC_IIR); + if (iir) { + I915_WRITE(GEN8_DE_MISC_IIR, iir); ret = IRQ_HANDLED; - if (tmp & GEN8_DE_MISC_GSE) + if (iir & GEN8_DE_MISC_GSE) intel_opregion_asle_intr(dev); else DRM_ERROR("Unexpected DE Misc interrupt\n"); @@ -2314,33 +2287,40 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg) } if (master_ctl & GEN8_DE_PORT_IRQ) { - tmp = I915_READ(GEN8_DE_PORT_IIR); - if (tmp) { + iir = I915_READ(GEN8_DE_PORT_IIR); + if (iir) { + u32 tmp_mask; bool found = false; - u32 hotplug_trigger = 0; - - if (IS_BROXTON(dev_priv)) - hotplug_trigger = tmp & BXT_DE_PORT_HOTPLUG_MASK; - else if (IS_BROADWELL(dev_priv)) - hotplug_trigger = tmp & GEN8_PORT_DP_A_HOTPLUG; - I915_WRITE(GEN8_DE_PORT_IIR, tmp); + I915_WRITE(GEN8_DE_PORT_IIR, iir); ret = IRQ_HANDLED; - if (tmp & aux_mask) { + tmp_mask = GEN8_AUX_CHANNEL_A; + if (INTEL_INFO(dev_priv)->gen >= 9) + tmp_mask |= GEN9_AUX_CHANNEL_B | + GEN9_AUX_CHANNEL_C | + GEN9_AUX_CHANNEL_D; + + if (iir & tmp_mask) { dp_aux_irq_handler(dev); found = true; } - if (hotplug_trigger) { - if (IS_BROXTON(dev)) - bxt_hpd_irq_handler(dev, hotplug_trigger, hpd_bxt); - else - ilk_hpd_irq_handler(dev, hotplug_trigger, hpd_bdw); - found = true; + if (IS_BROXTON(dev_priv)) { + tmp_mask = iir & BXT_DE_PORT_HOTPLUG_MASK; + if (tmp_mask) { + bxt_hpd_irq_handler(dev, tmp_mask, hpd_bxt); + found = true; + } + } else if (IS_BROADWELL(dev_priv)) { + tmp_mask = iir & GEN8_PORT_DP_A_HOTPLUG; + if (tmp_mask) { + ilk_hpd_irq_handler(dev, tmp_mask, hpd_bdw); + found = true; + } } - if (IS_BROXTON(dev) && (tmp & BXT_DE_PORT_GMBUS)) { + if (IS_BROXTON(dev) && (iir & BXT_DE_PORT_GMBUS)) { gmbus_irq_handler(dev); found = true; } @@ -2353,49 +2333,51 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg) } for_each_pipe(dev_priv, pipe) { - uint32_t pipe_iir, flip_done = 0, fault_errors = 0; + u32 flip_done, fault_errors; if (!(master_ctl & GEN8_DE_PIPE_IRQ(pipe))) continue; - pipe_iir = I915_READ(GEN8_DE_PIPE_IIR(pipe)); - if (pipe_iir) { - ret = IRQ_HANDLED; - I915_WRITE(GEN8_DE_PIPE_IIR(pipe), pipe_iir); + iir = I915_READ(GEN8_DE_PIPE_IIR(pipe)); + if (!iir) { + DRM_ERROR("The master control interrupt lied (DE PIPE)!\n"); + continue; + } - if (pipe_iir & GEN8_PIPE_VBLANK && - intel_pipe_handle_vblank(dev, pipe)) - intel_check_page_flip(dev, pipe); + ret = IRQ_HANDLED; + I915_WRITE(GEN8_DE_PIPE_IIR(pipe), iir); - if (INTEL_INFO(dev_priv)->gen >= 9) - flip_done = pipe_iir & GEN9_PIPE_PLANE1_FLIP_DONE; - else - flip_done = pipe_iir & GEN8_PIPE_PRIMARY_FLIP_DONE; + if (iir & GEN8_PIPE_VBLANK && + intel_pipe_handle_vblank(dev, pipe)) + intel_check_page_flip(dev, pipe); - if (flip_done) { - intel_prepare_page_flip(dev, pipe); - intel_finish_page_flip_plane(dev, pipe); - } + flip_done = iir; + if (INTEL_INFO(dev_priv)->gen >= 9) + flip_done &= GEN9_PIPE_PLANE1_FLIP_DONE; + else + flip_done &= GEN8_PIPE_PRIMARY_FLIP_DONE; - if (pipe_iir & GEN8_PIPE_CDCLK_CRC_DONE) - hsw_pipe_crc_irq_handler(dev, pipe); + if (flip_done) { + intel_prepare_page_flip(dev, pipe); + intel_finish_page_flip_plane(dev, pipe); + } - if (pipe_iir & GEN8_PIPE_FIFO_UNDERRUN) - intel_cpu_fifo_underrun_irq_handler(dev_priv, - pipe); + if (iir & GEN8_PIPE_CDCLK_CRC_DONE) + hsw_pipe_crc_irq_handler(dev, pipe); + if (iir & GEN8_PIPE_FIFO_UNDERRUN) + intel_cpu_fifo_underrun_irq_handler(dev_priv, pipe); - if (INTEL_INFO(dev_priv)->gen >= 9) - fault_errors = pipe_iir & GEN9_DE_PIPE_IRQ_FAULT_ERRORS; - else - fault_errors = pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS; + fault_errors = iir; + if (INTEL_INFO(dev_priv)->gen >= 9) + fault_errors &= GEN9_DE_PIPE_IRQ_FAULT_ERRORS; + else + fault_errors &= GEN8_DE_PIPE_IRQ_FAULT_ERRORS; - if (fault_errors) - DRM_ERROR("Fault errors on pipe %c\n: 0x%08x", - pipe_name(pipe), - pipe_iir & GEN8_DE_PIPE_IRQ_FAULT_ERRORS); - } else - DRM_ERROR("The master control interrupt lied (DE PIPE)!\n"); + if (fault_errors) + DRM_ERROR("Fault errors on pipe %c\n: 0x%08x", + pipe_name(pipe), + fault_errors); } if (HAS_PCH_SPLIT(dev) && !HAS_PCH_NOP(dev) && @@ -2405,15 +2387,15 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg) * scheme also closed the SDE interrupt handling race we've seen * on older pch-split platforms. But this needs testing. */ - u32 pch_iir = I915_READ(SDEIIR); - if (pch_iir) { - I915_WRITE(SDEIIR, pch_iir); + iir = I915_READ(SDEIIR); + if (iir) { + I915_WRITE(SDEIIR, iir); ret = IRQ_HANDLED; if (HAS_PCH_SPT(dev_priv)) - spt_irq_handler(dev, pch_iir); + spt_irq_handler(dev, iir); else - cpt_irq_handler(dev, pch_iir); + cpt_irq_handler(dev, iir); } else { /* * Like on previous PCH there seems to be something @@ -2423,10 +2405,36 @@ static irqreturn_t gen8_irq_handler(int irq, void *arg) } } + return ret; +} + +static irqreturn_t gen8_irq_handler(int irq, void *arg) +{ + struct drm_device *dev = arg; + struct drm_i915_private *dev_priv = dev->dev_private; + u32 master_ctl; + irqreturn_t ret; + + if (!intel_irqs_enabled(dev_priv)) + return IRQ_NONE; + + master_ctl = I915_READ_FW(GEN8_MASTER_IRQ); + master_ctl &= ~GEN8_MASTER_IRQ_CONTROL; + if (!master_ctl) + return IRQ_NONE; + + I915_WRITE_FW(GEN8_MASTER_IRQ, 0); + + /* IRQs are synced during runtime_suspend, we don't require a wakeref */ + disable_rpm_wakeref_asserts(dev_priv); + + /* Find, clear, then process each source of interrupt */ + ret = gen8_gt_irq_handler(dev_priv, master_ctl); + ret |= gen8_de_irq_handler(dev_priv, master_ctl); + I915_WRITE_FW(GEN8_MASTER_IRQ, GEN8_MASTER_IRQ_CONTROL); POSTING_READ_FW(GEN8_MASTER_IRQ); -out: enable_rpm_wakeref_asserts(dev_priv); return ret; @@ -2949,14 +2957,44 @@ static void semaphore_clear_deadlocks(struct drm_i915_private *dev_priv) ring->hangcheck.deadlock = 0; } -static enum intel_ring_hangcheck_action -ring_stuck(struct intel_engine_cs *ring, u64 acthd) +static bool subunits_stuck(struct intel_engine_cs *ring) { - struct drm_device *dev = ring->dev; - struct drm_i915_private *dev_priv = dev->dev_private; - u32 tmp; + u32 instdone[I915_NUM_INSTDONE_REG]; + bool stuck; + int i; + + if (ring->id != RCS) + return true; + + i915_get_extra_instdone(ring->dev, instdone); + + /* There might be unstable subunit states even when + * actual head is not moving. Filter out the unstable ones by + * accumulating the undone -> done transitions and only + * consider those as progress. + */ + stuck = true; + for (i = 0; i < I915_NUM_INSTDONE_REG; i++) { + const u32 tmp = instdone[i] | ring->hangcheck.instdone[i]; + + if (tmp != ring->hangcheck.instdone[i]) + stuck = false; + + ring->hangcheck.instdone[i] |= tmp; + } + return stuck; +} + +static enum intel_ring_hangcheck_action +head_stuck(struct intel_engine_cs *ring, u64 acthd) +{ if (acthd != ring->hangcheck.acthd) { + + /* Clear subunit states on head movement */ + memset(ring->hangcheck.instdone, 0, + sizeof(ring->hangcheck.instdone)); + if (acthd > ring->hangcheck.max_acthd) { ring->hangcheck.max_acthd = acthd; return HANGCHECK_ACTIVE; @@ -2965,6 +3003,24 @@ ring_stuck(struct intel_engine_cs *ring, u64 acthd) return HANGCHECK_ACTIVE_LOOP; } + if (!subunits_stuck(ring)) + return HANGCHECK_ACTIVE; + + return HANGCHECK_HUNG; +} + +static enum intel_ring_hangcheck_action +ring_stuck(struct intel_engine_cs *ring, u64 acthd) +{ + struct drm_device *dev = ring->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + enum intel_ring_hangcheck_action ha; + u32 tmp; + + ha = head_stuck(ring, acthd); + if (ha != HANGCHECK_HUNG) + return ha; + if (IS_GEN2(dev)) return HANGCHECK_HUNG; @@ -3032,6 +3088,12 @@ static void i915_hangcheck_elapsed(struct work_struct *work) */ DISABLE_RPM_WAKEREF_ASSERTS(dev_priv); + /* As enabling the GPU requires fairly extensive mmio access, + * periodically arm the mmio checker to see if we are triggering + * any invalid access. + */ + intel_uncore_arm_unclaimed_mmio_detection(dev_priv); + for_each_ring(ring, dev_priv, i) { u64 acthd; u32 seqno; @@ -3106,7 +3168,11 @@ static void i915_hangcheck_elapsed(struct work_struct *work) if (ring->hangcheck.score > 0) ring->hangcheck.score--; + /* Clear head and subunit states on seqno movement */ ring->hangcheck.acthd = ring->hangcheck.max_acthd = 0; + + memset(ring->hangcheck.instdone, 0, + sizeof(ring->hangcheck.instdone)); } ring->hangcheck.seqno = seqno; diff --git a/drivers/gpu/drm/i915/i915_params.c b/drivers/gpu/drm/i915/i915_params.c index 835d6099c769..8d90c256520a 100644 --- a/drivers/gpu/drm/i915/i915_params.c +++ b/drivers/gpu/drm/i915/i915_params.c @@ -22,6 +22,7 @@ * IN THE SOFTWARE. */ +#include "i915_params.h" #include "i915_drv.h" struct i915_params i915 __read_mostly = { diff --git a/drivers/gpu/drm/i915/i915_params.h b/drivers/gpu/drm/i915/i915_params.h new file mode 100644 index 000000000000..529929073120 --- /dev/null +++ b/drivers/gpu/drm/i915/i915_params.h @@ -0,0 +1,68 @@ +/* + * Copyright © 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + */ + +#ifndef _I915_PARAMS_H_ +#define _I915_PARAMS_H_ + +#include <linux/cache.h> /* for __read_mostly */ + +struct i915_params { + int modeset; + int panel_ignore_lid; + int semaphores; + int lvds_channel_mode; + int panel_use_ssc; + int vbt_sdvo_panel_type; + int enable_rc6; + int enable_dc; + int enable_fbc; + int enable_ppgtt; + int enable_execlists; + int enable_psr; + unsigned int preliminary_hw_support; + int disable_power_well; + int enable_ips; + int invert_brightness; + int enable_cmd_parser; + int guc_log_level; + int use_mmio_flip; + int mmio_debug; + int edp_vswing; + /* leave bools at the end to not create holes */ + bool enable_hangcheck; + bool fastboot; + bool prefault_disable; + bool load_detect_test; + bool reset; + bool disable_display; + bool disable_vtd_wa; + bool enable_guc_submission; + bool verbose_state_checks; + bool nuclear_pageflip; +}; + +extern struct i915_params i915 __read_mostly; + +#endif + diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 007ae83a4086..0a988895165f 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -1711,6 +1711,11 @@ enum skl_disp_power_wells { #define FPGA_DBG _MMIO(0x42300) #define FPGA_DBG_RM_NOCLAIM (1<<31) +#define CLAIM_ER _MMIO(VLV_DISPLAY_BASE + 0x2028) +#define CLAIM_ER_CLR (1 << 31) +#define CLAIM_ER_OVERFLOW (1 << 16) +#define CLAIM_ER_CTR_MASK 0xffff + #define DERRMR _MMIO(0x44050) /* Note that HBLANK events are reserved on bdw+ */ #define DERRMR_PIPEA_SCANLINE (1<<0) diff --git a/drivers/gpu/drm/i915/i915_sysfs.c b/drivers/gpu/drm/i915/i915_sysfs.c index 37e3f0ddf8e0..c6188dddb341 100644 --- a/drivers/gpu/drm/i915/i915_sysfs.c +++ b/drivers/gpu/drm/i915/i915_sysfs.c @@ -164,7 +164,7 @@ i915_l3_read(struct file *filp, struct kobject *kobj, struct bin_attribute *attr, char *buf, loff_t offset, size_t count) { - struct device *dev = container_of(kobj, struct device, kobj); + struct device *dev = kobj_to_dev(kobj); struct drm_minor *dminor = dev_to_drm_minor(dev); struct drm_device *drm_dev = dminor->dev; struct drm_i915_private *dev_priv = drm_dev->dev_private; @@ -200,7 +200,7 @@ i915_l3_write(struct file *filp, struct kobject *kobj, struct bin_attribute *attr, char *buf, loff_t offset, size_t count) { - struct device *dev = container_of(kobj, struct device, kobj); + struct device *dev = kobj_to_dev(kobj); struct drm_minor *dminor = dev_to_drm_minor(dev); struct drm_device *drm_dev = dminor->dev; struct drm_i915_private *dev_priv = drm_dev->dev_private; @@ -521,7 +521,7 @@ static ssize_t error_state_read(struct file *filp, struct kobject *kobj, loff_t off, size_t count) { - struct device *kdev = container_of(kobj, struct device, kobj); + struct device *kdev = kobj_to_dev(kobj); struct drm_minor *minor = dev_to_drm_minor(kdev); struct drm_device *dev = minor->dev; struct i915_error_state_file_priv error_priv; @@ -556,7 +556,7 @@ static ssize_t error_state_write(struct file *file, struct kobject *kobj, struct bin_attribute *attr, char *buf, loff_t off, size_t count) { - struct device *kdev = container_of(kobj, struct device, kobj); + struct device *kdev = kobj_to_dev(kobj); struct drm_minor *minor = dev_to_drm_minor(kdev); struct drm_device *dev = minor->dev; int ret; diff --git a/drivers/gpu/drm/i915/intel_atomic.c b/drivers/gpu/drm/i915/intel_atomic.c index d0b1c9afa35e..4625f8a9ba12 100644 --- a/drivers/gpu/drm/i915/intel_atomic.c +++ b/drivers/gpu/drm/i915/intel_atomic.c @@ -308,5 +308,5 @@ void intel_atomic_state_clear(struct drm_atomic_state *s) { struct intel_atomic_state *state = to_intel_atomic_state(s); drm_atomic_state_default_clear(&state->base); - state->dpll_set = false; + state->dpll_set = state->modeset = false; } diff --git a/drivers/gpu/drm/i915/intel_atomic_plane.c b/drivers/gpu/drm/i915/intel_atomic_plane.c index c6bb0fc1edfb..e0b851a0004a 100644 --- a/drivers/gpu/drm/i915/intel_atomic_plane.c +++ b/drivers/gpu/drm/i915/intel_atomic_plane.c @@ -152,9 +152,9 @@ static int intel_plane_atomic_check(struct drm_plane *plane, intel_state->clip.x1 = 0; intel_state->clip.y1 = 0; intel_state->clip.x2 = - crtc_state->base.active ? crtc_state->pipe_src_w : 0; + crtc_state->base.enable ? crtc_state->pipe_src_w : 0; intel_state->clip.y2 = - crtc_state->base.active ? crtc_state->pipe_src_h : 0; + crtc_state->base.enable ? crtc_state->pipe_src_h : 0; if (state->fb && intel_rotation_90_or_270(state->rotation)) { if (!(state->fb->modifier[0] == I915_FORMAT_MOD_Y_TILED || @@ -194,8 +194,16 @@ static void intel_plane_atomic_update(struct drm_plane *plane, struct intel_plane *intel_plane = to_intel_plane(plane); struct intel_plane_state *intel_state = to_intel_plane_state(plane->state); + struct drm_crtc *crtc = plane->state->crtc ?: old_state->crtc; + struct drm_crtc_state *crtc_state = + drm_atomic_get_existing_crtc_state(old_state->state, crtc); - intel_plane->commit_plane(plane, intel_state); + if (intel_state->visible) + intel_plane->update_plane(plane, + to_intel_crtc_state(crtc_state), + intel_state); + else + intel_plane->disable_plane(plane, crtc); } const struct drm_plane_helper_funcs intel_plane_helper_funcs = { diff --git a/drivers/gpu/drm/i915/intel_bios.c b/drivers/gpu/drm/i915/intel_bios.c index eba3e0f87181..bf62a19c8f69 100644 --- a/drivers/gpu/drm/i915/intel_bios.c +++ b/drivers/gpu/drm/i915/intel_bios.c @@ -31,11 +31,49 @@ #include "i915_drv.h" #include "intel_bios.h" +/** + * DOC: Video BIOS Table (VBT) + * + * The Video BIOS Table, or VBT, provides platform and board specific + * configuration information to the driver that is not discoverable or available + * through other means. The configuration is mostly related to display + * hardware. The VBT is available via the ACPI OpRegion or, on older systems, in + * the PCI ROM. + * + * The VBT consists of a VBT Header (defined as &struct vbt_header), a BDB + * Header (&struct bdb_header), and a number of BIOS Data Blocks (BDB) that + * contain the actual configuration information. The VBT Header, and thus the + * VBT, begins with "$VBT" signature. The VBT Header contains the offset of the + * BDB Header. The data blocks are concatenated after the BDB Header. The data + * blocks have a 1-byte Block ID, 2-byte Block Size, and Block Size bytes of + * data. (Block 53, the MIPI Sequence Block is an exception.) + * + * The driver parses the VBT during load. The relevant information is stored in + * driver private data for ease of use, and the actual VBT is not read after + * that. + */ + #define SLAVE_ADDR1 0x70 #define SLAVE_ADDR2 0x72 static int panel_type; +/* Get BDB block size given a pointer to Block ID. */ +static u32 _get_blocksize(const u8 *block_base) +{ + /* The MIPI Sequence Block v3+ has a separate size field. */ + if (*block_base == BDB_MIPI_SEQUENCE && *(block_base + 3) >= 3) + return *((const u32 *)(block_base + 4)); + else + return *((const u16 *)(block_base + 1)); +} + +/* Get BDB block size give a pointer to data after Block ID and Block Size. */ +static u32 get_blocksize(const void *block_data) +{ + return _get_blocksize(block_data - 3); +} + static const void * find_section(const void *_bdb, int section_id) { @@ -52,14 +90,8 @@ find_section(const void *_bdb, int section_id) /* walk the sections looking for section_id */ while (index + 3 < total) { current_id = *(base + index); - index++; - - current_size = *((const u16 *)(base + index)); - index += 2; - - /* The MIPI Sequence Block v3+ has a separate size field. */ - if (current_id == BDB_MIPI_SEQUENCE && *(base + index) >= 3) - current_size = *((const u32 *)(base + index + 1)); + current_size = _get_blocksize(base + index); + index += 3; if (index + current_size > total) return NULL; @@ -73,16 +105,6 @@ find_section(const void *_bdb, int section_id) return NULL; } -static u16 -get_blocksize(const void *p) -{ - u16 *block_ptr, block_size; - - block_ptr = (u16 *)((char *)p - 2); - block_size = *block_ptr; - return block_size; -} - static void fill_detail_timing_data(struct drm_display_mode *panel_fixed_mode, const struct lvds_dvo_timing *dvo_timing) @@ -675,84 +697,13 @@ parse_psr(struct drm_i915_private *dev_priv, const struct bdb_header *bdb) dev_priv->vbt.psr.tp2_tp3_wakeup_time = psr_table->tp2_tp3_wakeup_time; } -static u8 *goto_next_sequence(u8 *data, int *size) -{ - u16 len; - int tmp = *size; - - if (--tmp < 0) - return NULL; - - /* goto first element */ - data++; - while (1) { - switch (*data) { - case MIPI_SEQ_ELEM_SEND_PKT: - /* - * skip by this element payload size - * skip elem id, command flag and data type - */ - tmp -= 5; - if (tmp < 0) - return NULL; - - data += 3; - len = *((u16 *)data); - - tmp -= len; - if (tmp < 0) - return NULL; - - /* skip by len */ - data = data + 2 + len; - break; - case MIPI_SEQ_ELEM_DELAY: - /* skip by elem id, and delay is 4 bytes */ - tmp -= 5; - if (tmp < 0) - return NULL; - - data += 5; - break; - case MIPI_SEQ_ELEM_GPIO: - tmp -= 3; - if (tmp < 0) - return NULL; - - data += 3; - break; - default: - DRM_ERROR("Unknown element\n"); - return NULL; - } - - /* end of sequence ? */ - if (*data == 0) - break; - } - - /* goto next sequence or end of block byte */ - if (--tmp < 0) - return NULL; - - data++; - - /* update amount of data left for the sequence block to be parsed */ - *size = tmp; - return data; -} - static void -parse_mipi(struct drm_i915_private *dev_priv, const struct bdb_header *bdb) +parse_mipi_config(struct drm_i915_private *dev_priv, + const struct bdb_header *bdb) { const struct bdb_mipi_config *start; - const struct bdb_mipi_sequence *sequence; const struct mipi_config *config; const struct mipi_pps_data *pps; - u8 *data; - const u8 *seq_data; - int i, panel_id, seq_size; - u16 block_size; /* parse MIPI blocks only if LFP type is MIPI */ if (!dev_priv->vbt.has_mipi) @@ -798,104 +749,233 @@ parse_mipi(struct drm_i915_private *dev_priv, const struct bdb_header *bdb) /* We have mandatory mipi config blocks. Initialize as generic panel */ dev_priv->vbt.dsi.panel_id = MIPI_DSI_GENERIC_PANEL_ID; +} - /* Check if we have sequence block as well */ - sequence = find_section(bdb, BDB_MIPI_SEQUENCE); - if (!sequence) { - DRM_DEBUG_KMS("No MIPI Sequence found, parsing complete\n"); - return; +/* Find the sequence block and size for the given panel. */ +static const u8 * +find_panel_sequence_block(const struct bdb_mipi_sequence *sequence, + u16 panel_id, u32 *seq_size) +{ + u32 total = get_blocksize(sequence); + const u8 *data = &sequence->data[0]; + u8 current_id; + u32 current_size; + int header_size = sequence->version >= 3 ? 5 : 3; + int index = 0; + int i; + + /* skip new block size */ + if (sequence->version >= 3) + data += 4; + + for (i = 0; i < MAX_MIPI_CONFIGURATIONS && index < total; i++) { + if (index + header_size > total) { + DRM_ERROR("Invalid sequence block (header)\n"); + return NULL; + } + + current_id = *(data + index); + if (sequence->version >= 3) + current_size = *((const u32 *)(data + index + 1)); + else + current_size = *((const u16 *)(data + index + 1)); + + index += header_size; + + if (index + current_size > total) { + DRM_ERROR("Invalid sequence block\n"); + return NULL; + } + + if (current_id == panel_id) { + *seq_size = current_size; + return data + index; + } + + index += current_size; } - /* Fail gracefully for forward incompatible sequence block. */ - if (sequence->version >= 3) { - DRM_ERROR("Unable to parse MIPI Sequence Block v3+\n"); - return; + DRM_ERROR("Sequence block detected but no valid configuration\n"); + + return NULL; +} + +static int goto_next_sequence(const u8 *data, int index, int total) +{ + u16 len; + + /* Skip Sequence Byte. */ + for (index = index + 1; index < total; index += len) { + u8 operation_byte = *(data + index); + index++; + + switch (operation_byte) { + case MIPI_SEQ_ELEM_END: + return index; + case MIPI_SEQ_ELEM_SEND_PKT: + if (index + 4 > total) + return 0; + + len = *((const u16 *)(data + index + 2)) + 4; + break; + case MIPI_SEQ_ELEM_DELAY: + len = 4; + break; + case MIPI_SEQ_ELEM_GPIO: + len = 2; + break; + case MIPI_SEQ_ELEM_I2C: + if (index + 7 > total) + return 0; + len = *(data + index + 6) + 7; + break; + default: + DRM_ERROR("Unknown operation byte\n"); + return 0; + } } - DRM_DEBUG_DRIVER("Found MIPI sequence block\n"); + return 0; +} - block_size = get_blocksize(sequence); +static int goto_next_sequence_v3(const u8 *data, int index, int total) +{ + int seq_end; + u16 len; + u32 size_of_sequence; /* - * parse the sequence block for individual sequences + * Could skip sequence based on Size of Sequence alone, but also do some + * checking on the structure. */ - dev_priv->vbt.dsi.seq_version = sequence->version; + if (total < 5) { + DRM_ERROR("Too small sequence size\n"); + return 0; + } - seq_data = &sequence->data[0]; + /* Skip Sequence Byte. */ + index++; /* - * sequence block is variable length and hence we need to parse and - * get the sequence data for specific panel id + * Size of Sequence. Excludes the Sequence Byte and the size itself, + * includes MIPI_SEQ_ELEM_END byte, excludes the final MIPI_SEQ_END + * byte. */ - for (i = 0; i < MAX_MIPI_CONFIGURATIONS; i++) { - panel_id = *seq_data; - seq_size = *((u16 *) (seq_data + 1)); - if (panel_id == panel_type) - break; + size_of_sequence = *((const uint32_t *)(data + index)); + index += 4; - /* skip the sequence including seq header of 3 bytes */ - seq_data = seq_data + 3 + seq_size; - if ((seq_data - &sequence->data[0]) > block_size) { - DRM_ERROR("Sequence start is beyond sequence block size, corrupted sequence block\n"); - return; + seq_end = index + size_of_sequence; + if (seq_end > total) { + DRM_ERROR("Invalid sequence size\n"); + return 0; + } + + for (; index < total; index += len) { + u8 operation_byte = *(data + index); + index++; + + if (operation_byte == MIPI_SEQ_ELEM_END) { + if (index != seq_end) { + DRM_ERROR("Invalid element structure\n"); + return 0; + } + return index; + } + + len = *(data + index); + index++; + + /* + * FIXME: Would be nice to check elements like for v1/v2 in + * goto_next_sequence() above. + */ + switch (operation_byte) { + case MIPI_SEQ_ELEM_SEND_PKT: + case MIPI_SEQ_ELEM_DELAY: + case MIPI_SEQ_ELEM_GPIO: + case MIPI_SEQ_ELEM_I2C: + case MIPI_SEQ_ELEM_SPI: + case MIPI_SEQ_ELEM_PMIC: + break; + default: + DRM_ERROR("Unknown operation byte %u\n", + operation_byte); + break; } } - if (i == MAX_MIPI_CONFIGURATIONS) { - DRM_ERROR("Sequence block detected but no valid configuration\n"); + return 0; +} + +static void +parse_mipi_sequence(struct drm_i915_private *dev_priv, + const struct bdb_header *bdb) +{ + const struct bdb_mipi_sequence *sequence; + const u8 *seq_data; + u32 seq_size; + u8 *data; + int index = 0; + + /* Only our generic panel driver uses the sequence block. */ + if (dev_priv->vbt.dsi.panel_id != MIPI_DSI_GENERIC_PANEL_ID) + return; + + sequence = find_section(bdb, BDB_MIPI_SEQUENCE); + if (!sequence) { + DRM_DEBUG_KMS("No MIPI Sequence found, parsing complete\n"); return; } - /* check if found sequence is completely within the sequence block - * just being paranoid */ - if (seq_size > block_size) { - DRM_ERROR("Corrupted sequence/size, bailing out\n"); + /* Fail gracefully for forward incompatible sequence block. */ + if (sequence->version >= 4) { + DRM_ERROR("Unable to parse MIPI Sequence Block v%u\n", + sequence->version); return; } - /* skip the panel id(1 byte) and seq size(2 bytes) */ - dev_priv->vbt.dsi.data = kmemdup(seq_data + 3, seq_size, GFP_KERNEL); - if (!dev_priv->vbt.dsi.data) + DRM_DEBUG_DRIVER("Found MIPI sequence block v%u\n", sequence->version); + + seq_data = find_panel_sequence_block(sequence, panel_type, &seq_size); + if (!seq_data) return; - /* - * loop into the sequence data and split into multiple sequneces - * There are only 5 types of sequences as of now - */ - data = dev_priv->vbt.dsi.data; - dev_priv->vbt.dsi.size = seq_size; + data = kmemdup(seq_data, seq_size, GFP_KERNEL); + if (!data) + return; - /* two consecutive 0x00 indicate end of all sequences */ - while (1) { - int seq_id = *data; - if (MIPI_SEQ_MAX > seq_id && seq_id > MIPI_SEQ_UNDEFINED) { - dev_priv->vbt.dsi.sequence[seq_id] = data; - DRM_DEBUG_DRIVER("Found mipi sequence - %d\n", seq_id); - } else { - DRM_ERROR("undefined sequence\n"); + /* Parse the sequences, store pointers to each sequence. */ + for (;;) { + u8 seq_id = *(data + index); + if (seq_id == MIPI_SEQ_END) + break; + + if (seq_id >= MIPI_SEQ_MAX) { + DRM_ERROR("Unknown sequence %u\n", seq_id); goto err; } - /* partial parsing to skip elements */ - data = goto_next_sequence(data, &seq_size); + dev_priv->vbt.dsi.sequence[seq_id] = data + index; - if (data == NULL) { - DRM_ERROR("Sequence elements going beyond block itself. Sequence block parsing failed\n"); + if (sequence->version >= 3) + index = goto_next_sequence_v3(data, index, seq_size); + else + index = goto_next_sequence(data, index, seq_size); + if (!index) { + DRM_ERROR("Invalid sequence %u\n", seq_id); goto err; } - - if (*data == 0) - break; /* end of sequence reached */ } - DRM_DEBUG_DRIVER("MIPI related vbt parsing complete\n"); + dev_priv->vbt.dsi.data = data; + dev_priv->vbt.dsi.size = seq_size; + dev_priv->vbt.dsi.seq_version = sequence->version; + + DRM_DEBUG_DRIVER("MIPI related VBT parsing complete\n"); return; -err: - kfree(dev_priv->vbt.dsi.data); - dev_priv->vbt.dsi.data = NULL; - /* error during parsing so set all pointers to null - * because of partial parsing */ +err: + kfree(data); memset(dev_priv->vbt.dsi.sequence, 0, sizeof(dev_priv->vbt.dsi.sequence)); } @@ -1088,7 +1168,12 @@ parse_device_mapping(struct drm_i915_private *dev_priv, DRM_DEBUG_KMS("No general definition block is found, no devices defined.\n"); return; } - if (bdb->version < 195) { + if (bdb->version < 106) { + expected_size = 22; + } else if (bdb->version < 109) { + expected_size = 27; + } else if (bdb->version < 195) { + BUILD_BUG_ON(sizeof(struct old_child_dev_config) != 33); expected_size = sizeof(struct old_child_dev_config); } else if (bdb->version == 195) { expected_size = 37; @@ -1101,18 +1186,18 @@ parse_device_mapping(struct drm_i915_private *dev_priv, bdb->version, expected_size); } - /* The legacy sized child device config is the minimum we need. */ - if (p_defs->child_dev_size < sizeof(struct old_child_dev_config)) { - DRM_ERROR("Child device config size %u is too small.\n", - p_defs->child_dev_size); - return; - } - /* Flag an error for unexpected size, but continue anyway. */ if (p_defs->child_dev_size != expected_size) DRM_ERROR("Unexpected child device config size %u (expected %u for VBT version %u)\n", p_defs->child_dev_size, expected_size, bdb->version); + /* The legacy sized child device config is the minimum we need. */ + if (p_defs->child_dev_size < sizeof(struct old_child_dev_config)) { + DRM_DEBUG_KMS("Child device config size %u is too small.\n", + p_defs->child_dev_size); + return; + } + /* get the block size of general definitions */ block_size = get_blocksize(p_defs); /* get the number of child device */ @@ -1285,7 +1370,7 @@ static const struct vbt_header *find_vbt(void __iomem *bios, size_t size) /** * intel_bios_init - find VBT and initialize settings from the BIOS - * @dev: DRM device + * @dev_priv: i915 device instance * * Loads the Video BIOS and checks that the VBT exists. Sets scratch registers * to appropriate values. @@ -1337,7 +1422,8 @@ intel_bios_init(struct drm_i915_private *dev_priv) parse_driver_features(dev_priv, bdb); parse_edp(dev_priv, bdb); parse_psr(dev_priv, bdb); - parse_mipi(dev_priv, bdb); + parse_mipi_config(dev_priv, bdb); + parse_mipi_sequence(dev_priv, bdb); parse_ddi_ports(dev_priv, bdb); if (bios) diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h index 54eac1003a1e..350d4e0f75a4 100644 --- a/drivers/gpu/drm/i915/intel_bios.h +++ b/drivers/gpu/drm/i915/intel_bios.h @@ -25,25 +25,43 @@ * */ -#ifndef _I830_BIOS_H_ -#define _I830_BIOS_H_ - +#ifndef _INTEL_BIOS_H_ +#define _INTEL_BIOS_H_ + +/** + * struct vbt_header - VBT Header structure + * @signature: VBT signature, always starts with "$VBT" + * @version: Version of this structure + * @header_size: Size of this structure + * @vbt_size: Size of VBT (VBT Header, BDB Header and data blocks) + * @vbt_checksum: Checksum + * @reserved0: Reserved + * @bdb_offset: Offset of &struct bdb_header from beginning of VBT + * @aim_offset: Offsets of add-in data blocks from beginning of VBT + */ struct vbt_header { - u8 signature[20]; /**< Always starts with 'VBT$' */ - u16 version; /**< decimal */ - u16 header_size; /**< in bytes */ - u16 vbt_size; /**< in bytes */ + u8 signature[20]; + u16 version; + u16 header_size; + u16 vbt_size; u8 vbt_checksum; u8 reserved0; - u32 bdb_offset; /**< from beginning of VBT */ - u32 aim_offset[4]; /**< from beginning of VBT */ + u32 bdb_offset; + u32 aim_offset[4]; } __packed; +/** + * struct bdb_header - BDB Header structure + * @signature: BDB signature "BIOS_DATA_BLOCK" + * @version: Version of the data block definitions + * @header_size: Size of this structure + * @bdb_size: Size of BDB (BDB Header and data blocks) + */ struct bdb_header { - u8 signature[16]; /**< Always 'BIOS_DATA_BLOCK' */ - u16 version; /**< decimal */ - u16 header_size; /**< in bytes */ - u16 bdb_size; /**< in bytes */ + u8 signature[16]; + u16 version; + u16 header_size; + u16 bdb_size; } __packed; /* strictly speaking, this is a "skip" block, but it has interesting info */ @@ -936,21 +954,29 @@ struct bdb_mipi_sequence { /* MIPI Sequnece Block definitions */ enum mipi_seq { - MIPI_SEQ_UNDEFINED = 0, + MIPI_SEQ_END = 0, MIPI_SEQ_ASSERT_RESET, MIPI_SEQ_INIT_OTP, MIPI_SEQ_DISPLAY_ON, MIPI_SEQ_DISPLAY_OFF, MIPI_SEQ_DEASSERT_RESET, + MIPI_SEQ_BACKLIGHT_ON, /* sequence block v2+ */ + MIPI_SEQ_BACKLIGHT_OFF, /* sequence block v2+ */ + MIPI_SEQ_TEAR_ON, /* sequence block v2+ */ + MIPI_SEQ_TEAR_OFF, /* sequence block v3+ */ + MIPI_SEQ_POWER_ON, /* sequence block v3+ */ + MIPI_SEQ_POWER_OFF, /* sequence block v3+ */ MIPI_SEQ_MAX }; enum mipi_seq_element { - MIPI_SEQ_ELEM_UNDEFINED = 0, + MIPI_SEQ_ELEM_END = 0, MIPI_SEQ_ELEM_SEND_PKT, MIPI_SEQ_ELEM_DELAY, MIPI_SEQ_ELEM_GPIO, - MIPI_SEQ_ELEM_STATUS, + MIPI_SEQ_ELEM_I2C, /* sequence block v2+ */ + MIPI_SEQ_ELEM_SPI, /* sequence block v3+ */ + MIPI_SEQ_ELEM_PMIC, /* sequence block v3+ */ MIPI_SEQ_ELEM_MAX }; @@ -965,4 +991,4 @@ enum mipi_gpio_pin_index { MIPI_GPIO_MAX }; -#endif /* _I830_BIOS_H_ */ +#endif /* _INTEL_BIOS_H_ */ diff --git a/drivers/gpu/drm/i915/intel_csr.c b/drivers/gpu/drm/i915/intel_csr.c index 9bb63a85997a..5c2f9a40c81b 100644 --- a/drivers/gpu/drm/i915/intel_csr.c +++ b/drivers/gpu/drm/i915/intel_csr.c @@ -44,6 +44,8 @@ #define I915_CSR_SKL "i915/skl_dmc_ver1.bin" #define I915_CSR_BXT "i915/bxt_dmc_ver1.bin" +#define FIRMWARE_URL "https://01.org/linuxgraphics/intel-linux-graphics-firmwares" + MODULE_FIRMWARE(I915_CSR_SKL); MODULE_FIRMWARE(I915_CSR_BXT); @@ -278,10 +280,11 @@ static uint32_t *parse_csr_fw(struct drm_i915_private *dev_priv, csr->version = css_header->version; - if (IS_SKYLAKE(dev) && csr->version < SKL_CSR_VERSION_REQUIRED) { + if ((IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) && + csr->version < SKL_CSR_VERSION_REQUIRED) { DRM_INFO("Refusing to load old Skylake DMC firmware v%u.%u," " please upgrade to v%u.%u or later" - " [https://01.org/linuxgraphics/intel-linux-graphics-firmwares].\n", + " [" FIRMWARE_URL "].\n", CSR_VERSION_MAJOR(csr->version), CSR_VERSION_MINOR(csr->version), CSR_VERSION_MAJOR(SKL_CSR_VERSION_REQUIRED), @@ -399,7 +402,10 @@ out: CSR_VERSION_MAJOR(csr->version), CSR_VERSION_MINOR(csr->version)); } else { - DRM_ERROR("Failed to load DMC firmware, disabling rpm\n"); + dev_notice(dev_priv->dev->dev, + "Failed to load DMC firmware" + " [" FIRMWARE_URL "]," + " disabling runtime power management.\n"); } release_firmware(fw); @@ -421,7 +427,7 @@ void intel_csr_ucode_init(struct drm_i915_private *dev_priv) if (!HAS_CSR(dev_priv)) return; - if (IS_SKYLAKE(dev_priv)) + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) csr->fw_path = I915_CSR_SKL; else if (IS_BROXTON(dev_priv)) csr->fw_path = I915_CSR_BXT; diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index e6408e5583d7..1f9a3687b540 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -133,38 +133,38 @@ static const struct ddi_buf_trans skl_ddi_translations_dp[] = { { 0x00002016, 0x000000A0, 0x0 }, { 0x00005012, 0x0000009B, 0x0 }, { 0x00007011, 0x00000088, 0x0 }, - { 0x80009010, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80009010, 0x000000C0, 0x1 }, { 0x00002016, 0x0000009B, 0x0 }, { 0x00005012, 0x00000088, 0x0 }, - { 0x80007011, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80007011, 0x000000C0, 0x1 }, { 0x00002016, 0x000000DF, 0x0 }, - { 0x80005012, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80005012, 0x000000C0, 0x1 }, }; /* Skylake U */ static const struct ddi_buf_trans skl_u_ddi_translations_dp[] = { { 0x0000201B, 0x000000A2, 0x0 }, { 0x00005012, 0x00000088, 0x0 }, - { 0x00007011, 0x00000087, 0x0 }, - { 0x80009010, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80007011, 0x000000CD, 0x0 }, + { 0x80009010, 0x000000C0, 0x1 }, { 0x0000201B, 0x0000009D, 0x0 }, - { 0x80005012, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ - { 0x80007011, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80005012, 0x000000C0, 0x1 }, + { 0x80007011, 0x000000C0, 0x1 }, { 0x00002016, 0x00000088, 0x0 }, - { 0x80005012, 0x000000C0, 0x1 }, /* Uses I_boost level 0x1 */ + { 0x80005012, 0x000000C0, 0x1 }, }; /* Skylake Y */ static const struct ddi_buf_trans skl_y_ddi_translations_dp[] = { { 0x00000018, 0x000000A2, 0x0 }, { 0x00005012, 0x00000088, 0x0 }, - { 0x00007011, 0x00000087, 0x0 }, - { 0x80009010, 0x000000C0, 0x3 }, /* Uses I_boost level 0x3 */ + { 0x80007011, 0x000000CD, 0x0 }, + { 0x80009010, 0x000000C0, 0x3 }, { 0x00000018, 0x0000009D, 0x0 }, - { 0x80005012, 0x000000C0, 0x3 }, /* Uses I_boost level 0x3 */ - { 0x80007011, 0x000000C0, 0x3 }, /* Uses I_boost level 0x3 */ + { 0x80005012, 0x000000C0, 0x3 }, + { 0x80007011, 0x000000C0, 0x3 }, { 0x00000018, 0x00000088, 0x0 }, - { 0x80005012, 0x000000C0, 0x3 }, /* Uses I_boost level 0x3 */ + { 0x80005012, 0x000000C0, 0x3 }, }; /* @@ -226,26 +226,26 @@ static const struct ddi_buf_trans skl_ddi_translations_hdmi[] = { { 0x00000018, 0x000000A1, 0x0 }, { 0x00000018, 0x00000098, 0x0 }, { 0x00004013, 0x00000088, 0x0 }, - { 0x00006012, 0x00000087, 0x0 }, + { 0x80006012, 0x000000CD, 0x1 }, { 0x00000018, 0x000000DF, 0x0 }, - { 0x00003015, 0x00000087, 0x0 }, /* Default */ - { 0x00003015, 0x000000C7, 0x0 }, - { 0x00000018, 0x000000C7, 0x0 }, + { 0x80003015, 0x000000CD, 0x1 }, /* Default */ + { 0x80003015, 0x000000C0, 0x1 }, + { 0x80000018, 0x000000C0, 0x1 }, }; /* Skylake Y */ static const struct ddi_buf_trans skl_y_ddi_translations_hdmi[] = { { 0x00000018, 0x000000A1, 0x0 }, { 0x00005012, 0x000000DF, 0x0 }, - { 0x00007011, 0x00000084, 0x0 }, + { 0x80007011, 0x000000CB, 0x3 }, { 0x00000018, 0x000000A4, 0x0 }, { 0x00000018, 0x0000009D, 0x0 }, { 0x00004013, 0x00000080, 0x0 }, - { 0x00006013, 0x000000C7, 0x0 }, + { 0x80006013, 0x000000C0, 0x3 }, { 0x00000018, 0x0000008A, 0x0 }, - { 0x00003015, 0x000000C7, 0x0 }, /* Default */ - { 0x80003015, 0x000000C7, 0x7 }, /* Uses I_boost level 0x7 */ - { 0x00000018, 0x000000C7, 0x0 }, + { 0x80003015, 0x000000C0, 0x3 }, /* Default */ + { 0x80003015, 0x000000C0, 0x3 }, + { 0x80000018, 0x000000C0, 0x3 }, }; struct bxt_ddi_buf_trans { @@ -301,8 +301,8 @@ static const struct bxt_ddi_buf_trans bxt_ddi_translations_hdmi[] = { { 154, 0x9A, 1, 128, true }, /* 9: 1200 0 */ }; -static void bxt_ddi_vswing_sequence(struct drm_device *dev, u32 level, - enum port port, int type); +static void bxt_ddi_vswing_sequence(struct drm_i915_private *dev_priv, + u32 level, enum port port, int type); static void ddi_get_encoder_port(struct intel_encoder *intel_encoder, struct intel_digital_port **dig_port, @@ -342,81 +342,50 @@ enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder) return port; } -static bool -intel_dig_port_supports_hdmi(const struct intel_digital_port *intel_dig_port) -{ - return i915_mmio_reg_valid(intel_dig_port->hdmi.hdmi_reg); -} - -static const struct ddi_buf_trans *skl_get_buf_trans_dp(struct drm_device *dev, - int *n_entries) +static const struct ddi_buf_trans * +skl_get_buf_trans_dp(struct drm_i915_private *dev_priv, int *n_entries) { - const struct ddi_buf_trans *ddi_translations; - - if (IS_SKL_ULX(dev) || IS_KBL_ULX(dev)) { - ddi_translations = skl_y_ddi_translations_dp; + if (IS_SKL_ULX(dev_priv) || IS_KBL_ULX(dev_priv)) { *n_entries = ARRAY_SIZE(skl_y_ddi_translations_dp); - } else if (IS_SKL_ULT(dev) || IS_KBL_ULT(dev)) { - ddi_translations = skl_u_ddi_translations_dp; + return skl_y_ddi_translations_dp; + } else if (IS_SKL_ULT(dev_priv) || IS_KBL_ULT(dev_priv)) { *n_entries = ARRAY_SIZE(skl_u_ddi_translations_dp); + return skl_u_ddi_translations_dp; } else { - ddi_translations = skl_ddi_translations_dp; *n_entries = ARRAY_SIZE(skl_ddi_translations_dp); + return skl_ddi_translations_dp; } - - return ddi_translations; } -static const struct ddi_buf_trans *skl_get_buf_trans_edp(struct drm_device *dev, - int *n_entries) +static const struct ddi_buf_trans * +skl_get_buf_trans_edp(struct drm_i915_private *dev_priv, int *n_entries) { - struct drm_i915_private *dev_priv = dev->dev_private; - const struct ddi_buf_trans *ddi_translations; - - if (IS_SKL_ULX(dev) || IS_KBL_ULX(dev)) { - if (dev_priv->edp_low_vswing) { - ddi_translations = skl_y_ddi_translations_edp; + if (dev_priv->edp_low_vswing) { + if (IS_SKL_ULX(dev_priv) || IS_KBL_ULX(dev_priv)) { *n_entries = ARRAY_SIZE(skl_y_ddi_translations_edp); - } else { - ddi_translations = skl_y_ddi_translations_dp; - *n_entries = ARRAY_SIZE(skl_y_ddi_translations_dp); - } - } else if (IS_SKL_ULT(dev) || IS_KBL_ULT(dev)) { - if (dev_priv->edp_low_vswing) { - ddi_translations = skl_u_ddi_translations_edp; + return skl_y_ddi_translations_edp; + } else if (IS_SKL_ULT(dev_priv) || IS_KBL_ULT(dev_priv)) { *n_entries = ARRAY_SIZE(skl_u_ddi_translations_edp); + return skl_u_ddi_translations_edp; } else { - ddi_translations = skl_u_ddi_translations_dp; - *n_entries = ARRAY_SIZE(skl_u_ddi_translations_dp); - } - } else { - if (dev_priv->edp_low_vswing) { - ddi_translations = skl_ddi_translations_edp; *n_entries = ARRAY_SIZE(skl_ddi_translations_edp); - } else { - ddi_translations = skl_ddi_translations_dp; - *n_entries = ARRAY_SIZE(skl_ddi_translations_dp); + return skl_ddi_translations_edp; } } - return ddi_translations; + return skl_get_buf_trans_dp(dev_priv, n_entries); } static const struct ddi_buf_trans * -skl_get_buf_trans_hdmi(struct drm_device *dev, - int *n_entries) +skl_get_buf_trans_hdmi(struct drm_i915_private *dev_priv, int *n_entries) { - const struct ddi_buf_trans *ddi_translations; - - if (IS_SKL_ULX(dev) || IS_KBL_ULX(dev)) { - ddi_translations = skl_y_ddi_translations_hdmi; + if (IS_SKL_ULX(dev_priv) || IS_KBL_ULX(dev_priv)) { *n_entries = ARRAY_SIZE(skl_y_ddi_translations_hdmi); + return skl_y_ddi_translations_hdmi; } else { - ddi_translations = skl_ddi_translations_hdmi; *n_entries = ARRAY_SIZE(skl_ddi_translations_hdmi); + return skl_ddi_translations_hdmi; } - - return ddi_translations; } /* @@ -426,42 +395,52 @@ skl_get_buf_trans_hdmi(struct drm_device *dev, * in either FDI or DP modes only, as HDMI connections will work with both * of those */ -static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, - bool supports_hdmi) +void intel_prepare_ddi_buffer(struct intel_encoder *encoder) { - struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(encoder->base.dev); u32 iboost_bit = 0; int i, n_hdmi_entries, n_dp_entries, n_edp_entries, hdmi_default_entry, size; - int hdmi_level = dev_priv->vbt.ddi_port_info[port].hdmi_level_shift; + int hdmi_level; + enum port port; const struct ddi_buf_trans *ddi_translations_fdi; const struct ddi_buf_trans *ddi_translations_dp; const struct ddi_buf_trans *ddi_translations_edp; const struct ddi_buf_trans *ddi_translations_hdmi; const struct ddi_buf_trans *ddi_translations; - if (IS_BROXTON(dev)) { - if (!supports_hdmi) + port = intel_ddi_get_encoder_port(encoder); + hdmi_level = dev_priv->vbt.ddi_port_info[port].hdmi_level_shift; + + if (IS_BROXTON(dev_priv)) { + if (encoder->type != INTEL_OUTPUT_HDMI) return; /* Vswing programming for HDMI */ - bxt_ddi_vswing_sequence(dev, hdmi_level, port, + bxt_ddi_vswing_sequence(dev_priv, hdmi_level, port, INTEL_OUTPUT_HDMI); return; - } else if (IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) { + } + + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) { ddi_translations_fdi = NULL; ddi_translations_dp = - skl_get_buf_trans_dp(dev, &n_dp_entries); + skl_get_buf_trans_dp(dev_priv, &n_dp_entries); ddi_translations_edp = - skl_get_buf_trans_edp(dev, &n_edp_entries); + skl_get_buf_trans_edp(dev_priv, &n_edp_entries); ddi_translations_hdmi = - skl_get_buf_trans_hdmi(dev, &n_hdmi_entries); + skl_get_buf_trans_hdmi(dev_priv, &n_hdmi_entries); hdmi_default_entry = 8; /* If we're boosting the current, set bit 31 of trans1 */ if (dev_priv->vbt.ddi_port_info[port].hdmi_boost_level || dev_priv->vbt.ddi_port_info[port].dp_boost_level) iboost_bit = 1<<31; - } else if (IS_BROADWELL(dev)) { + + if (WARN_ON(encoder->type == INTEL_OUTPUT_EDP && + port != PORT_A && port != PORT_E && + n_edp_entries > 9)) + n_edp_entries = 9; + } else if (IS_BROADWELL(dev_priv)) { ddi_translations_fdi = bdw_ddi_translations_fdi; ddi_translations_dp = bdw_ddi_translations_dp; ddi_translations_edp = bdw_ddi_translations_edp; @@ -470,7 +449,7 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, n_dp_entries = ARRAY_SIZE(bdw_ddi_translations_dp); n_hdmi_entries = ARRAY_SIZE(bdw_ddi_translations_hdmi); hdmi_default_entry = 7; - } else if (IS_HASWELL(dev)) { + } else if (IS_HASWELL(dev_priv)) { ddi_translations_fdi = hsw_ddi_translations_fdi; ddi_translations_dp = hsw_ddi_translations_dp; ddi_translations_edp = hsw_ddi_translations_dp; @@ -490,30 +469,18 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, hdmi_default_entry = 7; } - switch (port) { - case PORT_A: + switch (encoder->type) { + case INTEL_OUTPUT_EDP: ddi_translations = ddi_translations_edp; size = n_edp_entries; break; - case PORT_B: - case PORT_C: + case INTEL_OUTPUT_DISPLAYPORT: + case INTEL_OUTPUT_HDMI: ddi_translations = ddi_translations_dp; size = n_dp_entries; break; - case PORT_D: - if (intel_dp_is_edp(dev, PORT_D)) { - ddi_translations = ddi_translations_edp; - size = n_edp_entries; - } else { - ddi_translations = ddi_translations_dp; - size = n_dp_entries; - } - break; - case PORT_E: - if (ddi_translations_fdi) - ddi_translations = ddi_translations_fdi; - else - ddi_translations = ddi_translations_dp; + case INTEL_OUTPUT_ANALOG: + ddi_translations = ddi_translations_fdi; size = n_dp_entries; break; default: @@ -527,7 +494,7 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, ddi_translations[i].trans2); } - if (!supports_hdmi) + if (encoder->type != INTEL_OUTPUT_HDMI) return; /* Choose a good default if VBT is badly populated */ @@ -542,37 +509,6 @@ static void intel_prepare_ddi_buffers(struct drm_device *dev, enum port port, ddi_translations_hdmi[hdmi_level].trans2); } -/* Program DDI buffers translations for DP. By default, program ports A-D in DP - * mode and port E for FDI. - */ -void intel_prepare_ddi(struct drm_device *dev) -{ - struct intel_encoder *intel_encoder; - bool visited[I915_MAX_PORTS] = { 0, }; - - if (!HAS_DDI(dev)) - return; - - for_each_intel_encoder(dev, intel_encoder) { - struct intel_digital_port *intel_dig_port; - enum port port; - bool supports_hdmi; - - if (intel_encoder->type == INTEL_OUTPUT_DSI) - continue; - - ddi_get_encoder_port(intel_encoder, &intel_dig_port, &port); - if (visited[port]) - continue; - - supports_hdmi = intel_dig_port && - intel_dig_port_supports_hdmi(intel_dig_port); - - intel_prepare_ddi_buffers(dev, port, supports_hdmi); - visited[port] = true; - } -} - static void intel_wait_ddi_buf_idle(struct drm_i915_private *dev_priv, enum port port) { @@ -601,8 +537,14 @@ void hsw_fdi_link_train(struct drm_crtc *crtc) struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + struct intel_encoder *encoder; u32 temp, i, rx_ctl_val; + for_each_encoder_on_crtc(dev, crtc, encoder) { + WARN_ON(encoder->type != INTEL_OUTPUT_ANALOG); + intel_prepare_ddi_buffer(encoder); + } + /* Set the FDI_RX_MISC pwrdn lanes and the 2 workarounds listed at the * mode set "sequence for CRT port" document: * - TP1 to TP2 time with the default value @@ -2085,10 +2027,9 @@ void intel_ddi_disable_pipe_clock(struct intel_crtc *intel_crtc) TRANS_CLK_SEL_DISABLED); } -static void skl_ddi_set_iboost(struct drm_device *dev, u32 level, - enum port port, int type) +static void skl_ddi_set_iboost(struct drm_i915_private *dev_priv, + u32 level, enum port port, int type) { - struct drm_i915_private *dev_priv = dev->dev_private; const struct ddi_buf_trans *ddi_translations; uint8_t iboost; uint8_t dp_iboost, hdmi_iboost; @@ -2103,21 +2044,26 @@ static void skl_ddi_set_iboost(struct drm_device *dev, u32 level, if (dp_iboost) { iboost = dp_iboost; } else { - ddi_translations = skl_get_buf_trans_dp(dev, &n_entries); + ddi_translations = skl_get_buf_trans_dp(dev_priv, &n_entries); iboost = ddi_translations[level].i_boost; } } else if (type == INTEL_OUTPUT_EDP) { if (dp_iboost) { iboost = dp_iboost; } else { - ddi_translations = skl_get_buf_trans_edp(dev, &n_entries); + ddi_translations = skl_get_buf_trans_edp(dev_priv, &n_entries); + + if (WARN_ON(port != PORT_A && + port != PORT_E && n_entries > 9)) + n_entries = 9; + iboost = ddi_translations[level].i_boost; } } else if (type == INTEL_OUTPUT_HDMI) { if (hdmi_iboost) { iboost = hdmi_iboost; } else { - ddi_translations = skl_get_buf_trans_hdmi(dev, &n_entries); + ddi_translations = skl_get_buf_trans_hdmi(dev_priv, &n_entries); iboost = ddi_translations[level].i_boost; } } else { @@ -2142,10 +2088,9 @@ static void skl_ddi_set_iboost(struct drm_device *dev, u32 level, I915_WRITE(DISPIO_CR_TX_BMU_CR0, reg); } -static void bxt_ddi_vswing_sequence(struct drm_device *dev, u32 level, - enum port port, int type) +static void bxt_ddi_vswing_sequence(struct drm_i915_private *dev_priv, + u32 level, enum port port, int type) { - struct drm_i915_private *dev_priv = dev->dev_private; const struct bxt_ddi_buf_trans *ddi_translations; u32 n_entries, i; uint32_t val; @@ -2260,7 +2205,7 @@ static uint32_t translate_signal_level(int signal_levels) uint32_t ddi_signal_levels(struct intel_dp *intel_dp) { struct intel_digital_port *dport = dp_to_dig_port(intel_dp); - struct drm_device *dev = dport->base.base.dev; + struct drm_i915_private *dev_priv = to_i915(dport->base.base.dev); struct intel_encoder *encoder = &dport->base; uint8_t train_set = intel_dp->train_set[0]; int signal_levels = train_set & (DP_TRAIN_VOLTAGE_SWING_MASK | @@ -2270,10 +2215,10 @@ uint32_t ddi_signal_levels(struct intel_dp *intel_dp) level = translate_signal_level(signal_levels); - if (IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) - skl_ddi_set_iboost(dev, level, port, encoder->type); - else if (IS_BROXTON(dev)) - bxt_ddi_vswing_sequence(dev, level, port, encoder->type); + if (IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) + skl_ddi_set_iboost(dev_priv, level, port, encoder->type); + else if (IS_BROXTON(dev_priv)) + bxt_ddi_vswing_sequence(dev_priv, level, port, encoder->type); return DDI_BUF_TRANS_SELECT(level); } @@ -2325,12 +2270,12 @@ void intel_ddi_clk_select(struct intel_encoder *encoder, static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder) { struct drm_encoder *encoder = &intel_encoder->base; - struct drm_device *dev = encoder->dev; - struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_i915_private *dev_priv = to_i915(encoder->dev); struct intel_crtc *crtc = to_intel_crtc(encoder->crtc); enum port port = intel_ddi_get_encoder_port(intel_encoder); int type = intel_encoder->type; - int hdmi_level; + + intel_prepare_ddi_buffer(intel_encoder); if (type == INTEL_OUTPUT_EDP) { struct intel_dp *intel_dp = enc_to_intel_dp(encoder); @@ -2348,17 +2293,11 @@ static void intel_ddi_pre_enable(struct intel_encoder *intel_encoder) intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_ON); intel_dp_start_link_train(intel_dp); - if (port != PORT_A || INTEL_INFO(dev)->gen >= 9) + if (port != PORT_A || INTEL_INFO(dev_priv)->gen >= 9) intel_dp_stop_link_train(intel_dp); } else if (type == INTEL_OUTPUT_HDMI) { struct intel_hdmi *intel_hdmi = enc_to_intel_hdmi(encoder); - if (IS_BROXTON(dev)) { - hdmi_level = dev_priv->vbt. - ddi_port_info[port].hdmi_level_shift; - bxt_ddi_vswing_sequence(dev, hdmi_level, port, - INTEL_OUTPUT_HDMI); - } intel_hdmi->set_infoframes(encoder, crtc->config->has_hdmi_sink, &crtc->config->base.adjusted_mode); @@ -3282,6 +3221,33 @@ void intel_ddi_init(struct drm_device *dev, enum port port) struct intel_encoder *intel_encoder; struct drm_encoder *encoder; bool init_hdmi, init_dp; + int max_lanes; + + if (I915_READ(DDI_BUF_CTL(PORT_A)) & DDI_A_4_LANES) { + switch (port) { + case PORT_A: + max_lanes = 4; + break; + case PORT_E: + max_lanes = 0; + break; + default: + max_lanes = 4; + break; + } + } else { + switch (port) { + case PORT_A: + max_lanes = 2; + break; + case PORT_E: + max_lanes = 2; + break; + default: + max_lanes = 4; + break; + } + } init_hdmi = (dev_priv->vbt.ddi_port_info[port].supports_dvi || dev_priv->vbt.ddi_port_info[port].supports_hdmi); @@ -3315,6 +3281,7 @@ void intel_ddi_init(struct drm_device *dev, enum port port) intel_dig_port->saved_port_bits = I915_READ(DDI_BUF_CTL(port)) & (DDI_BUF_PORT_REVERSAL | DDI_A_4_LANES); + intel_dig_port->max_lanes = max_lanes; /* * Bspec says that DDI_A_4_LANES is the only supported configuration diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c index 5feb65725c04..e1e7cdee3bbd 100644 --- a/drivers/gpu/drm/i915/intel_display.c +++ b/drivers/gpu/drm/i915/intel_display.c @@ -85,8 +85,6 @@ static const uint32_t intel_cursor_formats[] = { DRM_FORMAT_ARGB8888, }; -static void intel_crtc_update_cursor(struct drm_crtc *crtc, bool on); - static void i9xx_crtc_clock_get(struct intel_crtc *crtc, struct intel_crtc_state *pipe_config); static void ironlake_pch_clock_get(struct intel_crtc *crtc, @@ -1152,11 +1150,6 @@ static void intel_wait_for_pipe_off(struct intel_crtc *crtc) } } -static const char *state_string(bool enabled) -{ - return enabled ? "on" : "off"; -} - /* Only for pre-ILK configs */ void assert_pll(struct drm_i915_private *dev_priv, enum pipe pipe, bool state) @@ -1168,7 +1161,7 @@ void assert_pll(struct drm_i915_private *dev_priv, cur_state = !!(val & DPLL_VCO_ENABLE); I915_STATE_WARN(cur_state != state, "PLL state assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } /* XXX: the dsi pll is shared between MIPI DSI ports */ @@ -1184,7 +1177,7 @@ static void assert_dsi_pll(struct drm_i915_private *dev_priv, bool state) cur_state = val & DSI_PLL_VCO_EN; I915_STATE_WARN(cur_state != state, "DSI PLL state assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } #define assert_dsi_pll_enabled(d) assert_dsi_pll(d, true) #define assert_dsi_pll_disabled(d) assert_dsi_pll(d, false) @@ -1208,14 +1201,13 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv, bool cur_state; struct intel_dpll_hw_state hw_state; - if (WARN (!pll, - "asserting DPLL %s with no DPLL\n", state_string(state))) + if (WARN(!pll, "asserting DPLL %s with no DPLL\n", onoff(state))) return; cur_state = pll->get_hw_state(dev_priv, pll, &hw_state); I915_STATE_WARN(cur_state != state, "%s assertion failure (expected %s, current %s)\n", - pll->name, state_string(state), state_string(cur_state)); + pll->name, onoff(state), onoff(cur_state)); } static void assert_fdi_tx(struct drm_i915_private *dev_priv, @@ -1235,7 +1227,7 @@ static void assert_fdi_tx(struct drm_i915_private *dev_priv, } I915_STATE_WARN(cur_state != state, "FDI TX state assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } #define assert_fdi_tx_enabled(d, p) assert_fdi_tx(d, p, true) #define assert_fdi_tx_disabled(d, p) assert_fdi_tx(d, p, false) @@ -1250,7 +1242,7 @@ static void assert_fdi_rx(struct drm_i915_private *dev_priv, cur_state = !!(val & FDI_RX_ENABLE); I915_STATE_WARN(cur_state != state, "FDI RX state assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } #define assert_fdi_rx_enabled(d, p) assert_fdi_rx(d, p, true) #define assert_fdi_rx_disabled(d, p) assert_fdi_rx(d, p, false) @@ -1282,7 +1274,7 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv, cur_state = !!(val & FDI_RX_PLL_ENABLE); I915_STATE_WARN(cur_state != state, "FDI RX PLL assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } void assert_panel_unlocked(struct drm_i915_private *dev_priv, @@ -1340,7 +1332,7 @@ static void assert_cursor(struct drm_i915_private *dev_priv, I915_STATE_WARN(cur_state != state, "cursor on pipe %c assertion failure (expected %s, current %s)\n", - pipe_name(pipe), state_string(state), state_string(cur_state)); + pipe_name(pipe), onoff(state), onoff(cur_state)); } #define assert_cursor_enabled(d, p) assert_cursor(d, p, true) #define assert_cursor_disabled(d, p) assert_cursor(d, p, false) @@ -1367,7 +1359,7 @@ void assert_pipe(struct drm_i915_private *dev_priv, I915_STATE_WARN(cur_state != state, "pipe %c assertion failure (expected %s, current %s)\n", - pipe_name(pipe), state_string(state), state_string(cur_state)); + pipe_name(pipe), onoff(state), onoff(cur_state)); } static void assert_plane(struct drm_i915_private *dev_priv, @@ -1380,7 +1372,7 @@ static void assert_plane(struct drm_i915_private *dev_priv, cur_state = !!(val & DISPLAY_PLANE_ENABLE); I915_STATE_WARN(cur_state != state, "plane %c assertion failure (expected %s, current %s)\n", - plane_name(plane), state_string(state), state_string(cur_state)); + plane_name(plane), onoff(state), onoff(cur_state)); } #define assert_plane_enabled(d, p) assert_plane(d, p, true) @@ -2153,6 +2145,17 @@ static void intel_enable_pipe(struct intel_crtc *crtc) I915_WRITE(reg, val | PIPECONF_ENABLE); POSTING_READ(reg); + + /* + * Until the pipe starts DSL will read as 0, which would cause + * an apparent vblank timestamp jump, which messes up also the + * frame count when it's derived from the timestamps. So let's + * wait for the pipe to start properly before we call + * drm_crtc_vblank_on() + */ + if (dev->max_vblank_count == 0 && + wait_for(intel_get_crtc_scanline(crtc) != crtc->scanline_offset, 50)) + DRM_ERROR("pipe %c didn't start\n", pipe_name(pipe)); } /** @@ -2214,67 +2217,75 @@ static bool need_vtd_wa(struct drm_device *dev) return false; } -unsigned int -intel_tile_height(struct drm_device *dev, uint32_t pixel_format, - uint64_t fb_format_modifier, unsigned int plane) +static unsigned int intel_tile_size(const struct drm_i915_private *dev_priv) { - unsigned int tile_height; - uint32_t pixel_bytes; + return IS_GEN2(dev_priv) ? 2048 : 4096; +} - switch (fb_format_modifier) { +static unsigned int intel_tile_width(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier, unsigned int cpp) +{ + switch (fb_modifier) { case DRM_FORMAT_MOD_NONE: - tile_height = 1; - break; + return cpp; case I915_FORMAT_MOD_X_TILED: - tile_height = IS_GEN2(dev) ? 16 : 8; - break; + if (IS_GEN2(dev_priv)) + return 128; + else + return 512; case I915_FORMAT_MOD_Y_TILED: - tile_height = 32; - break; + if (IS_GEN2(dev_priv) || HAS_128_BYTE_Y_TILING(dev_priv)) + return 128; + else + return 512; case I915_FORMAT_MOD_Yf_TILED: - pixel_bytes = drm_format_plane_cpp(pixel_format, plane); - switch (pixel_bytes) { - default: + switch (cpp) { case 1: - tile_height = 64; - break; + return 64; case 2: case 4: - tile_height = 32; - break; + return 128; case 8: - tile_height = 16; - break; case 16: - WARN_ONCE(1, - "128-bit pixels are not supported for display!"); - tile_height = 16; - break; + return 256; + default: + MISSING_CASE(cpp); + return cpp; } break; default: - MISSING_CASE(fb_format_modifier); - tile_height = 1; - break; + MISSING_CASE(fb_modifier); + return cpp; } +} - return tile_height; +unsigned int intel_tile_height(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier, unsigned int cpp) +{ + if (fb_modifier == DRM_FORMAT_MOD_NONE) + return 1; + else + return intel_tile_size(dev_priv) / + intel_tile_width(dev_priv, fb_modifier, cpp); } unsigned int intel_fb_align_height(struct drm_device *dev, unsigned int height, - uint32_t pixel_format, uint64_t fb_format_modifier) + uint32_t pixel_format, uint64_t fb_modifier) { - return ALIGN(height, intel_tile_height(dev, pixel_format, - fb_format_modifier, 0)); + unsigned int cpp = drm_format_plane_cpp(pixel_format, 0); + unsigned int tile_height = intel_tile_height(to_i915(dev), fb_modifier, cpp); + + return ALIGN(height, tile_height); } static void intel_fill_fb_ggtt_view(struct i915_ggtt_view *view, struct drm_framebuffer *fb, const struct drm_plane_state *plane_state) { + struct drm_i915_private *dev_priv = to_i915(fb->dev); struct intel_rotation_info *info = &view->params.rotation_info; - unsigned int tile_height, tile_pitch; + unsigned int tile_size, tile_width, tile_height, cpp; *view = i915_ggtt_view_normal; @@ -2292,26 +2303,28 @@ intel_fill_fb_ggtt_view(struct i915_ggtt_view *view, struct drm_framebuffer *fb, info->uv_offset = fb->offsets[1]; info->fb_modifier = fb->modifier[0]; - tile_height = intel_tile_height(fb->dev, fb->pixel_format, - fb->modifier[0], 0); - tile_pitch = PAGE_SIZE / tile_height; - info->width_pages = DIV_ROUND_UP(fb->pitches[0], tile_pitch); + tile_size = intel_tile_size(dev_priv); + + cpp = drm_format_plane_cpp(fb->pixel_format, 0); + tile_width = intel_tile_width(dev_priv, cpp, fb->modifier[0]); + tile_height = tile_size / tile_width; + + info->width_pages = DIV_ROUND_UP(fb->pitches[0], tile_width); info->height_pages = DIV_ROUND_UP(fb->height, tile_height); - info->size = info->width_pages * info->height_pages * PAGE_SIZE; + info->size = info->width_pages * info->height_pages * tile_size; if (info->pixel_format == DRM_FORMAT_NV12) { - tile_height = intel_tile_height(fb->dev, fb->pixel_format, - fb->modifier[0], 1); - tile_pitch = PAGE_SIZE / tile_height; - info->width_pages_uv = DIV_ROUND_UP(fb->pitches[0], tile_pitch); - info->height_pages_uv = DIV_ROUND_UP(fb->height / 2, - tile_height); - info->size_uv = info->width_pages_uv * info->height_pages_uv * - PAGE_SIZE; + cpp = drm_format_plane_cpp(fb->pixel_format, 1); + tile_width = intel_tile_width(dev_priv, fb->modifier[1], cpp); + tile_height = tile_size / tile_width; + + info->width_pages_uv = DIV_ROUND_UP(fb->pitches[1], tile_width); + info->height_pages_uv = DIV_ROUND_UP(fb->height / 2, tile_height); + info->size_uv = info->width_pages_uv * info->height_pages_uv * tile_size; } } -static unsigned int intel_linear_alignment(struct drm_i915_private *dev_priv) +static unsigned int intel_linear_alignment(const struct drm_i915_private *dev_priv) { if (INTEL_INFO(dev_priv)->gen >= 9) return 256 * 1024; @@ -2324,6 +2337,25 @@ static unsigned int intel_linear_alignment(struct drm_i915_private *dev_priv) return 0; } +static unsigned int intel_surf_alignment(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier) +{ + switch (fb_modifier) { + case DRM_FORMAT_MOD_NONE: + return intel_linear_alignment(dev_priv); + case I915_FORMAT_MOD_X_TILED: + if (INTEL_INFO(dev_priv)->gen >= 9) + return 256 * 1024; + return 0; + case I915_FORMAT_MOD_Y_TILED: + case I915_FORMAT_MOD_Yf_TILED: + return 1 * 1024 * 1024; + default: + MISSING_CASE(fb_modifier); + return 0; + } +} + int intel_pin_and_fence_fb_obj(struct drm_plane *plane, struct drm_framebuffer *fb, @@ -2338,29 +2370,7 @@ intel_pin_and_fence_fb_obj(struct drm_plane *plane, WARN_ON(!mutex_is_locked(&dev->struct_mutex)); - switch (fb->modifier[0]) { - case DRM_FORMAT_MOD_NONE: - alignment = intel_linear_alignment(dev_priv); - break; - case I915_FORMAT_MOD_X_TILED: - if (INTEL_INFO(dev)->gen >= 9) - alignment = 256 * 1024; - else { - /* pin() will align the object as required by fence */ - alignment = 0; - } - break; - case I915_FORMAT_MOD_Y_TILED: - case I915_FORMAT_MOD_Yf_TILED: - if (WARN_ONCE(INTEL_INFO(dev)->gen < 9, - "Y tiling bo slipped through, driver bug!\n")) - return -EINVAL; - alignment = 1 * 1024 * 1024; - break; - default: - MISSING_CASE(fb->modifier[0]); - return -EINVAL; - } + alignment = intel_surf_alignment(dev_priv, fb->modifier[0]); intel_fill_fb_ggtt_view(&view, fb, plane_state); @@ -2438,22 +2448,27 @@ static void intel_unpin_fb_obj(struct drm_framebuffer *fb, /* Computes the linear offset to the base tile and adjusts x, y. bytes per pixel * is assumed to be a power-of-two. */ -unsigned long intel_gen4_compute_page_offset(struct drm_i915_private *dev_priv, - int *x, int *y, - unsigned int tiling_mode, - unsigned int cpp, - unsigned int pitch) -{ - if (tiling_mode != I915_TILING_NONE) { +unsigned long intel_compute_tile_offset(struct drm_i915_private *dev_priv, + int *x, int *y, + uint64_t fb_modifier, + unsigned int cpp, + unsigned int pitch) +{ + if (fb_modifier != DRM_FORMAT_MOD_NONE) { + unsigned int tile_size, tile_width, tile_height; unsigned int tile_rows, tiles; - tile_rows = *y / 8; - *y %= 8; + tile_size = intel_tile_size(dev_priv); + tile_width = intel_tile_width(dev_priv, fb_modifier, cpp); + tile_height = tile_size / tile_width; + + tile_rows = *y / tile_height; + *y %= tile_height; - tiles = *x / (512/cpp); - *x %= 512/cpp; + tiles = *x / (tile_width/cpp); + *x %= tile_width/cpp; - return tile_rows * pitch * 8 + tiles * 4096; + return tile_rows * pitch * tile_height + tiles * tile_size; } else { unsigned int alignment = intel_linear_alignment(dev_priv) - 1; unsigned int offset; @@ -2598,6 +2613,8 @@ intel_find_initial_plane_obj(struct intel_crtc *intel_crtc, struct drm_plane_state *plane_state = primary->state; struct drm_crtc_state *crtc_state = intel_crtc->base.state; struct intel_plane *intel_plane = to_intel_plane(primary); + struct intel_plane_state *intel_state = + to_intel_plane_state(plane_state); struct drm_framebuffer *fb; if (!plane_config->fb) @@ -2659,6 +2676,15 @@ valid_fb: plane_state->crtc_w = fb->width; plane_state->crtc_h = fb->height; + intel_state->src.x1 = plane_state->src_x; + intel_state->src.y1 = plane_state->src_y; + intel_state->src.x2 = plane_state->src_x + plane_state->src_w; + intel_state->src.y2 = plane_state->src_y + plane_state->src_h; + intel_state->dst.x1 = plane_state->crtc_x; + intel_state->dst.y1 = plane_state->crtc_y; + intel_state->dst.x2 = plane_state->crtc_x + plane_state->crtc_w; + intel_state->dst.y2 = plane_state->crtc_y + plane_state->crtc_h; + obj = intel_fb_obj(fb); if (obj->tiling_mode != I915_TILING_NONE) dev_priv->preserve_bios_swizzle = true; @@ -2670,36 +2696,23 @@ valid_fb: obj->frontbuffer_bits |= to_intel_plane(primary)->frontbuffer_bit; } -static void i9xx_update_primary_plane(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y) +static void i9xx_update_primary_plane(struct drm_plane *primary, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { - struct drm_device *dev = crtc->dev; + struct drm_device *dev = primary->dev; struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_crtc *intel_crtc = to_intel_crtc(crtc); - struct drm_plane *primary = crtc->primary; - bool visible = to_intel_plane_state(primary->state)->visible; - struct drm_i915_gem_object *obj; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_framebuffer *fb = plane_state->base.fb; + struct drm_i915_gem_object *obj = intel_fb_obj(fb); int plane = intel_crtc->plane; unsigned long linear_offset; + int x = plane_state->src.x1 >> 16; + int y = plane_state->src.y1 >> 16; u32 dspcntr; i915_reg_t reg = DSPCNTR(plane); int pixel_size; - if (!visible || !fb) { - I915_WRITE(reg, 0); - if (INTEL_INFO(dev)->gen >= 4) - I915_WRITE(DSPSURF(plane), 0); - else - I915_WRITE(DSPADDR(plane), 0); - POSTING_READ(reg); - return; - } - - obj = intel_fb_obj(fb); - if (WARN_ON(obj == NULL)) - return; - pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); dspcntr = DISPPLANE_GAMMA_ENABLE; @@ -2714,13 +2727,13 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc, * which should always be the user's requested size. */ I915_WRITE(DSPSIZE(plane), - ((intel_crtc->config->pipe_src_h - 1) << 16) | - (intel_crtc->config->pipe_src_w - 1)); + ((crtc_state->pipe_src_h - 1) << 16) | + (crtc_state->pipe_src_w - 1)); I915_WRITE(DSPPOS(plane), 0); } else if (IS_CHERRYVIEW(dev) && plane == PLANE_B) { I915_WRITE(PRIMSIZE(plane), - ((intel_crtc->config->pipe_src_h - 1) << 16) | - (intel_crtc->config->pipe_src_w - 1)); + ((crtc_state->pipe_src_h - 1) << 16) | + (crtc_state->pipe_src_w - 1)); I915_WRITE(PRIMPOS(plane), 0); I915_WRITE(PRIMCNSTALPHA(plane), 0); } @@ -2762,26 +2775,26 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc, if (INTEL_INFO(dev)->gen >= 4) { intel_crtc->dspaddr_offset = - intel_gen4_compute_page_offset(dev_priv, - &x, &y, obj->tiling_mode, - pixel_size, - fb->pitches[0]); + intel_compute_tile_offset(dev_priv, &x, &y, + fb->modifier[0], + pixel_size, + fb->pitches[0]); linear_offset -= intel_crtc->dspaddr_offset; } else { intel_crtc->dspaddr_offset = linear_offset; } - if (crtc->primary->state->rotation == BIT(DRM_ROTATE_180)) { + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) { dspcntr |= DISPPLANE_ROTATE_180; - x += (intel_crtc->config->pipe_src_w - 1); - y += (intel_crtc->config->pipe_src_h - 1); + x += (crtc_state->pipe_src_w - 1); + y += (crtc_state->pipe_src_h - 1); /* Finding the last pixel of the last line of the display data and adding to linear_offset*/ linear_offset += - (intel_crtc->config->pipe_src_h - 1) * fb->pitches[0] + - (intel_crtc->config->pipe_src_w - 1) * pixel_size; + (crtc_state->pipe_src_h - 1) * fb->pitches[0] + + (crtc_state->pipe_src_w - 1) * pixel_size; } intel_crtc->adjusted_x = x; @@ -2800,37 +2813,40 @@ static void i9xx_update_primary_plane(struct drm_crtc *crtc, POSTING_READ(reg); } -static void ironlake_update_primary_plane(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y) +static void i9xx_disable_primary_plane(struct drm_plane *primary, + struct drm_crtc *crtc) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); - struct drm_plane *primary = crtc->primary; - bool visible = to_intel_plane_state(primary->state)->visible; - struct drm_i915_gem_object *obj; int plane = intel_crtc->plane; - unsigned long linear_offset; - u32 dspcntr; - i915_reg_t reg = DSPCNTR(plane); - int pixel_size; - if (!visible || !fb) { - I915_WRITE(reg, 0); + I915_WRITE(DSPCNTR(plane), 0); + if (INTEL_INFO(dev_priv)->gen >= 4) I915_WRITE(DSPSURF(plane), 0); - POSTING_READ(reg); - return; - } - - obj = intel_fb_obj(fb); - if (WARN_ON(obj == NULL)) - return; + else + I915_WRITE(DSPADDR(plane), 0); + POSTING_READ(DSPCNTR(plane)); +} - pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); +static void ironlake_update_primary_plane(struct drm_plane *primary, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) +{ + struct drm_device *dev = primary->dev; + struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_framebuffer *fb = plane_state->base.fb; + struct drm_i915_gem_object *obj = intel_fb_obj(fb); + int plane = intel_crtc->plane; + unsigned long linear_offset; + u32 dspcntr; + i915_reg_t reg = DSPCNTR(plane); + int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); + int x = plane_state->src.x1 >> 16; + int y = plane_state->src.y1 >> 16; dspcntr = DISPPLANE_GAMMA_ENABLE; - dspcntr |= DISPLAY_PLANE_ENABLE; if (IS_HASWELL(dev) || IS_BROADWELL(dev)) @@ -2867,23 +2883,23 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc, linear_offset = y * fb->pitches[0] + x * pixel_size; intel_crtc->dspaddr_offset = - intel_gen4_compute_page_offset(dev_priv, - &x, &y, obj->tiling_mode, - pixel_size, - fb->pitches[0]); + intel_compute_tile_offset(dev_priv, &x, &y, + fb->modifier[0], + pixel_size, + fb->pitches[0]); linear_offset -= intel_crtc->dspaddr_offset; - if (crtc->primary->state->rotation == BIT(DRM_ROTATE_180)) { + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) { dspcntr |= DISPPLANE_ROTATE_180; if (!IS_HASWELL(dev) && !IS_BROADWELL(dev)) { - x += (intel_crtc->config->pipe_src_w - 1); - y += (intel_crtc->config->pipe_src_h - 1); + x += (crtc_state->pipe_src_w - 1); + y += (crtc_state->pipe_src_h - 1); /* Finding the last pixel of the last line of the display data and adding to linear_offset*/ linear_offset += - (intel_crtc->config->pipe_src_h - 1) * fb->pitches[0] + - (intel_crtc->config->pipe_src_w - 1) * pixel_size; + (crtc_state->pipe_src_h - 1) * fb->pitches[0] + + (crtc_state->pipe_src_w - 1) * pixel_size; } } @@ -2904,37 +2920,15 @@ static void ironlake_update_primary_plane(struct drm_crtc *crtc, POSTING_READ(reg); } -u32 intel_fb_stride_alignment(struct drm_device *dev, uint64_t fb_modifier, - uint32_t pixel_format) +u32 intel_fb_stride_alignment(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier, uint32_t pixel_format) { - u32 bits_per_pixel = drm_format_plane_cpp(pixel_format, 0) * 8; - - /* - * The stride is either expressed as a multiple of 64 bytes - * chunks for linear buffers or in number of tiles for tiled - * buffers. - */ - switch (fb_modifier) { - case DRM_FORMAT_MOD_NONE: - return 64; - case I915_FORMAT_MOD_X_TILED: - if (INTEL_INFO(dev)->gen == 2) - return 128; - return 512; - case I915_FORMAT_MOD_Y_TILED: - /* No need to check for old gens and Y tiling since this is - * about the display engine and those will be blocked before - * we get here. - */ - return 128; - case I915_FORMAT_MOD_Yf_TILED: - if (bits_per_pixel == 8) - return 64; - else - return 128; - default: - MISSING_CASE(fb_modifier); + if (fb_modifier == DRM_FORMAT_MOD_NONE) { return 64; + } else { + int cpp = drm_format_plane_cpp(pixel_format, 0); + + return intel_tile_width(dev_priv, fb_modifier, cpp); } } @@ -3074,36 +3068,30 @@ u32 skl_plane_ctl_rotation(unsigned int rotation) return 0; } -static void skylake_update_primary_plane(struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int x, int y) +static void skylake_update_primary_plane(struct drm_plane *plane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { - struct drm_device *dev = crtc->dev; + struct drm_device *dev = plane->dev; struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_crtc *intel_crtc = to_intel_crtc(crtc); - struct drm_plane *plane = crtc->primary; - bool visible = to_intel_plane_state(plane->state)->visible; - struct drm_i915_gem_object *obj; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc_state->base.crtc); + struct drm_framebuffer *fb = plane_state->base.fb; + struct drm_i915_gem_object *obj = intel_fb_obj(fb); int pipe = intel_crtc->pipe; u32 plane_ctl, stride_div, stride; u32 tile_height, plane_offset, plane_size; - unsigned int rotation; + unsigned int rotation = plane_state->base.rotation; int x_offset, y_offset; u32 surf_addr; - struct intel_crtc_state *crtc_state = intel_crtc->config; - struct intel_plane_state *plane_state; - int src_x = 0, src_y = 0, src_w = 0, src_h = 0; - int dst_x = 0, dst_y = 0, dst_w = 0, dst_h = 0; - int scaler_id = -1; - - plane_state = to_intel_plane_state(plane->state); - - if (!visible || !fb) { - I915_WRITE(PLANE_CTL(pipe, 0), 0); - I915_WRITE(PLANE_SURF(pipe, 0), 0); - POSTING_READ(PLANE_CTL(pipe, 0)); - return; - } + int scaler_id = plane_state->scaler_id; + int src_x = plane_state->src.x1 >> 16; + int src_y = plane_state->src.y1 >> 16; + int src_w = drm_rect_width(&plane_state->src) >> 16; + int src_h = drm_rect_height(&plane_state->src) >> 16; + int dst_x = plane_state->dst.x1; + int dst_y = plane_state->dst.y1; + int dst_w = drm_rect_width(&plane_state->dst); + int dst_h = drm_rect_height(&plane_state->dst); plane_ctl = PLANE_CTL_ENABLE | PLANE_CTL_PIPE_GAMMA_ENABLE | @@ -3112,41 +3100,27 @@ static void skylake_update_primary_plane(struct drm_crtc *crtc, plane_ctl |= skl_plane_ctl_format(fb->pixel_format); plane_ctl |= skl_plane_ctl_tiling(fb->modifier[0]); plane_ctl |= PLANE_CTL_PLANE_GAMMA_DISABLE; - - rotation = plane->state->rotation; plane_ctl |= skl_plane_ctl_rotation(rotation); - obj = intel_fb_obj(fb); - stride_div = intel_fb_stride_alignment(dev, fb->modifier[0], + stride_div = intel_fb_stride_alignment(dev_priv, fb->modifier[0], fb->pixel_format); surf_addr = intel_plane_obj_offset(to_intel_plane(plane), obj, 0); WARN_ON(drm_rect_width(&plane_state->src) == 0); - scaler_id = plane_state->scaler_id; - src_x = plane_state->src.x1 >> 16; - src_y = plane_state->src.y1 >> 16; - src_w = drm_rect_width(&plane_state->src) >> 16; - src_h = drm_rect_height(&plane_state->src) >> 16; - dst_x = plane_state->dst.x1; - dst_y = plane_state->dst.y1; - dst_w = drm_rect_width(&plane_state->dst); - dst_h = drm_rect_height(&plane_state->dst); - - WARN_ON(x != src_x || y != src_y); - if (intel_rotation_90_or_270(rotation)) { + int cpp = drm_format_plane_cpp(fb->pixel_format, 0); + /* stride = Surface height in tiles */ - tile_height = intel_tile_height(dev, fb->pixel_format, - fb->modifier[0], 0); + tile_height = intel_tile_height(dev_priv, fb->modifier[0], cpp); stride = DIV_ROUND_UP(fb->height, tile_height); - x_offset = stride * tile_height - y - src_h; - y_offset = x; + x_offset = stride * tile_height - src_y - src_h; + y_offset = src_x; plane_size = (src_w - 1) << 16 | (src_h - 1); } else { stride = fb->pitches[0] / stride_div; - x_offset = x; - y_offset = y; + x_offset = src_x; + y_offset = src_y; plane_size = (src_h - 1) << 16 | (src_w - 1); } plane_offset = y_offset << 16 | x_offset; @@ -3179,20 +3153,30 @@ static void skylake_update_primary_plane(struct drm_crtc *crtc, POSTING_READ(PLANE_SURF(pipe, 0)); } -/* Assume fb object is pinned & idle & fenced and just update base pointers */ -static int -intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb, - int x, int y, enum mode_set_atomic state) +static void skylake_disable_primary_plane(struct drm_plane *primary, + struct drm_crtc *crtc) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; + int pipe = to_intel_crtc(crtc)->pipe; if (dev_priv->fbc.deactivate) dev_priv->fbc.deactivate(dev_priv); - dev_priv->display.update_primary_plane(crtc, fb, x, y); + I915_WRITE(PLANE_CTL(pipe, 0), 0); + I915_WRITE(PLANE_SURF(pipe, 0), 0); + POSTING_READ(PLANE_SURF(pipe, 0)); +} + +/* Assume fb object is pinned & idle & fenced and just update base pointers */ +static int +intel_pipe_set_base_atomic(struct drm_crtc *crtc, struct drm_framebuffer *fb, + int x, int y, enum mode_set_atomic state) +{ + /* Support for kgdboc is disabled, this needs a major rework. */ + DRM_ERROR("legacy panic handler not supported any more.\n"); - return 0; + return -ENODEV; } static void intel_complete_page_flips(struct drm_device *dev) @@ -3219,8 +3203,10 @@ static void intel_update_primary_planes(struct drm_device *dev) drm_modeset_lock_crtc(crtc, &plane->base); plane_state = to_intel_plane_state(plane->base.state); - if (crtc->state->active && plane_state->base.fb) - plane->commit_plane(&plane->base, plane_state); + if (plane_state->visible) + plane->update_plane(&plane->base, + to_intel_crtc_state(crtc->state), + plane_state); drm_modeset_unlock_crtc(crtc); } @@ -4452,7 +4438,7 @@ int skl_update_scaler_crtc(struct intel_crtc_state *state) intel_crtc->base.base.id, intel_crtc->pipe, SKL_CRTC_INDEX); return skl_update_scaler(state, !state->base.active, SKL_CRTC_INDEX, - &state->scaler_state.scaler_id, DRM_ROTATE_0, + &state->scaler_state.scaler_id, BIT(DRM_ROTATE_0), state->pipe_src_w, state->pipe_src_h, adjusted_mode->crtc_hdisplay, adjusted_mode->crtc_vdisplay); } @@ -5370,6 +5356,7 @@ static void modeset_put_power_domains(struct drm_i915_private *dev_priv, static void modeset_update_crtc_power_domains(struct drm_atomic_state *state) { + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); struct drm_device *dev = state->dev; struct drm_i915_private *dev_priv = dev->dev_private; unsigned long put_domains[I915_MAX_PIPES] = {}; @@ -5383,13 +5370,9 @@ static void modeset_update_crtc_power_domains(struct drm_atomic_state *state) modeset_get_crtc_power_domains(crtc); } - if (dev_priv->display.modeset_commit_cdclk) { - unsigned int cdclk = to_intel_atomic_state(state)->cdclk; - - if (cdclk != dev_priv->cdclk_freq && - !WARN_ON(!state->allow_modeset)) - dev_priv->display.modeset_commit_cdclk(state); - } + if (dev_priv->display.modeset_commit_cdclk && + intel_state->dev_cdclk != dev_priv->cdclk_freq) + dev_priv->display.modeset_commit_cdclk(state); for (i = 0; i < I915_MAX_PIPES; i++) if (put_domains[i]) @@ -6063,22 +6046,31 @@ static int broxton_calc_cdclk(struct drm_i915_private *dev_priv, static int intel_mode_max_pixclk(struct drm_device *dev, struct drm_atomic_state *state) { - struct intel_crtc *intel_crtc; - struct intel_crtc_state *crtc_state; - int max_pixclk = 0; + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + struct drm_i915_private *dev_priv = dev->dev_private; + struct drm_crtc *crtc; + struct drm_crtc_state *crtc_state; + unsigned max_pixclk = 0, i; + enum pipe pipe; - for_each_intel_crtc(dev, intel_crtc) { - crtc_state = intel_atomic_get_crtc_state(state, intel_crtc); - if (IS_ERR(crtc_state)) - return PTR_ERR(crtc_state); + memcpy(intel_state->min_pixclk, dev_priv->min_pixclk, + sizeof(intel_state->min_pixclk)); - if (!crtc_state->base.enable) - continue; + for_each_crtc_in_state(state, crtc, crtc_state, i) { + int pixclk = 0; + + if (crtc_state->enable) + pixclk = crtc_state->adjusted_mode.crtc_clock; - max_pixclk = max(max_pixclk, - crtc_state->base.adjusted_mode.crtc_clock); + intel_state->min_pixclk[i] = pixclk; } + if (!intel_state->active_crtcs) + return 0; + + for_each_pipe(dev_priv, pipe) + max_pixclk = max(intel_state->min_pixclk[pipe], max_pixclk); + return max_pixclk; } @@ -6087,13 +6079,18 @@ static int valleyview_modeset_calc_cdclk(struct drm_atomic_state *state) struct drm_device *dev = state->dev; struct drm_i915_private *dev_priv = dev->dev_private; int max_pixclk = intel_mode_max_pixclk(dev, state); + struct intel_atomic_state *intel_state = + to_intel_atomic_state(state); if (max_pixclk < 0) return max_pixclk; - to_intel_atomic_state(state)->cdclk = + intel_state->cdclk = intel_state->dev_cdclk = valleyview_calc_cdclk(dev_priv, max_pixclk); + if (!intel_state->active_crtcs) + intel_state->dev_cdclk = valleyview_calc_cdclk(dev_priv, 0); + return 0; } @@ -6102,13 +6099,18 @@ static int broxton_modeset_calc_cdclk(struct drm_atomic_state *state) struct drm_device *dev = state->dev; struct drm_i915_private *dev_priv = dev->dev_private; int max_pixclk = intel_mode_max_pixclk(dev, state); + struct intel_atomic_state *intel_state = + to_intel_atomic_state(state); if (max_pixclk < 0) return max_pixclk; - to_intel_atomic_state(state)->cdclk = + intel_state->cdclk = intel_state->dev_cdclk = broxton_calc_cdclk(dev_priv, max_pixclk); + if (!intel_state->active_crtcs) + intel_state->dev_cdclk = broxton_calc_cdclk(dev_priv, 0); + return 0; } @@ -6151,8 +6153,10 @@ static void vlv_program_pfi_credits(struct drm_i915_private *dev_priv) static void valleyview_modeset_commit_cdclk(struct drm_atomic_state *old_state) { struct drm_device *dev = old_state->dev; - unsigned int req_cdclk = to_intel_atomic_state(old_state)->cdclk; struct drm_i915_private *dev_priv = dev->dev_private; + struct intel_atomic_state *old_intel_state = + to_intel_atomic_state(old_state); + unsigned req_cdclk = old_intel_state->dev_cdclk; /* * FIXME: We can end up here with all power domains off, yet @@ -6383,6 +6387,9 @@ static void intel_crtc_disable_noatomic(struct drm_crtc *crtc) for_each_power_domain(domain, domains) intel_display_power_put(dev_priv, domain); intel_crtc->enabled_power_domains = 0; + + dev_priv->active_crtcs &= ~(1 << intel_crtc->pipe); + dev_priv->min_pixclk[intel_crtc->pipe] = 0; } /* @@ -7593,26 +7600,34 @@ static void chv_prepare_pll(struct intel_crtc *crtc, * in cases where we need the PLL enabled even when @pipe is not going to * be enabled. */ -void vlv_force_pll_on(struct drm_device *dev, enum pipe pipe, - const struct dpll *dpll) +int vlv_force_pll_on(struct drm_device *dev, enum pipe pipe, + const struct dpll *dpll) { struct intel_crtc *crtc = to_intel_crtc(intel_get_crtc_for_pipe(dev, pipe)); - struct intel_crtc_state pipe_config = { - .base.crtc = &crtc->base, - .pixel_multiplier = 1, - .dpll = *dpll, - }; + struct intel_crtc_state *pipe_config; + + pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL); + if (!pipe_config) + return -ENOMEM; + + pipe_config->base.crtc = &crtc->base; + pipe_config->pixel_multiplier = 1; + pipe_config->dpll = *dpll; if (IS_CHERRYVIEW(dev)) { - chv_compute_dpll(crtc, &pipe_config); - chv_prepare_pll(crtc, &pipe_config); - chv_enable_pll(crtc, &pipe_config); + chv_compute_dpll(crtc, pipe_config); + chv_prepare_pll(crtc, pipe_config); + chv_enable_pll(crtc, pipe_config); } else { - vlv_compute_dpll(crtc, &pipe_config); - vlv_prepare_pll(crtc, &pipe_config); - vlv_enable_pll(crtc, &pipe_config); + vlv_compute_dpll(crtc, pipe_config); + vlv_prepare_pll(crtc, pipe_config); + vlv_enable_pll(crtc, pipe_config); } + + kfree(pipe_config); + + return 0; } /** @@ -9246,7 +9261,7 @@ skylake_get_initial_plane_config(struct intel_crtc *crtc, fb->width = ((val >> 0) & 0x1fff) + 1; val = I915_READ(PLANE_STRIDE(pipe, 0)); - stride_mult = intel_fb_stride_alignment(dev, fb->modifier[0], + stride_mult = intel_fb_stride_alignment(dev_priv, fb->modifier[0], fb->pixel_format); fb->pitches[0] = (val & 0x3ff) * stride_mult; @@ -9662,14 +9677,14 @@ void hsw_disable_pc8(struct drm_i915_private *dev_priv) val |= PCH_LP_PARTITION_LEVEL_DISABLE; I915_WRITE(SOUTH_DSPCLK_GATE_D, val); } - - intel_prepare_ddi(dev); } static void broxton_modeset_commit_cdclk(struct drm_atomic_state *old_state) { struct drm_device *dev = old_state->dev; - unsigned int req_cdclk = to_intel_atomic_state(old_state)->cdclk; + struct intel_atomic_state *old_intel_state = + to_intel_atomic_state(old_state); + unsigned int req_cdclk = old_intel_state->dev_cdclk; broxton_set_cdclk(dev, req_cdclk); } @@ -9677,29 +9692,41 @@ static void broxton_modeset_commit_cdclk(struct drm_atomic_state *old_state) /* compute the max rate for new configuration */ static int ilk_max_pixel_rate(struct drm_atomic_state *state) { - struct intel_crtc *intel_crtc; + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + struct drm_i915_private *dev_priv = state->dev->dev_private; + struct drm_crtc *crtc; + struct drm_crtc_state *cstate; struct intel_crtc_state *crtc_state; - int max_pixel_rate = 0; + unsigned max_pixel_rate = 0, i; + enum pipe pipe; - for_each_intel_crtc(state->dev, intel_crtc) { - int pixel_rate; + memcpy(intel_state->min_pixclk, dev_priv->min_pixclk, + sizeof(intel_state->min_pixclk)); - crtc_state = intel_atomic_get_crtc_state(state, intel_crtc); - if (IS_ERR(crtc_state)) - return PTR_ERR(crtc_state); + for_each_crtc_in_state(state, crtc, cstate, i) { + int pixel_rate; - if (!crtc_state->base.enable) + crtc_state = to_intel_crtc_state(cstate); + if (!crtc_state->base.enable) { + intel_state->min_pixclk[i] = 0; continue; + } pixel_rate = ilk_pipe_pixel_rate(crtc_state); /* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */ - if (IS_BROADWELL(state->dev) && crtc_state->ips_enabled) + if (IS_BROADWELL(dev_priv) && crtc_state->ips_enabled) pixel_rate = DIV_ROUND_UP(pixel_rate * 100, 95); - max_pixel_rate = max(max_pixel_rate, pixel_rate); + intel_state->min_pixclk[i] = pixel_rate; } + if (!intel_state->active_crtcs) + return 0; + + for_each_pipe(dev_priv, pipe) + max_pixel_rate = max(intel_state->min_pixclk[pipe], max_pixel_rate); + return max_pixel_rate; } @@ -9783,6 +9810,7 @@ static void broadwell_set_cdclk(struct drm_device *dev, int cdclk) static int broadwell_modeset_calc_cdclk(struct drm_atomic_state *state) { struct drm_i915_private *dev_priv = to_i915(state->dev); + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); int max_pixclk = ilk_max_pixel_rate(state); int cdclk; @@ -9805,7 +9833,9 @@ static int broadwell_modeset_calc_cdclk(struct drm_atomic_state *state) return -EINVAL; } - to_intel_atomic_state(state)->cdclk = cdclk; + intel_state->cdclk = intel_state->dev_cdclk = cdclk; + if (!intel_state->active_crtcs) + intel_state->dev_cdclk = 337500; return 0; } @@ -9813,7 +9843,9 @@ static int broadwell_modeset_calc_cdclk(struct drm_atomic_state *state) static void broadwell_modeset_commit_cdclk(struct drm_atomic_state *old_state) { struct drm_device *dev = old_state->dev; - unsigned int req_cdclk = to_intel_atomic_state(old_state)->cdclk; + struct intel_atomic_state *old_intel_state = + to_intel_atomic_state(old_state); + unsigned req_cdclk = old_intel_state->dev_cdclk; broadwell_set_cdclk(dev, req_cdclk); } @@ -10026,16 +10058,17 @@ static bool haswell_get_pipe_config(struct intel_crtc *crtc, return true; } -static void i845_update_cursor(struct drm_crtc *crtc, u32 base, bool on) +static void i845_update_cursor(struct drm_crtc *crtc, u32 base, + const struct intel_plane_state *plane_state) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); uint32_t cntl = 0, size = 0; - if (on) { - unsigned int width = intel_crtc->base.cursor->state->crtc_w; - unsigned int height = intel_crtc->base.cursor->state->crtc_h; + if (plane_state && plane_state->visible) { + unsigned int width = plane_state->base.crtc_w; + unsigned int height = plane_state->base.crtc_h; unsigned int stride = roundup_pow_of_two(width) * 4; switch (stride) { @@ -10088,7 +10121,8 @@ static void i845_update_cursor(struct drm_crtc *crtc, u32 base, bool on) } } -static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base, bool on) +static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base, + const struct intel_plane_state *plane_state) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; @@ -10096,9 +10130,9 @@ static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base, bool on) int pipe = intel_crtc->pipe; uint32_t cntl = 0; - if (on) { + if (plane_state && plane_state->visible) { cntl = MCURSOR_GAMMA_ENABLE; - switch (intel_crtc->base.cursor->state->crtc_w) { + switch (plane_state->base.crtc_w) { case 64: cntl |= CURSOR_MODE_64_ARGB_AX; break; @@ -10109,17 +10143,17 @@ static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base, bool on) cntl |= CURSOR_MODE_256_ARGB_AX; break; default: - MISSING_CASE(intel_crtc->base.cursor->state->crtc_w); + MISSING_CASE(plane_state->base.crtc_w); return; } cntl |= pipe << 28; /* Connect to correct pipe */ if (HAS_DDI(dev)) cntl |= CURSOR_PIPE_CSC_ENABLE; - } - if (crtc->cursor->state->rotation == BIT(DRM_ROTATE_180)) - cntl |= CURSOR_ROTATE_180; + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) + cntl |= CURSOR_ROTATE_180; + } if (intel_crtc->cursor_cntl != cntl) { I915_WRITE(CURCNTR(pipe), cntl); @@ -10136,56 +10170,45 @@ static void i9xx_update_cursor(struct drm_crtc *crtc, u32 base, bool on) /* If no-part of the cursor is visible on the framebuffer, then the GPU may hang... */ static void intel_crtc_update_cursor(struct drm_crtc *crtc, - bool on) + const struct intel_plane_state *plane_state) { struct drm_device *dev = crtc->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_crtc *intel_crtc = to_intel_crtc(crtc); int pipe = intel_crtc->pipe; - struct drm_plane_state *cursor_state = crtc->cursor->state; - int x = cursor_state->crtc_x; - int y = cursor_state->crtc_y; - u32 base = 0, pos = 0; - - base = intel_crtc->cursor_addr; - - if (x >= intel_crtc->config->pipe_src_w) - on = false; - - if (y >= intel_crtc->config->pipe_src_h) - on = false; + u32 base = intel_crtc->cursor_addr; + u32 pos = 0; - if (x < 0) { - if (x + cursor_state->crtc_w <= 0) - on = false; + if (plane_state) { + int x = plane_state->base.crtc_x; + int y = plane_state->base.crtc_y; - pos |= CURSOR_POS_SIGN << CURSOR_X_SHIFT; - x = -x; - } - pos |= x << CURSOR_X_SHIFT; + if (x < 0) { + pos |= CURSOR_POS_SIGN << CURSOR_X_SHIFT; + x = -x; + } + pos |= x << CURSOR_X_SHIFT; - if (y < 0) { - if (y + cursor_state->crtc_h <= 0) - on = false; + if (y < 0) { + pos |= CURSOR_POS_SIGN << CURSOR_Y_SHIFT; + y = -y; + } + pos |= y << CURSOR_Y_SHIFT; - pos |= CURSOR_POS_SIGN << CURSOR_Y_SHIFT; - y = -y; + /* ILK+ do this automagically */ + if (HAS_GMCH_DISPLAY(dev) && + plane_state->base.rotation == BIT(DRM_ROTATE_180)) { + base += (plane_state->base.crtc_h * + plane_state->base.crtc_w - 1) * 4; + } } - pos |= y << CURSOR_Y_SHIFT; I915_WRITE(CURPOS(pipe), pos); - /* ILK+ do this automagically */ - if (HAS_GMCH_DISPLAY(dev) && - crtc->cursor->state->rotation == BIT(DRM_ROTATE_180)) { - base += (cursor_state->crtc_h * - cursor_state->crtc_w - 1) * 4; - } - if (IS_845G(dev) || IS_I865G(dev)) - i845_update_cursor(crtc, base, on); + i845_update_cursor(crtc, base, plane_state); else - i9xx_update_cursor(crtc, base, on); + i9xx_update_cursor(crtc, base, plane_state); } static bool cursor_size_ok(struct drm_device *dev, @@ -10498,7 +10521,6 @@ retry: } connector_state->crtc = crtc; - connector_state->best_encoder = &intel_encoder->base; crtc_state = intel_atomic_get_crtc_state(state, intel_crtc); if (IS_ERR(crtc_state)) { @@ -10594,7 +10616,6 @@ void intel_release_load_detect_pipe(struct drm_connector *connector, if (IS_ERR(crtc_state)) goto fail; - connector_state->best_encoder = NULL; connector_state->crtc = NULL; crtc_state->base.enable = crtc_state->base.active = false; @@ -10778,7 +10799,7 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev, struct intel_crtc *intel_crtc = to_intel_crtc(crtc); enum transcoder cpu_transcoder = intel_crtc->config->cpu_transcoder; struct drm_display_mode *mode; - struct intel_crtc_state pipe_config; + struct intel_crtc_state *pipe_config; int htot = I915_READ(HTOTAL(cpu_transcoder)); int hsync = I915_READ(HSYNC(cpu_transcoder)); int vtot = I915_READ(VTOTAL(cpu_transcoder)); @@ -10789,6 +10810,12 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev, if (!mode) return NULL; + pipe_config = kzalloc(sizeof(*pipe_config), GFP_KERNEL); + if (!pipe_config) { + kfree(mode); + return NULL; + } + /* * Construct a pipe_config sufficient for getting the clock info * back out of crtc_clock_get. @@ -10796,14 +10823,14 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev, * Note, if LVDS ever uses a non-1 pixel multiplier, we'll need * to use a real value here instead. */ - pipe_config.cpu_transcoder = (enum transcoder) pipe; - pipe_config.pixel_multiplier = 1; - pipe_config.dpll_hw_state.dpll = I915_READ(DPLL(pipe)); - pipe_config.dpll_hw_state.fp0 = I915_READ(FP0(pipe)); - pipe_config.dpll_hw_state.fp1 = I915_READ(FP1(pipe)); - i9xx_crtc_clock_get(intel_crtc, &pipe_config); - - mode->clock = pipe_config.port_clock / pipe_config.pixel_multiplier; + pipe_config->cpu_transcoder = (enum transcoder) pipe; + pipe_config->pixel_multiplier = 1; + pipe_config->dpll_hw_state.dpll = I915_READ(DPLL(pipe)); + pipe_config->dpll_hw_state.fp0 = I915_READ(FP0(pipe)); + pipe_config->dpll_hw_state.fp1 = I915_READ(FP1(pipe)); + i9xx_crtc_clock_get(intel_crtc, pipe_config); + + mode->clock = pipe_config->port_clock / pipe_config->pixel_multiplier; mode->hdisplay = (htot & 0xffff) + 1; mode->htotal = ((htot & 0xffff0000) >> 16) + 1; mode->hsync_start = (hsync & 0xffff) + 1; @@ -10815,6 +10842,8 @@ struct drm_display_mode *intel_crtc_mode_get(struct drm_device *dev, drm_mode_set_name(mode); + kfree(pipe_config); + return mode; } @@ -11319,13 +11348,12 @@ static void skl_do_mmio_flip(struct intel_crtc *intel_crtc, */ if (intel_rotation_90_or_270(rotation)) { /* stride = Surface height in tiles */ - tile_height = intel_tile_height(dev, fb->pixel_format, - fb->modifier[0], 0); + tile_height = intel_tile_height(dev_priv, fb->modifier[0], 0); stride = DIV_ROUND_UP(fb->height, tile_height); } else { stride = fb->pitches[0] / - intel_fb_stride_alignment(dev, fb->modifier[0], - fb->pixel_format); + intel_fb_stride_alignment(dev_priv, fb->modifier[0], + fb->pixel_format); } /* @@ -11660,9 +11688,11 @@ static int intel_crtc_page_flip(struct drm_crtc *crtc, obj->last_write_req); } else { if (!request) { - ret = i915_gem_request_alloc(ring, ring->default_context, &request); - if (ret) + request = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(request)) { + ret = PTR_ERR(request); goto cleanup_unpin; + } } ret = dev_priv->display.queue_flip(dev, crtc, fb, obj, request, @@ -11831,8 +11861,13 @@ int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, if (!was_crtc_enabled && WARN_ON(was_visible)) was_visible = false; - if (!is_crtc_enabled && WARN_ON(visible)) - visible = false; + /* + * Visibility is calculated as if the crtc was on, but + * after scaler setup everything depends on it being off + * when the crtc isn't active. + */ + if (!is_crtc_enabled) + to_intel_plane_state(plane_state)->visible = visible = false; if (!was_visible && !visible) return 0; @@ -12529,19 +12564,22 @@ intel_compare_m_n(unsigned int m, unsigned int n, BUILD_BUG_ON(DATA_LINK_M_N_MASK > INT_MAX); - if (m > m2) { - while (m > m2) { + if (n > n2) { + while (n > n2) { m2 <<= 1; n2 <<= 1; } - } else if (m < m2) { - while (m < m2) { + } else if (n < n2) { + while (n < n2) { m <<= 1; n <<= 1; } } - return m == m2 && n == n2; + if (n != n2) + return false; + + return intel_fuzzy_clock_check(m, m2); } static bool @@ -13216,15 +13254,27 @@ static int intel_modeset_all_pipes(struct drm_atomic_state *state) static int intel_modeset_checks(struct drm_atomic_state *state) { - struct drm_device *dev = state->dev; - struct drm_i915_private *dev_priv = dev->dev_private; - int ret; + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); + struct drm_i915_private *dev_priv = state->dev->dev_private; + struct drm_crtc *crtc; + struct drm_crtc_state *crtc_state; + int ret = 0, i; if (!check_digital_port_conflicts(state)) { DRM_DEBUG_KMS("rejecting conflicting digital port configuration\n"); return -EINVAL; } + intel_state->modeset = true; + intel_state->active_crtcs = dev_priv->active_crtcs; + + for_each_crtc_in_state(state, crtc, crtc_state, i) { + if (crtc_state->active) + intel_state->active_crtcs |= 1 << i; + else + intel_state->active_crtcs &= ~(1 << i); + } + /* * See if the config requires any additional preparation, e.g. * to adjust global state with pipes off. We need to do this @@ -13233,22 +13283,19 @@ static int intel_modeset_checks(struct drm_atomic_state *state) * adjusted_mode bits in the crtc directly. */ if (dev_priv->display.modeset_calc_cdclk) { - unsigned int cdclk; - ret = dev_priv->display.modeset_calc_cdclk(state); - cdclk = to_intel_atomic_state(state)->cdclk; - if (!ret && cdclk != dev_priv->cdclk_freq) + if (!ret && intel_state->dev_cdclk != dev_priv->cdclk_freq) ret = intel_modeset_all_pipes(state); if (ret < 0) return ret; } else - to_intel_atomic_state(state)->cdclk = dev_priv->cdclk_freq; + to_intel_atomic_state(state)->cdclk = dev_priv->atomic_cdclk_freq; intel_modeset_clear_plls(state); - if (IS_HASWELL(dev)) + if (IS_HASWELL(dev_priv)) return haswell_mode_set_planes_workaround(state); return 0; @@ -13466,12 +13513,12 @@ static int intel_atomic_commit(struct drm_device *dev, struct drm_atomic_state *state, bool async) { + struct intel_atomic_state *intel_state = to_intel_atomic_state(state); struct drm_i915_private *dev_priv = dev->dev_private; struct drm_crtc_state *crtc_state; struct drm_crtc *crtc; - int ret = 0; - int i; - bool any_ms = false; + int ret = 0, i; + bool hw_check = intel_state->modeset; ret = intel_atomic_prepare_commit(dev, state, async); if (ret) { @@ -13482,13 +13529,19 @@ static int intel_atomic_commit(struct drm_device *dev, drm_atomic_helper_swap_state(dev, state); dev_priv->wm.config = to_intel_atomic_state(state)->wm_config; + if (intel_state->modeset) { + memcpy(dev_priv->min_pixclk, intel_state->min_pixclk, + sizeof(intel_state->min_pixclk)); + dev_priv->active_crtcs = intel_state->active_crtcs; + dev_priv->atomic_cdclk_freq = intel_state->cdclk; + } + for_each_crtc_in_state(state, crtc, crtc_state, i) { struct intel_crtc *intel_crtc = to_intel_crtc(crtc); if (!needs_modeset(crtc->state)) continue; - any_ms = true; intel_pre_plane_update(intel_crtc); if (crtc_state->active) { @@ -13513,7 +13566,7 @@ static int intel_atomic_commit(struct drm_device *dev, * update the the output configuration. */ intel_modeset_update_crtc_state(state); - if (any_ms) { + if (intel_state->modeset) { intel_shared_dpll_commit(state); drm_atomic_helper_update_legacy_modeset_state(state->dev, state); @@ -13540,7 +13593,7 @@ static int intel_atomic_commit(struct drm_device *dev, put_domains = modeset_get_crtc_power_domains(crtc); /* make sure intel_modeset_check_state runs */ - any_ms = true; + hw_check = true; } if (!modeset) @@ -13567,11 +13620,24 @@ static int intel_atomic_commit(struct drm_device *dev, drm_atomic_helper_cleanup_planes(dev, state); mutex_unlock(&dev->struct_mutex); - if (any_ms) + if (hw_check) intel_modeset_check_state(dev, state); drm_atomic_state_free(state); + /* As one of the primary mmio accessors, KMS has a high likelihood + * of triggering bugs in unclaimed access. After we finish + * modesetting, see if an error has been flagged, and if so + * enable debugging for the next modeset - and hope we catch + * the culprit. + * + * XXX note that we assume display power is on at this point. + * This might hold true now but we need to add pm helper to check + * unclaimed only when the hardware is on, as atomic commits + * can happen also when the device is completely off. + */ + intel_uncore_arm_unclaimed_mmio_detection(dev_priv); + return 0; } @@ -13860,7 +13926,7 @@ skl_max_scale(struct intel_crtc *intel_crtc, struct intel_crtc_state *crtc_state struct drm_i915_private *dev_priv; int crtc_clock, cdclk; - if (!intel_crtc || !crtc_state) + if (!intel_crtc || !crtc_state->base.enable) return DRM_PLANE_HELPER_NO_SCALING; dev = intel_crtc->base.dev; @@ -13909,32 +13975,6 @@ intel_check_primary_plane(struct drm_plane *plane, &state->visible); } -static void -intel_commit_primary_plane(struct drm_plane *plane, - struct intel_plane_state *state) -{ - struct drm_crtc *crtc = state->base.crtc; - struct drm_framebuffer *fb = state->base.fb; - struct drm_device *dev = plane->dev; - struct drm_i915_private *dev_priv = dev->dev_private; - - crtc = crtc ? crtc : plane->crtc; - - dev_priv->display.update_primary_plane(crtc, fb, - state->src.x1 >> 16, - state->src.y1 >> 16); -} - -static void -intel_disable_primary_plane(struct drm_plane *plane, - struct drm_crtc *crtc) -{ - struct drm_device *dev = plane->dev; - struct drm_i915_private *dev_priv = dev->dev_private; - - dev_priv->display.update_primary_plane(crtc, NULL, 0, 0); -} - static void intel_begin_crtc_commit(struct drm_crtc *crtc, struct drm_crtc_state *old_crtc_state) { @@ -14019,20 +14059,33 @@ static struct drm_plane *intel_primary_plane_create(struct drm_device *dev, primary->plane = pipe; primary->frontbuffer_bit = INTEL_FRONTBUFFER_PRIMARY(pipe); primary->check_plane = intel_check_primary_plane; - primary->commit_plane = intel_commit_primary_plane; - primary->disable_plane = intel_disable_primary_plane; if (HAS_FBC(dev) && INTEL_INFO(dev)->gen < 4) primary->plane = !pipe; if (INTEL_INFO(dev)->gen >= 9) { intel_primary_formats = skl_primary_formats; num_formats = ARRAY_SIZE(skl_primary_formats); + + primary->update_plane = skylake_update_primary_plane; + primary->disable_plane = skylake_disable_primary_plane; + } else if (HAS_PCH_SPLIT(dev)) { + intel_primary_formats = i965_primary_formats; + num_formats = ARRAY_SIZE(i965_primary_formats); + + primary->update_plane = ironlake_update_primary_plane; + primary->disable_plane = i9xx_disable_primary_plane; } else if (INTEL_INFO(dev)->gen >= 4) { intel_primary_formats = i965_primary_formats; num_formats = ARRAY_SIZE(i965_primary_formats); + + primary->update_plane = i9xx_update_primary_plane; + primary->disable_plane = i9xx_disable_primary_plane; } else { intel_primary_formats = i8xx_primary_formats; num_formats = ARRAY_SIZE(i8xx_primary_formats); + + primary->update_plane = i9xx_update_primary_plane; + primary->disable_plane = i9xx_disable_primary_plane; } drm_universal_plane_init(dev, &primary->base, 0, @@ -14131,22 +14184,23 @@ static void intel_disable_cursor_plane(struct drm_plane *plane, struct drm_crtc *crtc) { - intel_crtc_update_cursor(crtc, false); + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); + + intel_crtc->cursor_addr = 0; + intel_crtc_update_cursor(crtc, NULL); } static void -intel_commit_cursor_plane(struct drm_plane *plane, - struct intel_plane_state *state) +intel_update_cursor_plane(struct drm_plane *plane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *state) { - struct drm_crtc *crtc = state->base.crtc; + struct drm_crtc *crtc = crtc_state->base.crtc; + struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct drm_device *dev = plane->dev; - struct intel_crtc *intel_crtc; struct drm_i915_gem_object *obj = intel_fb_obj(state->base.fb); uint32_t addr; - crtc = crtc ? crtc : plane->crtc; - intel_crtc = to_intel_crtc(crtc); - if (!obj) addr = 0; else if (!INTEL_INFO(dev)->cursor_needs_physical) @@ -14155,9 +14209,7 @@ intel_commit_cursor_plane(struct drm_plane *plane, addr = obj->phys_handle->busaddr; intel_crtc->cursor_addr = addr; - - if (crtc->state->active) - intel_crtc_update_cursor(crtc, state->visible); + intel_crtc_update_cursor(crtc, state); } static struct drm_plane *intel_cursor_plane_create(struct drm_device *dev, @@ -14183,7 +14235,7 @@ static struct drm_plane *intel_cursor_plane_create(struct drm_device *dev, cursor->plane = pipe; cursor->frontbuffer_bit = INTEL_FRONTBUFFER_CURSOR(pipe); cursor->check_plane = intel_check_cursor_plane; - cursor->commit_plane = intel_commit_cursor_plane; + cursor->update_plane = intel_update_cursor_plane; cursor->disable_plane = intel_disable_cursor_plane; drm_universal_plane_init(dev, &cursor->base, 0, @@ -14657,6 +14709,7 @@ static int intel_framebuffer_init(struct drm_device *dev, struct drm_mode_fb_cmd2 *mode_cmd, struct drm_i915_gem_object *obj) { + struct drm_i915_private *dev_priv = to_i915(dev); unsigned int aligned_height; int ret; u32 pitch_limit, stride_alignment; @@ -14698,7 +14751,8 @@ static int intel_framebuffer_init(struct drm_device *dev, return -EINVAL; } - stride_alignment = intel_fb_stride_alignment(dev, mode_cmd->modifier[0], + stride_alignment = intel_fb_stride_alignment(dev_priv, + mode_cmd->modifier[0], mode_cmd->pixel_format); if (mode_cmd->pitches[0] & (stride_alignment - 1)) { DRM_DEBUG("pitch (%d) must be at least %u byte aligned\n", @@ -14790,7 +14844,6 @@ static int intel_framebuffer_init(struct drm_device *dev, drm_helper_mode_fill_fb_struct(&intel_fb->base, mode_cmd); intel_fb->obj = obj; - intel_fb->obj->framebuffer_references++; ret = drm_framebuffer_init(dev, &intel_fb->base, &intel_fb_funcs); if (ret) { @@ -14798,6 +14851,8 @@ static int intel_framebuffer_init(struct drm_device *dev, return ret; } + intel_fb->obj->framebuffer_references++; + return 0; } @@ -14861,8 +14916,6 @@ static void intel_init_display(struct drm_device *dev) haswell_crtc_compute_clock; dev_priv->display.crtc_enable = haswell_crtc_enable; dev_priv->display.crtc_disable = haswell_crtc_disable; - dev_priv->display.update_primary_plane = - skylake_update_primary_plane; } else if (HAS_DDI(dev)) { dev_priv->display.get_pipe_config = haswell_get_pipe_config; dev_priv->display.get_initial_plane_config = @@ -14871,8 +14924,6 @@ static void intel_init_display(struct drm_device *dev) haswell_crtc_compute_clock; dev_priv->display.crtc_enable = haswell_crtc_enable; dev_priv->display.crtc_disable = haswell_crtc_disable; - dev_priv->display.update_primary_plane = - ironlake_update_primary_plane; } else if (HAS_PCH_SPLIT(dev)) { dev_priv->display.get_pipe_config = ironlake_get_pipe_config; dev_priv->display.get_initial_plane_config = @@ -14881,8 +14932,6 @@ static void intel_init_display(struct drm_device *dev) ironlake_crtc_compute_clock; dev_priv->display.crtc_enable = ironlake_crtc_enable; dev_priv->display.crtc_disable = ironlake_crtc_disable; - dev_priv->display.update_primary_plane = - ironlake_update_primary_plane; } else if (IS_VALLEYVIEW(dev) || IS_CHERRYVIEW(dev)) { dev_priv->display.get_pipe_config = i9xx_get_pipe_config; dev_priv->display.get_initial_plane_config = @@ -14890,8 +14939,6 @@ static void intel_init_display(struct drm_device *dev) dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock; dev_priv->display.crtc_enable = valleyview_crtc_enable; dev_priv->display.crtc_disable = i9xx_crtc_disable; - dev_priv->display.update_primary_plane = - i9xx_update_primary_plane; } else { dev_priv->display.get_pipe_config = i9xx_get_pipe_config; dev_priv->display.get_initial_plane_config = @@ -14899,8 +14946,6 @@ static void intel_init_display(struct drm_device *dev) dev_priv->display.crtc_compute_clock = i9xx_crtc_compute_clock; dev_priv->display.crtc_enable = i9xx_crtc_enable; dev_priv->display.crtc_disable = i9xx_crtc_disable; - dev_priv->display.update_primary_plane = - i9xx_update_primary_plane; } /* Returns the core display clock speed */ @@ -15206,12 +15251,89 @@ static void i915_disable_vga(struct drm_device *dev) void intel_modeset_init_hw(struct drm_device *dev) { + struct drm_i915_private *dev_priv = dev->dev_private; + intel_update_cdclk(dev); - intel_prepare_ddi(dev); + + dev_priv->atomic_cdclk_freq = dev_priv->cdclk_freq; + intel_init_clock_gating(dev); intel_enable_gt_powersave(dev); } +/* + * Calculate what we think the watermarks should be for the state we've read + * out of the hardware and then immediately program those watermarks so that + * we ensure the hardware settings match our internal state. + * + * We can calculate what we think WM's should be by creating a duplicate of the + * current state (which was constructed during hardware readout) and running it + * through the atomic check code to calculate new watermark values in the + * state object. + */ +static void sanitize_watermarks(struct drm_device *dev) +{ + struct drm_i915_private *dev_priv = to_i915(dev); + struct drm_atomic_state *state; + struct drm_crtc *crtc; + struct drm_crtc_state *cstate; + struct drm_modeset_acquire_ctx ctx; + int ret; + int i; + + /* Only supported on platforms that use atomic watermark design */ + if (!dev_priv->display.program_watermarks) + return; + + /* + * We need to hold connection_mutex before calling duplicate_state so + * that the connector loop is protected. + */ + drm_modeset_acquire_init(&ctx, 0); +retry: + ret = drm_modeset_lock_all_ctx(dev, &ctx); + if (ret == -EDEADLK) { + drm_modeset_backoff(&ctx); + goto retry; + } else if (WARN_ON(ret)) { + goto fail; + } + + state = drm_atomic_helper_duplicate_state(dev, &ctx); + if (WARN_ON(IS_ERR(state))) + goto fail; + + ret = intel_atomic_check(dev, state); + if (ret) { + /* + * If we fail here, it means that the hardware appears to be + * programmed in a way that shouldn't be possible, given our + * understanding of watermark requirements. This might mean a + * mistake in the hardware readout code or a mistake in the + * watermark calculations for a given platform. Raise a WARN + * so that this is noticeable. + * + * If this actually happens, we'll have to just leave the + * BIOS-programmed watermarks untouched and hope for the best. + */ + WARN(true, "Could not determine valid watermarks for inherited state\n"); + goto fail; + } + + /* Write calculated watermark values back */ + to_i915(dev)->wm.config = to_intel_atomic_state(state)->wm_config; + for_each_crtc_in_state(state, crtc, cstate, i) { + struct intel_crtc_state *cs = to_intel_crtc_state(cstate); + + dev_priv->display.program_watermarks(cs); + } + + drm_atomic_state_free(state); +fail: + drm_modeset_drop_locks(&ctx); + drm_modeset_acquire_fini(&ctx); +} + void intel_modeset_init(struct drm_device *dev) { struct drm_i915_private *dev_priv = dev->dev_private; @@ -15332,6 +15454,13 @@ void intel_modeset_init(struct drm_device *dev) */ intel_find_initial_plane_obj(crtc, &plane_config); } + + /* + * Make sure hardware watermarks really match the state we read out. + * Note that we need to do this after reconstructing the BIOS fb's + * since the watermark calculation done here will use pstate->fb. + */ + sanitize_watermarks(dev); } static void intel_enable_pipe_a(struct drm_device *dev) @@ -15462,6 +15591,7 @@ static void intel_sanitize_crtc(struct intel_crtc *crtc) crtc->base.state->active = crtc->active; crtc->base.enabled = crtc->active; crtc->base.state->connector_mask = 0; + crtc->base.state->encoder_mask = 0; /* Because we only establish the connector -> encoder -> * crtc links if something is active, this means the @@ -15604,16 +15734,40 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) struct intel_connector *connector; int i; + dev_priv->active_crtcs = 0; + for_each_intel_crtc(dev, crtc) { - __drm_atomic_helper_crtc_destroy_state(&crtc->base, crtc->base.state); - memset(crtc->config, 0, sizeof(*crtc->config)); - crtc->config->base.crtc = &crtc->base; + struct intel_crtc_state *crtc_state = crtc->config; + int pixclk = 0; - crtc->active = dev_priv->display.get_pipe_config(crtc, - crtc->config); + __drm_atomic_helper_crtc_destroy_state(&crtc->base, &crtc_state->base); + memset(crtc_state, 0, sizeof(*crtc_state)); + crtc_state->base.crtc = &crtc->base; - crtc->base.state->active = crtc->active; - crtc->base.enabled = crtc->active; + crtc_state->base.active = crtc_state->base.enable = + dev_priv->display.get_pipe_config(crtc, crtc_state); + + crtc->base.enabled = crtc_state->base.enable; + crtc->active = crtc_state->base.active; + + if (crtc_state->base.active) { + dev_priv->active_crtcs |= 1 << crtc->pipe; + + if (IS_BROADWELL(dev_priv)) { + pixclk = ilk_pipe_pixel_rate(crtc_state); + + /* pixel rate mustn't exceed 95% of cdclk with IPS on BDW */ + if (crtc_state->ips_enabled) + pixclk = DIV_ROUND_UP(pixclk * 100, 95); + } else if (IS_VALLEYVIEW(dev_priv) || + IS_CHERRYVIEW(dev_priv) || + IS_BROXTON(dev_priv)) + pixclk = crtc_state->base.adjusted_mode.crtc_clock; + else + WARN_ON(dev_priv->display.modeset_calc_cdclk); + } + + dev_priv->min_pixclk[crtc->pipe] = pixclk; readout_plane_state(crtc); @@ -15677,6 +15831,8 @@ static void intel_modeset_readout_hw_state(struct drm_device *dev) */ encoder->base.crtc->state->connector_mask |= 1 << drm_connector_index(&connector->base); + encoder->base.crtc->state->encoder_mask |= + 1 << drm_encoder_index(&encoder->base); } } else { @@ -16117,7 +16273,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m, for_each_pipe(dev_priv, i) { err_printf(m, "Pipe [%d]:\n", i); err_printf(m, " Power: %s\n", - error->pipe[i].power_domain_on ? "on" : "off"); + onoff(error->pipe[i].power_domain_on)); err_printf(m, " SRC: %08x\n", error->pipe[i].source); err_printf(m, " STAT: %08x\n", error->pipe[i].stat); @@ -16145,7 +16301,7 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m, err_printf(m, "CPU transcoder: %c\n", transcoder_name(error->transcoder[i].cpu_transcoder)); err_printf(m, " Power: %s\n", - error->transcoder[i].power_domain_on ? "on" : "off"); + onoff(error->transcoder[i].power_domain_on)); err_printf(m, " CONF: %08x\n", error->transcoder[i].conf); err_printf(m, " HTOTAL: %08x\n", error->transcoder[i].htotal); err_printf(m, " HBLANK: %08x\n", error->transcoder[i].hblank); @@ -16155,24 +16311,3 @@ intel_display_print_error_state(struct drm_i915_error_state_buf *m, err_printf(m, " VSYNC: %08x\n", error->transcoder[i].vsync); } } - -void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct intel_crtc *crtc; - - for_each_intel_crtc(dev, crtc) { - struct intel_unpin_work *work; - - spin_lock_irq(&dev->event_lock); - - work = crtc->unpin_work; - - if (work && work->event && - work->event->base.file_priv == file) { - kfree(work->event); - work->event = NULL; - } - - spin_unlock_irq(&dev->event_lock); - } -} diff --git a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c index 796e3d313cb9..da704c6ee5cb 100644 --- a/drivers/gpu/drm/i915/intel_dp.c +++ b/drivers/gpu/drm/i915/intel_dp.c @@ -157,14 +157,9 @@ intel_dp_max_link_bw(struct intel_dp *intel_dp) static u8 intel_dp_max_lane_count(struct intel_dp *intel_dp) { struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); - struct drm_device *dev = intel_dig_port->base.base.dev; u8 source_max, sink_max; - source_max = 4; - if (HAS_DDI(dev) && intel_dig_port->port == PORT_A && - (intel_dig_port->saved_port_bits & DDI_A_4_LANES) == 0) - source_max = 2; - + source_max = intel_dig_port->max_lanes; sink_max = drm_dp_max_lane_count(intel_dp->dpcd); return min(source_max, sink_max); @@ -340,8 +335,12 @@ vlv_power_sequencer_kick(struct intel_dp *intel_dp) release_cl_override = IS_CHERRYVIEW(dev) && !chv_phy_powergate_ch(dev_priv, phy, ch, true); - vlv_force_pll_on(dev, pipe, IS_CHERRYVIEW(dev) ? - &chv_dpll[0].dpll : &vlv_dpll[0].dpll); + if (vlv_force_pll_on(dev, pipe, IS_CHERRYVIEW(dev) ? + &chv_dpll[0].dpll : &vlv_dpll[0].dpll)) { + DRM_ERROR("Failed to force on pll for pipe %c!\n", + pipe_name(pipe)); + return; + } } /* @@ -1189,7 +1188,6 @@ intel_dp_aux_fini(struct intel_dp *intel_dp) static int intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) { - struct drm_device *dev = intel_dp_to_dev(intel_dp); struct intel_digital_port *intel_dig_port = dp_to_dig_port(intel_dp); enum port port = intel_dig_port->port; int ret; @@ -1200,7 +1198,7 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) if (!intel_dp->aux.name) return -ENOMEM; - intel_dp->aux.dev = dev->dev; + intel_dp->aux.dev = connector->base.kdev; intel_dp->aux.transfer = intel_dp_aux_transfer; DRM_DEBUG_KMS("registering %s bus for %s\n", @@ -1215,16 +1213,6 @@ intel_dp_aux_init(struct intel_dp *intel_dp, struct intel_connector *connector) return ret; } - ret = sysfs_create_link(&connector->base.kdev->kobj, - &intel_dp->aux.ddc.dev.kobj, - intel_dp->aux.ddc.dev.kobj.name); - if (ret < 0) { - DRM_ERROR("sysfs_create_link() for %s failed (%d)\n", - intel_dp->aux.name, ret); - intel_dp_aux_fini(intel_dp); - return ret; - } - return 0; } @@ -1233,9 +1221,7 @@ intel_dp_connector_unregister(struct intel_connector *intel_connector) { struct intel_dp *intel_dp = intel_attached_dp(&intel_connector->base); - if (!intel_connector->mst_port) - sysfs_remove_link(&intel_connector->base.kdev->kobj, - intel_dp->aux.ddc.dev.kobj.name); + intel_dp_aux_fini(intel_dp); intel_connector_unregister(intel_connector); } @@ -2243,11 +2229,6 @@ static void intel_edp_backlight_power(struct intel_connector *connector, _intel_edp_backlight_off(intel_dp); } -static const char *state_string(bool enabled) -{ - return enabled ? "on" : "off"; -} - static void assert_dp_port(struct intel_dp *intel_dp, bool state) { struct intel_digital_port *dig_port = dp_to_dig_port(intel_dp); @@ -2257,7 +2238,7 @@ static void assert_dp_port(struct intel_dp *intel_dp, bool state) I915_STATE_WARN(cur_state != state, "DP port %c state assertion failure (expected %s, current %s)\n", port_name(dig_port->port), - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } #define assert_dp_port_disabled(d) assert_dp_port((d), false) @@ -2267,7 +2248,7 @@ static void assert_edp_pll(struct drm_i915_private *dev_priv, bool state) I915_STATE_WARN(cur_state != state, "eDP PLL state assertion failure (expected %s, current %s)\n", - state_string(state), state_string(cur_state)); + onoff(state), onoff(cur_state)); } #define assert_edp_pll_enabled(d) assert_edp_pll((d), true) #define assert_edp_pll_disabled(d) assert_edp_pll((d), false) @@ -4874,7 +4855,6 @@ void intel_dp_encoder_destroy(struct drm_encoder *encoder) struct intel_digital_port *intel_dig_port = enc_to_dig_port(encoder); struct intel_dp *intel_dp = &intel_dig_port->dp; - intel_dp_aux_fini(intel_dp); intel_dp_mst_encoder_cleanup(intel_dig_port); if (is_edp(intel_dp)) { cancel_delayed_work_sync(&intel_dp->panel_vdd_work); @@ -5839,6 +5819,11 @@ intel_dp_init_connector(struct intel_digital_port *intel_dig_port, enum port port = intel_dig_port->port; int type, ret; + if (WARN(intel_dig_port->max_lanes < 1, + "Not enough lanes (%d) for DP on port %c\n", + intel_dig_port->max_lanes, port_name(port))) + return false; + intel_dp->pps_pipe = INVALID_PIPE; /* intel_dp vfuncs */ @@ -6037,6 +6022,7 @@ intel_dp_init(struct drm_device *dev, intel_dig_port->port = port; dev_priv->dig_port_map[port] = intel_encoder; intel_dig_port->dp.output_reg = output_reg; + intel_dig_port->max_lanes = 4; intel_encoder->type = INTEL_OUTPUT_DISPLAYPORT; if (IS_CHERRYVIEW(dev)) { diff --git a/drivers/gpu/drm/i915/intel_dp_mst.c b/drivers/gpu/drm/i915/intel_dp_mst.c index fa0dabf578dc..2a2ab306ad84 100644 --- a/drivers/gpu/drm/i915/intel_dp_mst.c +++ b/drivers/gpu/drm/i915/intel_dp_mst.c @@ -184,7 +184,9 @@ static void intel_mst_pre_enable_dp(struct intel_encoder *encoder) intel_mst->port = found->port; if (intel_dp->active_mst_links == 0) { - intel_ddi_clk_select(encoder, intel_crtc->config); + intel_prepare_ddi_buffer(&intel_dig_port->base); + + intel_ddi_clk_select(&intel_dig_port->base, intel_crtc->config); intel_dp_set_link_params(intel_dp, intel_crtc->config); diff --git a/drivers/gpu/drm/i915/intel_drv.h b/drivers/gpu/drm/i915/intel_drv.h index ea5415851c6e..bf6f98134b50 100644 --- a/drivers/gpu/drm/i915/intel_drv.h +++ b/drivers/gpu/drm/i915/intel_drv.h @@ -246,7 +246,18 @@ struct intel_atomic_state { struct drm_atomic_state base; unsigned int cdclk; - bool dpll_set; + + /* + * Calculated device cdclk, can be different from cdclk + * only when all crtc's are DPMS off. + */ + unsigned int dev_cdclk; + + bool dpll_set, modeset; + + unsigned int active_crtcs; + unsigned int min_pixclk[I915_MAX_PIPES]; + struct intel_shared_dpll_config shared_dpll[I915_NUM_PLLS]; struct intel_wm_config wm_config; }; @@ -647,23 +658,17 @@ struct intel_plane { /* * NOTE: Do not place new plane state fields here (e.g., when adding * new plane properties). New runtime state should now be placed in - * the intel_plane_state structure and accessed via drm_plane->state. + * the intel_plane_state structure and accessed via plane_state. */ void (*update_plane)(struct drm_plane *plane, - struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t x, uint32_t y, - uint32_t src_w, uint32_t src_h); + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state); void (*disable_plane)(struct drm_plane *plane, struct drm_crtc *crtc); int (*check_plane)(struct drm_plane *plane, struct intel_crtc_state *crtc_state, struct intel_plane_state *state); - void (*commit_plane)(struct drm_plane *plane, - struct intel_plane_state *state); }; struct intel_watermark_params { @@ -817,6 +822,7 @@ struct intel_digital_port { struct intel_hdmi hdmi; enum irqreturn (*hpd_pulse)(struct intel_digital_port *, bool); bool release_cl2_override; + uint8_t max_lanes; /* for communication with audio component; protected by av_mutex */ const struct drm_connector *audio_connector; }; @@ -996,7 +1002,7 @@ void intel_crt_init(struct drm_device *dev); /* intel_ddi.c */ void intel_ddi_clk_select(struct intel_encoder *encoder, const struct intel_crtc_state *pipe_config); -void intel_prepare_ddi(struct drm_device *dev); +void intel_prepare_ddi_buffer(struct intel_encoder *encoder); void hsw_fdi_link_train(struct drm_crtc *crtc); void intel_ddi_init(struct drm_device *dev, enum port port); enum port intel_ddi_get_encoder_port(struct intel_encoder *intel_encoder); @@ -1041,8 +1047,8 @@ unsigned int intel_fb_align_height(struct drm_device *dev, uint64_t fb_format_modifier); void intel_fb_obj_flush(struct drm_i915_gem_object *obj, bool retire, enum fb_op_origin origin); -u32 intel_fb_stride_alignment(struct drm_device *dev, uint64_t fb_modifier, - uint32_t pixel_format); +u32 intel_fb_stride_alignment(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier, uint32_t pixel_format); /* intel_audio.c */ void intel_init_audio(struct drm_device *dev); @@ -1126,9 +1132,8 @@ int intel_plane_atomic_set_property(struct drm_plane *plane, int intel_plane_atomic_calc_changes(struct drm_crtc_state *crtc_state, struct drm_plane_state *plane_state); -unsigned int -intel_tile_height(struct drm_device *dev, uint32_t pixel_format, - uint64_t fb_format_modifier, unsigned int plane); +unsigned int intel_tile_height(const struct drm_i915_private *dev_priv, + uint64_t fb_modifier, unsigned int cpp); static inline bool intel_rotation_90_or_270(unsigned int rotation) @@ -1149,8 +1154,8 @@ void assert_shared_dpll(struct drm_i915_private *dev_priv, struct intel_shared_dpll *intel_get_shared_dpll(struct intel_crtc *crtc, struct intel_crtc_state *state); -void vlv_force_pll_on(struct drm_device *dev, enum pipe pipe, - const struct dpll *dpll); +int vlv_force_pll_on(struct drm_device *dev, enum pipe pipe, + const struct dpll *dpll); void vlv_force_pll_off(struct drm_device *dev, enum pipe pipe); /* modesetting asserts */ @@ -1167,11 +1172,11 @@ void assert_fdi_rx_pll(struct drm_i915_private *dev_priv, void assert_pipe(struct drm_i915_private *dev_priv, enum pipe pipe, bool state); #define assert_pipe_enabled(d, p) assert_pipe(d, p, true) #define assert_pipe_disabled(d, p) assert_pipe(d, p, false) -unsigned long intel_gen4_compute_page_offset(struct drm_i915_private *dev_priv, - int *x, int *y, - unsigned int tiling_mode, - unsigned int bpp, - unsigned int pitch); +unsigned long intel_compute_tile_offset(struct drm_i915_private *dev_priv, + int *x, int *y, + uint64_t fb_modifier, + unsigned int cpp, + unsigned int pitch); void intel_prepare_reset(struct drm_device *dev); void intel_finish_reset(struct drm_device *dev); void hsw_enable_pc8(struct drm_i915_private *dev_priv); @@ -1207,7 +1212,6 @@ enum intel_display_power_domain intel_display_port_aux_power_domain(struct intel_encoder *intel_encoder); void intel_mode_from_pipe_config(struct drm_display_mode *mode, struct intel_crtc_state *pipe_config); -void intel_modeset_preclose(struct drm_device *dev, struct drm_file *file); int skl_update_scaler_crtc(struct intel_crtc_state *crtc_state); int skl_max_scale(struct intel_crtc *crtc, struct intel_crtc_state *crtc_state); diff --git a/drivers/gpu/drm/i915/intel_dsi.c b/drivers/gpu/drm/i915/intel_dsi.c index 44742fa2f616..91cef3525c93 100644 --- a/drivers/gpu/drm/i915/intel_dsi.c +++ b/drivers/gpu/drm/i915/intel_dsi.c @@ -702,7 +702,7 @@ static bool intel_dsi_get_hw_state(struct intel_encoder *encoder, static void intel_dsi_get_config(struct intel_encoder *encoder, struct intel_crtc_state *pipe_config) { - u32 pclk = 0; + u32 pclk; DRM_DEBUG_KMS("\n"); pipe_config->has_dsi_encoder = true; @@ -713,12 +713,7 @@ static void intel_dsi_get_config(struct intel_encoder *encoder, */ pipe_config->dpll_hw_state.dpll_md = 0; - if (IS_BROXTON(encoder->base.dev)) - pclk = bxt_get_dsi_pclk(encoder, pipe_config->pipe_bpp); - else if (IS_VALLEYVIEW(encoder->base.dev) || - IS_CHERRYVIEW(encoder->base.dev)) - pclk = vlv_get_dsi_pclk(encoder, pipe_config->pipe_bpp); - + pclk = intel_dsi_get_pclk(encoder, pipe_config->pipe_bpp); if (!pclk) return; diff --git a/drivers/gpu/drm/i915/intel_dsi.h b/drivers/gpu/drm/i915/intel_dsi.h index 02551ff228c2..de7be7f3fb42 100644 --- a/drivers/gpu/drm/i915/intel_dsi.h +++ b/drivers/gpu/drm/i915/intel_dsi.h @@ -126,8 +126,7 @@ static inline struct intel_dsi *enc_to_intel_dsi(struct drm_encoder *encoder) extern void intel_enable_dsi_pll(struct intel_encoder *encoder); extern void intel_disable_dsi_pll(struct intel_encoder *encoder); -extern u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp); -extern u32 bxt_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp); +extern u32 intel_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp); extern void intel_dsi_reset_clocks(struct intel_encoder *encoder, enum port port); diff --git a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c index a5e99ac305da..1d43e6f37fc1 100644 --- a/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c +++ b/drivers/gpu/drm/i915/intel_dsi_panel_vbt.c @@ -229,14 +229,18 @@ static const u8 *mipi_exec_gpio(struct intel_dsi *intel_dsi, const u8 *data) return data; } +static const u8 *mipi_exec_i2c_skip(struct intel_dsi *intel_dsi, const u8 *data) +{ + return data + *(data + 6) + 7; +} + typedef const u8 * (*fn_mipi_elem_exec)(struct intel_dsi *intel_dsi, const u8 *data); static const fn_mipi_elem_exec exec_elem[] = { - NULL, /* reserved */ - mipi_exec_send_packet, - mipi_exec_delay, - mipi_exec_gpio, - NULL, /* status read; later */ + [MIPI_SEQ_ELEM_SEND_PKT] = mipi_exec_send_packet, + [MIPI_SEQ_ELEM_DELAY] = mipi_exec_delay, + [MIPI_SEQ_ELEM_GPIO] = mipi_exec_gpio, + [MIPI_SEQ_ELEM_I2C] = mipi_exec_i2c_skip, }; /* @@ -246,107 +250,114 @@ static const fn_mipi_elem_exec exec_elem[] = { */ static const char * const seq_name[] = { - "UNDEFINED", - "MIPI_SEQ_ASSERT_RESET", - "MIPI_SEQ_INIT_OTP", - "MIPI_SEQ_DISPLAY_ON", - "MIPI_SEQ_DISPLAY_OFF", - "MIPI_SEQ_DEASSERT_RESET" + [MIPI_SEQ_ASSERT_RESET] = "MIPI_SEQ_ASSERT_RESET", + [MIPI_SEQ_INIT_OTP] = "MIPI_SEQ_INIT_OTP", + [MIPI_SEQ_DISPLAY_ON] = "MIPI_SEQ_DISPLAY_ON", + [MIPI_SEQ_DISPLAY_OFF] = "MIPI_SEQ_DISPLAY_OFF", + [MIPI_SEQ_DEASSERT_RESET] = "MIPI_SEQ_DEASSERT_RESET", + [MIPI_SEQ_BACKLIGHT_ON] = "MIPI_SEQ_BACKLIGHT_ON", + [MIPI_SEQ_BACKLIGHT_OFF] = "MIPI_SEQ_BACKLIGHT_OFF", + [MIPI_SEQ_TEAR_ON] = "MIPI_SEQ_TEAR_ON", + [MIPI_SEQ_TEAR_OFF] = "MIPI_SEQ_TEAR_OFF", + [MIPI_SEQ_POWER_ON] = "MIPI_SEQ_POWER_ON", + [MIPI_SEQ_POWER_OFF] = "MIPI_SEQ_POWER_OFF", }; -static void generic_exec_sequence(struct intel_dsi *intel_dsi, const u8 *data) +static const char *sequence_name(enum mipi_seq seq_id) { + if (seq_id < ARRAY_SIZE(seq_name) && seq_name[seq_id]) + return seq_name[seq_id]; + else + return "(unknown)"; +} + +static void generic_exec_sequence(struct drm_panel *panel, enum mipi_seq seq_id) +{ + struct vbt_panel *vbt_panel = to_vbt_panel(panel); + struct intel_dsi *intel_dsi = vbt_panel->intel_dsi; + struct drm_i915_private *dev_priv = to_i915(intel_dsi->base.base.dev); + const u8 *data; fn_mipi_elem_exec mipi_elem_exec; - int index; - if (!data) + if (WARN_ON(seq_id >= ARRAY_SIZE(dev_priv->vbt.dsi.sequence))) + return; + + data = dev_priv->vbt.dsi.sequence[seq_id]; + if (!data) { + DRM_DEBUG_KMS("MIPI sequence %d - %s not available\n", + seq_id, sequence_name(seq_id)); return; + } - DRM_DEBUG_DRIVER("Starting MIPI sequence - %s\n", seq_name[*data]); + WARN_ON(*data != seq_id); - /* go to the first element of the sequence */ - data++; + DRM_DEBUG_KMS("Starting MIPI sequence %d - %s\n", + seq_id, sequence_name(seq_id)); - /* parse each byte till we reach end of sequence byte - 0x00 */ - while (1) { - index = *data; - mipi_elem_exec = exec_elem[index]; - if (!mipi_elem_exec) { - DRM_ERROR("Unsupported MIPI element, skipping sequence execution\n"); - return; - } + /* Skip Sequence Byte. */ + data++; - /* goto element payload */ - data++; + /* Skip Size of Sequence. */ + if (dev_priv->vbt.dsi.seq_version >= 3) + data += 4; - /* execute the element specific rotines */ - data = mipi_elem_exec(intel_dsi, data); + while (1) { + u8 operation_byte = *data++; + u8 operation_size = 0; - /* - * After processing the element, data should point to - * next element or end of sequence - * check if have we reached end of sequence - */ - if (*data == 0x00) + if (operation_byte == MIPI_SEQ_ELEM_END) break; + + if (operation_byte < ARRAY_SIZE(exec_elem)) + mipi_elem_exec = exec_elem[operation_byte]; + else + mipi_elem_exec = NULL; + + /* Size of Operation. */ + if (dev_priv->vbt.dsi.seq_version >= 3) + operation_size = *data++; + + if (mipi_elem_exec) { + data = mipi_elem_exec(intel_dsi, data); + } else if (operation_size) { + /* We have size, skip. */ + DRM_DEBUG_KMS("Unsupported MIPI operation byte %u\n", + operation_byte); + data += operation_size; + } else { + /* No size, can't skip without parsing. */ + DRM_ERROR("Unsupported MIPI operation byte %u\n", + operation_byte); + return; + } } } static int vbt_panel_prepare(struct drm_panel *panel) { - struct vbt_panel *vbt_panel = to_vbt_panel(panel); - struct intel_dsi *intel_dsi = vbt_panel->intel_dsi; - struct drm_device *dev = intel_dsi->base.base.dev; - struct drm_i915_private *dev_priv = dev->dev_private; - const u8 *sequence; - - sequence = dev_priv->vbt.dsi.sequence[MIPI_SEQ_ASSERT_RESET]; - generic_exec_sequence(intel_dsi, sequence); - - sequence = dev_priv->vbt.dsi.sequence[MIPI_SEQ_INIT_OTP]; - generic_exec_sequence(intel_dsi, sequence); + generic_exec_sequence(panel, MIPI_SEQ_ASSERT_RESET); + generic_exec_sequence(panel, MIPI_SEQ_INIT_OTP); return 0; } static int vbt_panel_unprepare(struct drm_panel *panel) { - struct vbt_panel *vbt_panel = to_vbt_panel(panel); - struct intel_dsi *intel_dsi = vbt_panel->intel_dsi; - struct drm_device *dev = intel_dsi->base.base.dev; - struct drm_i915_private *dev_priv = dev->dev_private; - const u8 *sequence; - - sequence = dev_priv->vbt.dsi.sequence[MIPI_SEQ_DEASSERT_RESET]; - generic_exec_sequence(intel_dsi, sequence); + generic_exec_sequence(panel, MIPI_SEQ_DEASSERT_RESET); return 0; } static int vbt_panel_enable(struct drm_panel *panel) { - struct vbt_panel *vbt_panel = to_vbt_panel(panel); - struct intel_dsi *intel_dsi = vbt_panel->intel_dsi; - struct drm_device *dev = intel_dsi->base.base.dev; - struct drm_i915_private *dev_priv = dev->dev_private; - const u8 *sequence; - - sequence = dev_priv->vbt.dsi.sequence[MIPI_SEQ_DISPLAY_ON]; - generic_exec_sequence(intel_dsi, sequence); + generic_exec_sequence(panel, MIPI_SEQ_DISPLAY_ON); return 0; } static int vbt_panel_disable(struct drm_panel *panel) { - struct vbt_panel *vbt_panel = to_vbt_panel(panel); - struct intel_dsi *intel_dsi = vbt_panel->intel_dsi; - struct drm_device *dev = intel_dsi->base.base.dev; - struct drm_i915_private *dev_priv = dev->dev_private; - const u8 *sequence; - - sequence = dev_priv->vbt.dsi.sequence[MIPI_SEQ_DISPLAY_OFF]; - generic_exec_sequence(intel_dsi, sequence); + generic_exec_sequence(panel, MIPI_SEQ_DISPLAY_OFF); return 0; } @@ -666,6 +677,8 @@ struct drm_panel *vbt_panel_init(struct intel_dsi *intel_dsi, u16 panel_id) /* This is cheating a bit with the cleanup. */ vbt_panel = devm_kzalloc(dev->dev, sizeof(*vbt_panel), GFP_KERNEL); + if (!vbt_panel) + return NULL; vbt_panel->intel_dsi = intel_dsi; drm_panel_init(&vbt_panel->panel); diff --git a/drivers/gpu/drm/i915/intel_dsi_pll.c b/drivers/gpu/drm/i915/intel_dsi_pll.c index fbd2b51810ca..bb5e95a1a453 100644 --- a/drivers/gpu/drm/i915/intel_dsi_pll.c +++ b/drivers/gpu/drm/i915/intel_dsi_pll.c @@ -30,14 +30,6 @@ #include "i915_drv.h" #include "intel_dsi.h" -#define DSI_HSS_PACKET_SIZE 4 -#define DSI_HSE_PACKET_SIZE 4 -#define DSI_HSA_PACKET_EXTRA_SIZE 6 -#define DSI_HBP_PACKET_EXTRA_SIZE 6 -#define DSI_HACTIVE_PACKET_EXTRA_SIZE 6 -#define DSI_HFP_PACKET_EXTRA_SIZE 6 -#define DSI_EOTP_PACKET_SIZE 4 - static int dsi_pixel_format_bpp(int pixel_format) { int bpp; @@ -71,77 +63,6 @@ static const u32 lfsr_converts[] = { 71, 35, 273, 136, 324, 418, 465, 488, 500, 506 /* 91 - 100 */ }; -#ifdef DSI_CLK_FROM_RR - -static u32 dsi_rr_formula(const struct drm_display_mode *mode, - int pixel_format, int video_mode_format, - int lane_count, bool eotp) -{ - u32 bpp; - u32 hactive, vactive, hfp, hsync, hbp, vfp, vsync, vbp; - u32 hsync_bytes, hbp_bytes, hactive_bytes, hfp_bytes; - u32 bytes_per_line, bytes_per_frame; - u32 num_frames; - u32 bytes_per_x_frames, bytes_per_x_frames_x_lanes; - u32 dsi_bit_clock_hz; - u32 dsi_clk; - - bpp = dsi_pixel_format_bpp(pixel_format); - - hactive = mode->hdisplay; - vactive = mode->vdisplay; - hfp = mode->hsync_start - mode->hdisplay; - hsync = mode->hsync_end - mode->hsync_start; - hbp = mode->htotal - mode->hsync_end; - - vfp = mode->vsync_start - mode->vdisplay; - vsync = mode->vsync_end - mode->vsync_start; - vbp = mode->vtotal - mode->vsync_end; - - hsync_bytes = DIV_ROUND_UP(hsync * bpp, 8); - hbp_bytes = DIV_ROUND_UP(hbp * bpp, 8); - hactive_bytes = DIV_ROUND_UP(hactive * bpp, 8); - hfp_bytes = DIV_ROUND_UP(hfp * bpp, 8); - - bytes_per_line = DSI_HSS_PACKET_SIZE + hsync_bytes + - DSI_HSA_PACKET_EXTRA_SIZE + DSI_HSE_PACKET_SIZE + - hbp_bytes + DSI_HBP_PACKET_EXTRA_SIZE + - hactive_bytes + DSI_HACTIVE_PACKET_EXTRA_SIZE + - hfp_bytes + DSI_HFP_PACKET_EXTRA_SIZE; - - /* - * XXX: Need to accurately calculate LP to HS transition timeout and add - * it to bytes_per_line/bytes_per_frame. - */ - - if (eotp && video_mode_format == VIDEO_MODE_BURST) - bytes_per_line += DSI_EOTP_PACKET_SIZE; - - bytes_per_frame = vsync * bytes_per_line + vbp * bytes_per_line + - vactive * bytes_per_line + vfp * bytes_per_line; - - if (eotp && - (video_mode_format == VIDEO_MODE_NON_BURST_WITH_SYNC_PULSE || - video_mode_format == VIDEO_MODE_NON_BURST_WITH_SYNC_EVENTS)) - bytes_per_frame += DSI_EOTP_PACKET_SIZE; - - num_frames = drm_mode_vrefresh(mode); - bytes_per_x_frames = num_frames * bytes_per_frame; - - bytes_per_x_frames_x_lanes = bytes_per_x_frames / lane_count; - - /* the dsi clock is divided by 2 in the hardware to get dsi ddr clock */ - dsi_bit_clock_hz = bytes_per_x_frames_x_lanes * 8; - dsi_clk = dsi_bit_clock_hz / 1000; - - if (eotp && video_mode_format == VIDEO_MODE_BURST) - dsi_clk *= 2; - - return dsi_clk; -} - -#else - /* Get DSI clock from pixel clock */ static u32 dsi_clk_from_pclk(u32 pclk, int pixel_format, int lane_count) { @@ -155,8 +76,6 @@ static u32 dsi_clk_from_pclk(u32 pclk, int pixel_format, int lane_count) return dsi_clk_khz; } -#endif - static int dsi_calc_mnp(struct drm_i915_private *dev_priv, struct dsi_mnp *dsi_mnp, int target_dsi_clk) { @@ -322,7 +241,7 @@ static void assert_bpp_mismatch(int pixel_format, int pipe_bpp) bpp, pipe_bpp); } -u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp) +static u32 vlv_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp) { struct drm_i915_private *dev_priv = encoder->base.dev->dev_private; struct intel_dsi *intel_dsi = enc_to_intel_dsi(&encoder->base); @@ -384,7 +303,7 @@ u32 vlv_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp) return pclk; } -u32 bxt_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp) +static u32 bxt_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp) { u32 pclk; u32 dsi_clk; @@ -419,6 +338,14 @@ u32 bxt_get_dsi_pclk(struct intel_encoder *encoder, int pipe_bpp) return pclk; } +u32 intel_dsi_get_pclk(struct intel_encoder *encoder, int pipe_bpp) +{ + if (IS_BROXTON(encoder->base.dev)) + return bxt_dsi_get_pclk(encoder, pipe_bpp); + else + return vlv_dsi_get_pclk(encoder, pipe_bpp); +} + static void vlv_dsi_reset_clocks(struct intel_encoder *encoder, enum port port) { u32 temp; diff --git a/drivers/gpu/drm/i915/intel_fbdev.c b/drivers/gpu/drm/i915/intel_fbdev.c index bea75cafc623..09840f4380f9 100644 --- a/drivers/gpu/drm/i915/intel_fbdev.c +++ b/drivers/gpu/drm/i915/intel_fbdev.c @@ -119,7 +119,7 @@ static int intelfb_alloc(struct drm_fb_helper *helper, { struct intel_fbdev *ifbdev = container_of(helper, struct intel_fbdev, helper); - struct drm_framebuffer *fb = NULL; + struct drm_framebuffer *fb; struct drm_device *dev = helper->dev; struct drm_i915_private *dev_priv = to_i915(dev); struct drm_mode_fb_cmd2 mode_cmd = {}; @@ -171,8 +171,6 @@ static int intelfb_alloc(struct drm_fb_helper *helper, out: mutex_unlock(&dev->struct_mutex); - if (!IS_ERR_OR_NULL(fb)) - drm_framebuffer_unreference(fb); return ret; } diff --git a/drivers/gpu/drm/i915/intel_guc.h b/drivers/gpu/drm/i915/intel_guc.h index 822952235dcf..045b1491ff7a 100644 --- a/drivers/gpu/drm/i915/intel_guc.h +++ b/drivers/gpu/drm/i915/intel_guc.h @@ -43,6 +43,7 @@ struct i915_guc_client { uint32_t wq_offset; uint32_t wq_size; uint32_t wq_tail; + uint32_t wq_head; /* GuC submission statistics & status */ uint64_t submissions[I915_NUM_RINGS]; @@ -88,6 +89,8 @@ struct intel_guc { uint32_t log_flags; struct drm_i915_gem_object *log_obj; + struct drm_i915_gem_object *ads_obj; + struct drm_i915_gem_object *ctx_pool_obj; struct ida ctx_ids; @@ -122,5 +125,6 @@ int i915_guc_submit(struct i915_guc_client *client, struct drm_i915_gem_request *rq); void i915_guc_submission_disable(struct drm_device *dev); void i915_guc_submission_fini(struct drm_device *dev); +int i915_guc_wq_check_space(struct i915_guc_client *client); #endif diff --git a/drivers/gpu/drm/i915/intel_guc_fwif.h b/drivers/gpu/drm/i915/intel_guc_fwif.h index 40b2ea572e16..1856a4740b83 100644 --- a/drivers/gpu/drm/i915/intel_guc_fwif.h +++ b/drivers/gpu/drm/i915/intel_guc_fwif.h @@ -39,6 +39,7 @@ #define GUC_CTX_PRIORITY_HIGH 1 #define GUC_CTX_PRIORITY_KMD_NORMAL 2 #define GUC_CTX_PRIORITY_NORMAL 3 +#define GUC_CTX_PRIORITY_NUM 4 #define GUC_MAX_GPU_CONTEXTS 1024 #define GUC_INVALID_CTX_ID GUC_MAX_GPU_CONTEXTS @@ -81,11 +82,14 @@ #define GUC_CTL_CTXINFO 0 #define GUC_CTL_CTXNUM_IN16_SHIFT 0 #define GUC_CTL_BASE_ADDR_SHIFT 12 + #define GUC_CTL_ARAT_HIGH 1 #define GUC_CTL_ARAT_LOW 2 + #define GUC_CTL_DEVICE_INFO 3 #define GUC_CTL_GTTYPE_SHIFT 0 #define GUC_CTL_COREFAMILY_SHIFT 7 + #define GUC_CTL_LOG_PARAMS 4 #define GUC_LOG_VALID (1 << 0) #define GUC_LOG_NOTIFY_ON_HALF_FULL (1 << 1) @@ -97,9 +101,12 @@ #define GUC_LOG_ISR_PAGES 3 #define GUC_LOG_ISR_SHIFT 9 #define GUC_LOG_BUF_ADDR_SHIFT 12 + #define GUC_CTL_PAGE_FAULT_CONTROL 5 + #define GUC_CTL_WA 6 #define GUC_CTL_WA_UK_BY_DRIVER (1 << 3) + #define GUC_CTL_FEATURE 7 #define GUC_CTL_VCS2_ENABLED (1 << 0) #define GUC_CTL_KERNEL_SUBMISSIONS (1 << 1) @@ -109,6 +116,7 @@ #define GUC_CTL_PREEMPTION_LOG (1 << 5) #define GUC_CTL_ENABLE_SLPC (1 << 7) #define GUC_CTL_RESET_ON_PREMPT_FAILURE (1 << 8) + #define GUC_CTL_DEBUG 8 #define GUC_LOG_VERBOSITY_SHIFT 0 #define GUC_LOG_VERBOSITY_LOW (0 << GUC_LOG_VERBOSITY_SHIFT) @@ -118,9 +126,19 @@ /* Verbosity range-check limits, without the shift */ #define GUC_LOG_VERBOSITY_MIN 0 #define GUC_LOG_VERBOSITY_MAX 3 +#define GUC_LOG_VERBOSITY_MASK 0x0000000f +#define GUC_LOG_DESTINATION_MASK (3 << 4) +#define GUC_LOG_DISABLED (1 << 6) +#define GUC_PROFILE_ENABLED (1 << 7) +#define GUC_WQ_TRACK_ENABLED (1 << 8) +#define GUC_ADS_ENABLED (1 << 9) +#define GUC_DEBUG_RESERVED (1 << 10) +#define GUC_ADS_ADDR_SHIFT 11 +#define GUC_ADS_ADDR_MASK 0xfffff800 + #define GUC_CTL_RSRVD 9 -#define GUC_CTL_MAX_DWORDS (GUC_CTL_RSRVD + 1) +#define GUC_CTL_MAX_DWORDS (SOFT_SCRATCH_COUNT - 2) /* [1..14] */ /** * DOC: GuC Firmware Layout @@ -299,6 +317,99 @@ struct guc_context_desc { #define GUC_POWER_D2 3 #define GUC_POWER_D3 4 +/* Scheduling policy settings */ + +/* Reset engine upon preempt failure */ +#define POLICY_RESET_ENGINE (1<<0) +/* Preempt to idle on quantum expiry */ +#define POLICY_PREEMPT_TO_IDLE (1<<1) + +#define POLICY_MAX_NUM_WI 15 + +struct guc_policy { + /* Time for one workload to execute. (in micro seconds) */ + u32 execution_quantum; + u32 reserved1; + + /* Time to wait for a preemption request to completed before issuing a + * reset. (in micro seconds). */ + u32 preemption_time; + + /* How much time to allow to run after the first fault is observed. + * Then preempt afterwards. (in micro seconds) */ + u32 fault_time; + + u32 policy_flags; + u32 reserved[2]; +} __packed; + +struct guc_policies { + struct guc_policy policy[GUC_CTX_PRIORITY_NUM][I915_NUM_RINGS]; + + /* In micro seconds. How much time to allow before DPC processing is + * called back via interrupt (to prevent DPC queue drain starving). + * Typically 1000s of micro seconds (example only, not granularity). */ + u32 dpc_promote_time; + + /* Must be set to take these new values. */ + u32 is_valid; + + /* Max number of WIs to process per call. A large value may keep CS + * idle. */ + u32 max_num_work_items; + + u32 reserved[19]; +} __packed; + +/* GuC MMIO reg state struct */ + +#define GUC_REGSET_FLAGS_NONE 0x0 +#define GUC_REGSET_POWERCYCLE 0x1 +#define GUC_REGSET_MASKED 0x2 +#define GUC_REGSET_ENGINERESET 0x4 +#define GUC_REGSET_SAVE_DEFAULT_VALUE 0x8 +#define GUC_REGSET_SAVE_CURRENT_VALUE 0x10 + +#define GUC_REGSET_MAX_REGISTERS 25 +#define GUC_MMIO_WHITE_LIST_START 0x24d0 +#define GUC_MMIO_WHITE_LIST_MAX 12 +#define GUC_S3_SAVE_SPACE_PAGES 10 + +struct guc_mmio_regset { + struct __packed { + u32 offset; + u32 value; + u32 flags; + } registers[GUC_REGSET_MAX_REGISTERS]; + + u32 values_valid; + u32 number_of_registers; +} __packed; + +struct guc_mmio_reg_state { + struct guc_mmio_regset global_reg; + struct guc_mmio_regset engine_reg[I915_NUM_RINGS]; + + /* MMIO registers that are set as non privileged */ + struct __packed { + u32 mmio_start; + u32 offsets[GUC_MMIO_WHITE_LIST_MAX]; + u32 count; + } mmio_white_list[I915_NUM_RINGS]; +} __packed; + +/* GuC Additional Data Struct */ + +struct guc_ads { + u32 reg_state_addr; + u32 reg_state_buffer; + u32 golden_context_lrca; + u32 scheduler_policies; + u32 reserved0[3]; + u32 eng_state_size[I915_NUM_RINGS]; + u32 reserved2[4]; +} __packed; + /* This Action will be programmed in C180 - SOFT_SCRATCH_O_REG */ enum host2guc_action { HOST2GUC_ACTION_DEFAULT = 0x0, diff --git a/drivers/gpu/drm/i915/intel_guc_loader.c b/drivers/gpu/drm/i915/intel_guc_loader.c index 550921f2ef7d..3accd914490f 100644 --- a/drivers/gpu/drm/i915/intel_guc_loader.c +++ b/drivers/gpu/drm/i915/intel_guc_loader.c @@ -165,6 +165,13 @@ static void set_guc_init_params(struct drm_i915_private *dev_priv) i915.guc_log_level << GUC_LOG_VERBOSITY_SHIFT; } + if (guc->ads_obj) { + u32 ads = (u32)i915_gem_obj_ggtt_offset(guc->ads_obj) + >> PAGE_SHIFT; + params[GUC_CTL_DEBUG] |= ads << GUC_ADS_ADDR_SHIFT; + params[GUC_CTL_DEBUG] |= GUC_ADS_ENABLED; + } + /* If GuC submission is enabled, set up additional parameters here */ if (i915.enable_guc_submission) { u32 pgs = i915_gem_obj_ggtt_offset(dev_priv->guc.ctx_pool_obj); @@ -438,6 +445,7 @@ fail: direct_interrupts_to_host(dev_priv); i915_guc_submission_disable(dev); + i915_guc_submission_fini(dev); return err; } @@ -554,10 +562,12 @@ fail: DRM_ERROR("Failed to fetch GuC firmware from %s (error %d)\n", guc_fw->guc_fw_path, err); + mutex_lock(&dev->struct_mutex); obj = guc_fw->guc_fw_obj; if (obj) drm_gem_object_unreference(&obj->base); guc_fw->guc_fw_obj = NULL; + mutex_unlock(&dev->struct_mutex); release_firmware(fw); /* OK even if fw is NULL */ guc_fw->guc_fw_fetch_status = GUC_FIRMWARE_FAIL; @@ -624,10 +634,11 @@ void intel_guc_ucode_fini(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; struct intel_guc_fw *guc_fw = &dev_priv->guc.guc_fw; + mutex_lock(&dev->struct_mutex); direct_interrupts_to_host(dev_priv); + i915_guc_submission_disable(dev); i915_guc_submission_fini(dev); - mutex_lock(&dev->struct_mutex); if (guc_fw->guc_fw_obj) drm_gem_object_unreference(&guc_fw->guc_fw_obj->base); guc_fw->guc_fw_obj = NULL; diff --git a/drivers/gpu/drm/i915/intel_hdmi.c b/drivers/gpu/drm/i915/intel_hdmi.c index 4a77639a489d..8698a643d027 100644..100755 --- a/drivers/gpu/drm/i915/intel_hdmi.c +++ b/drivers/gpu/drm/i915/intel_hdmi.c @@ -2033,6 +2033,11 @@ void intel_hdmi_init_connector(struct intel_digital_port *intel_dig_port, enum port port = intel_dig_port->port; uint8_t alternate_ddc_pin; + if (WARN(intel_dig_port->max_lanes < 4, + "Not enough lanes (%d) for HDMI on port %c\n", + intel_dig_port->max_lanes, port_name(port))) + return; + drm_connector_init(dev, connector, &intel_hdmi_connector_funcs, DRM_MODE_CONNECTOR_HDMIA); drm_connector_helper_add(connector, &intel_hdmi_connector_helper_funcs); @@ -2218,6 +2223,7 @@ void intel_hdmi_init(struct drm_device *dev, dev_priv->dig_port_map[port] = intel_encoder; intel_dig_port->hdmi.hdmi_reg = hdmi_reg; intel_dig_port->dp.output_reg = INVALID_MMIO_REG; + intel_dig_port->max_lanes = 4; intel_hdmi_init_connector(intel_dig_port, intel_connector); } diff --git a/drivers/gpu/drm/i915/intel_lrc.c b/drivers/gpu/drm/i915/intel_lrc.c index f1fa756c5d5d..73d4347429df 100644 --- a/drivers/gpu/drm/i915/intel_lrc.c +++ b/drivers/gpu/drm/i915/intel_lrc.c @@ -263,65 +263,92 @@ int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists return 0; } +static void +logical_ring_init_platform_invariants(struct intel_engine_cs *ring) +{ + struct drm_device *dev = ring->dev; + + ring->disable_lite_restore_wa = (IS_SKL_REVID(dev, 0, SKL_REVID_B0) || + IS_BXT_REVID(dev, 0, BXT_REVID_A1)) && + (ring->id == VCS || ring->id == VCS2); + + ring->ctx_desc_template = GEN8_CTX_VALID; + ring->ctx_desc_template |= GEN8_CTX_ADDRESSING_MODE(dev) << + GEN8_CTX_ADDRESSING_MODE_SHIFT; + if (IS_GEN8(dev)) + ring->ctx_desc_template |= GEN8_CTX_L3LLC_COHERENT; + ring->ctx_desc_template |= GEN8_CTX_PRIVILEGE; + + /* TODO: WaDisableLiteRestore when we start using semaphore + * signalling between Command Streamers */ + /* ring->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE; */ + + /* WaEnableForceRestoreInCtxtDescForVCS:skl */ + /* WaEnableForceRestoreInCtxtDescForVCS:bxt */ + if (ring->disable_lite_restore_wa) + ring->ctx_desc_template |= GEN8_CTX_FORCE_RESTORE; +} + /** - * intel_execlists_ctx_id() - get the Execlists Context ID - * @ctx_obj: Logical Ring Context backing object. + * intel_lr_context_descriptor_update() - calculate & cache the descriptor + * descriptor for a pinned context * - * Do not confuse with ctx->id! Unfortunately we have a name overload - * here: the old context ID we pass to userspace as a handler so that - * they can refer to a context, and the new context ID we pass to the - * ELSP so that the GPU can inform us of the context status via - * interrupts. + * @ctx: Context to work on + * @ring: Engine the descriptor will be used with * - * Return: 20-bits globally unique context ID. + * The context descriptor encodes various attributes of a context, + * including its GTT address and some flags. Because it's fairly + * expensive to calculate, we'll just do it once and cache the result, + * which remains valid until the context is unpinned. + * + * This is what a descriptor looks like, from LSB to MSB: + * bits 0-11: flags, GEN8_CTX_* (cached in ctx_desc_template) + * bits 12-31: LRCA, GTT address of (the HWSP of) this context + * bits 32-51: ctx ID, a globally unique tag (the LRCA again!) + * bits 52-63: reserved, may encode the engine ID (for GuC) */ -u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj) +static void +intel_lr_context_descriptor_update(struct intel_context *ctx, + struct intel_engine_cs *ring) { - u32 lrca = i915_gem_obj_ggtt_offset(ctx_obj) + - LRC_PPHWSP_PN * PAGE_SIZE; + uint64_t lrca, desc; - /* LRCA is required to be 4K aligned so the more significant 20 bits - * are globally unique */ - return lrca >> 12; -} + lrca = ctx->engine[ring->id].lrc_vma->node.start + + LRC_PPHWSP_PN * PAGE_SIZE; -static bool disable_lite_restore_wa(struct intel_engine_cs *ring) -{ - struct drm_device *dev = ring->dev; + desc = ring->ctx_desc_template; /* bits 0-11 */ + desc |= lrca; /* bits 12-31 */ + desc |= (lrca >> PAGE_SHIFT) << GEN8_CTX_ID_SHIFT; /* bits 32-51 */ - return (IS_SKL_REVID(dev, 0, SKL_REVID_B0) || - IS_BXT_REVID(dev, 0, BXT_REVID_A1)) && - (ring->id == VCS || ring->id == VCS2); + ctx->engine[ring->id].lrc_desc = desc; } uint64_t intel_lr_context_descriptor(struct intel_context *ctx, struct intel_engine_cs *ring) { - struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state; - uint64_t desc; - uint64_t lrca = i915_gem_obj_ggtt_offset(ctx_obj) + - LRC_PPHWSP_PN * PAGE_SIZE; - - WARN_ON(lrca & 0xFFFFFFFF00000FFFULL); - - desc = GEN8_CTX_VALID; - desc |= GEN8_CTX_ADDRESSING_MODE(dev) << GEN8_CTX_ADDRESSING_MODE_SHIFT; - if (IS_GEN8(ctx_obj->base.dev)) - desc |= GEN8_CTX_L3LLC_COHERENT; - desc |= GEN8_CTX_PRIVILEGE; - desc |= lrca; - desc |= (u64)intel_execlists_ctx_id(ctx_obj) << GEN8_CTX_ID_SHIFT; - - /* TODO: WaDisableLiteRestore when we start using semaphore - * signalling between Command Streamers */ - /* desc |= GEN8_CTX_FORCE_RESTORE; */ - - /* WaEnableForceRestoreInCtxtDescForVCS:skl */ - /* WaEnableForceRestoreInCtxtDescForVCS:bxt */ - if (disable_lite_restore_wa(ring)) - desc |= GEN8_CTX_FORCE_RESTORE; + return ctx->engine[ring->id].lrc_desc; +} - return desc; +/** + * intel_execlists_ctx_id() - get the Execlists Context ID + * @ctx: Context to get the ID for + * @ring: Engine to get the ID for + * + * Do not confuse with ctx->id! Unfortunately we have a name overload + * here: the old context ID we pass to userspace as a handler so that + * they can refer to a context, and the new context ID we pass to the + * ELSP so that the GPU can inform us of the context status via + * interrupts. + * + * The context ID is a portion of the context descriptor, so we can + * just extract the required part from the cached descriptor. + * + * Return: 20-bits globally unique context ID. + */ +u32 intel_execlists_ctx_id(struct intel_context *ctx, + struct intel_engine_cs *ring) +{ + return intel_lr_context_descriptor(ctx, ring) >> GEN8_CTX_ID_SHIFT; } static void execlists_elsp_write(struct drm_i915_gem_request *rq0, @@ -363,20 +390,10 @@ static int execlists_update_context(struct drm_i915_gem_request *rq) { struct intel_engine_cs *ring = rq->ring; struct i915_hw_ppgtt *ppgtt = rq->ctx->ppgtt; - struct drm_i915_gem_object *ctx_obj = rq->ctx->engine[ring->id].state; - struct drm_i915_gem_object *rb_obj = rq->ringbuf->obj; - struct page *page; - uint32_t *reg_state; - - BUG_ON(!ctx_obj); - WARN_ON(!i915_gem_obj_is_pinned(ctx_obj)); - WARN_ON(!i915_gem_obj_is_pinned(rb_obj)); - - page = i915_gem_object_get_dirty_page(ctx_obj, LRC_STATE_PN); - reg_state = kmap_atomic(page); + uint32_t *reg_state = rq->ctx->engine[ring->id].lrc_reg_state; reg_state[CTX_RING_TAIL+1] = rq->tail; - reg_state[CTX_RING_BUFFER_START+1] = i915_gem_obj_ggtt_offset(rb_obj); + reg_state[CTX_RING_BUFFER_START+1] = rq->ringbuf->vma->node.start; if (ppgtt && !USES_FULL_48BIT_PPGTT(ppgtt->base.dev)) { /* True 32b PPGTT with dynamic page allocation: update PDP @@ -390,8 +407,6 @@ static int execlists_update_context(struct drm_i915_gem_request *rq) ASSIGN_CTX_PDP(ppgtt, reg_state, 0); } - kunmap_atomic(reg_state); - return 0; } @@ -431,9 +446,8 @@ static void execlists_context_unqueue(struct intel_engine_cs *ring) /* Same ctx: ignore first request, as second request * will update tail past first request's workload */ cursor->elsp_submitted = req0->elsp_submitted; - list_del(&req0->execlist_link); - list_add_tail(&req0->execlist_link, - &ring->execlist_retired_req_list); + list_move_tail(&req0->execlist_link, + &ring->execlist_retired_req_list); req0 = cursor; } else { req1 = cursor; @@ -478,16 +492,13 @@ static bool execlists_check_remove_request(struct intel_engine_cs *ring, execlist_link); if (head_req != NULL) { - struct drm_i915_gem_object *ctx_obj = - head_req->ctx->engine[ring->id].state; - if (intel_execlists_ctx_id(ctx_obj) == request_id) { + if (intel_execlists_ctx_id(head_req->ctx, ring) == request_id) { WARN(head_req->elsp_submitted == 0, "Never submitted head request\n"); if (--head_req->elsp_submitted <= 0) { - list_del(&head_req->execlist_link); - list_add_tail(&head_req->execlist_link, - &ring->execlist_retired_req_list); + list_move_tail(&head_req->execlist_link, + &ring->execlist_retired_req_list); return true; } } @@ -496,6 +507,19 @@ static bool execlists_check_remove_request(struct intel_engine_cs *ring, return false; } +static void get_context_status(struct intel_engine_cs *ring, + u8 read_pointer, + u32 *status, u32 *context_id) +{ + struct drm_i915_private *dev_priv = ring->dev->dev_private; + + if (WARN_ON(read_pointer >= GEN8_CSB_ENTRIES)) + return; + + *status = I915_READ(RING_CONTEXT_STATUS_BUF_LO(ring, read_pointer)); + *context_id = I915_READ(RING_CONTEXT_STATUS_BUF_HI(ring, read_pointer)); +} + /** * intel_lrc_irq_handler() - handle Context Switch interrupts * @ring: Engine Command Streamer to handle. @@ -516,16 +540,16 @@ void intel_lrc_irq_handler(struct intel_engine_cs *ring) status_pointer = I915_READ(RING_CONTEXT_STATUS_PTR(ring)); read_pointer = ring->next_context_status_buffer; - write_pointer = status_pointer & GEN8_CSB_PTR_MASK; + write_pointer = GEN8_CSB_WRITE_PTR(status_pointer); if (read_pointer > write_pointer) write_pointer += GEN8_CSB_ENTRIES; spin_lock(&ring->execlist_lock); while (read_pointer < write_pointer) { - read_pointer++; - status = I915_READ(RING_CONTEXT_STATUS_BUF_LO(ring, read_pointer % GEN8_CSB_ENTRIES)); - status_id = I915_READ(RING_CONTEXT_STATUS_BUF_HI(ring, read_pointer % GEN8_CSB_ENTRIES)); + + get_context_status(ring, ++read_pointer % GEN8_CSB_ENTRIES, + &status, &status_id); if (status & GEN8_CTX_STATUS_IDLE_ACTIVE) continue; @@ -538,14 +562,14 @@ void intel_lrc_irq_handler(struct intel_engine_cs *ring) WARN(1, "Preemption without Lite Restore\n"); } - if ((status & GEN8_CTX_STATUS_ACTIVE_IDLE) || - (status & GEN8_CTX_STATUS_ELEMENT_SWITCH)) { + if ((status & GEN8_CTX_STATUS_ACTIVE_IDLE) || + (status & GEN8_CTX_STATUS_ELEMENT_SWITCH)) { if (execlists_check_remove_request(ring, status_id)) submit_contexts++; } } - if (disable_lite_restore_wa(ring)) { + if (ring->disable_lite_restore_wa) { /* Prevent a ctx to preempt itself */ if ((status & GEN8_CTX_STATUS_ACTIVE_IDLE) && (submit_contexts != 0)) @@ -556,13 +580,16 @@ void intel_lrc_irq_handler(struct intel_engine_cs *ring) spin_unlock(&ring->execlist_lock); - WARN(submit_contexts > 2, "More than two context complete events?\n"); + if (unlikely(submit_contexts > 2)) + DRM_ERROR("More than two context complete events?\n"); + ring->next_context_status_buffer = write_pointer % GEN8_CSB_ENTRIES; + /* Update the read pointer to the old write pointer. Manual ringbuffer + * management ftw </sarcasm> */ I915_WRITE(RING_CONTEXT_STATUS_PTR(ring), - _MASKED_FIELD(GEN8_CSB_PTR_MASK << 8, - ((u32)ring->next_context_status_buffer & - GEN8_CSB_PTR_MASK) << 8)); + _MASKED_FIELD(GEN8_CSB_READ_PTR_MASK, + ring->next_context_status_buffer << 8)); } static int execlists_context_queue(struct drm_i915_gem_request *request) @@ -571,7 +598,7 @@ static int execlists_context_queue(struct drm_i915_gem_request *request) struct drm_i915_gem_request *cursor; int num_elements = 0; - if (request->ctx != ring->default_context) + if (request->ctx != request->i915->kernel_context) intel_lr_context_pin(request); i915_gem_request_reference(request); @@ -592,9 +619,8 @@ static int execlists_context_queue(struct drm_i915_gem_request *request) if (request->ctx == tail_req->ctx) { WARN(tail_req->elsp_submitted != 0, "More than 2 already-submitted reqs queued\n"); - list_del(&tail_req->execlist_link); - list_add_tail(&tail_req->execlist_link, - &ring->execlist_retired_req_list); + list_move_tail(&tail_req->execlist_link, + &ring->execlist_retired_req_list); } } @@ -660,17 +686,27 @@ static int execlists_move_to_gpu(struct drm_i915_gem_request *req, int intel_logical_ring_alloc_request_extras(struct drm_i915_gem_request *request) { - int ret; + int ret = 0; request->ringbuf = request->ctx->engine[request->ring->id].ringbuf; - if (request->ctx != request->ring->default_context) { - ret = intel_lr_context_pin(request); + if (i915.enable_guc_submission) { + /* + * Check that the GuC has space for the request before + * going any further, as the i915_add_request() call + * later on mustn't fail ... + */ + struct intel_guc *guc = &request->i915->guc; + + ret = i915_guc_wq_check_space(guc->execbuf_client); if (ret) return ret; } - return 0; + if (request->ctx != request->i915->kernel_context) + ret = intel_lr_context_pin(request); + + return ret; } static int logical_ring_wait_for_space(struct drm_i915_gem_request *req, @@ -724,23 +760,34 @@ static int logical_ring_wait_for_space(struct drm_i915_gem_request *req, * on a queue waiting for the ELSP to be ready to accept a new context submission. At that * point, the tail *inside* the context is updated and the ELSP written to. */ -static void +static int intel_logical_ring_advance_and_submit(struct drm_i915_gem_request *request) { - struct intel_engine_cs *ring = request->ring; + struct intel_ringbuffer *ringbuf = request->ringbuf; struct drm_i915_private *dev_priv = request->i915; - intel_logical_ring_advance(request->ringbuf); + intel_logical_ring_advance(ringbuf); + request->tail = ringbuf->tail; - request->tail = request->ringbuf->tail; + /* + * Here we add two extra NOOPs as padding to avoid + * lite restore of a context with HEAD==TAIL. + * + * Caller must reserve WA_TAIL_DWORDS for us! + */ + intel_logical_ring_emit(ringbuf, MI_NOOP); + intel_logical_ring_emit(ringbuf, MI_NOOP); + intel_logical_ring_advance(ringbuf); - if (intel_ring_stopped(ring)) - return; + if (intel_ring_stopped(request->ring)) + return 0; if (dev_priv->guc.execbuf_client) i915_guc_submit(dev_priv->guc.execbuf_client, request); else execlists_context_queue(request); + + return 0; } static void __wrap_ring_buffer(struct intel_ringbuffer *ringbuf) @@ -967,7 +1014,7 @@ void intel_execlists_retire_requests(struct intel_engine_cs *ring) struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state; - if (ctx_obj && (ctx != ring->default_context)) + if (ctx_obj && (ctx != req->i915->kernel_context)) intel_lr_context_unpin(req); list_del(&req->execlist_link); i915_gem_request_unreference(req); @@ -1013,23 +1060,35 @@ int logical_ring_flush_all_caches(struct drm_i915_gem_request *req) } static int intel_lr_context_do_pin(struct intel_engine_cs *ring, - struct drm_i915_gem_object *ctx_obj, - struct intel_ringbuffer *ringbuf) + struct intel_context *ctx) { struct drm_device *dev = ring->dev; struct drm_i915_private *dev_priv = dev->dev_private; - int ret = 0; + struct drm_i915_gem_object *ctx_obj = ctx->engine[ring->id].state; + struct intel_ringbuffer *ringbuf = ctx->engine[ring->id].ringbuf; + struct page *lrc_state_page; + int ret; WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex)); + ret = i915_gem_obj_ggtt_pin(ctx_obj, GEN8_LR_CONTEXT_ALIGN, PIN_OFFSET_BIAS | GUC_WOPCM_TOP); if (ret) return ret; + lrc_state_page = i915_gem_object_get_dirty_page(ctx_obj, LRC_STATE_PN); + if (WARN_ON(!lrc_state_page)) { + ret = -ENODEV; + goto unpin_ctx_obj; + } + ret = intel_pin_and_map_ringbuffer_obj(ring->dev, ringbuf); if (ret) goto unpin_ctx_obj; + ctx->engine[ring->id].lrc_vma = i915_gem_obj_to_ggtt(ctx_obj); + intel_lr_context_descriptor_update(ctx, ring); + ctx->engine[ring->id].lrc_reg_state = kmap(lrc_state_page); ctx_obj->dirty = true; /* Invalidate GuC TLB. */ @@ -1048,11 +1107,9 @@ static int intel_lr_context_pin(struct drm_i915_gem_request *rq) { int ret = 0; struct intel_engine_cs *ring = rq->ring; - struct drm_i915_gem_object *ctx_obj = rq->ctx->engine[ring->id].state; - struct intel_ringbuffer *ringbuf = rq->ringbuf; if (rq->ctx->engine[ring->id].pin_count++ == 0) { - ret = intel_lr_context_do_pin(ring, ctx_obj, ringbuf); + ret = intel_lr_context_do_pin(ring, rq->ctx); if (ret) goto reset_pin_count; } @@ -1069,12 +1126,18 @@ void intel_lr_context_unpin(struct drm_i915_gem_request *rq) struct drm_i915_gem_object *ctx_obj = rq->ctx->engine[ring->id].state; struct intel_ringbuffer *ringbuf = rq->ringbuf; - if (ctx_obj) { - WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex)); - if (--rq->ctx->engine[ring->id].pin_count == 0) { - intel_unpin_ringbuffer_obj(ringbuf); - i915_gem_object_ggtt_unpin(ctx_obj); - } + WARN_ON(!mutex_is_locked(&ring->dev->struct_mutex)); + + if (!ctx_obj) + return; + + if (--rq->ctx->engine[ring->id].pin_count == 0) { + kunmap(kmap_to_page(rq->ctx->engine[ring->id].lrc_reg_state)); + intel_unpin_ringbuffer_obj(ringbuf); + i915_gem_object_ggtt_unpin(ctx_obj); + rq->ctx->engine[ring->id].lrc_vma = NULL; + rq->ctx->engine[ring->id].lrc_desc = 0; + rq->ctx->engine[ring->id].lrc_reg_state = NULL; } } @@ -1087,7 +1150,7 @@ static int intel_logical_ring_workarounds_emit(struct drm_i915_gem_request *req) struct drm_i915_private *dev_priv = dev->dev_private; struct i915_workarounds *w = &dev_priv->workarounds; - if (WARN_ON_ONCE(w->count == 0)) + if (w->count == 0) return 0; ring->gpu_caches_dirty = true; @@ -1474,7 +1537,7 @@ static int gen8_init_common_ring(struct intel_engine_cs *ring) u8 next_context_status_buffer_hw; lrc_setup_hardware_status_page(ring, - ring->default_context->engine[ring->id].state); + dev_priv->kernel_context->engine[ring->id].state); I915_WRITE_IMR(ring, ~(ring->irq_enable_mask | ring->irq_keep_mask)); I915_WRITE(RING_HWSTAM(ring->mmio_base), 0xffffffff); @@ -1493,9 +1556,11 @@ static int gen8_init_common_ring(struct intel_engine_cs *ring) * | Suspend-to-idle (freeze) | Suspend-to-RAM (mem) | * BDW | CSB regs not reset | CSB regs reset | * CHT | CSB regs not reset | CSB regs not reset | + * SKL | ? | ? | + * BXT | ? | ? | */ - next_context_status_buffer_hw = (I915_READ(RING_CONTEXT_STATUS_PTR(ring)) - & GEN8_CSB_PTR_MASK); + next_context_status_buffer_hw = + GEN8_CSB_WRITE_PTR(I915_READ(RING_CONTEXT_STATUS_PTR(ring))); /* * When the CSB registers are reset (also after power-up / gpu reset), @@ -1698,7 +1763,7 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request, struct intel_ringbuffer *ringbuf = request->ringbuf; struct intel_engine_cs *ring = ringbuf->ring; u32 scratch_addr = ring->scratch.gtt_offset + 2 * CACHELINE_BYTES; - bool vf_flush_wa; + bool vf_flush_wa = false; u32 flags = 0; int ret; @@ -1720,14 +1785,14 @@ static int gen8_emit_flush_render(struct drm_i915_gem_request *request, flags |= PIPE_CONTROL_STATE_CACHE_INVALIDATE; flags |= PIPE_CONTROL_QW_WRITE; flags |= PIPE_CONTROL_GLOBAL_GTT_IVB; - } - /* - * On GEN9+ Before VF_CACHE_INVALIDATE we need to emit a NULL pipe - * control. - */ - vf_flush_wa = INTEL_INFO(ring->dev)->gen >= 9 && - flags & PIPE_CONTROL_VF_CACHE_INVALIDATE; + /* + * On GEN9: before VF_CACHE_INVALIDATE we need to emit a NULL + * pipe control. + */ + if (IS_GEN9(ring->dev)) + vf_flush_wa = true; + } ret = intel_logical_ring_begin(request, vf_flush_wa ? 12 : 6); if (ret) @@ -1791,44 +1856,65 @@ static void bxt_a_set_seqno(struct intel_engine_cs *ring, u32 seqno) intel_flush_status_page(ring, I915_GEM_HWS_INDEX); } +/* + * Reserve space for 2 NOOPs at the end of each request to be + * used as a workaround for not being allowed to do lite + * restore with HEAD==TAIL (WaIdleLiteRestore). + */ +#define WA_TAIL_DWORDS 2 + +static inline u32 hws_seqno_address(struct intel_engine_cs *engine) +{ + return engine->status_page.gfx_addr + I915_GEM_HWS_INDEX_ADDR; +} + static int gen8_emit_request(struct drm_i915_gem_request *request) { struct intel_ringbuffer *ringbuf = request->ringbuf; - struct intel_engine_cs *ring = ringbuf->ring; - u32 cmd; int ret; - /* - * Reserve space for 2 NOOPs at the end of each request to be - * used as a workaround for not being allowed to do lite - * restore with HEAD==TAIL (WaIdleLiteRestore). - */ - ret = intel_logical_ring_begin(request, 8); + ret = intel_logical_ring_begin(request, 6 + WA_TAIL_DWORDS); if (ret) return ret; - cmd = MI_STORE_DWORD_IMM_GEN4; - cmd |= MI_GLOBAL_GTT; + /* w/a: bit 5 needs to be zero for MI_FLUSH_DW address. */ + BUILD_BUG_ON(I915_GEM_HWS_INDEX_ADDR & (1 << 5)); - intel_logical_ring_emit(ringbuf, cmd); intel_logical_ring_emit(ringbuf, - (ring->status_page.gfx_addr + - (I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT))); + (MI_FLUSH_DW + 1) | MI_FLUSH_DW_OP_STOREDW); + intel_logical_ring_emit(ringbuf, + hws_seqno_address(request->ring) | + MI_FLUSH_DW_USE_GTT); intel_logical_ring_emit(ringbuf, 0); intel_logical_ring_emit(ringbuf, i915_gem_request_get_seqno(request)); intel_logical_ring_emit(ringbuf, MI_USER_INTERRUPT); intel_logical_ring_emit(ringbuf, MI_NOOP); - intel_logical_ring_advance_and_submit(request); + return intel_logical_ring_advance_and_submit(request); +} - /* - * Here we add two extra NOOPs as padding to avoid - * lite restore of a context with HEAD==TAIL. - */ - intel_logical_ring_emit(ringbuf, MI_NOOP); - intel_logical_ring_emit(ringbuf, MI_NOOP); - intel_logical_ring_advance(ringbuf); +static int gen8_emit_request_render(struct drm_i915_gem_request *request) +{ + struct intel_ringbuffer *ringbuf = request->ringbuf; + int ret; - return 0; + ret = intel_logical_ring_begin(request, 6 + WA_TAIL_DWORDS); + if (ret) + return ret; + + /* w/a for post sync ops following a GPGPU operation we + * need a prior CS_STALL, which is emitted by the flush + * following the batch. + */ + intel_logical_ring_emit(ringbuf, GFX_OP_PIPE_CONTROL(5)); + intel_logical_ring_emit(ringbuf, + (PIPE_CONTROL_GLOBAL_GTT_IVB | + PIPE_CONTROL_CS_STALL | + PIPE_CONTROL_QW_WRITE)); + intel_logical_ring_emit(ringbuf, hws_seqno_address(request->ring)); + intel_logical_ring_emit(ringbuf, 0); + intel_logical_ring_emit(ringbuf, i915_gem_request_get_seqno(request)); + intel_logical_ring_emit(ringbuf, MI_USER_INTERRUPT); + return intel_logical_ring_advance_and_submit(request); } static int intel_lr_context_render_state_init(struct drm_i915_gem_request *req) @@ -1911,12 +1997,44 @@ void intel_logical_ring_cleanup(struct intel_engine_cs *ring) ring->status_page.obj = NULL; } + ring->disable_lite_restore_wa = false; + ring->ctx_desc_template = 0; + lrc_destroy_wa_ctx_obj(ring); ring->dev = NULL; } -static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *ring) +static void +logical_ring_default_vfuncs(struct drm_device *dev, + struct intel_engine_cs *ring) { + /* Default vfuncs which can be overriden by each engine. */ + ring->init_hw = gen8_init_common_ring; + ring->emit_request = gen8_emit_request; + ring->emit_flush = gen8_emit_flush; + ring->irq_get = gen8_logical_ring_get_irq; + ring->irq_put = gen8_logical_ring_put_irq; + ring->emit_bb_start = gen8_emit_bb_start; + if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { + ring->get_seqno = bxt_a_get_seqno; + ring->set_seqno = bxt_a_set_seqno; + } else { + ring->get_seqno = gen8_get_seqno; + ring->set_seqno = gen8_set_seqno; + } +} + +static inline void +logical_ring_default_irqs(struct intel_engine_cs *ring, unsigned shift) +{ + ring->irq_enable_mask = GT_RENDER_USER_INTERRUPT << shift; + ring->irq_keep_mask = GT_CONTEXT_SWITCH_INTERRUPT << shift; +} + +static int +logical_ring_init(struct drm_device *dev, struct intel_engine_cs *ring) +{ + struct intel_context *dctx = to_i915(dev)->kernel_context; int ret; /* Intentionally left blank. */ @@ -1933,19 +2051,18 @@ static int logical_ring_init(struct drm_device *dev, struct intel_engine_cs *rin INIT_LIST_HEAD(&ring->execlist_retired_req_list); spin_lock_init(&ring->execlist_lock); + logical_ring_init_platform_invariants(ring); + ret = i915_cmd_parser_init_ring(ring); if (ret) goto error; - ret = intel_lr_context_deferred_alloc(ring->default_context, ring); + ret = intel_lr_context_deferred_alloc(dctx, ring); if (ret) goto error; /* As this is the default context, always pin it */ - ret = intel_lr_context_do_pin( - ring, - ring->default_context->engine[ring->id].state, - ring->default_context->engine[ring->id].ringbuf); + ret = intel_lr_context_do_pin(ring, dctx); if (ret) { DRM_ERROR( "Failed to pin and map ringbuffer %s: %d\n", @@ -1968,32 +2085,24 @@ static int logical_render_ring_init(struct drm_device *dev) ring->name = "render ring"; ring->id = RCS; + ring->exec_id = I915_EXEC_RENDER; ring->mmio_base = RENDER_RING_BASE; - ring->irq_enable_mask = - GT_RENDER_USER_INTERRUPT << GEN8_RCS_IRQ_SHIFT; - ring->irq_keep_mask = - GT_CONTEXT_SWITCH_INTERRUPT << GEN8_RCS_IRQ_SHIFT; + + logical_ring_default_irqs(ring, GEN8_RCS_IRQ_SHIFT); if (HAS_L3_DPF(dev)) ring->irq_keep_mask |= GT_RENDER_L3_PARITY_ERROR_INTERRUPT; + logical_ring_default_vfuncs(dev, ring); + + /* Override some for render ring. */ if (INTEL_INFO(dev)->gen >= 9) ring->init_hw = gen9_init_render_ring; else ring->init_hw = gen8_init_render_ring; ring->init_context = gen8_init_rcs_context; ring->cleanup = intel_fini_pipe_control; - if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { - ring->get_seqno = bxt_a_get_seqno; - ring->set_seqno = bxt_a_set_seqno; - } else { - ring->get_seqno = gen8_get_seqno; - ring->set_seqno = gen8_set_seqno; - } - ring->emit_request = gen8_emit_request; ring->emit_flush = gen8_emit_flush_render; - ring->irq_get = gen8_logical_ring_get_irq; - ring->irq_put = gen8_logical_ring_put_irq; - ring->emit_bb_start = gen8_emit_bb_start; + ring->emit_request = gen8_emit_request_render; ring->dev = dev; @@ -2027,25 +2136,11 @@ static int logical_bsd_ring_init(struct drm_device *dev) ring->name = "bsd ring"; ring->id = VCS; + ring->exec_id = I915_EXEC_BSD; ring->mmio_base = GEN6_BSD_RING_BASE; - ring->irq_enable_mask = - GT_RENDER_USER_INTERRUPT << GEN8_VCS1_IRQ_SHIFT; - ring->irq_keep_mask = - GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS1_IRQ_SHIFT; - ring->init_hw = gen8_init_common_ring; - if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { - ring->get_seqno = bxt_a_get_seqno; - ring->set_seqno = bxt_a_set_seqno; - } else { - ring->get_seqno = gen8_get_seqno; - ring->set_seqno = gen8_set_seqno; - } - ring->emit_request = gen8_emit_request; - ring->emit_flush = gen8_emit_flush; - ring->irq_get = gen8_logical_ring_get_irq; - ring->irq_put = gen8_logical_ring_put_irq; - ring->emit_bb_start = gen8_emit_bb_start; + logical_ring_default_irqs(ring, GEN8_VCS1_IRQ_SHIFT); + logical_ring_default_vfuncs(dev, ring); return logical_ring_init(dev, ring); } @@ -2055,22 +2150,13 @@ static int logical_bsd2_ring_init(struct drm_device *dev) struct drm_i915_private *dev_priv = dev->dev_private; struct intel_engine_cs *ring = &dev_priv->ring[VCS2]; - ring->name = "bds2 ring"; + ring->name = "bsd2 ring"; ring->id = VCS2; + ring->exec_id = I915_EXEC_BSD; ring->mmio_base = GEN8_BSD2_RING_BASE; - ring->irq_enable_mask = - GT_RENDER_USER_INTERRUPT << GEN8_VCS2_IRQ_SHIFT; - ring->irq_keep_mask = - GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VCS2_IRQ_SHIFT; - ring->init_hw = gen8_init_common_ring; - ring->get_seqno = gen8_get_seqno; - ring->set_seqno = gen8_set_seqno; - ring->emit_request = gen8_emit_request; - ring->emit_flush = gen8_emit_flush; - ring->irq_get = gen8_logical_ring_get_irq; - ring->irq_put = gen8_logical_ring_put_irq; - ring->emit_bb_start = gen8_emit_bb_start; + logical_ring_default_irqs(ring, GEN8_VCS2_IRQ_SHIFT); + logical_ring_default_vfuncs(dev, ring); return logical_ring_init(dev, ring); } @@ -2082,25 +2168,11 @@ static int logical_blt_ring_init(struct drm_device *dev) ring->name = "blitter ring"; ring->id = BCS; + ring->exec_id = I915_EXEC_BLT; ring->mmio_base = BLT_RING_BASE; - ring->irq_enable_mask = - GT_RENDER_USER_INTERRUPT << GEN8_BCS_IRQ_SHIFT; - ring->irq_keep_mask = - GT_CONTEXT_SWITCH_INTERRUPT << GEN8_BCS_IRQ_SHIFT; - ring->init_hw = gen8_init_common_ring; - if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { - ring->get_seqno = bxt_a_get_seqno; - ring->set_seqno = bxt_a_set_seqno; - } else { - ring->get_seqno = gen8_get_seqno; - ring->set_seqno = gen8_set_seqno; - } - ring->emit_request = gen8_emit_request; - ring->emit_flush = gen8_emit_flush; - ring->irq_get = gen8_logical_ring_get_irq; - ring->irq_put = gen8_logical_ring_put_irq; - ring->emit_bb_start = gen8_emit_bb_start; + logical_ring_default_irqs(ring, GEN8_BCS_IRQ_SHIFT); + logical_ring_default_vfuncs(dev, ring); return logical_ring_init(dev, ring); } @@ -2112,25 +2184,11 @@ static int logical_vebox_ring_init(struct drm_device *dev) ring->name = "video enhancement ring"; ring->id = VECS; + ring->exec_id = I915_EXEC_VEBOX; ring->mmio_base = VEBOX_RING_BASE; - ring->irq_enable_mask = - GT_RENDER_USER_INTERRUPT << GEN8_VECS_IRQ_SHIFT; - ring->irq_keep_mask = - GT_CONTEXT_SWITCH_INTERRUPT << GEN8_VECS_IRQ_SHIFT; - ring->init_hw = gen8_init_common_ring; - if (IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { - ring->get_seqno = bxt_a_get_seqno; - ring->set_seqno = bxt_a_set_seqno; - } else { - ring->get_seqno = gen8_get_seqno; - ring->set_seqno = gen8_set_seqno; - } - ring->emit_request = gen8_emit_request; - ring->emit_flush = gen8_emit_flush; - ring->irq_get = gen8_logical_ring_get_irq; - ring->irq_put = gen8_logical_ring_put_irq; - ring->emit_bb_start = gen8_emit_bb_start; + logical_ring_default_irqs(ring, GEN8_VECS_IRQ_SHIFT); + logical_ring_default_vfuncs(dev, ring); return logical_ring_init(dev, ring); } @@ -2368,26 +2426,39 @@ void intel_lr_context_free(struct intel_context *ctx) { int i; - for (i = 0; i < I915_NUM_RINGS; i++) { + for (i = I915_NUM_RINGS; --i >= 0; ) { + struct intel_ringbuffer *ringbuf = ctx->engine[i].ringbuf; struct drm_i915_gem_object *ctx_obj = ctx->engine[i].state; - if (ctx_obj) { - struct intel_ringbuffer *ringbuf = - ctx->engine[i].ringbuf; - struct intel_engine_cs *ring = ringbuf->ring; + if (!ctx_obj) + continue; - if (ctx == ring->default_context) { - intel_unpin_ringbuffer_obj(ringbuf); - i915_gem_object_ggtt_unpin(ctx_obj); - } - WARN_ON(ctx->engine[ring->id].pin_count); - intel_ringbuffer_free(ringbuf); - drm_gem_object_unreference(&ctx_obj->base); + if (ctx == ctx->i915->kernel_context) { + intel_unpin_ringbuffer_obj(ringbuf); + i915_gem_object_ggtt_unpin(ctx_obj); } + + WARN_ON(ctx->engine[i].pin_count); + intel_ringbuffer_free(ringbuf); + drm_gem_object_unreference(&ctx_obj->base); } } -static uint32_t get_lr_context_size(struct intel_engine_cs *ring) +/** + * intel_lr_context_size() - return the size of the context for an engine + * @ring: which engine to find the context size for + * + * Each engine may require a different amount of space for a context image, + * so when allocating (or copying) an image, this function can be used to + * find the right size for the specific engine. + * + * Return: size (in bytes) of an engine-specific context image + * + * Note: this size includes the HWSP, which is part of the context image + * in LRC mode, but does not include the "shared data page" used with + * GuC submission. The caller should account for this if using the GuC. + */ +uint32_t intel_lr_context_size(struct intel_engine_cs *ring) { int ret = 0; @@ -2444,7 +2515,7 @@ static void lrc_setup_hardware_status_page(struct intel_engine_cs *ring, */ int intel_lr_context_deferred_alloc(struct intel_context *ctx, - struct intel_engine_cs *ring) + struct intel_engine_cs *ring) { struct drm_device *dev = ring->dev; struct drm_i915_gem_object *ctx_obj; @@ -2455,7 +2526,7 @@ int intel_lr_context_deferred_alloc(struct intel_context *ctx, WARN_ON(ctx->legacy_hw_ctx.rcs_state != NULL); WARN_ON(ctx->engine[ring->id].state); - context_size = round_up(get_lr_context_size(ring), 4096); + context_size = round_up(intel_lr_context_size(ring), 4096); /* One extra page as the sharing data between driver and GuC */ context_size += PAGE_SIZE * LRC_PPHWSP_PN; @@ -2481,14 +2552,13 @@ int intel_lr_context_deferred_alloc(struct intel_context *ctx, ctx->engine[ring->id].ringbuf = ringbuf; ctx->engine[ring->id].state = ctx_obj; - if (ctx != ring->default_context && ring->init_context) { + if (ctx != ctx->i915->kernel_context && ring->init_context) { struct drm_i915_gem_request *req; - ret = i915_gem_request_alloc(ring, - ctx, &req); - if (ret) { - DRM_ERROR("ring create req: %d\n", - ret); + req = i915_gem_request_alloc(ring, ctx); + if (IS_ERR(req)) { + ret = PTR_ERR(req); + DRM_ERROR("ring create req: %d\n", ret); goto error_ringbuf; } diff --git a/drivers/gpu/drm/i915/intel_lrc.h b/drivers/gpu/drm/i915/intel_lrc.h index 0b821b91723a..49af638f6213 100644 --- a/drivers/gpu/drm/i915/intel_lrc.h +++ b/drivers/gpu/drm/i915/intel_lrc.h @@ -25,8 +25,6 @@ #define _INTEL_LRC_H_ #define GEN8_LR_CONTEXT_ALIGN 4096 -#define GEN8_CSB_ENTRIES 6 -#define GEN8_CSB_PTR_MASK 0x07 /* Execlists regs */ #define RING_ELSP(ring) _MMIO((ring)->mmio_base + 0x230) @@ -40,6 +38,22 @@ #define RING_CONTEXT_STATUS_BUF_HI(ring, i) _MMIO((ring)->mmio_base + 0x370 + (i) * 8 + 4) #define RING_CONTEXT_STATUS_PTR(ring) _MMIO((ring)->mmio_base + 0x3a0) +/* The docs specify that the write pointer wraps around after 5h, "After status + * is written out to the last available status QW at offset 5h, this pointer + * wraps to 0." + * + * Therefore, one must infer than even though there are 3 bits available, 6 and + * 7 appear to be * reserved. + */ +#define GEN8_CSB_ENTRIES 6 +#define GEN8_CSB_PTR_MASK 0x7 +#define GEN8_CSB_READ_PTR_MASK (GEN8_CSB_PTR_MASK << 8) +#define GEN8_CSB_WRITE_PTR_MASK (GEN8_CSB_PTR_MASK << 0) +#define GEN8_CSB_WRITE_PTR(csb_status) \ + (((csb_status) & GEN8_CSB_WRITE_PTR_MASK) >> 0) +#define GEN8_CSB_READ_PTR(csb_status) \ + (((csb_status) & GEN8_CSB_READ_PTR_MASK) >> 8) + /* Logical Rings */ int intel_logical_ring_alloc_request_extras(struct drm_i915_gem_request *request); int intel_logical_ring_reserve_space(struct drm_i915_gem_request *request); @@ -84,6 +98,7 @@ static inline void intel_logical_ring_emit_reg(struct intel_ringbuffer *ringbuf, #define LRC_STATE_PN (LRC_PPHWSP_PN + 1) void intel_lr_context_free(struct intel_context *ctx); +uint32_t intel_lr_context_size(struct intel_engine_cs *ring); int intel_lr_context_deferred_alloc(struct intel_context *ctx, struct intel_engine_cs *ring); void intel_lr_context_unpin(struct drm_i915_gem_request *req); @@ -92,13 +107,15 @@ void intel_lr_context_reset(struct drm_device *dev, uint64_t intel_lr_context_descriptor(struct intel_context *ctx, struct intel_engine_cs *ring); +u32 intel_execlists_ctx_id(struct intel_context *ctx, + struct intel_engine_cs *ring); + /* Execlists */ int intel_sanitize_enable_execlists(struct drm_device *dev, int enable_execlists); struct i915_execbuffer_params; int intel_execlists_submission(struct i915_execbuffer_params *params, struct drm_i915_gem_execbuffer2 *args, struct list_head *vmas); -u32 intel_execlists_ctx_id(struct drm_i915_gem_object *ctx_obj); void intel_lrc_irq_handler(struct intel_engine_cs *ring); void intel_execlists_retire_requests(struct intel_engine_cs *ring); diff --git a/drivers/gpu/drm/i915/intel_lvds.c b/drivers/gpu/drm/i915/intel_lvds.c index 0da0240caf81..811ddf7799f0 100644 --- a/drivers/gpu/drm/i915/intel_lvds.c +++ b/drivers/gpu/drm/i915/intel_lvds.c @@ -31,6 +31,7 @@ #include <linux/dmi.h> #include <linux/i2c.h> #include <linux/slab.h> +#include <linux/vga_switcheroo.h> #include <drm/drmP.h> #include <drm/drm_atomic_helper.h> #include <drm/drm_crtc.h> @@ -1080,7 +1081,12 @@ void intel_lvds_init(struct drm_device *dev) * preferred mode is the right one. */ mutex_lock(&dev->mode_config.mutex); - edid = drm_get_edid(connector, intel_gmbus_get_adapter(dev_priv, pin)); + if (vga_switcheroo_handler_flags() & VGA_SWITCHEROO_CAN_SWITCH_DDC) + edid = drm_get_edid_switcheroo(connector, + intel_gmbus_get_adapter(dev_priv, pin)); + else + edid = drm_get_edid(connector, + intel_gmbus_get_adapter(dev_priv, pin)); if (edid) { if (drm_add_edid_modes(connector, edid)) { drm_mode_connector_update_edid_property(connector, diff --git a/drivers/gpu/drm/i915/intel_overlay.c b/drivers/gpu/drm/i915/intel_overlay.c index 76f1980a7541..9168413fe204 100644 --- a/drivers/gpu/drm/i915/intel_overlay.c +++ b/drivers/gpu/drm/i915/intel_overlay.c @@ -240,9 +240,9 @@ static int intel_overlay_on(struct intel_overlay *overlay) WARN_ON(overlay->active); WARN_ON(IS_I830(dev) && !(dev_priv->quirks & QUIRK_PIPEA_FORCE)); - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) - return ret; + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); ret = intel_ring_begin(req, 4); if (ret) { @@ -283,9 +283,9 @@ static int intel_overlay_continue(struct intel_overlay *overlay, if (tmp & (1 << 17)) DRM_DEBUG("overlay underrun, DOVSTA: %x\n", tmp); - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) - return ret; + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); ret = intel_ring_begin(req, 2); if (ret) { @@ -349,9 +349,9 @@ static int intel_overlay_off(struct intel_overlay *overlay) * of the hw. Do it in both cases */ flip_addr |= OFC_UPDATE; - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) - return ret; + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); ret = intel_ring_begin(req, 6); if (ret) { @@ -423,9 +423,9 @@ static int intel_overlay_release_old_vid(struct intel_overlay *overlay) /* synchronous slowpath */ struct drm_i915_gem_request *req; - ret = i915_gem_request_alloc(ring, ring->default_context, &req); - if (ret) - return ret; + req = i915_gem_request_alloc(ring, NULL); + if (IS_ERR(req)) + return PTR_ERR(req); ret = intel_ring_begin(req, 2); if (ret) { diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c index eb5fa05cf476..20bf854eae8c 100644 --- a/drivers/gpu/drm/i915/intel_pm.c +++ b/drivers/gpu/drm/i915/intel_pm.c @@ -32,6 +32,8 @@ #include <linux/module.h> /** + * DOC: RC6 + * * RC6 is a special power stage which allows the GPU to enter an very * low-voltage mode when idle, using down to 0V while at this stage. This * stage is entered automatically when the GPU is idle when RC6 support is @@ -1672,6 +1674,9 @@ uint32_t ilk_pipe_pixel_rate(const struct intel_crtc_state *pipe_config) if (pipe_h < pfit_h) pipe_h = pfit_h; + if (WARN_ON(!pfit_w || !pfit_h)) + return pixel_rate; + pixel_rate = div_u64((uint64_t) pixel_rate * pipe_w * pipe_h, pfit_w * pfit_h); } @@ -1703,6 +1708,8 @@ static uint32_t ilk_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal, if (WARN(latency == 0, "Latency value missing\n")) return UINT_MAX; + if (WARN_ON(!pipe_htotal)) + return UINT_MAX; ret = (latency * pixel_rate) / (pipe_htotal * 10000); ret = (ret + 1) * horiz_pixels * bytes_per_pixel; @@ -1713,6 +1720,17 @@ static uint32_t ilk_wm_method2(uint32_t pixel_rate, uint32_t pipe_htotal, static uint32_t ilk_wm_fbc(uint32_t pri_val, uint32_t horiz_pixels, uint8_t bytes_per_pixel) { + /* + * Neither of these should be possible since this function shouldn't be + * called if the CRTC is off or the plane is invisible. But let's be + * extra paranoid to avoid a potential divide-by-zero if we screw up + * elsewhere in the driver. + */ + if (WARN_ON(!bytes_per_pixel)) + return 0; + if (WARN_ON(!horiz_pixels)) + return 0; + return DIV_ROUND_UP(pri_val * 64, horiz_pixels * bytes_per_pixel) + 2; } @@ -1998,14 +2016,19 @@ static void ilk_compute_wm_level(const struct drm_i915_private *dev_priv, } static uint32_t -hsw_compute_linetime_wm(struct drm_device *dev, struct drm_crtc *crtc) +hsw_compute_linetime_wm(struct drm_device *dev, + struct intel_crtc_state *cstate) { struct drm_i915_private *dev_priv = dev->dev_private; - struct intel_crtc *intel_crtc = to_intel_crtc(crtc); - const struct drm_display_mode *adjusted_mode = &intel_crtc->config->base.adjusted_mode; + const struct drm_display_mode *adjusted_mode = + &cstate->base.adjusted_mode; u32 linetime, ips_linetime; - if (!intel_crtc->active) + if (!cstate->base.active) + return 0; + if (WARN_ON(adjusted_mode->crtc_clock == 0)) + return 0; + if (WARN_ON(dev_priv->cdclk_freq == 0)) return 0; /* The WM are computed with base on how long it takes to fill a single @@ -2277,6 +2300,7 @@ static int ilk_compute_pipe_wm(struct intel_crtc *intel_crtc, return PTR_ERR(cstate); pipe_wm = &cstate->wm.optimal.ilk; + memset(pipe_wm, 0, sizeof(*pipe_wm)); for_each_intel_plane_on_crtc(dev, intel_crtc, intel_plane) { ps = drm_atomic_get_plane_state(state, @@ -2313,8 +2337,7 @@ static int ilk_compute_pipe_wm(struct intel_crtc *intel_crtc, pristate, sprstate, curstate, &pipe_wm->wm[0]); if (IS_HASWELL(dev) || IS_BROADWELL(dev)) - pipe_wm->linetime = hsw_compute_linetime_wm(dev, - &intel_crtc->base); + pipe_wm->linetime = hsw_compute_linetime_wm(dev, cstate); /* LP0 watermarks always use 1/2 DDB partitioning */ ilk_compute_wm_maximums(dev, 0, &config, INTEL_DDB_PART_1_2, &max); @@ -3597,23 +3620,45 @@ static void skl_update_wm(struct drm_crtc *crtc) dev_priv->wm.skl_hw = *results; } -static void ilk_program_watermarks(struct drm_i915_private *dev_priv) +static void ilk_compute_wm_config(struct drm_device *dev, + struct intel_wm_config *config) { - struct drm_device *dev = dev_priv->dev; + struct intel_crtc *crtc; + + /* Compute the currently _active_ config */ + for_each_intel_crtc(dev, crtc) { + const struct intel_pipe_wm *wm = &crtc->wm.active.ilk; + + if (!wm->pipe_enabled) + continue; + + config->sprites_enabled |= wm->sprites_enabled; + config->sprites_scaled |= wm->sprites_scaled; + config->num_pipes_active++; + } +} + +static void ilk_program_watermarks(struct intel_crtc_state *cstate) +{ + struct drm_crtc *crtc = cstate->base.crtc; + struct drm_device *dev = crtc->dev; + struct drm_i915_private *dev_priv = to_i915(dev); struct intel_pipe_wm lp_wm_1_2 = {}, lp_wm_5_6 = {}, *best_lp_wm; struct ilk_wm_maximums max; - struct intel_wm_config *config = &dev_priv->wm.config; + struct intel_wm_config config = {}; struct ilk_wm_values results = {}; enum intel_ddb_partitioning partitioning; - ilk_compute_wm_maximums(dev, 1, config, INTEL_DDB_PART_1_2, &max); - ilk_wm_merge(dev, config, &max, &lp_wm_1_2); + ilk_compute_wm_config(dev, &config); + + ilk_compute_wm_maximums(dev, 1, &config, INTEL_DDB_PART_1_2, &max); + ilk_wm_merge(dev, &config, &max, &lp_wm_1_2); /* 5/6 split only in single pipe config on IVB+ */ if (INTEL_INFO(dev)->gen >= 7 && - config->num_pipes_active == 1 && config->sprites_enabled) { - ilk_compute_wm_maximums(dev, 1, config, INTEL_DDB_PART_5_6, &max); - ilk_wm_merge(dev, config, &max, &lp_wm_5_6); + config.num_pipes_active == 1 && config.sprites_enabled) { + ilk_compute_wm_maximums(dev, 1, &config, INTEL_DDB_PART_5_6, &max); + ilk_wm_merge(dev, &config, &max, &lp_wm_5_6); best_lp_wm = ilk_find_best_result(dev, &lp_wm_1_2, &lp_wm_5_6); } else { @@ -3630,7 +3675,6 @@ static void ilk_program_watermarks(struct drm_i915_private *dev_priv) static void ilk_update_wm(struct drm_crtc *crtc) { - struct drm_i915_private *dev_priv = to_i915(crtc->dev); struct intel_crtc *intel_crtc = to_intel_crtc(crtc); struct intel_crtc_state *cstate = to_intel_crtc_state(crtc->state); @@ -3650,7 +3694,7 @@ static void ilk_update_wm(struct drm_crtc *crtc) intel_crtc->wm.active.ilk = cstate->wm.optimal.ilk; - ilk_program_watermarks(dev_priv); + ilk_program_watermarks(cstate); } static void skl_pipe_wm_active_state(uint32_t val, @@ -4036,7 +4080,7 @@ void intel_update_watermarks(struct drm_crtc *crtc) dev_priv->display.update_wm(crtc); } -/** +/* * Lock protecting IPS related data structures */ DEFINE_SPINLOCK(mchdev_lock); @@ -4509,13 +4553,13 @@ static void intel_print_rc6_info(struct drm_device *dev, u32 mode) } if (HAS_RC6p(dev)) DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s RC6p %s RC6pp %s\n", - (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off", - (mode & GEN6_RC_CTL_RC6p_ENABLE) ? "on" : "off", - (mode & GEN6_RC_CTL_RC6pp_ENABLE) ? "on" : "off"); + onoff(mode & GEN6_RC_CTL_RC6_ENABLE), + onoff(mode & GEN6_RC_CTL_RC6p_ENABLE), + onoff(mode & GEN6_RC_CTL_RC6pp_ENABLE)); else DRM_DEBUG_KMS("Enabling RC6 states: RC6 %s\n", - (mode & GEN6_RC_CTL_RC6_ENABLE) ? "on" : "off"); + onoff(mode & GEN6_RC_CTL_RC6_ENABLE)); } static int sanitize_rc6_option(const struct drm_device *dev, int enable_rc6) @@ -4693,8 +4737,7 @@ static void gen9_enable_rc6(struct drm_device *dev) /* 3a: Enable RC6 */ if (intel_enable_rc6(dev) & INTEL_RC6_ENABLE) rc6_mask = GEN6_RC_CTL_RC6_ENABLE; - DRM_INFO("RC6 %s\n", (rc6_mask & GEN6_RC_CTL_RC6_ENABLE) ? - "on" : "off"); + DRM_INFO("RC6 %s\n", onoff(rc6_mask & GEN6_RC_CTL_RC6_ENABLE)); /* WaRsUseTimeoutMode */ if (IS_SKL_REVID(dev, 0, SKL_REVID_D0) || IS_BXT_REVID(dev, 0, BXT_REVID_A1)) { @@ -4713,8 +4756,7 @@ static void gen9_enable_rc6(struct drm_device *dev) * 3b: Enable Coarse Power Gating only when RC6 is enabled. * WaRsDisableCoarsePowerGating:skl,bxt - Render/Media PG need to be disabled with RC6. */ - if ((IS_BROXTON(dev) && (INTEL_REVID(dev) < BXT_REVID_B0)) || - ((IS_SKL_GT3(dev) || IS_SKL_GT4(dev)) && (INTEL_REVID(dev) <= SKL_REVID_F0))) + if (NEEDS_WaRsDisableCoarsePowerGating(dev)) I915_WRITE(GEN9_PG_ENABLE, 0); else I915_WRITE(GEN9_PG_ENABLE, (rc6_mask & GEN6_RC_CTL_RC6_ENABLE) ? @@ -6981,6 +7023,7 @@ void intel_init_pm(struct drm_device *dev) dev_priv->wm.spr_latency[0] && dev_priv->wm.cur_latency[0])) { dev_priv->display.update_wm = ilk_update_wm; dev_priv->display.compute_pipe_wm = ilk_compute_pipe_wm; + dev_priv->display.program_watermarks = ilk_program_watermarks; } else { DRM_DEBUG_KMS("Failed to read display plane latency. " "Disable CxSR\n"); diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.c b/drivers/gpu/drm/i915/intel_ringbuffer.c index 40c6aff57256..9030e2bca0c0 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.c +++ b/drivers/gpu/drm/i915/intel_ringbuffer.c @@ -1867,15 +1867,13 @@ i830_dispatch_execbuffer(struct drm_i915_gem_request *req, offset = cs_offset; } - ret = intel_ring_begin(req, 4); + ret = intel_ring_begin(req, 2); if (ret) return ret; - intel_ring_emit(ring, MI_BATCH_BUFFER); + intel_ring_emit(ring, MI_BATCH_BUFFER_START | MI_BATCH_GTT); intel_ring_emit(ring, offset | (dispatch_flags & I915_DISPATCH_SECURE ? 0 : MI_BATCH_NON_SECURE)); - intel_ring_emit(ring, offset + len - 8); - intel_ring_emit(ring, MI_NOOP); intel_ring_advance(ring); return 0; @@ -1901,6 +1899,17 @@ i915_dispatch_execbuffer(struct drm_i915_gem_request *req, return 0; } +static void cleanup_phys_status_page(struct intel_engine_cs *ring) +{ + struct drm_i915_private *dev_priv = to_i915(ring->dev); + + if (!dev_priv->status_page_dmah) + return; + + drm_pci_free(ring->dev, dev_priv->status_page_dmah); + ring->status_page.page_addr = NULL; +} + static void cleanup_status_page(struct intel_engine_cs *ring) { struct drm_i915_gem_object *obj; @@ -1917,9 +1926,9 @@ static void cleanup_status_page(struct intel_engine_cs *ring) static int init_status_page(struct intel_engine_cs *ring) { - struct drm_i915_gem_object *obj; + struct drm_i915_gem_object *obj = ring->status_page.obj; - if ((obj = ring->status_page.obj) == NULL) { + if (obj == NULL) { unsigned flags; int ret; @@ -1990,6 +1999,7 @@ void intel_unpin_ringbuffer_obj(struct intel_ringbuffer *ringbuf) else iounmap(ringbuf->virtual_start); ringbuf->virtual_start = NULL; + ringbuf->vma = NULL; i915_gem_object_ggtt_unpin(ringbuf->obj); } @@ -2056,6 +2066,8 @@ int intel_pin_and_map_ringbuffer_obj(struct drm_device *dev, } } + ringbuf->vma = i915_gem_obj_to_ggtt(obj); + return 0; } @@ -2164,7 +2176,7 @@ static int intel_init_ring_buffer(struct drm_device *dev, if (ret) goto error; } else { - BUG_ON(ring->id != RCS); + WARN_ON(ring->id != RCS); ret = init_phys_status_page(ring); if (ret) goto error; @@ -2210,7 +2222,12 @@ void intel_cleanup_ring_buffer(struct intel_engine_cs *ring) if (ring->cleanup) ring->cleanup(ring); - cleanup_status_page(ring); + if (I915_NEED_GFX_HWS(ring->dev)) { + cleanup_status_page(ring); + } else { + WARN_ON(ring->id != RCS); + cleanup_phys_status_page(ring); + } i915_cmd_parser_fini_ring(ring); i915_gem_batch_pool_fini(&ring->batch_pool); @@ -2666,6 +2683,7 @@ int intel_init_render_ring_buffer(struct drm_device *dev) ring->name = "render ring"; ring->id = RCS; + ring->exec_id = I915_EXEC_RENDER; ring->mmio_base = RENDER_RING_BASE; if (INTEL_INFO(dev)->gen >= 8) { @@ -2814,6 +2832,7 @@ int intel_init_bsd_ring_buffer(struct drm_device *dev) ring->name = "bsd ring"; ring->id = VCS; + ring->exec_id = I915_EXEC_BSD; ring->write_tail = ring_write_tail; if (INTEL_INFO(dev)->gen >= 6) { @@ -2890,6 +2909,7 @@ int intel_init_bsd2_ring_buffer(struct drm_device *dev) ring->name = "bsd2 ring"; ring->id = VCS2; + ring->exec_id = I915_EXEC_BSD; ring->write_tail = ring_write_tail; ring->mmio_base = GEN8_BSD2_RING_BASE; @@ -2920,6 +2940,7 @@ int intel_init_blt_ring_buffer(struct drm_device *dev) ring->name = "blitter ring"; ring->id = BCS; + ring->exec_id = I915_EXEC_BLT; ring->mmio_base = BLT_RING_BASE; ring->write_tail = ring_write_tail; @@ -2977,6 +2998,7 @@ int intel_init_vebox_ring_buffer(struct drm_device *dev) ring->name = "video enhancement ring"; ring->id = VECS; + ring->exec_id = I915_EXEC_VEBOX; ring->mmio_base = VEBOX_RING_BASE; ring->write_tail = ring_write_tail; diff --git a/drivers/gpu/drm/i915/intel_ringbuffer.h b/drivers/gpu/drm/i915/intel_ringbuffer.h index 49574ffe54bc..b12f2aabd104 100644 --- a/drivers/gpu/drm/i915/intel_ringbuffer.h +++ b/drivers/gpu/drm/i915/intel_ringbuffer.h @@ -93,11 +93,13 @@ struct intel_ring_hangcheck { int score; enum intel_ring_hangcheck_action action; int deadlock; + u32 instdone[I915_NUM_INSTDONE_REG]; }; struct intel_ringbuffer { struct drm_i915_gem_object *obj; void __iomem *virtual_start; + struct i915_vma *vma; struct intel_engine_cs *ring; struct list_head link; @@ -147,14 +149,15 @@ struct i915_ctx_workarounds { struct intel_engine_cs { const char *name; enum intel_ring_id { - RCS = 0x0, - VCS, + RCS = 0, BCS, - VECS, - VCS2 + VCS, + VCS2, /* Keep instances of the same type engine together. */ + VECS } id; #define I915_NUM_RINGS 5 -#define LAST_USER_RING (VECS + 1) +#define _VCS(n) (VCS + (n)) + unsigned int exec_id; u32 mmio_base; struct drm_device *dev; struct intel_ringbuffer *buffer; @@ -268,6 +271,8 @@ struct intel_engine_cs { struct list_head execlist_queue; struct list_head execlist_retired_req_list; u8 next_context_status_buffer; + bool disable_lite_restore_wa; + u32 ctx_desc_template; u32 irq_keep_mask; /* bitmask for interrupts that should not be masked */ int (*emit_request)(struct drm_i915_gem_request *request); int (*emit_flush)(struct drm_i915_gem_request *request, @@ -305,7 +310,6 @@ struct intel_engine_cs { wait_queue_head_t irq_queue; - struct intel_context *default_context; struct intel_context *last_context; struct intel_ring_hangcheck hangcheck; @@ -406,7 +410,7 @@ intel_write_status_page(struct intel_engine_cs *ring, ring->status_page.page_addr[reg] = value; } -/** +/* * Reads a dword out of the status page, which is written to from the command * queue by automatic updates, MI_REPORT_HEAD, MI_STORE_DATA_INDEX, or * MI_STORE_DATA_IMM. @@ -423,6 +427,7 @@ intel_write_status_page(struct intel_engine_cs *ring, * The area from dword 0x30 to 0x3ff is available for driver usage. */ #define I915_GEM_HWS_INDEX 0x30 +#define I915_GEM_HWS_INDEX_ADDR (I915_GEM_HWS_INDEX << MI_STORE_DWORD_INDEX_SHIFT) #define I915_GEM_HWS_SCRATCH_INDEX 0x40 #define I915_GEM_HWS_SCRATCH_ADDR (I915_GEM_HWS_SCRATCH_INDEX << MI_STORE_DWORD_INDEX_SHIFT) diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c index ddbdbffe829a..bbca527184d0 100644 --- a/drivers/gpu/drm/i915/intel_runtime_pm.c +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c @@ -532,7 +532,8 @@ static void assert_can_enable_dc5(struct drm_i915_private *dev_priv) bool pg2_enabled = intel_display_power_well_is_enabled(dev_priv, SKL_DISP_PW_2); - WARN_ONCE(!IS_SKYLAKE(dev), "Platform doesn't support DC5.\n"); + WARN_ONCE(!IS_SKYLAKE(dev) && !IS_KABYLAKE(dev), + "Platform doesn't support DC5.\n"); WARN_ONCE(!HAS_RUNTIME_PM(dev), "Runtime PM not enabled.\n"); WARN_ONCE(pg2_enabled, "PG2 not disabled to enable DC5.\n"); @@ -568,7 +569,8 @@ static void assert_can_enable_dc6(struct drm_i915_private *dev_priv) { struct drm_device *dev = dev_priv->dev; - WARN_ONCE(!IS_SKYLAKE(dev), "Platform doesn't support DC6.\n"); + WARN_ONCE(!IS_SKYLAKE(dev) && !IS_KABYLAKE(dev), + "Platform doesn't support DC6.\n"); WARN_ONCE(!HAS_RUNTIME_PM(dev), "Runtime PM not enabled.\n"); WARN_ONCE(I915_READ(UTIL_PIN_CTL) & UTIL_PIN_ENABLE, "Backlight is not disabled.\n"); @@ -595,7 +597,8 @@ static void gen9_disable_dc5_dc6(struct drm_i915_private *dev_priv) { assert_can_disable_dc5(dev_priv); - if (IS_SKYLAKE(dev_priv) && i915.enable_dc != 0 && i915.enable_dc != 1) + if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && + i915.enable_dc != 0 && i915.enable_dc != 1) assert_can_disable_dc6(dev_priv); gen9_set_dc_state(dev_priv, DC_STATE_DISABLE); @@ -623,7 +626,6 @@ void skl_disable_dc6(struct drm_i915_private *dev_priv) static void skl_set_power_well(struct drm_i915_private *dev_priv, struct i915_power_well *power_well, bool enable) { - struct drm_device *dev = dev_priv->dev; uint32_t tmp, fuse_status; uint32_t req_mask, state_mask; bool is_enabled, enable_requested, check_fuse_status = false; @@ -667,17 +669,6 @@ static void skl_set_power_well(struct drm_i915_private *dev_priv, !I915_READ(HSW_PWR_WELL_BIOS), "Invalid for power well status to be enabled, unless done by the BIOS, \ when request is to disable!\n"); - if (power_well->data == SKL_DISP_PW_2) { - /* - * DDI buffer programming unnecessary during - * driver-load/resume as it's already done - * during modeset initialization then. It's - * also invalid here as encoder list is still - * uninitialized. - */ - if (!dev_priv->power_domains.initializing) - intel_prepare_ddi(dev); - } I915_WRITE(HSW_PWR_WELL_DRIVER, tmp | req_mask); } @@ -783,7 +774,8 @@ static void gen9_dc_off_power_well_enable(struct drm_i915_private *dev_priv, static void gen9_dc_off_power_well_disable(struct drm_i915_private *dev_priv, struct i915_power_well *power_well) { - if (IS_SKYLAKE(dev_priv) && i915.enable_dc != 0 && i915.enable_dc != 1) + if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && + i915.enable_dc != 0 && i915.enable_dc != 1) skl_enable_dc6(dev_priv); else gen9_enable_dc5(dev_priv); @@ -795,7 +787,8 @@ static void gen9_dc_off_power_well_sync_hw(struct drm_i915_private *dev_priv, if (power_well->count > 0) { gen9_set_dc_state(dev_priv, DC_STATE_DISABLE); } else { - if (IS_SKYLAKE(dev_priv) && i915.enable_dc != 0 && + if ((IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv)) && + i915.enable_dc != 0 && i915.enable_dc != 1) gen9_set_dc_state(dev_priv, DC_STATE_EN_UPTO_DC6); else @@ -1851,7 +1844,7 @@ void skl_pw1_misc_io_init(struct drm_i915_private *dev_priv) { struct i915_power_well *well; - if (!IS_SKYLAKE(dev_priv)) + if (!(IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))) return; well = lookup_power_well(dev_priv, SKL_DISP_PW_1); @@ -1865,7 +1858,7 @@ void skl_pw1_misc_io_fini(struct drm_i915_private *dev_priv) { struct i915_power_well *well; - if (!IS_SKYLAKE(dev_priv)) + if (!(IS_SKYLAKE(dev_priv) || IS_KABYLAKE(dev_priv))) return; well = lookup_power_well(dev_priv, SKL_DISP_PW_1); diff --git a/drivers/gpu/drm/i915/intel_sdvo_regs.h b/drivers/gpu/drm/i915/intel_sdvo_regs.h index 2e2d4eb4a00d..db0ed499268a 100644 --- a/drivers/gpu/drm/i915/intel_sdvo_regs.h +++ b/drivers/gpu/drm/i915/intel_sdvo_regs.h @@ -24,8 +24,8 @@ * Eric Anholt <[email protected]> */ -/** - * @file SDVO command definitions and structures. +/* + * SDVO command definitions and structures. */ #define SDVO_OUTPUT_FIRST (0) @@ -66,39 +66,39 @@ struct intel_sdvo_caps { #define DTD_FLAG_VSYNC_POSITIVE (1 << 2) #define DTD_FLAG_INTERLACE (1 << 7) -/** This matches the EDID DTD structure, more or less */ +/* This matches the EDID DTD structure, more or less */ struct intel_sdvo_dtd { struct { - u16 clock; /**< pixel clock, in 10kHz units */ - u8 h_active; /**< lower 8 bits (pixels) */ - u8 h_blank; /**< lower 8 bits (pixels) */ - u8 h_high; /**< upper 4 bits each h_active, h_blank */ - u8 v_active; /**< lower 8 bits (lines) */ - u8 v_blank; /**< lower 8 bits (lines) */ - u8 v_high; /**< upper 4 bits each v_active, v_blank */ + u16 clock; /* pixel clock, in 10kHz units */ + u8 h_active; /* lower 8 bits (pixels) */ + u8 h_blank; /* lower 8 bits (pixels) */ + u8 h_high; /* upper 4 bits each h_active, h_blank */ + u8 v_active; /* lower 8 bits (lines) */ + u8 v_blank; /* lower 8 bits (lines) */ + u8 v_high; /* upper 4 bits each v_active, v_blank */ } part1; struct { - u8 h_sync_off; /**< lower 8 bits, from hblank start */ - u8 h_sync_width; /**< lower 8 bits (pixels) */ - /** lower 4 bits each vsync offset, vsync width */ + u8 h_sync_off; /* lower 8 bits, from hblank start */ + u8 h_sync_width; /* lower 8 bits (pixels) */ + /* lower 4 bits each vsync offset, vsync width */ u8 v_sync_off_width; - /** + /* * 2 high bits of hsync offset, 2 high bits of hsync width, * bits 4-5 of vsync offset, and 2 high bits of vsync width. */ u8 sync_off_width_high; u8 dtd_flags; u8 sdvo_flags; - /** bits 6-7 of vsync offset at bits 6-7 */ + /* bits 6-7 of vsync offset at bits 6-7 */ u8 v_sync_off_high; u8 reserved; } part2; } __packed; struct intel_sdvo_pixel_clock_range { - u16 min; /**< pixel clock, in 10kHz units */ - u16 max; /**< pixel clock, in 10kHz units */ + u16 min; /* pixel clock, in 10kHz units */ + u16 max; /* pixel clock, in 10kHz units */ } __packed; struct intel_sdvo_preferred_input_timing_args { @@ -144,7 +144,7 @@ struct intel_sdvo_preferred_input_timing_args { #define SDVO_CMD_RESET 0x01 -/** Returns a struct intel_sdvo_caps */ +/* Returns a struct intel_sdvo_caps */ #define SDVO_CMD_GET_DEVICE_CAPS 0x02 #define SDVO_CMD_GET_FIRMWARE_REV 0x86 @@ -152,7 +152,7 @@ struct intel_sdvo_preferred_input_timing_args { # define SDVO_DEVICE_FIRMWARE_MAJOR SDVO_I2C_RETURN_1 # define SDVO_DEVICE_FIRMWARE_PATCH SDVO_I2C_RETURN_2 -/** +/* * Reports which inputs are trained (managed to sync). * * Devices must have trained within 2 vsyncs of a mode change. @@ -164,10 +164,10 @@ struct intel_sdvo_get_trained_inputs_response { unsigned int pad:6; } __packed; -/** Returns a struct intel_sdvo_output_flags of active outputs. */ +/* Returns a struct intel_sdvo_output_flags of active outputs. */ #define SDVO_CMD_GET_ACTIVE_OUTPUTS 0x04 -/** +/* * Sets the current set of active outputs. * * Takes a struct intel_sdvo_output_flags. Must be preceded by a SET_IN_OUT_MAP @@ -175,7 +175,7 @@ struct intel_sdvo_get_trained_inputs_response { */ #define SDVO_CMD_SET_ACTIVE_OUTPUTS 0x05 -/** +/* * Returns the current mapping of SDVO inputs to outputs on the device. * * Returns two struct intel_sdvo_output_flags structures. @@ -185,29 +185,29 @@ struct intel_sdvo_in_out_map { u16 in0, in1; }; -/** +/* * Sets the current mapping of SDVO inputs to outputs on the device. * * Takes two struct i380_sdvo_output_flags structures. */ #define SDVO_CMD_SET_IN_OUT_MAP 0x07 -/** +/* * Returns a struct intel_sdvo_output_flags of attached displays. */ #define SDVO_CMD_GET_ATTACHED_DISPLAYS 0x0b -/** +/* * Returns a struct intel_sdvo_ouptut_flags of displays supporting hot plugging. */ #define SDVO_CMD_GET_HOT_PLUG_SUPPORT 0x0c -/** +/* * Takes a struct intel_sdvo_output_flags. */ #define SDVO_CMD_SET_ACTIVE_HOT_PLUG 0x0d -/** +/* * Returns a struct intel_sdvo_output_flags of displays with hot plug * interrupts enabled. */ @@ -221,7 +221,7 @@ struct intel_sdvo_get_interrupt_event_source_response { unsigned int pad:6; } __packed; -/** +/* * Selects which input is affected by future input commands. * * Commands affected include SET_INPUT_TIMINGS_PART[12], @@ -234,7 +234,7 @@ struct intel_sdvo_set_target_input_args { unsigned int pad:7; } __packed; -/** +/* * Takes a struct intel_sdvo_output_flags of which outputs are targeted by * future output commands. * @@ -280,7 +280,7 @@ struct intel_sdvo_set_target_input_args { # define SDVO_DTD_SDVO_FLAG_SCALING_SMOOTH (2 << 4) # define SDVO_DTD_VSYNC_OFF_HIGH SDVO_I2C_ARG_6 -/** +/* * Generates a DTD based on the given width, height, and flags. * * This will be supported by any device supporting scaling or interlaced @@ -300,24 +300,24 @@ struct intel_sdvo_set_target_input_args { #define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART1 0x1b #define SDVO_CMD_GET_PREFERRED_INPUT_TIMING_PART2 0x1c -/** Returns a struct intel_sdvo_pixel_clock_range */ +/* Returns a struct intel_sdvo_pixel_clock_range */ #define SDVO_CMD_GET_INPUT_PIXEL_CLOCK_RANGE 0x1d -/** Returns a struct intel_sdvo_pixel_clock_range */ +/* Returns a struct intel_sdvo_pixel_clock_range */ #define SDVO_CMD_GET_OUTPUT_PIXEL_CLOCK_RANGE 0x1e -/** Returns a byte bitfield containing SDVO_CLOCK_RATE_MULT_* flags */ +/* Returns a byte bitfield containing SDVO_CLOCK_RATE_MULT_* flags */ #define SDVO_CMD_GET_SUPPORTED_CLOCK_RATE_MULTS 0x1f -/** Returns a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ +/* Returns a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ #define SDVO_CMD_GET_CLOCK_RATE_MULT 0x20 -/** Takes a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ +/* Takes a byte containing a SDVO_CLOCK_RATE_MULT_* flag */ #define SDVO_CMD_SET_CLOCK_RATE_MULT 0x21 # define SDVO_CLOCK_RATE_MULT_1X (1 << 0) # define SDVO_CLOCK_RATE_MULT_2X (1 << 1) # define SDVO_CLOCK_RATE_MULT_4X (1 << 3) #define SDVO_CMD_GET_SUPPORTED_TV_FORMATS 0x27 -/** 6 bytes of bit flags for TV formats shared by all TV format functions */ +/* 6 bytes of bit flags for TV formats shared by all TV format functions */ struct intel_sdvo_tv_format { unsigned int ntsc_m:1; unsigned int ntsc_j:1; @@ -376,7 +376,7 @@ struct intel_sdvo_tv_format { #define SDVO_CMD_SET_TV_FORMAT 0x29 -/** Returns the resolutiosn that can be used with the given TV format */ +/* Returns the resolutiosn that can be used with the given TV format */ #define SDVO_CMD_GET_SDTV_RESOLUTION_SUPPORT 0x83 struct intel_sdvo_sdtv_resolution_request { unsigned int ntsc_m:1; @@ -539,7 +539,7 @@ struct intel_sdvo_hdtv_resolution_reply { #define SDVO_CMD_GET_MAX_PANEL_POWER_SEQUENCING 0x2d #define SDVO_CMD_GET_PANEL_POWER_SEQUENCING 0x2e #define SDVO_CMD_SET_PANEL_POWER_SEQUENCING 0x2f -/** +/* * The panel power sequencing parameters are in units of milliseconds. * The high fields are bits 8:9 of the 10-bit values. */ diff --git a/drivers/gpu/drm/i915/intel_sprite.c b/drivers/gpu/drm/i915/intel_sprite.c index 4ff7a1f4183e..0875c8e0ec0a 100644 --- a/drivers/gpu/drm/i915/intel_sprite.c +++ b/drivers/gpu/drm/i915/intel_sprite.c @@ -178,28 +178,33 @@ void intel_pipe_update_end(struct intel_crtc *crtc) } static void -skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t x, uint32_t y, - uint32_t src_w, uint32_t src_h) +skl_update_plane(struct drm_plane *drm_plane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { struct drm_device *dev = drm_plane->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_plane *intel_plane = to_intel_plane(drm_plane); + struct drm_framebuffer *fb = plane_state->base.fb; struct drm_i915_gem_object *obj = intel_fb_obj(fb); const int pipe = intel_plane->pipe; const int plane = intel_plane->plane + 1; u32 plane_ctl, stride_div, stride; - const struct drm_intel_sprite_colorkey *key = - &to_intel_plane_state(drm_plane->state)->ckey; + const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; u32 surf_addr; u32 tile_height, plane_offset, plane_size; unsigned int rotation; int x_offset, y_offset; - struct intel_crtc_state *crtc_state = to_intel_crtc(crtc)->config; - int scaler_id; + int crtc_x = plane_state->dst.x1; + int crtc_y = plane_state->dst.y1; + uint32_t crtc_w = drm_rect_width(&plane_state->dst); + uint32_t crtc_h = drm_rect_height(&plane_state->dst); + uint32_t x = plane_state->src.x1 >> 16; + uint32_t y = plane_state->src.y1 >> 16; + uint32_t src_w = drm_rect_width(&plane_state->src) >> 16; + uint32_t src_h = drm_rect_height(&plane_state->src) >> 16; + const struct intel_scaler *scaler = + &crtc_state->scaler_state.scalers[plane_state->scaler_id]; plane_ctl = PLANE_CTL_ENABLE | PLANE_CTL_PIPE_GAMMA_ENABLE | @@ -208,14 +213,12 @@ skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc, plane_ctl |= skl_plane_ctl_format(fb->pixel_format); plane_ctl |= skl_plane_ctl_tiling(fb->modifier[0]); - rotation = drm_plane->state->rotation; + rotation = plane_state->base.rotation; plane_ctl |= skl_plane_ctl_rotation(rotation); - stride_div = intel_fb_stride_alignment(dev, fb->modifier[0], + stride_div = intel_fb_stride_alignment(dev_priv, fb->modifier[0], fb->pixel_format); - scaler_id = to_intel_plane_state(drm_plane->state)->scaler_id; - /* Sizes are 0 based */ src_w--; src_h--; @@ -236,9 +239,10 @@ skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc, surf_addr = intel_plane_obj_offset(intel_plane, obj, 0); if (intel_rotation_90_or_270(rotation)) { + int cpp = drm_format_plane_cpp(fb->pixel_format, 0); + /* stride: Surface height in tiles */ - tile_height = intel_tile_height(dev, fb->pixel_format, - fb->modifier[0], 0); + tile_height = intel_tile_height(dev_priv, fb->modifier[0], cpp); stride = DIV_ROUND_UP(fb->height, tile_height); plane_size = (src_w << 16) | src_h; x_offset = stride * tile_height - y - (src_h + 1); @@ -256,13 +260,13 @@ skl_update_plane(struct drm_plane *drm_plane, struct drm_crtc *crtc, I915_WRITE(PLANE_SIZE(pipe, plane), plane_size); /* program plane scaler */ - if (scaler_id >= 0) { + if (plane_state->scaler_id >= 0) { uint32_t ps_ctrl = 0; + int scaler_id = plane_state->scaler_id; DRM_DEBUG_KMS("plane = %d PS_PLANE_SEL(plane) = 0x%x\n", plane, PS_PLANE_SEL(plane)); - ps_ctrl = PS_SCALER_EN | PS_PLANE_SEL(plane) | - crtc_state->scaler_state.scalers[scaler_id].mode; + ps_ctrl = PS_SCALER_EN | PS_PLANE_SEL(plane) | scaler->mode; I915_WRITE(SKL_PS_CTRL(pipe, scaler_id), ps_ctrl); I915_WRITE(SKL_PS_PWR_GATE(pipe, scaler_id), 0); I915_WRITE(SKL_PS_WIN_POS(pipe, scaler_id), (crtc_x << 16) | crtc_y); @@ -334,24 +338,29 @@ chv_update_csc(struct intel_plane *intel_plane, uint32_t format) } static void -vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t x, uint32_t y, - uint32_t src_w, uint32_t src_h) +vlv_update_plane(struct drm_plane *dplane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { struct drm_device *dev = dplane->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_plane *intel_plane = to_intel_plane(dplane); + struct drm_framebuffer *fb = plane_state->base.fb; struct drm_i915_gem_object *obj = intel_fb_obj(fb); int pipe = intel_plane->pipe; int plane = intel_plane->plane; u32 sprctl; unsigned long sprsurf_offset, linear_offset; int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); - const struct drm_intel_sprite_colorkey *key = - &to_intel_plane_state(dplane->state)->ckey; + const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; + int crtc_x = plane_state->dst.x1; + int crtc_y = plane_state->dst.y1; + uint32_t crtc_w = drm_rect_width(&plane_state->dst); + uint32_t crtc_h = drm_rect_height(&plane_state->dst); + uint32_t x = plane_state->src.x1 >> 16; + uint32_t y = plane_state->src.y1 >> 16; + uint32_t src_w = drm_rect_width(&plane_state->src) >> 16; + uint32_t src_h = drm_rect_height(&plane_state->src) >> 16; sprctl = SP_ENABLE; @@ -414,14 +423,13 @@ vlv_update_plane(struct drm_plane *dplane, struct drm_crtc *crtc, crtc_h--; linear_offset = y * fb->pitches[0] + x * pixel_size; - sprsurf_offset = intel_gen4_compute_page_offset(dev_priv, - &x, &y, - obj->tiling_mode, - pixel_size, - fb->pitches[0]); + sprsurf_offset = intel_compute_tile_offset(dev_priv, &x, &y, + fb->modifier[0], + pixel_size, + fb->pitches[0]); linear_offset -= sprsurf_offset; - if (dplane->state->rotation == BIT(DRM_ROTATE_180)) { + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) { sprctl |= SP_ROTATE_180; x += src_w; @@ -474,23 +482,28 @@ vlv_disable_plane(struct drm_plane *dplane, struct drm_crtc *crtc) } static void -ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t x, uint32_t y, - uint32_t src_w, uint32_t src_h) +ivb_update_plane(struct drm_plane *plane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { struct drm_device *dev = plane->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_plane *intel_plane = to_intel_plane(plane); + struct drm_framebuffer *fb = plane_state->base.fb; struct drm_i915_gem_object *obj = intel_fb_obj(fb); enum pipe pipe = intel_plane->pipe; u32 sprctl, sprscale = 0; unsigned long sprsurf_offset, linear_offset; int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); - const struct drm_intel_sprite_colorkey *key = - &to_intel_plane_state(plane->state)->ckey; + const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; + int crtc_x = plane_state->dst.x1; + int crtc_y = plane_state->dst.y1; + uint32_t crtc_w = drm_rect_width(&plane_state->dst); + uint32_t crtc_h = drm_rect_height(&plane_state->dst); + uint32_t x = plane_state->src.x1 >> 16; + uint32_t y = plane_state->src.y1 >> 16; + uint32_t src_w = drm_rect_width(&plane_state->src) >> 16; + uint32_t src_h = drm_rect_height(&plane_state->src) >> 16; sprctl = SPRITE_ENABLE; @@ -544,13 +557,13 @@ ivb_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, sprscale = SPRITE_SCALE_ENABLE | (src_w << 16) | src_h; linear_offset = y * fb->pitches[0] + x * pixel_size; - sprsurf_offset = - intel_gen4_compute_page_offset(dev_priv, - &x, &y, obj->tiling_mode, - pixel_size, fb->pitches[0]); + sprsurf_offset = intel_compute_tile_offset(dev_priv, &x, &y, + fb->modifier[0], + pixel_size, + fb->pitches[0]); linear_offset -= sprsurf_offset; - if (plane->state->rotation == BIT(DRM_ROTATE_180)) { + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) { sprctl |= SPRITE_ROTATE_180; /* HSW and BDW does this automagically in hardware */ @@ -612,23 +625,28 @@ ivb_disable_plane(struct drm_plane *plane, struct drm_crtc *crtc) } static void -ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, - struct drm_framebuffer *fb, - int crtc_x, int crtc_y, - unsigned int crtc_w, unsigned int crtc_h, - uint32_t x, uint32_t y, - uint32_t src_w, uint32_t src_h) +ilk_update_plane(struct drm_plane *plane, + const struct intel_crtc_state *crtc_state, + const struct intel_plane_state *plane_state) { struct drm_device *dev = plane->dev; struct drm_i915_private *dev_priv = dev->dev_private; struct intel_plane *intel_plane = to_intel_plane(plane); + struct drm_framebuffer *fb = plane_state->base.fb; struct drm_i915_gem_object *obj = intel_fb_obj(fb); int pipe = intel_plane->pipe; unsigned long dvssurf_offset, linear_offset; u32 dvscntr, dvsscale; int pixel_size = drm_format_plane_cpp(fb->pixel_format, 0); - const struct drm_intel_sprite_colorkey *key = - &to_intel_plane_state(plane->state)->ckey; + const struct drm_intel_sprite_colorkey *key = &plane_state->ckey; + int crtc_x = plane_state->dst.x1; + int crtc_y = plane_state->dst.y1; + uint32_t crtc_w = drm_rect_width(&plane_state->dst); + uint32_t crtc_h = drm_rect_height(&plane_state->dst); + uint32_t x = plane_state->src.x1 >> 16; + uint32_t y = plane_state->src.y1 >> 16; + uint32_t src_w = drm_rect_width(&plane_state->src) >> 16; + uint32_t src_h = drm_rect_height(&plane_state->src) >> 16; dvscntr = DVS_ENABLE; @@ -678,13 +696,13 @@ ilk_update_plane(struct drm_plane *plane, struct drm_crtc *crtc, dvsscale = DVS_SCALE_ENABLE | (src_w << 16) | src_h; linear_offset = y * fb->pitches[0] + x * pixel_size; - dvssurf_offset = - intel_gen4_compute_page_offset(dev_priv, - &x, &y, obj->tiling_mode, - pixel_size, fb->pitches[0]); + dvssurf_offset = intel_compute_tile_offset(dev_priv, &x, &y, + fb->modifier[0], + pixel_size, + fb->pitches[0]); linear_offset -= dvssurf_offset; - if (plane->state->rotation == BIT(DRM_ROTATE_180)) { + if (plane_state->base.rotation == BIT(DRM_ROTATE_180)) { dvscntr |= DVS_ROTATE_180; x += src_w; @@ -913,30 +931,6 @@ intel_check_sprite_plane(struct drm_plane *plane, return 0; } -static void -intel_commit_sprite_plane(struct drm_plane *plane, - struct intel_plane_state *state) -{ - struct drm_crtc *crtc = state->base.crtc; - struct intel_plane *intel_plane = to_intel_plane(plane); - struct drm_framebuffer *fb = state->base.fb; - - crtc = crtc ? crtc : plane->crtc; - - if (state->visible) { - intel_plane->update_plane(plane, crtc, fb, - state->dst.x1, state->dst.y1, - drm_rect_width(&state->dst), - drm_rect_height(&state->dst), - state->src.x1 >> 16, - state->src.y1 >> 16, - drm_rect_width(&state->src) >> 16, - drm_rect_height(&state->src) >> 16); - } else { - intel_plane->disable_plane(plane, crtc); - } -} - int intel_sprite_set_colorkey(struct drm_device *dev, void *data, struct drm_file *file_priv) { @@ -1118,7 +1112,6 @@ intel_plane_init(struct drm_device *dev, enum pipe pipe, int plane) intel_plane->plane = plane; intel_plane->frontbuffer_bit = INTEL_FRONTBUFFER_SPRITE(pipe, plane); intel_plane->check_plane = intel_check_sprite_plane; - intel_plane->commit_plane = intel_commit_sprite_plane; possible_crtcs = (1 << pipe); ret = drm_universal_plane_init(dev, &intel_plane->base, possible_crtcs, &intel_plane_funcs, diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c index 277e60ae0e47..bfa79e5c214e 100644 --- a/drivers/gpu/drm/i915/intel_uncore.c +++ b/drivers/gpu/drm/i915/intel_uncore.c @@ -327,13 +327,54 @@ static void intel_uncore_ellc_detect(struct drm_device *dev) } } +static bool +fpga_check_for_unclaimed_mmio(struct drm_i915_private *dev_priv) +{ + u32 dbg; + + dbg = __raw_i915_read32(dev_priv, FPGA_DBG); + if (likely(!(dbg & FPGA_DBG_RM_NOCLAIM))) + return false; + + __raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM); + + return true; +} + +static bool +vlv_check_for_unclaimed_mmio(struct drm_i915_private *dev_priv) +{ + u32 cer; + + cer = __raw_i915_read32(dev_priv, CLAIM_ER); + if (likely(!(cer & (CLAIM_ER_OVERFLOW | CLAIM_ER_CTR_MASK)))) + return false; + + __raw_i915_write32(dev_priv, CLAIM_ER, CLAIM_ER_CLR); + + return true; +} + +static bool +check_for_unclaimed_mmio(struct drm_i915_private *dev_priv) +{ + if (HAS_FPGA_DBG_UNCLAIMED(dev_priv)) + return fpga_check_for_unclaimed_mmio(dev_priv); + + if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv)) + return vlv_check_for_unclaimed_mmio(dev_priv); + + return false; +} + static void __intel_uncore_early_sanitize(struct drm_device *dev, bool restore_forcewake) { struct drm_i915_private *dev_priv = dev->dev_private; - if (HAS_FPGA_DBG_UNCLAIMED(dev)) - __raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM); + /* clear out unclaimed reg detection bit */ + if (check_for_unclaimed_mmio(dev_priv)) + DRM_DEBUG("unclaimed mmio detected on uncore init, clearing\n"); /* clear out old GT FIFO errors */ if (IS_GEN6(dev) || IS_GEN7(dev)) @@ -585,38 +626,38 @@ ilk_dummy_write(struct drm_i915_private *dev_priv) } static void -hsw_unclaimed_reg_debug(struct drm_i915_private *dev_priv, - i915_reg_t reg, bool read, bool before) +__unclaimed_reg_debug(struct drm_i915_private *dev_priv, + const i915_reg_t reg, + const bool read, + const bool before) { - const char *op = read ? "reading" : "writing to"; - const char *when = before ? "before" : "after"; - - if (!i915.mmio_debug) + /* XXX. We limit the auto arming traces for mmio + * debugs on these platforms. There are just too many + * revealed by these and CI/Bat suffers from the noise. + * Please fix and then re-enable the automatic traces. + */ + if (i915.mmio_debug < 2 && + (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))) return; - if (__raw_i915_read32(dev_priv, FPGA_DBG) & FPGA_DBG_RM_NOCLAIM) { - WARN(1, "Unclaimed register detected %s %s register 0x%x\n", - when, op, i915_mmio_reg_offset(reg)); - __raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM); + if (WARN(check_for_unclaimed_mmio(dev_priv), + "Unclaimed register detected %s %s register 0x%x\n", + before ? "before" : "after", + read ? "reading" : "writing to", + i915_mmio_reg_offset(reg))) i915.mmio_debug--; /* Only report the first N failures */ - } } -static void -hsw_unclaimed_reg_detect(struct drm_i915_private *dev_priv) +static inline void +unclaimed_reg_debug(struct drm_i915_private *dev_priv, + const i915_reg_t reg, + const bool read, + const bool before) { - static bool mmio_debug_once = true; - - if (i915.mmio_debug || !mmio_debug_once) + if (likely(!i915.mmio_debug)) return; - if (__raw_i915_read32(dev_priv, FPGA_DBG) & FPGA_DBG_RM_NOCLAIM) { - DRM_DEBUG("Unclaimed register detected, " - "enabling oneshot unclaimed register reporting. " - "Please use i915.mmio_debug=N for more information.\n"); - __raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM); - i915.mmio_debug = mmio_debug_once--; - } + __unclaimed_reg_debug(dev_priv, reg, read, before); } #define GEN2_READ_HEADER(x) \ @@ -664,9 +705,11 @@ __gen2_read(64) unsigned long irqflags; \ u##x val = 0; \ assert_rpm_wakelock_held(dev_priv); \ - spin_lock_irqsave(&dev_priv->uncore.lock, irqflags) + spin_lock_irqsave(&dev_priv->uncore.lock, irqflags); \ + unclaimed_reg_debug(dev_priv, reg, true, true) #define GEN6_READ_FOOTER \ + unclaimed_reg_debug(dev_priv, reg, true, false); \ spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags); \ trace_i915_reg_rw(false, reg, val, sizeof(val), trace); \ return val @@ -699,11 +742,9 @@ static inline void __force_wake_get(struct drm_i915_private *dev_priv, static u##x \ gen6_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \ GEN6_READ_HEADER(x); \ - hsw_unclaimed_reg_debug(dev_priv, reg, true, true); \ if (NEEDS_FORCE_WAKE(offset)) \ __force_wake_get(dev_priv, FORCEWAKE_RENDER); \ val = __raw_i915_read##x(dev_priv, reg); \ - hsw_unclaimed_reg_debug(dev_priv, reg, true, false); \ GEN6_READ_FOOTER; \ } @@ -751,7 +792,6 @@ static u##x \ gen9_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \ enum forcewake_domains fw_engine; \ GEN6_READ_HEADER(x); \ - hsw_unclaimed_reg_debug(dev_priv, reg, true, true); \ if (!SKL_NEEDS_FORCE_WAKE(offset)) \ fw_engine = 0; \ else if (FORCEWAKE_GEN9_RENDER_RANGE_OFFSET(offset)) \ @@ -765,7 +805,6 @@ gen9_read##x(struct drm_i915_private *dev_priv, i915_reg_t reg, bool trace) { \ if (fw_engine) \ __force_wake_get(dev_priv, fw_engine); \ val = __raw_i915_read##x(dev_priv, reg); \ - hsw_unclaimed_reg_debug(dev_priv, reg, true, false); \ GEN6_READ_FOOTER; \ } @@ -864,9 +903,11 @@ __gen2_write(64) unsigned long irqflags; \ trace_i915_reg_rw(true, reg, val, sizeof(val), trace); \ assert_rpm_wakelock_held(dev_priv); \ - spin_lock_irqsave(&dev_priv->uncore.lock, irqflags) + spin_lock_irqsave(&dev_priv->uncore.lock, irqflags); \ + unclaimed_reg_debug(dev_priv, reg, false, true) #define GEN6_WRITE_FOOTER \ + unclaimed_reg_debug(dev_priv, reg, false, false); \ spin_unlock_irqrestore(&dev_priv->uncore.lock, irqflags) #define __gen6_write(x) \ @@ -892,13 +933,10 @@ hsw_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool t if (NEEDS_FORCE_WAKE(offset)) { \ __fifo_ret = __gen6_gt_wait_for_fifo(dev_priv); \ } \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, true); \ __raw_i915_write##x(dev_priv, reg, val); \ if (unlikely(__fifo_ret)) { \ gen6_gt_check_fifodbg(dev_priv); \ } \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, false); \ - hsw_unclaimed_reg_detect(dev_priv); \ GEN6_WRITE_FOOTER; \ } @@ -928,12 +966,9 @@ static bool is_gen8_shadowed(struct drm_i915_private *dev_priv, static void \ gen8_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, bool trace) { \ GEN6_WRITE_HEADER; \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, true); \ if (NEEDS_FORCE_WAKE(offset) && !is_gen8_shadowed(dev_priv, reg)) \ __force_wake_get(dev_priv, FORCEWAKE_RENDER); \ __raw_i915_write##x(dev_priv, reg, val); \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, false); \ - hsw_unclaimed_reg_detect(dev_priv); \ GEN6_WRITE_FOOTER; \ } @@ -987,7 +1022,6 @@ gen9_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, \ bool trace) { \ enum forcewake_domains fw_engine; \ GEN6_WRITE_HEADER; \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, true); \ if (!SKL_NEEDS_FORCE_WAKE(offset) || \ is_gen9_shadowed(dev_priv, reg)) \ fw_engine = 0; \ @@ -1002,8 +1036,6 @@ gen9_write##x(struct drm_i915_private *dev_priv, i915_reg_t reg, u##x val, \ if (fw_engine) \ __force_wake_get(dev_priv, fw_engine); \ __raw_i915_write##x(dev_priv, reg, val); \ - hsw_unclaimed_reg_debug(dev_priv, reg, false, false); \ - hsw_unclaimed_reg_detect(dev_priv); \ GEN6_WRITE_FOOTER; \ } @@ -1223,6 +1255,8 @@ void intel_uncore_init(struct drm_device *dev) intel_uncore_fw_domains_init(dev); __intel_uncore_early_sanitize(dev, false); + dev_priv->uncore.unclaimed_mmio_check = 1; + switch (INTEL_INFO(dev)->gen) { default: case 9: @@ -1580,13 +1614,26 @@ bool intel_has_gpu_reset(struct drm_device *dev) return intel_get_gpu_reset(dev) != NULL; } -void intel_uncore_check_errors(struct drm_device *dev) +bool intel_uncore_unclaimed_mmio(struct drm_i915_private *dev_priv) { - struct drm_i915_private *dev_priv = dev->dev_private; + return check_for_unclaimed_mmio(dev_priv); +} + +bool +intel_uncore_arm_unclaimed_mmio_detection(struct drm_i915_private *dev_priv) +{ + if (unlikely(i915.mmio_debug || + dev_priv->uncore.unclaimed_mmio_check <= 0)) + return false; - if (HAS_FPGA_DBG_UNCLAIMED(dev) && - (__raw_i915_read32(dev_priv, FPGA_DBG) & FPGA_DBG_RM_NOCLAIM)) { - DRM_ERROR("Unclaimed register before interrupt\n"); - __raw_i915_write32(dev_priv, FPGA_DBG, FPGA_DBG_RM_NOCLAIM); + if (unlikely(intel_uncore_unclaimed_mmio(dev_priv))) { + DRM_DEBUG("Unclaimed register detected, " + "enabling oneshot unclaimed register reporting. " + "Please use i915.mmio_debug=N for more information.\n"); + i915.mmio_debug++; + dev_priv->uncore.unclaimed_mmio_check--; + return true; } + + return false; } diff --git a/drivers/gpu/drm/imx/dw_hdmi-imx.c b/drivers/gpu/drm/imx/dw_hdmi-imx.c index 063825fecbe2..21d615827837 100644 --- a/drivers/gpu/drm/imx/dw_hdmi-imx.c +++ b/drivers/gpu/drm/imx/dw_hdmi-imx.c @@ -109,13 +109,6 @@ static void dw_hdmi_imx_encoder_disable(struct drm_encoder *encoder) { } -static bool dw_hdmi_imx_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adj_mode) -{ - return true; -} - static void dw_hdmi_imx_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adj_mode) @@ -138,7 +131,6 @@ static void dw_hdmi_imx_encoder_prepare(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs dw_hdmi_imx_encoder_helper_funcs = { - .mode_fixup = dw_hdmi_imx_encoder_mode_fixup, .mode_set = dw_hdmi_imx_encoder_mode_set, .prepare = dw_hdmi_imx_encoder_prepare, .commit = dw_hdmi_imx_encoder_commit, diff --git a/drivers/gpu/drm/imx/imx-drm-core.c b/drivers/gpu/drm/imx/imx-drm-core.c index 2f57d7967417..7c4d1250e071 100644 --- a/drivers/gpu/drm/imx/imx-drm-core.c +++ b/drivers/gpu/drm/imx/imx-drm-core.c @@ -171,18 +171,6 @@ static void imx_drm_disable_vblank(struct drm_device *drm, unsigned int pipe) imx_drm_crtc->imx_drm_helper_funcs.disable_vblank(imx_drm_crtc->crtc); } -static void imx_drm_driver_preclose(struct drm_device *drm, - struct drm_file *file) -{ - int i; - - if (!file->is_master) - return; - - for (i = 0; i < MAX_CRTC; i++) - imx_drm_disable_vblank(drm, i); -} - static const struct file_operations imx_drm_driver_fops = { .owner = THIS_MODULE, .open = drm_open, @@ -463,7 +451,6 @@ static struct drm_driver imx_drm_driver = { .load = imx_drm_driver_load, .unload = imx_drm_driver_unload, .lastclose = imx_drm_driver_lastclose, - .preclose = imx_drm_driver_preclose, .set_busid = drm_platform_set_busid, .gem_free_object = drm_gem_cma_free_object, .gem_vm_ops = &drm_gem_cma_vm_ops, diff --git a/drivers/gpu/drm/imx/imx-ldb.c b/drivers/gpu/drm/imx/imx-ldb.c index 22ac482231ed..024d6134b848 100644 --- a/drivers/gpu/drm/imx/imx-ldb.c +++ b/drivers/gpu/drm/imx/imx-ldb.c @@ -139,13 +139,6 @@ static void imx_ldb_encoder_dpms(struct drm_encoder *encoder, int mode) { } -static bool imx_ldb_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void imx_ldb_set_clock(struct imx_ldb *ldb, int mux, int chno, unsigned long serial_clk, unsigned long di_clk) { @@ -376,7 +369,6 @@ static const struct drm_encoder_funcs imx_ldb_encoder_funcs = { static const struct drm_encoder_helper_funcs imx_ldb_encoder_helper_funcs = { .dpms = imx_ldb_encoder_dpms, - .mode_fixup = imx_ldb_encoder_mode_fixup, .prepare = imx_ldb_encoder_prepare, .commit = imx_ldb_encoder_commit, .mode_set = imx_ldb_encoder_mode_set, diff --git a/drivers/gpu/drm/imx/imx-tve.c b/drivers/gpu/drm/imx/imx-tve.c index 292349f0b132..ae7a9fb3b8a2 100644 --- a/drivers/gpu/drm/imx/imx-tve.c +++ b/drivers/gpu/drm/imx/imx-tve.c @@ -286,13 +286,6 @@ static void imx_tve_encoder_dpms(struct drm_encoder *encoder, int mode) dev_err(tve->dev, "failed to disable TVOUT: %d\n", ret); } -static bool imx_tve_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void imx_tve_encoder_prepare(struct drm_encoder *encoder) { struct imx_tve *tve = enc_to_tve(encoder); @@ -379,7 +372,6 @@ static const struct drm_encoder_funcs imx_tve_encoder_funcs = { static const struct drm_encoder_helper_funcs imx_tve_encoder_helper_funcs = { .dpms = imx_tve_encoder_dpms, - .mode_fixup = imx_tve_encoder_mode_fixup, .prepare = imx_tve_encoder_prepare, .mode_set = imx_tve_encoder_mode_set, .commit = imx_tve_encoder_commit, diff --git a/drivers/gpu/drm/imx/ipuv3-crtc.c b/drivers/gpu/drm/imx/ipuv3-crtc.c index 30a57185bdb4..846b5f558897 100644 --- a/drivers/gpu/drm/imx/ipuv3-crtc.c +++ b/drivers/gpu/drm/imx/ipuv3-crtc.c @@ -285,10 +285,6 @@ static int ipu_enable_vblank(struct drm_crtc *crtc) static void ipu_disable_vblank(struct drm_crtc *crtc) { - struct ipu_crtc *ipu_crtc = to_ipu_crtc(crtc); - - ipu_crtc->page_flip_event = NULL; - ipu_crtc->newfb = NULL; } static int ipu_set_interface_pix_fmt(struct drm_crtc *crtc, diff --git a/drivers/gpu/drm/imx/parallel-display.c b/drivers/gpu/drm/imx/parallel-display.c index 0ffef172afb4..363e2c7741e2 100644 --- a/drivers/gpu/drm/imx/parallel-display.c +++ b/drivers/gpu/drm/imx/parallel-display.c @@ -112,13 +112,6 @@ static void imx_pd_encoder_dpms(struct drm_encoder *encoder, int mode) drm_panel_enable(imxpd->panel); } -static bool imx_pd_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void imx_pd_encoder_prepare(struct drm_encoder *encoder) { struct imx_parallel_display *imxpd = enc_to_imxpd(encoder); @@ -166,7 +159,6 @@ static const struct drm_encoder_funcs imx_pd_encoder_funcs = { static const struct drm_encoder_helper_funcs imx_pd_encoder_helper_funcs = { .dpms = imx_pd_encoder_dpms, - .mode_fixup = imx_pd_encoder_mode_fixup, .prepare = imx_pd_encoder_prepare, .commit = imx_pd_encoder_commit, .mode_set = imx_pd_encoder_mode_set, diff --git a/drivers/gpu/drm/mgag200/mgag200_mode.c b/drivers/gpu/drm/mgag200/mgag200_mode.c index dc13c4857e6f..af8b4c19cf15 100644 --- a/drivers/gpu/drm/mgag200/mgag200_mode.c +++ b/drivers/gpu/drm/mgag200/mgag200_mode.c @@ -1479,13 +1479,6 @@ void mga_crtc_fb_gamma_get(struct drm_crtc *crtc, u16 *red, u16 *green, * These functions are analagous to those in the CRTC code, but are intended * to handle any encoder-specific limitations */ -static bool mga_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mga_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -1515,7 +1508,6 @@ static void mga_encoder_destroy(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs mga_encoder_helper_funcs = { .dpms = mga_encoder_dpms, - .mode_fixup = mga_encoder_mode_fixup, .mode_set = mga_encoder_mode_set, .prepare = mga_encoder_prepare, .commit = mga_encoder_commit, diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c index a3b54cc76495..0e1d0c57cd3d 100644 --- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c +++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c @@ -41,6 +41,9 @@ int adreno_get_param(struct msm_gpu *gpu, uint32_t param, uint64_t *value) (adreno_gpu->rev.major << 16) | (adreno_gpu->rev.core << 24); return 0; + case MSM_PARAM_MAX_FREQ: + *value = adreno_gpu->base.fast_rate; + return 0; default: DBG("%s: invalid param: %u", gpu->name, param); return -EINVAL; diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c index 28df397c3b04..909d74250de7 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_crtc.c @@ -575,13 +575,6 @@ uint32_t mdp4_crtc_vblank(struct drm_crtc *crtc) return mdp4_crtc->vblank.irqmask; } -void mdp4_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file) -{ - struct mdp4_crtc *mdp4_crtc = to_mdp4_crtc(crtc); - DBG("%s: cancel: %p", mdp4_crtc->name, file); - complete_flip(crtc, file); -} - /* set dma config, ie. the format the encoder wants. */ void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config) { diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dsi_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dsi_encoder.c index 2f57e9453b67..106f0e772595 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dsi_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dsi_encoder.c @@ -47,13 +47,6 @@ static const struct drm_encoder_funcs mdp4_dsi_encoder_funcs = { .destroy = mdp4_dsi_encoder_destroy, }; -static bool mdp4_dsi_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mdp4_dsi_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -163,7 +156,6 @@ static void mdp4_dsi_encoder_enable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs mdp4_dsi_encoder_helper_funcs = { - .mode_fixup = mdp4_dsi_encoder_mode_fixup, .mode_set = mdp4_dsi_encoder_mode_set, .disable = mdp4_dsi_encoder_disable, .enable = mdp4_dsi_encoder_enable, diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c index a21df54cb50f..35ad78a1dc1c 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_dtv_encoder.c @@ -94,13 +94,6 @@ static const struct drm_encoder_funcs mdp4_dtv_encoder_funcs = { .destroy = mdp4_dtv_encoder_destroy, }; -static bool mdp4_dtv_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mdp4_dtv_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -234,7 +227,6 @@ static void mdp4_dtv_encoder_enable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs mdp4_dtv_encoder_helper_funcs = { - .mode_fixup = mdp4_dtv_encoder_mode_fixup, .mode_set = mdp4_dtv_encoder_mode_set, .enable = mdp4_dtv_encoder_enable, .disable = mdp4_dtv_encoder_disable, diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c index 5a8e3d6bcbff..1c8e330f8d98 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.c @@ -179,16 +179,6 @@ static long mdp4_round_pixclk(struct msm_kms *kms, unsigned long rate, } } -static void mdp4_preclose(struct msm_kms *kms, struct drm_file *file) -{ - struct mdp4_kms *mdp4_kms = to_mdp4_kms(to_mdp_kms(kms)); - struct msm_drm_private *priv = mdp4_kms->dev->dev_private; - unsigned i; - - for (i = 0; i < priv->num_crtcs; i++) - mdp4_crtc_cancel_pending_flip(priv->crtcs[i], file); -} - static void mdp4_destroy(struct msm_kms *kms) { struct mdp4_kms *mdp4_kms = to_mdp4_kms(to_mdp_kms(kms)); @@ -213,7 +203,6 @@ static const struct mdp_kms_funcs kms_funcs = { .wait_for_crtc_commit_done = mdp4_wait_for_crtc_commit_done, .get_format = mdp_get_format, .round_pixclk = mdp4_round_pixclk, - .preclose = mdp4_preclose, .destroy = mdp4_destroy, }, .set_irqmask = mdp4_set_irqmask, diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h index d2c96ef431f4..9ec53b464662 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_kms.h @@ -199,7 +199,6 @@ struct drm_plane *mdp4_plane_init(struct drm_device *dev, enum mdp4_pipe pipe_id, bool private_plane); uint32_t mdp4_crtc_vblank(struct drm_crtc *crtc); -void mdp4_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file); void mdp4_crtc_set_config(struct drm_crtc *crtc, uint32_t config); void mdp4_crtc_set_intf(struct drm_crtc *crtc, enum mdp4_intf intf, int mixer); void mdp4_crtc_wait_for_commit_done(struct drm_crtc *crtc); diff --git a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c index cd63fedb67cc..bc3d8e719c6c 100644 --- a/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp4/mdp4_lcdc_encoder.c @@ -260,13 +260,6 @@ static void setup_phy(struct drm_encoder *encoder) mdp4_write(mdp4_kms, REG_MDP4_LVDS_PHY_CFG0, lvds_phy_cfg0); } -static bool mdp4_lcdc_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mdp4_lcdc_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -430,7 +423,6 @@ static void mdp4_lcdc_encoder_enable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs mdp4_lcdc_encoder_helper_funcs = { - .mode_fixup = mdp4_lcdc_encoder_mode_fixup, .mode_set = mdp4_lcdc_encoder_mode_set, .disable = mdp4_lcdc_encoder_disable, .enable = mdp4_lcdc_encoder_enable, diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c index 1aa21dba663d..69094cb28103 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_cmd_encoder.c @@ -188,13 +188,6 @@ static const struct drm_encoder_funcs mdp5_cmd_encoder_funcs = { .destroy = mdp5_cmd_encoder_destroy, }; -static bool mdp5_cmd_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mdp5_cmd_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -256,7 +249,6 @@ static void mdp5_cmd_encoder_enable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs mdp5_cmd_encoder_helper_funcs = { - .mode_fixup = mdp5_cmd_encoder_mode_fixup, .mode_set = mdp5_cmd_encoder_mode_set, .disable = mdp5_cmd_encoder_disable, .enable = mdp5_cmd_encoder_enable, @@ -340,4 +332,3 @@ fail: return ERR_PTR(ret); } - diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c index 20cee5ce4071..46682aa8870c 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_crtc.c @@ -721,12 +721,6 @@ uint32_t mdp5_crtc_vblank(struct drm_crtc *crtc) return mdp5_crtc->vblank.irqmask; } -void mdp5_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file) -{ - DBG("cancel: %p", file); - complete_flip(crtc, file); -} - void mdp5_crtc_set_pipeline(struct drm_crtc *crtc, struct mdp5_interface *intf, struct mdp5_ctl *ctl) { diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c index 0d737cad03a6..1d95f9fd9dc7 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_encoder.c @@ -112,13 +112,6 @@ static const struct drm_encoder_funcs mdp5_encoder_funcs = { .destroy = mdp5_encoder_destroy, }; -static bool mdp5_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void mdp5_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -287,7 +280,6 @@ static void mdp5_encoder_enable(struct drm_encoder *encoder) } static const struct drm_encoder_helper_funcs mdp5_encoder_helper_funcs = { - .mode_fixup = mdp5_encoder_mode_fixup, .mode_set = mdp5_encoder_mode_set, .disable = mdp5_encoder_disable, .enable = mdp5_encoder_enable, diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c index e115318402bd..5e4d16b399c7 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.c @@ -117,16 +117,6 @@ static int mdp5_set_split_display(struct msm_kms *kms, return mdp5_encoder_set_split_display(encoder, slave_encoder); } -static void mdp5_preclose(struct msm_kms *kms, struct drm_file *file) -{ - struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); - struct msm_drm_private *priv = mdp5_kms->dev->dev_private; - unsigned i; - - for (i = 0; i < priv->num_crtcs; i++) - mdp5_crtc_cancel_pending_flip(priv->crtcs[i], file); -} - static void mdp5_destroy(struct msm_kms *kms) { struct mdp5_kms *mdp5_kms = to_mdp5_kms(to_mdp_kms(kms)); @@ -164,7 +154,6 @@ static const struct mdp_kms_funcs kms_funcs = { .get_format = mdp_get_format, .round_pixclk = mdp5_round_pixclk, .set_split_display = mdp5_set_split_display, - .preclose = mdp5_preclose, .destroy = mdp5_destroy, }, .set_irqmask = mdp5_set_irqmask, diff --git a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h index 00730ba08a60..9a25898239d3 100644 --- a/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h +++ b/drivers/gpu/drm/msm/mdp/mdp5/mdp5_kms.h @@ -211,7 +211,6 @@ struct drm_plane *mdp5_plane_init(struct drm_device *dev, uint32_t mdp5_crtc_vblank(struct drm_crtc *crtc); int mdp5_crtc_get_lm(struct drm_crtc *crtc); -void mdp5_crtc_cancel_pending_flip(struct drm_crtc *crtc, struct drm_file *file); void mdp5_crtc_set_pipeline(struct drm_crtc *crtc, struct mdp5_interface *intf, struct mdp5_ctl *ctl); void mdp5_crtc_wait_for_commit_done(struct drm_crtc *crtc); diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c index d95af6eba602..d9759bf3482e 100644 --- a/drivers/gpu/drm/msm/msm_fbdev.c +++ b/drivers/gpu/drm/msm/msm_fbdev.c @@ -62,12 +62,8 @@ static int msm_fbdev_mmap(struct fb_info *info, struct vm_area_struct *vma) struct drm_fb_helper *helper = (struct drm_fb_helper *)info->par; struct msm_fbdev *fbdev = to_msm_fbdev(helper); struct drm_gem_object *drm_obj = fbdev->bo; - struct drm_device *dev = helper->dev; int ret = 0; - if (drm_device_is_unplugged(dev)) - return -ENODEV; - ret = drm_gem_mmap_obj(drm_obj, drm_obj->size, vma); if (ret) { pr_err("%s:drm_gem_mmap_obj fail\n", __func__); diff --git a/drivers/gpu/drm/nouveau/nouveau_acpi.c b/drivers/gpu/drm/nouveau/nouveau_acpi.c index d5e6938cc6bc..cdf522770cfa 100644 --- a/drivers/gpu/drm/nouveau/nouveau_acpi.c +++ b/drivers/gpu/drm/nouveau/nouveau_acpi.c @@ -314,7 +314,7 @@ void nouveau_register_dsm_handler(void) if (!r) return; - vga_switcheroo_register_handler(&nouveau_dsm_handler); + vga_switcheroo_register_handler(&nouveau_dsm_handler, 0); } /* Must be called for Optimus models before the card can be turned off */ diff --git a/drivers/gpu/drm/nouveau/nouveau_connector.c b/drivers/gpu/drm/nouveau/nouveau_connector.c index fcebfae5d426..ae96ebc490fb 100644 --- a/drivers/gpu/drm/nouveau/nouveau_connector.c +++ b/drivers/gpu/drm/nouveau/nouveau_connector.c @@ -27,6 +27,7 @@ #include <acpi/button.h> #include <linux/pm_runtime.h> +#include <linux/vga_switcheroo.h> #include <drm/drmP.h> #include <drm/drm_edid.h> @@ -153,6 +154,17 @@ nouveau_connector_ddc_detect(struct drm_connector *connector) if (ret == 0) break; } else + if ((vga_switcheroo_handler_flags() & + VGA_SWITCHEROO_CAN_SWITCH_DDC) && + nv_encoder->dcb->type == DCB_OUTPUT_LVDS && + nv_encoder->i2c) { + int ret; + vga_switcheroo_lock_ddc(dev->pdev); + ret = nvkm_probe_i2c(nv_encoder->i2c, 0x50); + vga_switcheroo_unlock_ddc(dev->pdev); + if (ret) + break; + } else if (nv_encoder->i2c) { if (nvkm_probe_i2c(nv_encoder->i2c, 0x50)) break; @@ -265,7 +277,14 @@ nouveau_connector_detect(struct drm_connector *connector, bool force) nv_encoder = nouveau_connector_ddc_detect(connector); if (nv_encoder && (i2c = nv_encoder->i2c) != NULL) { - nv_connector->edid = drm_get_edid(connector, i2c); + if ((vga_switcheroo_handler_flags() & + VGA_SWITCHEROO_CAN_SWITCH_DDC) && + nv_connector->type == DCB_CONNECTOR_LVDS) + nv_connector->edid = drm_get_edid_switcheroo(connector, + i2c); + else + nv_connector->edid = drm_get_edid(connector, i2c); + drm_mode_connector_update_edid_property(connector, nv_connector->edid); if (!nv_connector->edid) { diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index 2f2f252e3fb6..bb8498c9b13e 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -22,11 +22,13 @@ * Authors: Ben Skeggs */ +#include <linux/apple-gmux.h> #include <linux/console.h> #include <linux/delay.h> #include <linux/module.h> #include <linux/pci.h> #include <linux/pm_runtime.h> +#include <linux/vgaarb.h> #include <linux/vga_switcheroo.h> #include "drmP.h" @@ -312,6 +314,15 @@ static int nouveau_drm_probe(struct pci_dev *pdev, bool boot = false; int ret; + /* + * apple-gmux is needed on dual GPU MacBook Pro + * to probe the panel if we're the inactive GPU. + */ + if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) && + apple_gmux_present() && pdev != vga_default_device() && + !vga_switcheroo_handler_flags()) + return -EPROBE_DEFER; + /* remove conflicting drivers (vesafb, efifb etc) */ aper = alloc_apertures(3); if (!aper) diff --git a/drivers/gpu/drm/omapdrm/omap_crtc.c b/drivers/gpu/drm/omapdrm/omap_crtc.c index 2ed0754ed19e..d38fcbcc43a8 100644 --- a/drivers/gpu/drm/omapdrm/omap_crtc.c +++ b/drivers/gpu/drm/omapdrm/omap_crtc.c @@ -269,18 +269,7 @@ static void omap_crtc_complete_page_flip(struct drm_crtc *crtc) return; spin_lock_irqsave(&dev->event_lock, flags); - - list_del(&event->base.link); - - /* - * Queue the event for delivery if it's still linked to a file - * handle, otherwise just destroy it. - */ - if (event->base.file_priv) - drm_crtc_send_vblank_event(crtc, event); - else - event->base.destroy(&event->base); - + drm_crtc_send_vblank_event(crtc, event); spin_unlock_irqrestore(&dev->event_lock, flags); } diff --git a/drivers/gpu/drm/omapdrm/omap_drv.c b/drivers/gpu/drm/omapdrm/omap_drv.c index dfafdb602ad2..33370f42e4d7 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.c +++ b/drivers/gpu/drm/omapdrm/omap_drv.c @@ -142,7 +142,6 @@ static int omap_atomic_commit(struct drm_device *dev, { struct omap_drm_private *priv = dev->dev_private; struct omap_atomic_state_commit *commit; - unsigned long flags; unsigned int i; int ret; @@ -175,17 +174,6 @@ static int omap_atomic_commit(struct drm_device *dev, priv->commit.pending |= commit->crtcs; spin_unlock(&priv->commit.lock); - /* Keep track of all CRTC events to unlink them in preclose(). */ - spin_lock_irqsave(&dev->event_lock, flags); - for (i = 0; i < dev->mode_config.num_crtc; ++i) { - struct drm_crtc_state *cstate = state->crtc_states[i]; - - if (cstate && cstate->event) - list_add_tail(&cstate->event->base.link, - &priv->commit.events); - } - spin_unlock_irqrestore(&dev->event_lock, flags); - /* Swap the state, this is the point of no return. */ drm_atomic_helper_swap_state(dev, state); @@ -673,7 +661,6 @@ static int dev_load(struct drm_device *dev, unsigned long flags) priv->wq = alloc_ordered_workqueue("omapdrm", 0); init_waitqueue_head(&priv->commit.wait); spin_lock_init(&priv->commit.lock); - INIT_LIST_HEAD(&priv->commit.events); spin_lock_init(&priv->list_lock); INIT_LIST_HEAD(&priv->obj_list); @@ -787,33 +774,6 @@ static void dev_lastclose(struct drm_device *dev) } } -static void dev_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct omap_drm_private *priv = dev->dev_private; - struct drm_pending_event *event; - unsigned long flags; - - DBG("preclose: dev=%p", dev); - - /* - * Unlink all pending CRTC events to make sure they won't be queued up - * by a pending asynchronous commit. - */ - spin_lock_irqsave(&dev->event_lock, flags); - list_for_each_entry(event, &priv->commit.events, link) { - if (event->file_priv == file) { - file->event_space += event->event->length; - event->file_priv = NULL; - } - } - spin_unlock_irqrestore(&dev->event_lock, flags); -} - -static void dev_postclose(struct drm_device *dev, struct drm_file *file) -{ - DBG("postclose: dev=%p, file=%p", dev, file); -} - static const struct vm_operations_struct omap_gem_vm_ops = { .fault = omap_gem_fault, .open = drm_gem_vm_open, @@ -838,8 +798,6 @@ static struct drm_driver omap_drm_driver = { .unload = dev_unload, .open = dev_open, .lastclose = dev_lastclose, - .preclose = dev_preclose, - .postclose = dev_postclose, .set_busid = drm_platform_set_busid, .get_vblank_counter = drm_vblank_no_hw_counter, .enable_vblank = omap_irq_enable_vblank, diff --git a/drivers/gpu/drm/omapdrm/omap_drv.h b/drivers/gpu/drm/omapdrm/omap_drv.h index 9e0030731c37..c23cbe6fe9e4 100644 --- a/drivers/gpu/drm/omapdrm/omap_drv.h +++ b/drivers/gpu/drm/omapdrm/omap_drv.h @@ -106,7 +106,6 @@ struct omap_drm_private { /* atomic commit */ struct { - struct list_head events; wait_queue_head_t wait; u32 pending; spinlock_t lock; /* Protects commit.pending */ diff --git a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c index 27c297672076..aebae1c2dab2 100644 --- a/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c +++ b/drivers/gpu/drm/omapdrm/omap_gem_dmabuf.c @@ -79,7 +79,7 @@ static void omap_gem_dmabuf_release(struct dma_buf *buffer) static int omap_gem_dmabuf_begin_cpu_access(struct dma_buf *buffer, - size_t start, size_t len, enum dma_data_direction dir) + enum dma_data_direction dir) { struct drm_gem_object *obj = buffer->priv; struct page **pages; @@ -94,7 +94,7 @@ static int omap_gem_dmabuf_begin_cpu_access(struct dma_buf *buffer, } static void omap_gem_dmabuf_end_cpu_access(struct dma_buf *buffer, - size_t start, size_t len, enum dma_data_direction dir) + enum dma_data_direction dir) { struct drm_gem_object *obj = buffer->priv; omap_gem_put_pages(obj); diff --git a/drivers/gpu/drm/qxl/qxl_display.c b/drivers/gpu/drm/qxl/qxl_display.c index 86276519b2ef..43e5f503d1c5 100644 --- a/drivers/gpu/drm/qxl/qxl_display.c +++ b/drivers/gpu/drm/qxl/qxl_display.c @@ -734,14 +734,6 @@ static void qxl_enc_dpms(struct drm_encoder *encoder, int mode) DRM_DEBUG("\n"); } -static bool qxl_enc_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - DRM_DEBUG("\n"); - return true; -} - static void qxl_enc_prepare(struct drm_encoder *encoder) { DRM_DEBUG("\n"); @@ -864,7 +856,6 @@ static struct drm_encoder *qxl_best_encoder(struct drm_connector *connector) static const struct drm_encoder_helper_funcs qxl_enc_helper_funcs = { .dpms = qxl_enc_dpms, - .mode_fixup = qxl_enc_mode_fixup, .prepare = qxl_enc_prepare, .mode_set = qxl_enc_mode_set, .commit = qxl_enc_commit, diff --git a/drivers/gpu/drm/radeon/atombios_encoders.c b/drivers/gpu/drm/radeon/atombios_encoders.c index 01b20e14a247..1603751b1164 100644 --- a/drivers/gpu/drm/radeon/atombios_encoders.c +++ b/drivers/gpu/drm/radeon/atombios_encoders.c @@ -2623,16 +2623,8 @@ radeon_atom_ext_dpms(struct drm_encoder *encoder, int mode) } -static bool radeon_atom_ext_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static const struct drm_encoder_helper_funcs radeon_atom_ext_helper_funcs = { .dpms = radeon_atom_ext_dpms, - .mode_fixup = radeon_atom_ext_mode_fixup, .prepare = radeon_atom_ext_prepare, .mode_set = radeon_atom_ext_mode_set, .commit = radeon_atom_ext_commit, diff --git a/drivers/gpu/drm/radeon/cik.c b/drivers/gpu/drm/radeon/cik.c index 4c30d8c65558..06001400ce8b 100644 --- a/drivers/gpu/drm/radeon/cik.c +++ b/drivers/gpu/drm/radeon/cik.c @@ -4219,13 +4219,20 @@ int cik_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); return r; } - r = radeon_fence_wait(ib.fence, false); - if (r) { + r = radeon_fence_wait_timeout(ib.fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); radeon_scratch_free(rdev, scratch); radeon_ib_free(rdev, &ib); return r; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + radeon_scratch_free(rdev, scratch); + radeon_ib_free(rdev, &ib); + return -ETIMEDOUT; } + r = 0; for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) diff --git a/drivers/gpu/drm/radeon/cik_sdma.c b/drivers/gpu/drm/radeon/cik_sdma.c index d16f2eebd95e..9c351dc8a9e0 100644 --- a/drivers/gpu/drm/radeon/cik_sdma.c +++ b/drivers/gpu/drm/radeon/cik_sdma.c @@ -737,11 +737,16 @@ int cik_sdma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); return r; } - r = radeon_fence_wait(ib.fence, false); - if (r) { + r = radeon_fence_wait_timeout(ib.fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); return r; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + return -ETIMEDOUT; } + r = 0; for (i = 0; i < rdev->usec_timeout; i++) { tmp = le32_to_cpu(rdev->wb.wb[index/4]); if (tmp == 0xDEADBEEF) diff --git a/drivers/gpu/drm/radeon/r100.c b/drivers/gpu/drm/radeon/r100.c index 5eae0a88dd3e..6e478a248628 100644 --- a/drivers/gpu/drm/radeon/r100.c +++ b/drivers/gpu/drm/radeon/r100.c @@ -3732,11 +3732,17 @@ int r100_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); goto free_ib; } - r = radeon_fence_wait(ib.fence, false); - if (r) { + r = radeon_fence_wait_timeout(ib.fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); goto free_ib; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + r = -ETIMEDOUT; + goto free_ib; } + r = 0; for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) { diff --git a/drivers/gpu/drm/radeon/r600.c b/drivers/gpu/drm/radeon/r600.c index cc2fdf0be37a..ed121042247f 100644 --- a/drivers/gpu/drm/radeon/r600.c +++ b/drivers/gpu/drm/radeon/r600.c @@ -3381,11 +3381,17 @@ int r600_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); goto free_ib; } - r = radeon_fence_wait(ib.fence, false); - if (r) { + r = radeon_fence_wait_timeout(ib.fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); goto free_ib; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + r = -ETIMEDOUT; + goto free_ib; } + r = 0; for (i = 0; i < rdev->usec_timeout; i++) { tmp = RREG32(scratch); if (tmp == 0xDEADBEEF) diff --git a/drivers/gpu/drm/radeon/r600_dma.c b/drivers/gpu/drm/radeon/r600_dma.c index d2dd29ab24fa..fb65e6fb5c4f 100644 --- a/drivers/gpu/drm/radeon/r600_dma.c +++ b/drivers/gpu/drm/radeon/r600_dma.c @@ -368,11 +368,16 @@ int r600_dma_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) DRM_ERROR("radeon: failed to schedule ib (%d).\n", r); return r; } - r = radeon_fence_wait(ib.fence, false); - if (r) { + r = radeon_fence_wait_timeout(ib.fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); return r; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + return -ETIMEDOUT; } + r = 0; for (i = 0; i < rdev->usec_timeout; i++) { tmp = le32_to_cpu(rdev->wb.wb[index/4]); if (tmp == 0xDEADBEEF) diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h index 78a51b3eda10..007be29a0020 100644 --- a/drivers/gpu/drm/radeon/radeon.h +++ b/drivers/gpu/drm/radeon/radeon.h @@ -120,6 +120,7 @@ extern int radeon_mst; */ #define RADEON_MAX_USEC_TIMEOUT 100000 /* 100 ms */ #define RADEON_FENCE_JIFFIES_TIMEOUT (HZ / 2) +#define RADEON_USEC_IB_TEST_TIMEOUT 1000000 /* 1s */ /* RADEON_IB_POOL_SIZE must be a power of 2 */ #define RADEON_IB_POOL_SIZE 16 #define RADEON_DEBUGFS_MAX_COMPONENTS 32 @@ -382,6 +383,7 @@ void radeon_fence_driver_force_completion(struct radeon_device *rdev, int ring); int radeon_fence_emit(struct radeon_device *rdev, struct radeon_fence **fence, int ring); void radeon_fence_process(struct radeon_device *rdev, int ring); bool radeon_fence_signaled(struct radeon_fence *fence); +long radeon_fence_wait_timeout(struct radeon_fence *fence, bool interruptible, long timeout); int radeon_fence_wait(struct radeon_fence *fence, bool interruptible); int radeon_fence_wait_next(struct radeon_device *rdev, int ring); int radeon_fence_wait_empty(struct radeon_device *rdev, int ring); diff --git a/drivers/gpu/drm/radeon/radeon_atpx_handler.c b/drivers/gpu/drm/radeon/radeon_atpx_handler.c index c4b4f298a283..56482e35d43e 100644 --- a/drivers/gpu/drm/radeon/radeon_atpx_handler.c +++ b/drivers/gpu/drm/radeon/radeon_atpx_handler.c @@ -551,13 +551,14 @@ static bool radeon_atpx_detect(void) void radeon_register_atpx_handler(void) { bool r; + enum vga_switcheroo_handler_flags_t handler_flags = 0; /* detect if we have any ATPX + 2 VGA in the system */ r = radeon_atpx_detect(); if (!r) return; - vga_switcheroo_register_handler(&radeon_atpx_handler); + vga_switcheroo_register_handler(&radeon_atpx_handler, handler_flags); } /** diff --git a/drivers/gpu/drm/radeon/radeon_connectors.c b/drivers/gpu/drm/radeon/radeon_connectors.c index 340f3f549f29..cfcc099c537d 100644 --- a/drivers/gpu/drm/radeon/radeon_connectors.c +++ b/drivers/gpu/drm/radeon/radeon_connectors.c @@ -34,6 +34,7 @@ #include "atom.h" #include <linux/pm_runtime.h> +#include <linux/vga_switcheroo.h> static int radeon_dp_handle_hpd(struct drm_connector *connector) { @@ -344,6 +345,11 @@ static void radeon_connector_get_edid(struct drm_connector *connector) else if (radeon_connector->ddc_bus) radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); + } else if (vga_switcheroo_handler_flags() & VGA_SWITCHEROO_CAN_SWITCH_DDC && + connector->connector_type == DRM_MODE_CONNECTOR_LVDS && + radeon_connector->ddc_bus) { + radeon_connector->edid = drm_get_edid_switcheroo(&radeon_connector->base, + &radeon_connector->ddc_bus->adapter); } else if (radeon_connector->ddc_bus) { radeon_connector->edid = drm_get_edid(&radeon_connector->base, &radeon_connector->ddc_bus->adapter); diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c index 298ea1c453c3..a4674bfd979a 100644 --- a/drivers/gpu/drm/radeon/radeon_display.c +++ b/drivers/gpu/drm/radeon/radeon_display.c @@ -1686,6 +1686,9 @@ void radeon_modeset_fini(struct radeon_device *rdev) radeon_fbdev_fini(rdev); kfree(rdev->mode_info.bios_hardcoded_edid); + /* free i2c buses */ + radeon_i2c_fini(rdev); + if (rdev->mode_info.mode_config_initialized) { radeon_afmt_fini(rdev); drm_kms_helper_poll_fini(rdev->ddev); @@ -1693,8 +1696,6 @@ void radeon_modeset_fini(struct radeon_device *rdev) drm_mode_config_cleanup(rdev->ddev); rdev->mode_info.mode_config_initialized = false; } - /* free i2c buses */ - radeon_i2c_fini(rdev); } static bool is_hdtv_mode(const struct drm_display_mode *mode) diff --git a/drivers/gpu/drm/radeon/radeon_drv.c b/drivers/gpu/drm/radeon/radeon_drv.c index e266ffc520d2..cad25557650f 100644 --- a/drivers/gpu/drm/radeon/radeon_drv.c +++ b/drivers/gpu/drm/radeon/radeon_drv.c @@ -34,9 +34,11 @@ #include "radeon_drv.h" #include <drm/drm_pciids.h> +#include <linux/apple-gmux.h> #include <linux/console.h> #include <linux/module.h> #include <linux/pm_runtime.h> +#include <linux/vgaarb.h> #include <linux/vga_switcheroo.h> #include <drm/drm_gem.h> @@ -319,6 +321,15 @@ static int radeon_pci_probe(struct pci_dev *pdev, { int ret; + /* + * apple-gmux is needed on dual GPU MacBook Pro + * to probe the panel if we're the inactive GPU. + */ + if (IS_ENABLED(CONFIG_VGA_ARB) && IS_ENABLED(CONFIG_VGA_SWITCHEROO) && + apple_gmux_present() && pdev != vga_default_device() && + !vga_switcheroo_handler_flags()) + return -EPROBE_DEFER; + /* Get rid of things like offb */ ret = radeon_kick_out_firmware_fb(pdev); if (ret) diff --git a/drivers/gpu/drm/radeon/radeon_fence.c b/drivers/gpu/drm/radeon/radeon_fence.c index 05815c47b246..7ef075acde9c 100644 --- a/drivers/gpu/drm/radeon/radeon_fence.c +++ b/drivers/gpu/drm/radeon/radeon_fence.c @@ -527,7 +527,7 @@ static long radeon_fence_wait_seq_timeout(struct radeon_device *rdev, } /** - * radeon_fence_wait - wait for a fence to signal + * radeon_fence_wait_timeout - wait for a fence to signal with timeout * * @fence: radeon fence object * @intr: use interruptible sleep @@ -535,12 +535,15 @@ static long radeon_fence_wait_seq_timeout(struct radeon_device *rdev, * Wait for the requested fence to signal (all asics). * @intr selects whether to use interruptable (true) or non-interruptable * (false) sleep when waiting for the fence. - * Returns 0 if the fence has passed, error for all other cases. + * @timeout: maximum time to wait, or MAX_SCHEDULE_TIMEOUT for infinite wait + * Returns remaining time if the sequence number has passed, 0 when + * the wait timeout, or an error for all other cases. */ -int radeon_fence_wait(struct radeon_fence *fence, bool intr) +long radeon_fence_wait_timeout(struct radeon_fence *fence, bool intr, long timeout) { uint64_t seq[RADEON_NUM_RINGS] = {}; long r; + int r_sig; /* * This function should not be called on !radeon fences. @@ -552,15 +555,36 @@ int radeon_fence_wait(struct radeon_fence *fence, bool intr) return fence_wait(&fence->base, intr); seq[fence->ring] = fence->seq; - r = radeon_fence_wait_seq_timeout(fence->rdev, seq, intr, MAX_SCHEDULE_TIMEOUT); - if (r < 0) { + r = radeon_fence_wait_seq_timeout(fence->rdev, seq, intr, timeout); + if (r <= 0) { return r; } - r = fence_signal(&fence->base); - if (!r) + r_sig = fence_signal(&fence->base); + if (!r_sig) FENCE_TRACE(&fence->base, "signaled from fence_wait\n"); - return 0; + return r; +} + +/** + * radeon_fence_wait - wait for a fence to signal + * + * @fence: radeon fence object + * @intr: use interruptible sleep + * + * Wait for the requested fence to signal (all asics). + * @intr selects whether to use interruptable (true) or non-interruptable + * (false) sleep when waiting for the fence. + * Returns 0 if the fence has passed, error for all other cases. + */ +int radeon_fence_wait(struct radeon_fence *fence, bool intr) +{ + long r = radeon_fence_wait_timeout(fence, intr, MAX_SCHEDULE_TIMEOUT); + if (r > 0) { + return 0; + } else { + return r; + } } /** diff --git a/drivers/gpu/drm/radeon/radeon_vce.c b/drivers/gpu/drm/radeon/radeon_vce.c index 7eb1ae758906..566a1a01f6d1 100644 --- a/drivers/gpu/drm/radeon/radeon_vce.c +++ b/drivers/gpu/drm/radeon/radeon_vce.c @@ -810,11 +810,16 @@ int radeon_vce_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) goto error; } - r = radeon_fence_wait(fence, false); - if (r) { + r = radeon_fence_wait_timeout(fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + r = -ETIMEDOUT; } else { - DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + DRM_INFO("ib test on ring %d succeeded\n", ring->idx); + r = 0; } error: radeon_fence_unref(&fence); diff --git a/drivers/gpu/drm/radeon/uvd_v1_0.c b/drivers/gpu/drm/radeon/uvd_v1_0.c index c6b1cbca47fc..12ddcfa82e20 100644 --- a/drivers/gpu/drm/radeon/uvd_v1_0.c +++ b/drivers/gpu/drm/radeon/uvd_v1_0.c @@ -522,11 +522,17 @@ int uvd_v1_0_ib_test(struct radeon_device *rdev, struct radeon_ring *ring) goto error; } - r = radeon_fence_wait(fence, false); - if (r) { + r = radeon_fence_wait_timeout(fence, false, usecs_to_jiffies( + RADEON_USEC_IB_TEST_TIMEOUT)); + if (r < 0) { DRM_ERROR("radeon: fence wait failed (%d).\n", r); goto error; + } else if (r == 0) { + DRM_ERROR("radeon: fence wait timed out.\n"); + r = -ETIMEDOUT; + goto error; } + r = 0; DRM_INFO("ib test on ring %d succeeded\n", ring->idx); error: radeon_fence_unref(&fence); diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c index 88a4b706be16..4ec80ae1fa99 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.c @@ -282,26 +282,6 @@ static void rcar_du_crtc_update_planes(struct rcar_du_crtc *rcrtc) * Page Flip */ -void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc, - struct drm_file *file) -{ - struct drm_pending_vblank_event *event; - struct drm_device *dev = rcrtc->crtc.dev; - unsigned long flags; - - /* Destroy the pending vertical blanking event associated with the - * pending page flip, if any, and disable vertical blanking interrupts. - */ - spin_lock_irqsave(&dev->event_lock, flags); - event = rcrtc->event; - if (event && event->base.file_priv == file) { - rcrtc->event = NULL; - event->base.destroy(&event->base); - drm_crtc_vblank_put(&rcrtc->crtc); - } - spin_unlock_irqrestore(&dev->event_lock, flags); -} - static void rcar_du_crtc_finish_page_flip(struct rcar_du_crtc *rcrtc) { struct drm_pending_vblank_event *event; diff --git a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h index 4b95d9d08c49..2bbe3f5aab65 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_crtc.h +++ b/drivers/gpu/drm/rcar-du/rcar_du_crtc.h @@ -67,8 +67,6 @@ enum rcar_du_output { int rcar_du_crtc_create(struct rcar_du_group *rgrp, unsigned int index); void rcar_du_crtc_enable_vblank(struct rcar_du_crtc *rcrtc, bool enable); -void rcar_du_crtc_cancel_page_flip(struct rcar_du_crtc *rcrtc, - struct drm_file *file); void rcar_du_crtc_suspend(struct rcar_du_crtc *rcrtc); void rcar_du_crtc_resume(struct rcar_du_crtc *rcrtc); diff --git a/drivers/gpu/drm/rcar-du/rcar_du_drv.c b/drivers/gpu/drm/rcar-du/rcar_du_drv.c index 40422f6b645e..0bb2b31555bf 100644 --- a/drivers/gpu/drm/rcar-du/rcar_du_drv.c +++ b/drivers/gpu/drm/rcar-du/rcar_du_drv.c @@ -220,15 +220,6 @@ done: return ret; } -static void rcar_du_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct rcar_du_device *rcdu = dev->dev_private; - unsigned int i; - - for (i = 0; i < rcdu->num_crtcs; ++i) - rcar_du_crtc_cancel_page_flip(&rcdu->crtcs[i], file); -} - static void rcar_du_lastclose(struct drm_device *dev) { struct rcar_du_device *rcdu = dev->dev_private; @@ -271,7 +262,6 @@ static struct drm_driver rcar_du_driver = { | DRIVER_ATOMIC, .load = rcar_du_load, .unload = rcar_du_unload, - .preclose = rcar_du_preclose, .lastclose = rcar_du_lastclose, .set_busid = drm_platform_set_busid, .get_vblank_counter = drm_vblank_no_hw_counter, diff --git a/drivers/gpu/drm/rockchip/Kconfig b/drivers/gpu/drm/rockchip/Kconfig index 85739859dffc..76b3362c5e59 100644 --- a/drivers/gpu/drm/rockchip/Kconfig +++ b/drivers/gpu/drm/rockchip/Kconfig @@ -35,3 +35,11 @@ config ROCKCHIP_DW_MIPI_DSI for the Synopsys DesignWare HDMI driver. If you want to enable MIPI DSI on RK3288 based SoC, you should selet this option. + +config ROCKCHIP_INNO_HDMI + tristate "Rockchip specific extensions for Innosilicon HDMI" + depends on DRM_ROCKCHIP + help + This selects support for Rockchip SoC specific extensions + for the Innosilicon HDMI driver. If you want to enable + HDMI on RK3036 based SoC, you should select this option. diff --git a/drivers/gpu/drm/rockchip/Makefile b/drivers/gpu/drm/rockchip/Makefile index f6a809afceec..df8fbef17791 100644 --- a/drivers/gpu/drm/rockchip/Makefile +++ b/drivers/gpu/drm/rockchip/Makefile @@ -8,5 +8,6 @@ rockchipdrm-$(CONFIG_DRM_FBDEV_EMULATION) += rockchip_drm_fbdev.o obj-$(CONFIG_ROCKCHIP_DW_HDMI) += dw_hdmi-rockchip.o obj-$(CONFIG_ROCKCHIP_DW_MIPI_DSI) += dw-mipi-dsi.o +obj-$(CONFIG_ROCKCHIP_INNO_HDMI) += inno_hdmi.o obj-$(CONFIG_DRM_ROCKCHIP) += rockchipdrm.o rockchip_vop_reg.o diff --git a/drivers/gpu/drm/rockchip/dw-mipi-dsi.c b/drivers/gpu/drm/rockchip/dw-mipi-dsi.c index f8f8f29fb7c3..77662068bc11 100644 --- a/drivers/gpu/drm/rockchip/dw-mipi-dsi.c +++ b/drivers/gpu/drm/rockchip/dw-mipi-dsi.c @@ -875,13 +875,6 @@ static void dw_mipi_dsi_encoder_disable(struct drm_encoder *encoder) clk_disable_unprepare(dsi->pclk); } -static bool dw_mipi_dsi_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void dw_mipi_dsi_encoder_commit(struct drm_encoder *encoder) { struct dw_mipi_dsi *dsi = encoder_to_dsi(encoder); @@ -931,7 +924,6 @@ static void dw_mipi_dsi_encoder_commit(struct drm_encoder *encoder) static struct drm_encoder_helper_funcs dw_mipi_dsi_encoder_helper_funcs = { - .mode_fixup = dw_mipi_dsi_encoder_mode_fixup, .commit = dw_mipi_dsi_encoder_commit, .mode_set = dw_mipi_dsi_encoder_mode_set, .disable = dw_mipi_dsi_encoder_disable, diff --git a/drivers/gpu/drm/rockchip/inno_hdmi.c b/drivers/gpu/drm/rockchip/inno_hdmi.c new file mode 100644 index 000000000000..10d62fff22f1 --- /dev/null +++ b/drivers/gpu/drm/rockchip/inno_hdmi.c @@ -0,0 +1,938 @@ +/* + * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd + * Zheng Yang <[email protected]> + * Yakir Yang <[email protected]> + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#include <linux/irq.h> +#include <linux/clk.h> +#include <linux/delay.h> +#include <linux/err.h> +#include <linux/hdmi.h> +#include <linux/mfd/syscon.h> +#include <linux/module.h> +#include <linux/mutex.h> +#include <linux/of_device.h> + +#include <drm/drm_of.h> +#include <drm/drmP.h> +#include <drm/drm_atomic_helper.h> +#include <drm/drm_crtc_helper.h> +#include <drm/drm_edid.h> + +#include "rockchip_drm_drv.h" +#include "rockchip_drm_vop.h" + +#include "inno_hdmi.h" + +#define to_inno_hdmi(x) container_of(x, struct inno_hdmi, x) + +struct hdmi_data_info { + int vic; + bool sink_is_hdmi; + bool sink_has_audio; + unsigned int enc_in_format; + unsigned int enc_out_format; + unsigned int colorimetry; +}; + +struct inno_hdmi_i2c { + struct i2c_adapter adap; + + u8 ddc_addr; + u8 segment_addr; + + struct mutex lock; + struct completion cmp; +}; + +struct inno_hdmi { + struct device *dev; + struct drm_device *drm_dev; + + int irq; + struct clk *pclk; + void __iomem *regs; + + struct drm_connector connector; + struct drm_encoder encoder; + + struct inno_hdmi_i2c *i2c; + struct i2c_adapter *ddc; + + unsigned int tmds_rate; + + struct hdmi_data_info hdmi_data; + struct drm_display_mode previous_mode; +}; + +enum { + CSC_ITU601_16_235_TO_RGB_0_255_8BIT, + CSC_ITU601_0_255_TO_RGB_0_255_8BIT, + CSC_ITU709_16_235_TO_RGB_0_255_8BIT, + CSC_RGB_0_255_TO_ITU601_16_235_8BIT, + CSC_RGB_0_255_TO_ITU709_16_235_8BIT, + CSC_RGB_0_255_TO_RGB_16_235_8BIT, +}; + +static const char coeff_csc[][24] = { + /* + * YUV2RGB:601 SD mode(Y[16:235], UV[16:240], RGB[0:255]): + * R = 1.164*Y + 1.596*V - 204 + * G = 1.164*Y - 0.391*U - 0.813*V + 154 + * B = 1.164*Y + 2.018*U - 258 + */ + { + 0x04, 0xa7, 0x00, 0x00, 0x06, 0x62, 0x02, 0xcc, + 0x04, 0xa7, 0x11, 0x90, 0x13, 0x40, 0x00, 0x9a, + 0x04, 0xa7, 0x08, 0x12, 0x00, 0x00, 0x03, 0x02 + }, + /* + * YUV2RGB:601 SD mode(YUV[0:255],RGB[0:255]): + * R = Y + 1.402*V - 248 + * G = Y - 0.344*U - 0.714*V + 135 + * B = Y + 1.772*U - 227 + */ + { + 0x04, 0x00, 0x00, 0x00, 0x05, 0x9b, 0x02, 0xf8, + 0x04, 0x00, 0x11, 0x60, 0x12, 0xdb, 0x00, 0x87, + 0x04, 0x00, 0x07, 0x16, 0x00, 0x00, 0x02, 0xe3 + }, + /* + * YUV2RGB:709 HD mode(Y[16:235],UV[16:240],RGB[0:255]): + * R = 1.164*Y + 1.793*V - 248 + * G = 1.164*Y - 0.213*U - 0.534*V + 77 + * B = 1.164*Y + 2.115*U - 289 + */ + { + 0x04, 0xa7, 0x00, 0x00, 0x07, 0x2c, 0x02, 0xf8, + 0x04, 0xa7, 0x10, 0xda, 0x12, 0x22, 0x00, 0x4d, + 0x04, 0xa7, 0x08, 0x74, 0x00, 0x00, 0x03, 0x21 + }, + + /* + * RGB2YUV:601 SD mode: + * Cb = -0.291G - 0.148R + 0.439B + 128 + * Y = 0.504G + 0.257R + 0.098B + 16 + * Cr = -0.368G + 0.439R - 0.071B + 128 + */ + { + 0x11, 0x5f, 0x01, 0x82, 0x10, 0x23, 0x00, 0x80, + 0x02, 0x1c, 0x00, 0xa1, 0x00, 0x36, 0x00, 0x1e, + 0x11, 0x29, 0x10, 0x59, 0x01, 0x82, 0x00, 0x80 + }, + /* + * RGB2YUV:709 HD mode: + * Cb = - 0.338G - 0.101R + 0.439B + 128 + * Y = 0.614G + 0.183R + 0.062B + 16 + * Cr = - 0.399G + 0.439R - 0.040B + 128 + */ + { + 0x11, 0x98, 0x01, 0xc1, 0x10, 0x28, 0x00, 0x80, + 0x02, 0x74, 0x00, 0xbb, 0x00, 0x3f, 0x00, 0x10, + 0x11, 0x5a, 0x10, 0x67, 0x01, 0xc1, 0x00, 0x80 + }, + /* + * RGB[0:255]2RGB[16:235]: + * R' = R x (235-16)/255 + 16; + * G' = G x (235-16)/255 + 16; + * B' = B x (235-16)/255 + 16; + */ + { + 0x00, 0x00, 0x03, 0x6F, 0x00, 0x00, 0x00, 0x10, + 0x03, 0x6F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x10, + 0x00, 0x00, 0x00, 0x00, 0x03, 0x6F, 0x00, 0x10 + }, +}; + +static inline u8 hdmi_readb(struct inno_hdmi *hdmi, u16 offset) +{ + return readl_relaxed(hdmi->regs + (offset) * 0x04); +} + +static inline void hdmi_writeb(struct inno_hdmi *hdmi, u16 offset, u32 val) +{ + writel_relaxed(val, hdmi->regs + (offset) * 0x04); +} + +static inline void hdmi_modb(struct inno_hdmi *hdmi, u16 offset, + u32 msk, u32 val) +{ + u8 temp = hdmi_readb(hdmi, offset) & ~msk; + + temp |= val & msk; + hdmi_writeb(hdmi, offset, temp); +} + +static void inno_hdmi_i2c_init(struct inno_hdmi *hdmi) +{ + int ddc_bus_freq; + + ddc_bus_freq = (hdmi->tmds_rate >> 2) / HDMI_SCL_RATE; + + hdmi_writeb(hdmi, DDC_BUS_FREQ_L, ddc_bus_freq & 0xFF); + hdmi_writeb(hdmi, DDC_BUS_FREQ_H, (ddc_bus_freq >> 8) & 0xFF); + + /* Clear the EDID interrupt flag and mute the interrupt */ + hdmi_writeb(hdmi, HDMI_INTERRUPT_MASK1, 0); + hdmi_writeb(hdmi, HDMI_INTERRUPT_STATUS1, m_INT_EDID_READY); +} + +static void inno_hdmi_sys_power(struct inno_hdmi *hdmi, bool enable) +{ + if (enable) + hdmi_modb(hdmi, HDMI_SYS_CTRL, m_POWER, v_PWR_ON); + else + hdmi_modb(hdmi, HDMI_SYS_CTRL, m_POWER, v_PWR_OFF); +} + +static void inno_hdmi_set_pwr_mode(struct inno_hdmi *hdmi, int mode) +{ + switch (mode) { + case NORMAL: + inno_hdmi_sys_power(hdmi, false); + + hdmi_writeb(hdmi, HDMI_PHY_PRE_EMPHASIS, 0x6f); + hdmi_writeb(hdmi, HDMI_PHY_DRIVER, 0xbb); + + hdmi_writeb(hdmi, HDMI_PHY_SYS_CTL, 0x15); + hdmi_writeb(hdmi, HDMI_PHY_SYS_CTL, 0x14); + hdmi_writeb(hdmi, HDMI_PHY_SYS_CTL, 0x10); + hdmi_writeb(hdmi, HDMI_PHY_CHG_PWR, 0x0f); + hdmi_writeb(hdmi, HDMI_PHY_SYNC, 0x00); + hdmi_writeb(hdmi, HDMI_PHY_SYNC, 0x01); + + inno_hdmi_sys_power(hdmi, true); + break; + + case LOWER_PWR: + inno_hdmi_sys_power(hdmi, false); + hdmi_writeb(hdmi, HDMI_PHY_DRIVER, 0x00); + hdmi_writeb(hdmi, HDMI_PHY_PRE_EMPHASIS, 0x00); + hdmi_writeb(hdmi, HDMI_PHY_CHG_PWR, 0x00); + hdmi_writeb(hdmi, HDMI_PHY_SYS_CTL, 0x15); + + break; + + default: + dev_err(hdmi->dev, "Unknown power mode %d\n", mode); + } +} + +static void inno_hdmi_reset(struct inno_hdmi *hdmi) +{ + u32 val; + u32 msk; + + hdmi_modb(hdmi, HDMI_SYS_CTRL, m_RST_DIGITAL, v_NOT_RST_DIGITAL); + udelay(100); + + hdmi_modb(hdmi, HDMI_SYS_CTRL, m_RST_ANALOG, v_NOT_RST_ANALOG); + udelay(100); + + msk = m_REG_CLK_INV | m_REG_CLK_SOURCE | m_POWER | m_INT_POL; + val = v_REG_CLK_INV | v_REG_CLK_SOURCE_SYS | v_PWR_ON | v_INT_POL_HIGH; + hdmi_modb(hdmi, HDMI_SYS_CTRL, msk, val); + + inno_hdmi_set_pwr_mode(hdmi, NORMAL); +} + +static int inno_hdmi_upload_frame(struct inno_hdmi *hdmi, int setup_rc, + union hdmi_infoframe *frame, u32 frame_index, + u32 mask, u32 disable, u32 enable) +{ + if (mask) + hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, mask, disable); + + hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_BUF_INDEX, frame_index); + + if (setup_rc >= 0) { + u8 packed_frame[HDMI_MAXIMUM_INFO_FRAME_SIZE]; + ssize_t rc, i; + + rc = hdmi_infoframe_pack(frame, packed_frame, + sizeof(packed_frame)); + if (rc < 0) + return rc; + + for (i = 0; i < rc; i++) + hdmi_writeb(hdmi, HDMI_CONTROL_PACKET_ADDR + i, + packed_frame[i]); + + if (mask) + hdmi_modb(hdmi, HDMI_PACKET_SEND_AUTO, mask, enable); + } + + return setup_rc; +} + +static int inno_hdmi_config_video_vsi(struct inno_hdmi *hdmi, + struct drm_display_mode *mode) +{ + union hdmi_infoframe frame; + int rc; + + rc = drm_hdmi_vendor_infoframe_from_display_mode(&frame.vendor.hdmi, + mode); + + return inno_hdmi_upload_frame(hdmi, rc, &frame, INFOFRAME_VSI, + m_PACKET_VSI_EN, v_PACKET_VSI_EN(0), v_PACKET_VSI_EN(1)); +} + +static int inno_hdmi_config_video_avi(struct inno_hdmi *hdmi, + struct drm_display_mode *mode) +{ + union hdmi_infoframe frame; + int rc; + + rc = drm_hdmi_avi_infoframe_from_display_mode(&frame.avi, mode); + + if (hdmi->hdmi_data.enc_out_format == HDMI_COLORSPACE_YUV444) + frame.avi.colorspace = HDMI_COLORSPACE_YUV444; + else if (hdmi->hdmi_data.enc_out_format == HDMI_COLORSPACE_YUV422) + frame.avi.colorspace = HDMI_COLORSPACE_YUV422; + else + frame.avi.colorspace = HDMI_COLORSPACE_RGB; + + return inno_hdmi_upload_frame(hdmi, rc, &frame, INFOFRAME_AVI, 0, 0, 0); +} + +static int inno_hdmi_config_video_csc(struct inno_hdmi *hdmi) +{ + struct hdmi_data_info *data = &hdmi->hdmi_data; + int c0_c2_change = 0; + int csc_enable = 0; + int csc_mode = 0; + int auto_csc = 0; + int value; + int i; + + /* Input video mode is SDR RGB24bit, data enable signal from external */ + hdmi_writeb(hdmi, HDMI_VIDEO_CONTRL1, v_DE_EXTERNAL | + v_VIDEO_INPUT_FORMAT(VIDEO_INPUT_SDR_RGB444)); + + /* Input color hardcode to RGB, and output color hardcode to RGB888 */ + value = v_VIDEO_INPUT_BITS(VIDEO_INPUT_8BITS) | + v_VIDEO_OUTPUT_COLOR(0) | + v_VIDEO_INPUT_CSP(0); + hdmi_writeb(hdmi, HDMI_VIDEO_CONTRL2, value); + + if (data->enc_in_format == data->enc_out_format) { + if ((data->enc_in_format == HDMI_COLORSPACE_RGB) || + (data->enc_in_format >= HDMI_COLORSPACE_YUV444)) { + value = v_SOF_DISABLE | v_COLOR_DEPTH_NOT_INDICATED(1); + hdmi_writeb(hdmi, HDMI_VIDEO_CONTRL3, value); + + hdmi_modb(hdmi, HDMI_VIDEO_CONTRL, + m_VIDEO_AUTO_CSC | m_VIDEO_C0_C2_SWAP, + v_VIDEO_AUTO_CSC(AUTO_CSC_DISABLE) | + v_VIDEO_C0_C2_SWAP(C0_C2_CHANGE_DISABLE)); + return 0; + } + } + + if (data->colorimetry == HDMI_COLORIMETRY_ITU_601) { + if ((data->enc_in_format == HDMI_COLORSPACE_RGB) && + (data->enc_out_format == HDMI_COLORSPACE_YUV444)) { + csc_mode = CSC_RGB_0_255_TO_ITU601_16_235_8BIT; + auto_csc = AUTO_CSC_DISABLE; + c0_c2_change = C0_C2_CHANGE_DISABLE; + csc_enable = v_CSC_ENABLE; + } else if ((data->enc_in_format == HDMI_COLORSPACE_YUV444) && + (data->enc_out_format == HDMI_COLORSPACE_RGB)) { + csc_mode = CSC_ITU601_16_235_TO_RGB_0_255_8BIT; + auto_csc = AUTO_CSC_ENABLE; + c0_c2_change = C0_C2_CHANGE_DISABLE; + csc_enable = v_CSC_DISABLE; + } + } else { + if ((data->enc_in_format == HDMI_COLORSPACE_RGB) && + (data->enc_out_format == HDMI_COLORSPACE_YUV444)) { + csc_mode = CSC_RGB_0_255_TO_ITU709_16_235_8BIT; + auto_csc = AUTO_CSC_DISABLE; + c0_c2_change = C0_C2_CHANGE_DISABLE; + csc_enable = v_CSC_ENABLE; + } else if ((data->enc_in_format == HDMI_COLORSPACE_YUV444) && + (data->enc_out_format == HDMI_COLORSPACE_RGB)) { + csc_mode = CSC_ITU709_16_235_TO_RGB_0_255_8BIT; + auto_csc = AUTO_CSC_ENABLE; + c0_c2_change = C0_C2_CHANGE_DISABLE; + csc_enable = v_CSC_DISABLE; + } + } + + for (i = 0; i < 24; i++) + hdmi_writeb(hdmi, HDMI_VIDEO_CSC_COEF + i, + coeff_csc[csc_mode][i]); + + value = v_SOF_DISABLE | csc_enable | v_COLOR_DEPTH_NOT_INDICATED(1); + hdmi_writeb(hdmi, HDMI_VIDEO_CONTRL3, value); + hdmi_modb(hdmi, HDMI_VIDEO_CONTRL, m_VIDEO_AUTO_CSC | + m_VIDEO_C0_C2_SWAP, v_VIDEO_AUTO_CSC(auto_csc) | + v_VIDEO_C0_C2_SWAP(c0_c2_change)); + + return 0; +} + +static int inno_hdmi_config_video_timing(struct inno_hdmi *hdmi, + struct drm_display_mode *mode) +{ + int value; + + /* Set detail external video timing polarity and interlace mode */ + value = v_EXTERANL_VIDEO(1); + value |= mode->flags & DRM_MODE_FLAG_PHSYNC ? + v_HSYNC_POLARITY(1) : v_HSYNC_POLARITY(0); + value |= mode->flags & DRM_MODE_FLAG_PVSYNC ? + v_VSYNC_POLARITY(1) : v_VSYNC_POLARITY(0); + value |= mode->flags & DRM_MODE_FLAG_INTERLACE ? + v_INETLACE(1) : v_INETLACE(0); + hdmi_writeb(hdmi, HDMI_VIDEO_TIMING_CTL, value); + + /* Set detail external video timing */ + value = mode->htotal; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HTOTAL_L, value & 0xFF); + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HTOTAL_H, (value >> 8) & 0xFF); + + value = mode->htotal - mode->hdisplay; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HBLANK_L, value & 0xFF); + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HBLANK_H, (value >> 8) & 0xFF); + + value = mode->hsync_start - mode->hdisplay; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HDELAY_L, value & 0xFF); + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HDELAY_H, (value >> 8) & 0xFF); + + value = mode->hsync_end - mode->hsync_start; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HDURATION_L, value & 0xFF); + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_HDURATION_H, (value >> 8) & 0xFF); + + value = mode->vtotal; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_VTOTAL_L, value & 0xFF); + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_VTOTAL_H, (value >> 8) & 0xFF); + + value = mode->vtotal - mode->vdisplay; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_VBLANK, value & 0xFF); + + value = mode->vsync_start - mode->vdisplay; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_VDELAY, value & 0xFF); + + value = mode->vsync_end - mode->vsync_start; + hdmi_writeb(hdmi, HDMI_VIDEO_EXT_VDURATION, value & 0xFF); + + hdmi_writeb(hdmi, HDMI_PHY_PRE_DIV_RATIO, 0x1e); + hdmi_writeb(hdmi, HDMI_PHY_FEEDBACK_DIV_RATIO_LOW, 0x2c); + hdmi_writeb(hdmi, HDMI_PHY_FEEDBACK_DIV_RATIO_HIGH, 0x01); + + return 0; +} + +static int inno_hdmi_setup(struct inno_hdmi *hdmi, + struct drm_display_mode *mode) +{ + hdmi->hdmi_data.vic = drm_match_cea_mode(mode); + + hdmi->hdmi_data.enc_in_format = HDMI_COLORSPACE_RGB; + hdmi->hdmi_data.enc_out_format = HDMI_COLORSPACE_RGB; + + if ((hdmi->hdmi_data.vic == 6) || (hdmi->hdmi_data.vic == 7) || + (hdmi->hdmi_data.vic == 21) || (hdmi->hdmi_data.vic == 22) || + (hdmi->hdmi_data.vic == 2) || (hdmi->hdmi_data.vic == 3) || + (hdmi->hdmi_data.vic == 17) || (hdmi->hdmi_data.vic == 18)) + hdmi->hdmi_data.colorimetry = HDMI_COLORIMETRY_ITU_601; + else + hdmi->hdmi_data.colorimetry = HDMI_COLORIMETRY_ITU_709; + + /* Mute video and audio output */ + hdmi_modb(hdmi, HDMI_AV_MUTE, m_AUDIO_MUTE | m_VIDEO_BLACK, + v_AUDIO_MUTE(1) | v_VIDEO_MUTE(1)); + + /* Set HDMI Mode */ + hdmi_writeb(hdmi, HDMI_HDCP_CTRL, + v_HDMI_DVI(hdmi->hdmi_data.sink_is_hdmi)); + + inno_hdmi_config_video_timing(hdmi, mode); + + inno_hdmi_config_video_csc(hdmi); + + if (hdmi->hdmi_data.sink_is_hdmi) { + inno_hdmi_config_video_avi(hdmi, mode); + inno_hdmi_config_video_vsi(hdmi, mode); + } + + /* + * When IP controller have configured to an accurate video + * timing, then the TMDS clock source would be switched to + * DCLK_LCDC, so we need to init the TMDS rate to mode pixel + * clock rate, and reconfigure the DDC clock. + */ + hdmi->tmds_rate = mode->clock * 1000; + inno_hdmi_i2c_init(hdmi); + + /* Unmute video and audio output */ + hdmi_modb(hdmi, HDMI_AV_MUTE, m_AUDIO_MUTE | m_VIDEO_BLACK, + v_AUDIO_MUTE(0) | v_VIDEO_MUTE(0)); + + return 0; +} + +static void inno_hdmi_encoder_mode_set(struct drm_encoder *encoder, + struct drm_display_mode *mode, + struct drm_display_mode *adj_mode) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(encoder); + + inno_hdmi_setup(hdmi, adj_mode); + + /* Store the display mode for plugin/DPMS poweron events */ + memcpy(&hdmi->previous_mode, adj_mode, sizeof(hdmi->previous_mode)); +} + +static void inno_hdmi_encoder_enable(struct drm_encoder *encoder) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(encoder); + + rockchip_drm_crtc_mode_config(encoder->crtc, DRM_MODE_CONNECTOR_HDMIA, + ROCKCHIP_OUT_MODE_P888); + + inno_hdmi_set_pwr_mode(hdmi, NORMAL); +} + +static void inno_hdmi_encoder_disable(struct drm_encoder *encoder) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(encoder); + + inno_hdmi_set_pwr_mode(hdmi, LOWER_PWR); +} + +static bool inno_hdmi_encoder_mode_fixup(struct drm_encoder *encoder, + const struct drm_display_mode *mode, + struct drm_display_mode *adj_mode) +{ + return true; +} + +static struct drm_encoder_helper_funcs inno_hdmi_encoder_helper_funcs = { + .enable = inno_hdmi_encoder_enable, + .disable = inno_hdmi_encoder_disable, + .mode_fixup = inno_hdmi_encoder_mode_fixup, + .mode_set = inno_hdmi_encoder_mode_set, +}; + +static struct drm_encoder_funcs inno_hdmi_encoder_funcs = { + .destroy = drm_encoder_cleanup, +}; + +static enum drm_connector_status +inno_hdmi_connector_detect(struct drm_connector *connector, bool force) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(connector); + + return (hdmi_readb(hdmi, HDMI_STATUS) & m_HOTPLUG) ? + connector_status_connected : connector_status_disconnected; +} + +static int inno_hdmi_connector_get_modes(struct drm_connector *connector) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(connector); + struct edid *edid; + int ret = 0; + + if (!hdmi->ddc) + return 0; + + edid = drm_get_edid(connector, hdmi->ddc); + if (edid) { + hdmi->hdmi_data.sink_is_hdmi = drm_detect_hdmi_monitor(edid); + hdmi->hdmi_data.sink_has_audio = drm_detect_monitor_audio(edid); + drm_mode_connector_update_edid_property(connector, edid); + ret = drm_add_edid_modes(connector, edid); + kfree(edid); + } + + return ret; +} + +static enum drm_mode_status +inno_hdmi_connector_mode_valid(struct drm_connector *connector, + struct drm_display_mode *mode) +{ + return MODE_OK; +} + +static struct drm_encoder * +inno_hdmi_connector_best_encoder(struct drm_connector *connector) +{ + struct inno_hdmi *hdmi = to_inno_hdmi(connector); + + return &hdmi->encoder; +} + +static int +inno_hdmi_probe_single_connector_modes(struct drm_connector *connector, + uint32_t maxX, uint32_t maxY) +{ + return drm_helper_probe_single_connector_modes(connector, 1920, 1080); +} + +static void inno_hdmi_connector_destroy(struct drm_connector *connector) +{ + drm_connector_unregister(connector); + drm_connector_cleanup(connector); +} + +static struct drm_connector_funcs inno_hdmi_connector_funcs = { + .dpms = drm_atomic_helper_connector_dpms, + .fill_modes = inno_hdmi_probe_single_connector_modes, + .detect = inno_hdmi_connector_detect, + .destroy = inno_hdmi_connector_destroy, + .reset = drm_atomic_helper_connector_reset, + .atomic_duplicate_state = drm_atomic_helper_connector_duplicate_state, + .atomic_destroy_state = drm_atomic_helper_connector_destroy_state, +}; + +static struct drm_connector_helper_funcs inno_hdmi_connector_helper_funcs = { + .get_modes = inno_hdmi_connector_get_modes, + .mode_valid = inno_hdmi_connector_mode_valid, + .best_encoder = inno_hdmi_connector_best_encoder, +}; + +static int inno_hdmi_register(struct drm_device *drm, struct inno_hdmi *hdmi) +{ + struct drm_encoder *encoder = &hdmi->encoder; + struct device *dev = hdmi->dev; + + encoder->possible_crtcs = drm_of_find_possible_crtcs(drm, dev->of_node); + + /* + * If we failed to find the CRTC(s) which this encoder is + * supposed to be connected to, it's because the CRTC has + * not been registered yet. Defer probing, and hope that + * the required CRTC is added later. + */ + if (encoder->possible_crtcs == 0) + return -EPROBE_DEFER; + + drm_encoder_helper_add(encoder, &inno_hdmi_encoder_helper_funcs); + drm_encoder_init(drm, encoder, &inno_hdmi_encoder_funcs, + DRM_MODE_ENCODER_TMDS, NULL); + + hdmi->connector.polled = DRM_CONNECTOR_POLL_HPD; + + drm_connector_helper_add(&hdmi->connector, + &inno_hdmi_connector_helper_funcs); + drm_connector_init(drm, &hdmi->connector, &inno_hdmi_connector_funcs, + DRM_MODE_CONNECTOR_HDMIA); + + drm_mode_connector_attach_encoder(&hdmi->connector, encoder); + + return 0; +} + +static irqreturn_t inno_hdmi_i2c_irq(struct inno_hdmi *hdmi) +{ + struct inno_hdmi_i2c *i2c = hdmi->i2c; + u8 stat; + + stat = hdmi_readb(hdmi, HDMI_INTERRUPT_STATUS1); + if (!(stat & m_INT_EDID_READY)) + return IRQ_NONE; + + /* Clear HDMI EDID interrupt flag */ + hdmi_writeb(hdmi, HDMI_INTERRUPT_STATUS1, m_INT_EDID_READY); + + complete(&i2c->cmp); + + return IRQ_HANDLED; +} + +static irqreturn_t inno_hdmi_hardirq(int irq, void *dev_id) +{ + struct inno_hdmi *hdmi = dev_id; + irqreturn_t ret = IRQ_NONE; + u8 interrupt; + + if (hdmi->i2c) + ret = inno_hdmi_i2c_irq(hdmi); + + interrupt = hdmi_readb(hdmi, HDMI_STATUS); + if (interrupt & m_INT_HOTPLUG) { + hdmi_modb(hdmi, HDMI_STATUS, m_INT_HOTPLUG, m_INT_HOTPLUG); + ret = IRQ_WAKE_THREAD; + } + + return ret; +} + +static irqreturn_t inno_hdmi_irq(int irq, void *dev_id) +{ + struct inno_hdmi *hdmi = dev_id; + + drm_helper_hpd_irq_event(hdmi->connector.dev); + + return IRQ_HANDLED; +} + +static int inno_hdmi_i2c_read(struct inno_hdmi *hdmi, struct i2c_msg *msgs) +{ + int length = msgs->len; + u8 *buf = msgs->buf; + int ret; + + ret = wait_for_completion_timeout(&hdmi->i2c->cmp, HZ / 10); + if (!ret) + return -EAGAIN; + + while (length--) + *buf++ = hdmi_readb(hdmi, HDMI_EDID_FIFO_ADDR); + + return 0; +} + +static int inno_hdmi_i2c_write(struct inno_hdmi *hdmi, struct i2c_msg *msgs) +{ + /* + * The DDC module only support read EDID message, so + * we assume that each word write to this i2c adapter + * should be the offset of EDID word address. + */ + if ((msgs->len != 1) || + ((msgs->addr != DDC_ADDR) && (msgs->addr != DDC_SEGMENT_ADDR))) + return -EINVAL; + + reinit_completion(&hdmi->i2c->cmp); + + if (msgs->addr == DDC_SEGMENT_ADDR) + hdmi->i2c->segment_addr = msgs->buf[0]; + if (msgs->addr == DDC_ADDR) + hdmi->i2c->ddc_addr = msgs->buf[0]; + + /* Set edid fifo first addr */ + hdmi_writeb(hdmi, HDMI_EDID_FIFO_OFFSET, 0x00); + + /* Set edid word address 0x00/0x80 */ + hdmi_writeb(hdmi, HDMI_EDID_WORD_ADDR, hdmi->i2c->ddc_addr); + + /* Set edid segment pointer */ + hdmi_writeb(hdmi, HDMI_EDID_SEGMENT_POINTER, hdmi->i2c->segment_addr); + + return 0; +} + +static int inno_hdmi_i2c_xfer(struct i2c_adapter *adap, + struct i2c_msg *msgs, int num) +{ + struct inno_hdmi *hdmi = i2c_get_adapdata(adap); + struct inno_hdmi_i2c *i2c = hdmi->i2c; + int i, ret = 0; + + mutex_lock(&i2c->lock); + + /* Clear the EDID interrupt flag and unmute the interrupt */ + hdmi_writeb(hdmi, HDMI_INTERRUPT_MASK1, m_INT_EDID_READY); + hdmi_writeb(hdmi, HDMI_INTERRUPT_STATUS1, m_INT_EDID_READY); + + for (i = 0; i < num; i++) { + dev_dbg(hdmi->dev, "xfer: num: %d/%d, len: %d, flags: %#x\n", + i + 1, num, msgs[i].len, msgs[i].flags); + + if (msgs[i].flags & I2C_M_RD) + ret = inno_hdmi_i2c_read(hdmi, &msgs[i]); + else + ret = inno_hdmi_i2c_write(hdmi, &msgs[i]); + + if (ret < 0) + break; + } + + if (!ret) + ret = num; + + /* Mute HDMI EDID interrupt */ + hdmi_writeb(hdmi, HDMI_INTERRUPT_MASK1, 0); + + mutex_unlock(&i2c->lock); + + return ret; +} + +static u32 inno_hdmi_i2c_func(struct i2c_adapter *adapter) +{ + return I2C_FUNC_I2C | I2C_FUNC_SMBUS_EMUL; +} + +static const struct i2c_algorithm inno_hdmi_algorithm = { + .master_xfer = inno_hdmi_i2c_xfer, + .functionality = inno_hdmi_i2c_func, +}; + +static struct i2c_adapter *inno_hdmi_i2c_adapter(struct inno_hdmi *hdmi) +{ + struct i2c_adapter *adap; + struct inno_hdmi_i2c *i2c; + int ret; + + i2c = devm_kzalloc(hdmi->dev, sizeof(*i2c), GFP_KERNEL); + if (!i2c) + return ERR_PTR(-ENOMEM); + + mutex_init(&i2c->lock); + init_completion(&i2c->cmp); + + adap = &i2c->adap; + adap->class = I2C_CLASS_DDC; + adap->owner = THIS_MODULE; + adap->dev.parent = hdmi->dev; + adap->dev.of_node = hdmi->dev->of_node; + adap->algo = &inno_hdmi_algorithm; + strlcpy(adap->name, "Inno HDMI", sizeof(adap->name)); + i2c_set_adapdata(adap, hdmi); + + ret = i2c_add_adapter(adap); + if (ret) { + dev_warn(hdmi->dev, "cannot add %s I2C adapter\n", adap->name); + devm_kfree(hdmi->dev, i2c); + return ERR_PTR(ret); + } + + hdmi->i2c = i2c; + + dev_info(hdmi->dev, "registered %s I2C bus driver\n", adap->name); + + return adap; +} + +static int inno_hdmi_bind(struct device *dev, struct device *master, + void *data) +{ + struct platform_device *pdev = to_platform_device(dev); + struct drm_device *drm = data; + struct inno_hdmi *hdmi; + struct resource *iores; + int irq; + int ret; + + hdmi = devm_kzalloc(dev, sizeof(*hdmi), GFP_KERNEL); + if (!hdmi) + return -ENOMEM; + + hdmi->dev = dev; + hdmi->drm_dev = drm; + + iores = platform_get_resource(pdev, IORESOURCE_MEM, 0); + if (!iores) + return -ENXIO; + + hdmi->regs = devm_ioremap_resource(dev, iores); + if (IS_ERR(hdmi->regs)) + return PTR_ERR(hdmi->regs); + + hdmi->pclk = devm_clk_get(hdmi->dev, "pclk"); + if (IS_ERR(hdmi->pclk)) { + dev_err(hdmi->dev, "Unable to get HDMI pclk clk\n"); + return PTR_ERR(hdmi->pclk); + } + + ret = clk_prepare_enable(hdmi->pclk); + if (ret) { + dev_err(hdmi->dev, "Cannot enable HDMI pclk clock: %d\n", ret); + return ret; + } + + irq = platform_get_irq(pdev, 0); + if (irq < 0) + return irq; + + inno_hdmi_reset(hdmi); + + hdmi->ddc = inno_hdmi_i2c_adapter(hdmi); + if (IS_ERR(hdmi->ddc)) { + hdmi->ddc = NULL; + return PTR_ERR(hdmi->ddc); + } + + /* + * When IP controller haven't configured to an accurate video + * timing, then the TMDS clock source would be switched to + * PCLK_HDMI, so we need to init the TMDS rate to PCLK rate, + * and reconfigure the DDC clock. + */ + hdmi->tmds_rate = clk_get_rate(hdmi->pclk); + inno_hdmi_i2c_init(hdmi); + + ret = inno_hdmi_register(drm, hdmi); + if (ret) + return ret; + + dev_set_drvdata(dev, hdmi); + + /* Unmute hotplug interrupt */ + hdmi_modb(hdmi, HDMI_STATUS, m_MASK_INT_HOTPLUG, v_MASK_INT_HOTPLUG(1)); + + ret = devm_request_threaded_irq(dev, irq, inno_hdmi_hardirq, + inno_hdmi_irq, IRQF_SHARED, + dev_name(dev), hdmi); + + return ret; +} + +static void inno_hdmi_unbind(struct device *dev, struct device *master, + void *data) +{ + struct inno_hdmi *hdmi = dev_get_drvdata(dev); + + hdmi->connector.funcs->destroy(&hdmi->connector); + hdmi->encoder.funcs->destroy(&hdmi->encoder); + + clk_disable_unprepare(hdmi->pclk); + i2c_put_adapter(hdmi->ddc); +} + +static const struct component_ops inno_hdmi_ops = { + .bind = inno_hdmi_bind, + .unbind = inno_hdmi_unbind, +}; + +static int inno_hdmi_probe(struct platform_device *pdev) +{ + return component_add(&pdev->dev, &inno_hdmi_ops); +} + +static int inno_hdmi_remove(struct platform_device *pdev) +{ + component_del(&pdev->dev, &inno_hdmi_ops); + + return 0; +} + +static const struct of_device_id inno_hdmi_dt_ids[] = { + { .compatible = "rockchip,rk3036-inno-hdmi", + }, + {}, +}; +MODULE_DEVICE_TABLE(of, inno_hdmi_dt_ids); + +static struct platform_driver inno_hdmi_driver = { + .probe = inno_hdmi_probe, + .remove = inno_hdmi_remove, + .driver = { + .name = "innohdmi-rockchip", + .of_match_table = inno_hdmi_dt_ids, + }, +}; + +module_platform_driver(inno_hdmi_driver); + +MODULE_AUTHOR("Zheng Yang <[email protected]>"); +MODULE_AUTHOR("Yakir Yang <[email protected]>"); +MODULE_DESCRIPTION("Rockchip Specific INNO-HDMI Driver"); +MODULE_LICENSE("GPL v2"); +MODULE_ALIAS("platform:innohdmi-rockchip"); diff --git a/drivers/gpu/drm/rockchip/inno_hdmi.h b/drivers/gpu/drm/rockchip/inno_hdmi.h new file mode 100644 index 000000000000..aa7c415f8cc1 --- /dev/null +++ b/drivers/gpu/drm/rockchip/inno_hdmi.h @@ -0,0 +1,362 @@ +/* + * Copyright (C) Fuzhou Rockchip Electronics Co.Ltd + * Zheng Yang <[email protected]> + * Yakir Yang <[email protected]> + * + * This software is licensed under the terms of the GNU General Public + * License version 2, as published by the Free Software Foundation, and + * may be copied, distributed, and modified under those terms. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + */ + +#ifndef __INNO_HDMI_H__ +#define __INNO_HDMI_H__ + +#define DDC_SEGMENT_ADDR 0x30 + +enum PWR_MODE { + NORMAL, + LOWER_PWR, +}; + +#define HDMI_SCL_RATE (100*1000) +#define DDC_BUS_FREQ_L 0x4b +#define DDC_BUS_FREQ_H 0x4c + +#define HDMI_SYS_CTRL 0x00 +#define m_RST_ANALOG (1 << 6) +#define v_RST_ANALOG (0 << 6) +#define v_NOT_RST_ANALOG (1 << 6) +#define m_RST_DIGITAL (1 << 5) +#define v_RST_DIGITAL (0 << 5) +#define v_NOT_RST_DIGITAL (1 << 5) +#define m_REG_CLK_INV (1 << 4) +#define v_REG_CLK_NOT_INV (0 << 4) +#define v_REG_CLK_INV (1 << 4) +#define m_VCLK_INV (1 << 3) +#define v_VCLK_NOT_INV (0 << 3) +#define v_VCLK_INV (1 << 3) +#define m_REG_CLK_SOURCE (1 << 2) +#define v_REG_CLK_SOURCE_TMDS (0 << 2) +#define v_REG_CLK_SOURCE_SYS (1 << 2) +#define m_POWER (1 << 1) +#define v_PWR_ON (0 << 1) +#define v_PWR_OFF (1 << 1) +#define m_INT_POL (1 << 0) +#define v_INT_POL_HIGH 1 +#define v_INT_POL_LOW 0 + +#define HDMI_VIDEO_CONTRL1 0x01 +#define m_VIDEO_INPUT_FORMAT (7 << 1) +#define m_DE_SOURCE (1 << 0) +#define v_VIDEO_INPUT_FORMAT(n) (n << 1) +#define v_DE_EXTERNAL 1 +#define v_DE_INTERNAL 0 +enum { + VIDEO_INPUT_SDR_RGB444 = 0, + VIDEO_INPUT_DDR_RGB444 = 5, + VIDEO_INPUT_DDR_YCBCR422 = 6 +}; + +#define HDMI_VIDEO_CONTRL2 0x02 +#define m_VIDEO_OUTPUT_COLOR (3 << 6) +#define m_VIDEO_INPUT_BITS (3 << 4) +#define m_VIDEO_INPUT_CSP (1 << 0) +#define v_VIDEO_OUTPUT_COLOR(n) (((n) & 0x3) << 6) +#define v_VIDEO_INPUT_BITS(n) (n << 4) +#define v_VIDEO_INPUT_CSP(n) (n << 0) +enum { + VIDEO_INPUT_12BITS = 0, + VIDEO_INPUT_10BITS = 1, + VIDEO_INPUT_REVERT = 2, + VIDEO_INPUT_8BITS = 3, +}; + +#define HDMI_VIDEO_CONTRL 0x03 +#define m_VIDEO_AUTO_CSC (1 << 7) +#define v_VIDEO_AUTO_CSC(n) (n << 7) +#define m_VIDEO_C0_C2_SWAP (1 << 0) +#define v_VIDEO_C0_C2_SWAP(n) (n << 0) +enum { + C0_C2_CHANGE_ENABLE = 0, + C0_C2_CHANGE_DISABLE = 1, + AUTO_CSC_DISABLE = 0, + AUTO_CSC_ENABLE = 1, +}; + +#define HDMI_VIDEO_CONTRL3 0x04 +#define m_COLOR_DEPTH_NOT_INDICATED (1 << 4) +#define m_SOF (1 << 3) +#define m_COLOR_RANGE (1 << 2) +#define m_CSC (1 << 0) +#define v_COLOR_DEPTH_NOT_INDICATED(n) ((n) << 4) +#define v_SOF_ENABLE (0 << 3) +#define v_SOF_DISABLE (1 << 3) +#define v_COLOR_RANGE_FULL (1 << 2) +#define v_COLOR_RANGE_LIMITED (0 << 2) +#define v_CSC_ENABLE 1 +#define v_CSC_DISABLE 0 + +#define HDMI_AV_MUTE 0x05 +#define m_AVMUTE_CLEAR (1 << 7) +#define m_AVMUTE_ENABLE (1 << 6) +#define m_AUDIO_MUTE (1 << 1) +#define m_VIDEO_BLACK (1 << 0) +#define v_AVMUTE_CLEAR(n) (n << 7) +#define v_AVMUTE_ENABLE(n) (n << 6) +#define v_AUDIO_MUTE(n) (n << 1) +#define v_VIDEO_MUTE(n) (n << 0) + +#define HDMI_VIDEO_TIMING_CTL 0x08 +#define v_HSYNC_POLARITY(n) (n << 3) +#define v_VSYNC_POLARITY(n) (n << 2) +#define v_INETLACE(n) (n << 1) +#define v_EXTERANL_VIDEO(n) (n << 0) + +#define HDMI_VIDEO_EXT_HTOTAL_L 0x09 +#define HDMI_VIDEO_EXT_HTOTAL_H 0x0a +#define HDMI_VIDEO_EXT_HBLANK_L 0x0b +#define HDMI_VIDEO_EXT_HBLANK_H 0x0c +#define HDMI_VIDEO_EXT_HDELAY_L 0x0d +#define HDMI_VIDEO_EXT_HDELAY_H 0x0e +#define HDMI_VIDEO_EXT_HDURATION_L 0x0f +#define HDMI_VIDEO_EXT_HDURATION_H 0x10 +#define HDMI_VIDEO_EXT_VTOTAL_L 0x11 +#define HDMI_VIDEO_EXT_VTOTAL_H 0x12 +#define HDMI_VIDEO_EXT_VBLANK 0x13 +#define HDMI_VIDEO_EXT_VDELAY 0x14 +#define HDMI_VIDEO_EXT_VDURATION 0x15 + +#define HDMI_VIDEO_CSC_COEF 0x18 + +#define HDMI_AUDIO_CTRL1 0x35 +enum { + CTS_SOURCE_INTERNAL = 0, + CTS_SOURCE_EXTERNAL = 1, +}; +#define v_CTS_SOURCE(n) (n << 7) + +enum { + DOWNSAMPLE_DISABLE = 0, + DOWNSAMPLE_1_2 = 1, + DOWNSAMPLE_1_4 = 2, +}; +#define v_DOWN_SAMPLE(n) (n << 5) + +enum { + AUDIO_SOURCE_IIS = 0, + AUDIO_SOURCE_SPDIF = 1, +}; +#define v_AUDIO_SOURCE(n) (n << 3) + +#define v_MCLK_ENABLE(n) (n << 2) +enum { + MCLK_128FS = 0, + MCLK_256FS = 1, + MCLK_384FS = 2, + MCLK_512FS = 3, +}; +#define v_MCLK_RATIO(n) (n) + +#define AUDIO_SAMPLE_RATE 0x37 +enum { + AUDIO_32K = 0x3, + AUDIO_441K = 0x0, + AUDIO_48K = 0x2, + AUDIO_882K = 0x8, + AUDIO_96K = 0xa, + AUDIO_1764K = 0xc, + AUDIO_192K = 0xe, +}; + +#define AUDIO_I2S_MODE 0x38 +enum { + I2S_CHANNEL_1_2 = 1, + I2S_CHANNEL_3_4 = 3, + I2S_CHANNEL_5_6 = 7, + I2S_CHANNEL_7_8 = 0xf +}; +#define v_I2S_CHANNEL(n) ((n) << 2) +enum { + I2S_STANDARD = 0, + I2S_LEFT_JUSTIFIED = 1, + I2S_RIGHT_JUSTIFIED = 2, +}; +#define v_I2S_MODE(n) (n) + +#define AUDIO_I2S_MAP 0x39 +#define AUDIO_I2S_SWAPS_SPDIF 0x3a +#define v_SPIDF_FREQ(n) (n) + +#define N_32K 0x1000 +#define N_441K 0x1880 +#define N_882K 0x3100 +#define N_1764K 0x6200 +#define N_48K 0x1800 +#define N_96K 0x3000 +#define N_192K 0x6000 + +#define HDMI_AUDIO_CHANNEL_STATUS 0x3e +#define m_AUDIO_STATUS_NLPCM (1 << 7) +#define m_AUDIO_STATUS_USE (1 << 6) +#define m_AUDIO_STATUS_COPYRIGHT (1 << 5) +#define m_AUDIO_STATUS_ADDITION (3 << 2) +#define m_AUDIO_STATUS_CLK_ACCURACY (2 << 0) +#define v_AUDIO_STATUS_NLPCM(n) ((n & 1) << 7) +#define AUDIO_N_H 0x3f +#define AUDIO_N_M 0x40 +#define AUDIO_N_L 0x41 + +#define HDMI_AUDIO_CTS_H 0x45 +#define HDMI_AUDIO_CTS_M 0x46 +#define HDMI_AUDIO_CTS_L 0x47 + +#define HDMI_DDC_CLK_L 0x4b +#define HDMI_DDC_CLK_H 0x4c + +#define HDMI_EDID_SEGMENT_POINTER 0x4d +#define HDMI_EDID_WORD_ADDR 0x4e +#define HDMI_EDID_FIFO_OFFSET 0x4f +#define HDMI_EDID_FIFO_ADDR 0x50 + +#define HDMI_PACKET_SEND_MANUAL 0x9c +#define HDMI_PACKET_SEND_AUTO 0x9d +#define m_PACKET_GCP_EN (1 << 7) +#define m_PACKET_MSI_EN (1 << 6) +#define m_PACKET_SDI_EN (1 << 5) +#define m_PACKET_VSI_EN (1 << 4) +#define v_PACKET_GCP_EN(n) ((n & 1) << 7) +#define v_PACKET_MSI_EN(n) ((n & 1) << 6) +#define v_PACKET_SDI_EN(n) ((n & 1) << 5) +#define v_PACKET_VSI_EN(n) ((n & 1) << 4) + +#define HDMI_CONTROL_PACKET_BUF_INDEX 0x9f +enum { + INFOFRAME_VSI = 0x05, + INFOFRAME_AVI = 0x06, + INFOFRAME_AAI = 0x08, +}; + +#define HDMI_CONTROL_PACKET_ADDR 0xa0 +#define HDMI_MAXIMUM_INFO_FRAME_SIZE 0x11 +enum { + AVI_COLOR_MODE_RGB = 0, + AVI_COLOR_MODE_YCBCR422 = 1, + AVI_COLOR_MODE_YCBCR444 = 2, + AVI_COLORIMETRY_NO_DATA = 0, + + AVI_COLORIMETRY_SMPTE_170M = 1, + AVI_COLORIMETRY_ITU709 = 2, + AVI_COLORIMETRY_EXTENDED = 3, + + AVI_CODED_FRAME_ASPECT_NO_DATA = 0, + AVI_CODED_FRAME_ASPECT_4_3 = 1, + AVI_CODED_FRAME_ASPECT_16_9 = 2, + + ACTIVE_ASPECT_RATE_SAME_AS_CODED_FRAME = 0x08, + ACTIVE_ASPECT_RATE_4_3 = 0x09, + ACTIVE_ASPECT_RATE_16_9 = 0x0A, + ACTIVE_ASPECT_RATE_14_9 = 0x0B, +}; + +#define HDMI_HDCP_CTRL 0x52 +#define m_HDMI_DVI (1 << 1) +#define v_HDMI_DVI(n) (n << 1) + +#define HDMI_INTERRUPT_MASK1 0xc0 +#define HDMI_INTERRUPT_STATUS1 0xc1 +#define m_INT_ACTIVE_VSYNC (1 << 5) +#define m_INT_EDID_READY (1 << 2) + +#define HDMI_INTERRUPT_MASK2 0xc2 +#define HDMI_INTERRUPT_STATUS2 0xc3 +#define m_INT_HDCP_ERR (1 << 7) +#define m_INT_BKSV_FLAG (1 << 6) +#define m_INT_HDCP_OK (1 << 4) + +#define HDMI_STATUS 0xc8 +#define m_HOTPLUG (1 << 7) +#define m_MASK_INT_HOTPLUG (1 << 5) +#define m_INT_HOTPLUG (1 << 1) +#define v_MASK_INT_HOTPLUG(n) ((n & 0x1) << 5) + +#define HDMI_COLORBAR 0xc9 + +#define HDMI_PHY_SYNC 0xce +#define HDMI_PHY_SYS_CTL 0xe0 +#define m_TMDS_CLK_SOURCE (1 << 5) +#define v_TMDS_FROM_PLL (0 << 5) +#define v_TMDS_FROM_GEN (1 << 5) +#define m_PHASE_CLK (1 << 4) +#define v_DEFAULT_PHASE (0 << 4) +#define v_SYNC_PHASE (1 << 4) +#define m_TMDS_CURRENT_PWR (1 << 3) +#define v_TURN_ON_CURRENT (0 << 3) +#define v_CAT_OFF_CURRENT (1 << 3) +#define m_BANDGAP_PWR (1 << 2) +#define v_BANDGAP_PWR_UP (0 << 2) +#define v_BANDGAP_PWR_DOWN (1 << 2) +#define m_PLL_PWR (1 << 1) +#define v_PLL_PWR_UP (0 << 1) +#define v_PLL_PWR_DOWN (1 << 1) +#define m_TMDS_CHG_PWR (1 << 0) +#define v_TMDS_CHG_PWR_UP (0 << 0) +#define v_TMDS_CHG_PWR_DOWN (1 << 0) + +#define HDMI_PHY_CHG_PWR 0xe1 +#define v_CLK_CHG_PWR(n) ((n & 1) << 3) +#define v_DATA_CHG_PWR(n) ((n & 7) << 0) + +#define HDMI_PHY_DRIVER 0xe2 +#define v_CLK_MAIN_DRIVER(n) (n << 4) +#define v_DATA_MAIN_DRIVER(n) (n << 0) + +#define HDMI_PHY_PRE_EMPHASIS 0xe3 +#define v_PRE_EMPHASIS(n) ((n & 7) << 4) +#define v_CLK_PRE_DRIVER(n) ((n & 3) << 2) +#define v_DATA_PRE_DRIVER(n) ((n & 3) << 0) + +#define HDMI_PHY_FEEDBACK_DIV_RATIO_LOW 0xe7 +#define v_FEEDBACK_DIV_LOW(n) (n & 0xff) +#define HDMI_PHY_FEEDBACK_DIV_RATIO_HIGH 0xe8 +#define v_FEEDBACK_DIV_HIGH(n) (n & 1) + +#define HDMI_PHY_PRE_DIV_RATIO 0xed +#define v_PRE_DIV_RATIO(n) (n & 0x1f) + +#define HDMI_CEC_CTRL 0xd0 +#define m_ADJUST_FOR_HISENSE (1 << 6) +#define m_REJECT_RX_BROADCAST (1 << 5) +#define m_BUSFREETIME_ENABLE (1 << 2) +#define m_REJECT_RX (1 << 1) +#define m_START_TX (1 << 0) + +#define HDMI_CEC_DATA 0xd1 +#define HDMI_CEC_TX_OFFSET 0xd2 +#define HDMI_CEC_RX_OFFSET 0xd3 +#define HDMI_CEC_CLK_H 0xd4 +#define HDMI_CEC_CLK_L 0xd5 +#define HDMI_CEC_TX_LENGTH 0xd6 +#define HDMI_CEC_RX_LENGTH 0xd7 +#define HDMI_CEC_TX_INT_MASK 0xd8 +#define m_TX_DONE (1 << 3) +#define m_TX_NOACK (1 << 2) +#define m_TX_BROADCAST_REJ (1 << 1) +#define m_TX_BUSNOTFREE (1 << 0) + +#define HDMI_CEC_RX_INT_MASK 0xd9 +#define m_RX_LA_ERR (1 << 4) +#define m_RX_GLITCH (1 << 3) +#define m_RX_DONE (1 << 0) + +#define HDMI_CEC_TX_INT 0xda +#define HDMI_CEC_RX_INT 0xdb +#define HDMI_CEC_BUSFREETIME_L 0xdc +#define HDMI_CEC_BUSFREETIME_H 0xdd +#define HDMI_CEC_LOGICADDR 0xde + +#endif /* __INNO_HDMI_H__ */ diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c index db0763794edc..27342fd76e90 100644 --- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.c +++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.c @@ -438,26 +438,6 @@ static const struct drm_crtc_helper_funcs crtc_helper_funcs = { .mode_set_base = shmob_drm_crtc_mode_set_base, }; -void shmob_drm_crtc_cancel_page_flip(struct shmob_drm_crtc *scrtc, - struct drm_file *file) -{ - struct drm_pending_vblank_event *event; - struct drm_device *dev = scrtc->crtc.dev; - unsigned long flags; - - /* Destroy the pending vertical blanking event associated with the - * pending page flip, if any, and disable vertical blanking interrupts. - */ - spin_lock_irqsave(&dev->event_lock, flags); - event = scrtc->event; - if (event && event->base.file_priv == file) { - scrtc->event = NULL; - event->base.destroy(&event->base); - drm_vblank_put(dev, 0); - } - spin_unlock_irqrestore(&dev->event_lock, flags); -} - void shmob_drm_crtc_finish_page_flip(struct shmob_drm_crtc *scrtc) { struct drm_pending_vblank_event *event; diff --git a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h index eddad6dcc88a..38ed4ff8aaf2 100644 --- a/drivers/gpu/drm/shmobile/shmob_drm_crtc.h +++ b/drivers/gpu/drm/shmobile/shmob_drm_crtc.h @@ -47,8 +47,6 @@ struct shmob_drm_connector { int shmob_drm_crtc_create(struct shmob_drm_device *sdev); void shmob_drm_crtc_enable_vblank(struct shmob_drm_device *sdev, bool enable); -void shmob_drm_crtc_cancel_page_flip(struct shmob_drm_crtc *scrtc, - struct drm_file *file); void shmob_drm_crtc_finish_page_flip(struct shmob_drm_crtc *scrtc); void shmob_drm_crtc_suspend(struct shmob_drm_crtc *scrtc); void shmob_drm_crtc_resume(struct shmob_drm_crtc *scrtc); diff --git a/drivers/gpu/drm/shmobile/shmob_drm_drv.c b/drivers/gpu/drm/shmobile/shmob_drm_drv.c index 04e66e3751b4..7700ff172079 100644 --- a/drivers/gpu/drm/shmobile/shmob_drm_drv.c +++ b/drivers/gpu/drm/shmobile/shmob_drm_drv.c @@ -200,13 +200,6 @@ done: return ret; } -static void shmob_drm_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct shmob_drm_device *sdev = dev->dev_private; - - shmob_drm_crtc_cancel_page_flip(&sdev->crtc, file); -} - static irqreturn_t shmob_drm_irq(int irq, void *arg) { struct drm_device *dev = arg; @@ -266,7 +259,6 @@ static struct drm_driver shmob_drm_driver = { | DRIVER_PRIME, .load = shmob_drm_load, .unload = shmob_drm_unload, - .preclose = shmob_drm_preclose, .set_busid = drm_platform_set_busid, .irq_handler = shmob_drm_irq, .get_vblank_counter = drm_vblank_no_hw_counter, diff --git a/drivers/gpu/drm/sti/sti_tvout.c b/drivers/gpu/drm/sti/sti_tvout.c index f2afcf5438b8..24a3735b88fd 100644 --- a/drivers/gpu/drm/sti/sti_tvout.c +++ b/drivers/gpu/drm/sti/sti_tvout.c @@ -440,13 +440,6 @@ static void sti_tvout_encoder_dpms(struct drm_encoder *encoder, int mode) { } -static bool sti_tvout_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void sti_tvout_encoder_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -486,7 +479,6 @@ static void sti_dvo_encoder_disable(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs sti_dvo_encoder_helper_funcs = { .dpms = sti_tvout_encoder_dpms, - .mode_fixup = sti_tvout_encoder_mode_fixup, .mode_set = sti_tvout_encoder_mode_set, .prepare = sti_tvout_encoder_prepare, .commit = sti_dvo_encoder_commit, @@ -540,7 +532,6 @@ static void sti_hda_encoder_disable(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs sti_hda_encoder_helper_funcs = { .dpms = sti_tvout_encoder_dpms, - .mode_fixup = sti_tvout_encoder_mode_fixup, .mode_set = sti_tvout_encoder_mode_set, .prepare = sti_tvout_encoder_prepare, .commit = sti_hda_encoder_commit, @@ -589,7 +580,6 @@ static void sti_hdmi_encoder_disable(struct drm_encoder *encoder) static const struct drm_encoder_helper_funcs sti_hdmi_encoder_helper_funcs = { .dpms = sti_tvout_encoder_dpms, - .mode_fixup = sti_tvout_encoder_mode_fixup, .mode_set = sti_tvout_encoder_mode_set, .prepare = sti_tvout_encoder_prepare, .commit = sti_hdmi_encoder_commit, diff --git a/drivers/gpu/drm/tegra/dc.c b/drivers/gpu/drm/tegra/dc.c index dde6f208c347..fb2b4b0271a2 100644 --- a/drivers/gpu/drm/tegra/dc.c +++ b/drivers/gpu/drm/tegra/dc.c @@ -988,23 +988,6 @@ static void tegra_dc_finish_page_flip(struct tegra_dc *dc) spin_unlock_irqrestore(&drm->event_lock, flags); } -void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file) -{ - struct tegra_dc *dc = to_tegra_dc(crtc); - struct drm_device *drm = crtc->dev; - unsigned long flags; - - spin_lock_irqsave(&drm->event_lock, flags); - - if (dc->event && dc->event->base.file_priv == file) { - dc->event->base.destroy(&dc->event->base); - drm_crtc_vblank_put(crtc); - dc->event = NULL; - } - - spin_unlock_irqrestore(&drm->event_lock, flags); -} - static void tegra_dc_destroy(struct drm_crtc *crtc) { drm_crtc_cleanup(crtc); diff --git a/drivers/gpu/drm/tegra/drm.c b/drivers/gpu/drm/tegra/drm.c index c5c856a0879d..8e6b18caa706 100644 --- a/drivers/gpu/drm/tegra/drm.c +++ b/drivers/gpu/drm/tegra/drm.c @@ -858,10 +858,6 @@ static void tegra_drm_preclose(struct drm_device *drm, struct drm_file *file) { struct tegra_drm_file *fpriv = file->driver_priv; struct tegra_drm_context *context, *tmp; - struct drm_crtc *crtc; - - list_for_each_entry(crtc, &drm->mode_config.crtc_list, head) - tegra_dc_cancel_page_flip(crtc, file); list_for_each_entry_safe(context, tmp, &fpriv->contexts, list) tegra_drm_context_free(context); diff --git a/drivers/gpu/drm/tegra/drm.h b/drivers/gpu/drm/tegra/drm.h index c088f2f67eda..8a10f5b7d9dc 100644 --- a/drivers/gpu/drm/tegra/drm.h +++ b/drivers/gpu/drm/tegra/drm.h @@ -195,7 +195,6 @@ struct tegra_dc_window { u32 tegra_dc_get_vblank_counter(struct tegra_dc *dc); void tegra_dc_enable_vblank(struct tegra_dc *dc); void tegra_dc_disable_vblank(struct tegra_dc *dc); -void tegra_dc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file); void tegra_dc_commit(struct tegra_dc *dc); int tegra_dc_state_setup_clock(struct tegra_dc *dc, struct drm_crtc_state *crtc_state, diff --git a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c index 7d07733bdc86..4802da8e6d6f 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_crtc.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_crtc.c @@ -662,26 +662,6 @@ irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc) return IRQ_HANDLED; } -void tilcdc_crtc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file) -{ - struct tilcdc_crtc *tilcdc_crtc = to_tilcdc_crtc(crtc); - struct drm_pending_vblank_event *event; - struct drm_device *dev = crtc->dev; - unsigned long flags; - - /* Destroy the pending vertical blanking event associated with the - * pending page flip, if any, and disable vertical blanking interrupts. - */ - spin_lock_irqsave(&dev->event_lock, flags); - event = tilcdc_crtc->event; - if (event && event->base.file_priv == file) { - tilcdc_crtc->event = NULL; - event->base.destroy(&event->base); - drm_vblank_put(dev, 0); - } - spin_unlock_irqrestore(&dev->event_lock, flags); -} - struct drm_crtc *tilcdc_crtc_create(struct drm_device *dev) { struct tilcdc_crtc *tilcdc_crtc; diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.c b/drivers/gpu/drm/tilcdc/tilcdc_drv.c index d7f5b897c6c5..8190ac3b1b32 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_drv.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.c @@ -350,13 +350,6 @@ fail_free_priv: return ret; } -static void tilcdc_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct tilcdc_drm_private *priv = dev->dev_private; - - tilcdc_crtc_cancel_page_flip(priv->crtc, file); -} - static void tilcdc_lastclose(struct drm_device *dev) { struct tilcdc_drm_private *priv = dev->dev_private; @@ -557,7 +550,6 @@ static struct drm_driver tilcdc_driver = { .driver_features = DRIVER_HAVE_IRQ | DRIVER_GEM | DRIVER_MODESET, .load = tilcdc_load, .unload = tilcdc_unload, - .preclose = tilcdc_preclose, .lastclose = tilcdc_lastclose, .set_busid = drm_platform_set_busid, .irq_handler = tilcdc_irq, diff --git a/drivers/gpu/drm/tilcdc/tilcdc_drv.h b/drivers/gpu/drm/tilcdc/tilcdc_drv.h index e863ad0d26fe..66105d8dc620 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_drv.h +++ b/drivers/gpu/drm/tilcdc/tilcdc_drv.h @@ -163,7 +163,6 @@ struct tilcdc_panel_info { #define DBG(fmt, ...) DRM_DEBUG(fmt"\n", ##__VA_ARGS__) struct drm_crtc *tilcdc_crtc_create(struct drm_device *dev); -void tilcdc_crtc_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file); irqreturn_t tilcdc_crtc_irq(struct drm_crtc *crtc); void tilcdc_crtc_update_clk(struct drm_crtc *crtc); void tilcdc_crtc_set_panel_info(struct drm_crtc *crtc, diff --git a/drivers/gpu/drm/tilcdc/tilcdc_panel.c b/drivers/gpu/drm/tilcdc/tilcdc_panel.c index 4dda6e2f464b..8dcf02a79b23 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_panel.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_panel.c @@ -70,14 +70,6 @@ static void panel_encoder_dpms(struct drm_encoder *encoder, int mode) mode == DRM_MODE_DPMS_ON ? 1 : 0); } -static bool panel_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - /* nothing needed */ - return true; -} - static void panel_encoder_prepare(struct drm_encoder *encoder) { struct panel_encoder *panel_encoder = to_panel_encoder(encoder); @@ -103,7 +95,6 @@ static const struct drm_encoder_funcs panel_encoder_funcs = { static const struct drm_encoder_helper_funcs panel_encoder_helper_funcs = { .dpms = panel_encoder_dpms, - .mode_fixup = panel_encoder_mode_fixup, .prepare = panel_encoder_prepare, .commit = panel_encoder_commit, .mode_set = panel_encoder_mode_set, diff --git a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c index 5052a8af7ecb..1c230172b402 100644 --- a/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c +++ b/drivers/gpu/drm/tilcdc/tilcdc_tfp410.c @@ -80,14 +80,6 @@ static void tfp410_encoder_dpms(struct drm_encoder *encoder, int mode) tfp410_encoder->dpms = mode; } -static bool tfp410_encoder_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - /* nothing needed */ - return true; -} - static void tfp410_encoder_prepare(struct drm_encoder *encoder) { tfp410_encoder_dpms(encoder, DRM_MODE_DPMS_OFF); @@ -112,7 +104,6 @@ static const struct drm_encoder_funcs tfp410_encoder_funcs = { static const struct drm_encoder_helper_funcs tfp410_encoder_helper_funcs = { .dpms = tfp410_encoder_dpms, - .mode_fixup = tfp410_encoder_mode_fixup, .prepare = tfp410_encoder_prepare, .commit = tfp410_encoder_commit, .mode_set = tfp410_encoder_mode_set, diff --git a/drivers/gpu/drm/udl/udl_drv.c b/drivers/gpu/drm/udl/udl_drv.c index d5728ec85254..772ec9e1f590 100644 --- a/drivers/gpu/drm/udl/udl_drv.c +++ b/drivers/gpu/drm/udl/udl_drv.c @@ -125,17 +125,5 @@ static struct usb_driver udl_driver = { .disconnect = udl_usb_disconnect, .id_table = id_table, }; - -static int __init udl_init(void) -{ - return usb_register(&udl_driver); -} - -static void __exit udl_exit(void) -{ - usb_deregister(&udl_driver); -} - -module_init(udl_init); -module_exit(udl_exit); +module_usb_driver(udl_driver); MODULE_LICENSE("GPL"); diff --git a/drivers/gpu/drm/udl/udl_encoder.c b/drivers/gpu/drm/udl/udl_encoder.c index a181a647fcf9..59a4b34e87ed 100644 --- a/drivers/gpu/drm/udl/udl_encoder.c +++ b/drivers/gpu/drm/udl/udl_encoder.c @@ -26,13 +26,6 @@ static void udl_encoder_disable(struct drm_encoder *encoder) { } -static bool udl_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void udl_encoder_prepare(struct drm_encoder *encoder) { } @@ -54,7 +47,6 @@ udl_encoder_dpms(struct drm_encoder *encoder, int mode) static const struct drm_encoder_helper_funcs udl_helper_funcs = { .dpms = udl_encoder_dpms, - .mode_fixup = udl_mode_fixup, .prepare = udl_encoder_prepare, .mode_set = udl_encoder_mode_set, .commit = udl_encoder_commit, diff --git a/drivers/gpu/drm/udl/udl_fb.c b/drivers/gpu/drm/udl/udl_fb.c index 200419d4d43c..c427499133d6 100644 --- a/drivers/gpu/drm/udl/udl_fb.c +++ b/drivers/gpu/drm/udl/udl_fb.c @@ -409,7 +409,6 @@ static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb, if (ufb->obj->base.import_attach) { ret = dma_buf_begin_cpu_access(ufb->obj->base.import_attach->dmabuf, - 0, ufb->obj->base.size, DMA_FROM_DEVICE); if (ret) goto unlock; @@ -425,7 +424,6 @@ static int udl_user_framebuffer_dirty(struct drm_framebuffer *fb, if (ufb->obj->base.import_attach) { dma_buf_end_cpu_access(ufb->obj->base.import_attach->dmabuf, - 0, ufb->obj->base.size, DMA_FROM_DEVICE); } diff --git a/drivers/gpu/drm/vc4/vc4_crtc.c b/drivers/gpu/drm/vc4/vc4_crtc.c index 018145e0b87d..619dc781c517 100644 --- a/drivers/gpu/drm/vc4/vc4_crtc.c +++ b/drivers/gpu/drm/vc4/vc4_crtc.c @@ -49,22 +49,27 @@ struct vc4_crtc { /* Which HVS channel we're using for our CRTC. */ int channel; - /* Pointer to the actual hardware display list memory for the - * crtc. - */ - u32 __iomem *dlist; - - u32 dlist_size; /* in dwords */ - struct drm_pending_vblank_event *event; }; +struct vc4_crtc_state { + struct drm_crtc_state base; + /* Dlist area for this CRTC configuration. */ + struct drm_mm_node mm; +}; + static inline struct vc4_crtc * to_vc4_crtc(struct drm_crtc *crtc) { return (struct vc4_crtc *)crtc; } +static inline struct vc4_crtc_state * +to_vc4_crtc_state(struct drm_crtc_state *crtc_state) +{ + return (struct vc4_crtc_state *)crtc_state; +} + struct vc4_crtc_data { /* Which channel of the HVS this pixelvalve sources from. */ int hvs_channel; @@ -319,11 +324,13 @@ static void vc4_crtc_enable(struct drm_crtc *crtc) static int vc4_crtc_atomic_check(struct drm_crtc *crtc, struct drm_crtc_state *state) { + struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(state); struct drm_device *dev = crtc->dev; struct vc4_dev *vc4 = to_vc4_dev(dev); struct drm_plane *plane; - struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc); + unsigned long flags; u32 dlist_count = 0; + int ret; /* The pixelvalve can only feed one encoder (and encoders are * 1:1 with connectors.) @@ -346,18 +353,12 @@ static int vc4_crtc_atomic_check(struct drm_crtc *crtc, dlist_count++; /* Account for SCALER_CTL0_END. */ - if (!vc4_crtc->dlist || dlist_count > vc4_crtc->dlist_size) { - vc4_crtc->dlist = ((u32 __iomem *)vc4->hvs->dlist + - HVS_BOOTLOADER_DLIST_END); - vc4_crtc->dlist_size = ((SCALER_DLIST_SIZE >> 2) - - HVS_BOOTLOADER_DLIST_END); - - if (dlist_count > vc4_crtc->dlist_size) { - DRM_DEBUG_KMS("dlist too large for CRTC (%d > %d).\n", - dlist_count, vc4_crtc->dlist_size); - return -EINVAL; - } - } + spin_lock_irqsave(&vc4->hvs->mm_lock, flags); + ret = drm_mm_insert_node(&vc4->hvs->dlist_mm, &vc4_state->mm, + dlist_count, 1, 0); + spin_unlock_irqrestore(&vc4->hvs->mm_lock, flags); + if (ret) + return ret; return 0; } @@ -368,47 +369,29 @@ static void vc4_crtc_atomic_flush(struct drm_crtc *crtc, struct drm_device *dev = crtc->dev; struct vc4_dev *vc4 = to_vc4_dev(dev); struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc); + struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(crtc->state); struct drm_plane *plane; bool debug_dump_regs = false; - u32 __iomem *dlist_next = vc4_crtc->dlist; + u32 __iomem *dlist_start = vc4->hvs->dlist + vc4_state->mm.start; + u32 __iomem *dlist_next = dlist_start; if (debug_dump_regs) { DRM_INFO("CRTC %d HVS before:\n", drm_crtc_index(crtc)); vc4_hvs_dump_state(dev); } - /* Copy all the active planes' dlist contents to the hardware dlist. - * - * XXX: If the new display list was large enough that it - * overlapped a currently-read display list, we need to do - * something like disable scanout before putting in the new - * list. For now, we're safe because we only have the two - * planes. - */ + /* Copy all the active planes' dlist contents to the hardware dlist. */ drm_atomic_crtc_for_each_plane(plane, crtc) { dlist_next += vc4_plane_write_dlist(plane, dlist_next); } - if (dlist_next == vc4_crtc->dlist) { - /* If no planes were enabled, use the SCALER_CTL0_END - * at the start of the display list memory (in the - * bootloader section). We'll rewrite that - * SCALER_CTL0_END, just in case, though. - */ - writel(SCALER_CTL0_END, vc4->hvs->dlist); - HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), 0); - } else { - writel(SCALER_CTL0_END, dlist_next); - dlist_next++; - - HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), - (u32 __iomem *)vc4_crtc->dlist - - (u32 __iomem *)vc4->hvs->dlist); - - /* Make the next display list start after ours. */ - vc4_crtc->dlist_size -= (dlist_next - vc4_crtc->dlist); - vc4_crtc->dlist = dlist_next; - } + writel(SCALER_CTL0_END, dlist_next); + dlist_next++; + + WARN_ON_ONCE(dlist_next - dlist_start != vc4_state->mm.size); + + HVS_WRITE(SCALER_DISPLISTX(vc4_crtc->channel), + vc4_state->mm.start); if (debug_dump_regs) { DRM_INFO("CRTC %d HVS after:\n", drm_crtc_index(crtc)); @@ -573,6 +556,36 @@ static int vc4_page_flip(struct drm_crtc *crtc, return drm_atomic_helper_page_flip(crtc, fb, event, flags); } +static struct drm_crtc_state *vc4_crtc_duplicate_state(struct drm_crtc *crtc) +{ + struct vc4_crtc_state *vc4_state; + + vc4_state = kzalloc(sizeof(*vc4_state), GFP_KERNEL); + if (!vc4_state) + return NULL; + + __drm_atomic_helper_crtc_duplicate_state(crtc, &vc4_state->base); + return &vc4_state->base; +} + +static void vc4_crtc_destroy_state(struct drm_crtc *crtc, + struct drm_crtc_state *state) +{ + struct vc4_dev *vc4 = to_vc4_dev(crtc->dev); + struct vc4_crtc_state *vc4_state = to_vc4_crtc_state(state); + + if (vc4_state->mm.allocated) { + unsigned long flags; + + spin_lock_irqsave(&vc4->hvs->mm_lock, flags); + drm_mm_remove_node(&vc4_state->mm); + spin_unlock_irqrestore(&vc4->hvs->mm_lock, flags); + + } + + __drm_atomic_helper_crtc_destroy_state(crtc, state); +} + static const struct drm_crtc_funcs vc4_crtc_funcs = { .set_config = drm_atomic_helper_set_config, .destroy = vc4_crtc_destroy, @@ -581,8 +594,8 @@ static const struct drm_crtc_funcs vc4_crtc_funcs = { .cursor_set = NULL, /* handled by drm_mode_cursor_universal */ .cursor_move = NULL, /* handled by drm_mode_cursor_universal */ .reset = drm_atomic_helper_crtc_reset, - .atomic_duplicate_state = drm_atomic_helper_crtc_duplicate_state, - .atomic_destroy_state = drm_atomic_helper_crtc_destroy_state, + .atomic_duplicate_state = vc4_crtc_duplicate_state, + .atomic_destroy_state = vc4_crtc_destroy_state, }; static const struct drm_crtc_helper_funcs vc4_crtc_helper_funcs = { @@ -593,26 +606,6 @@ static const struct drm_crtc_helper_funcs vc4_crtc_helper_funcs = { .atomic_flush = vc4_crtc_atomic_flush, }; -/* Frees the page flip event when the DRM device is closed with the - * event still outstanding. - */ -void vc4_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file) -{ - struct vc4_crtc *vc4_crtc = to_vc4_crtc(crtc); - struct drm_device *dev = crtc->dev; - unsigned long flags; - - spin_lock_irqsave(&dev->event_lock, flags); - - if (vc4_crtc->event && vc4_crtc->event->base.file_priv == file) { - vc4_crtc->event->base.destroy(&vc4_crtc->event->base); - drm_crtc_vblank_put(crtc); - vc4_crtc->event = NULL; - } - - spin_unlock_irqrestore(&dev->event_lock, flags); -} - static const struct vc4_crtc_data pv0_data = { .hvs_channel = 0, .encoder0_type = VC4_ENCODER_TYPE_DSI0, @@ -664,9 +657,9 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data) struct vc4_dev *vc4 = to_vc4_dev(drm); struct vc4_crtc *vc4_crtc; struct drm_crtc *crtc; - struct drm_plane *primary_plane, *cursor_plane; + struct drm_plane *primary_plane, *cursor_plane, *destroy_plane, *temp; const struct of_device_id *match; - int ret; + int ret, i; vc4_crtc = devm_kzalloc(dev, sizeof(*vc4_crtc), GFP_KERNEL); if (!vc4_crtc) @@ -695,27 +688,49 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data) goto err; } - cursor_plane = vc4_plane_init(drm, DRM_PLANE_TYPE_CURSOR); - if (IS_ERR(cursor_plane)) { - dev_err(dev, "failed to construct cursor plane\n"); - ret = PTR_ERR(cursor_plane); - goto err_primary; - } - - drm_crtc_init_with_planes(drm, crtc, primary_plane, cursor_plane, + drm_crtc_init_with_planes(drm, crtc, primary_plane, NULL, &vc4_crtc_funcs, NULL); drm_crtc_helper_add(crtc, &vc4_crtc_helper_funcs); primary_plane->crtc = crtc; - cursor_plane->crtc = crtc; vc4->crtc[drm_crtc_index(crtc)] = vc4_crtc; vc4_crtc->channel = vc4_crtc->data->hvs_channel; + /* Set up some arbitrary number of planes. We're not limited + * by a set number of physical registers, just the space in + * the HVS (16k) and how small an plane can be (28 bytes). + * However, each plane we set up takes up some memory, and + * increases the cost of looping over planes, which atomic + * modesetting does quite a bit. As a result, we pick a + * modest number of planes to expose, that should hopefully + * still cover any sane usecase. + */ + for (i = 0; i < 8; i++) { + struct drm_plane *plane = + vc4_plane_init(drm, DRM_PLANE_TYPE_OVERLAY); + + if (IS_ERR(plane)) + continue; + + plane->possible_crtcs = 1 << drm_crtc_index(crtc); + } + + /* Set up the legacy cursor after overlay initialization, + * since we overlay planes on the CRTC in the order they were + * initialized. + */ + cursor_plane = vc4_plane_init(drm, DRM_PLANE_TYPE_CURSOR); + if (!IS_ERR(cursor_plane)) { + cursor_plane->possible_crtcs = 1 << drm_crtc_index(crtc); + cursor_plane->crtc = crtc; + crtc->cursor = cursor_plane; + } + CRTC_WRITE(PV_INTEN, 0); CRTC_WRITE(PV_INTSTAT, PV_INT_VFP_START); ret = devm_request_irq(dev, platform_get_irq(pdev, 0), vc4_crtc_irq_handler, 0, "vc4 crtc", vc4_crtc); if (ret) - goto err_cursor; + goto err_destroy_planes; vc4_set_crtc_possible_masks(drm, crtc); @@ -723,10 +738,12 @@ static int vc4_crtc_bind(struct device *dev, struct device *master, void *data) return 0; -err_cursor: - cursor_plane->funcs->destroy(cursor_plane); -err_primary: - primary_plane->funcs->destroy(primary_plane); +err_destroy_planes: + list_for_each_entry_safe(destroy_plane, temp, + &drm->mode_config.plane_list, head) { + if (destroy_plane->possible_crtcs == 1 << drm_crtc_index(crtc)) + destroy_plane->funcs->destroy(destroy_plane); + } err: return ret; } diff --git a/drivers/gpu/drm/vc4/vc4_drv.c b/drivers/gpu/drm/vc4/vc4_drv.c index f1655fff8425..b7d2ff0e6e1f 100644 --- a/drivers/gpu/drm/vc4/vc4_drv.c +++ b/drivers/gpu/drm/vc4/vc4_drv.c @@ -43,14 +43,6 @@ void __iomem *vc4_ioremap_regs(struct platform_device *dev, int index) return map; } -static void vc4_drm_preclose(struct drm_device *dev, struct drm_file *file) -{ - struct drm_crtc *crtc; - - list_for_each_entry(crtc, &dev->mode_config.crtc_list, head) - vc4_cancel_page_flip(crtc, file); -} - static void vc4_lastclose(struct drm_device *dev) { struct vc4_dev *vc4 = to_vc4_dev(dev); @@ -91,8 +83,6 @@ static struct drm_driver vc4_drm_driver = { DRIVER_HAVE_IRQ | DRIVER_PRIME), .lastclose = vc4_lastclose, - .preclose = vc4_drm_preclose, - .irq_handler = vc4_irq, .irq_preinstall = vc4_irq_preinstall, .irq_postinstall = vc4_irq_postinstall, diff --git a/drivers/gpu/drm/vc4/vc4_drv.h b/drivers/gpu/drm/vc4/vc4_drv.h index 080865ec2bae..3d1df6b1c4d3 100644 --- a/drivers/gpu/drm/vc4/vc4_drv.h +++ b/drivers/gpu/drm/vc4/vc4_drv.h @@ -149,7 +149,17 @@ struct vc4_v3d { struct vc4_hvs { struct platform_device *pdev; void __iomem *regs; - void __iomem *dlist; + u32 __iomem *dlist; + + /* Memory manager for CRTCs to allocate space in the display + * list. Units are dwords. + */ + struct drm_mm dlist_mm; + /* Memory manager for the LBM memory used by HVS scaling. */ + struct drm_mm lbm_mm; + spinlock_t mm_lock; + + struct drm_mm_node mitchell_netravali_filter; }; struct vc4_plane { @@ -376,7 +386,6 @@ int vc4_bo_stats_debugfs(struct seq_file *m, void *arg); extern struct platform_driver vc4_crtc_driver; int vc4_enable_vblank(struct drm_device *dev, unsigned int crtc_id); void vc4_disable_vblank(struct drm_device *dev, unsigned int crtc_id); -void vc4_cancel_page_flip(struct drm_crtc *crtc, struct drm_file *file); int vc4_crtc_debugfs_regs(struct seq_file *m, void *arg); /* vc4_debugfs.c */ diff --git a/drivers/gpu/drm/vc4/vc4_hvs.c b/drivers/gpu/drm/vc4/vc4_hvs.c index 8098c5b21ba4..6fbab1c82cb1 100644 --- a/drivers/gpu/drm/vc4/vc4_hvs.c +++ b/drivers/gpu/drm/vc4/vc4_hvs.c @@ -100,12 +100,76 @@ int vc4_hvs_debugfs_regs(struct seq_file *m, void *unused) } #endif +/* The filter kernel is composed of dwords each containing 3 9-bit + * signed integers packed next to each other. + */ +#define VC4_INT_TO_COEFF(coeff) (coeff & 0x1ff) +#define VC4_PPF_FILTER_WORD(c0, c1, c2) \ + ((((c0) & 0x1ff) << 0) | \ + (((c1) & 0x1ff) << 9) | \ + (((c2) & 0x1ff) << 18)) + +/* The whole filter kernel is arranged as the coefficients 0-16 going + * up, then a pad, then 17-31 going down and reversed within the + * dwords. This means that a linear phase kernel (where it's + * symmetrical at the boundary between 15 and 16) has the last 5 + * dwords matching the first 5, but reversed. + */ +#define VC4_LINEAR_PHASE_KERNEL(c0, c1, c2, c3, c4, c5, c6, c7, c8, \ + c9, c10, c11, c12, c13, c14, c15) \ + {VC4_PPF_FILTER_WORD(c0, c1, c2), \ + VC4_PPF_FILTER_WORD(c3, c4, c5), \ + VC4_PPF_FILTER_WORD(c6, c7, c8), \ + VC4_PPF_FILTER_WORD(c9, c10, c11), \ + VC4_PPF_FILTER_WORD(c12, c13, c14), \ + VC4_PPF_FILTER_WORD(c15, c15, 0)} + +#define VC4_LINEAR_PHASE_KERNEL_DWORDS 6 +#define VC4_KERNEL_DWORDS (VC4_LINEAR_PHASE_KERNEL_DWORDS * 2 - 1) + +/* Recommended B=1/3, C=1/3 filter choice from Mitchell/Netravali. + * http://www.cs.utexas.edu/~fussell/courses/cs384g/lectures/mitchell/Mitchell.pdf + */ +static const u32 mitchell_netravali_1_3_1_3_kernel[] = + VC4_LINEAR_PHASE_KERNEL(0, -2, -6, -8, -10, -8, -3, 2, 18, + 50, 82, 119, 155, 187, 213, 227); + +static int vc4_hvs_upload_linear_kernel(struct vc4_hvs *hvs, + struct drm_mm_node *space, + const u32 *kernel) +{ + int ret, i; + u32 __iomem *dst_kernel; + + ret = drm_mm_insert_node(&hvs->dlist_mm, space, VC4_KERNEL_DWORDS, 1, + 0); + if (ret) { + DRM_ERROR("Failed to allocate space for filter kernel: %d\n", + ret); + return ret; + } + + dst_kernel = hvs->dlist + space->start; + + for (i = 0; i < VC4_KERNEL_DWORDS; i++) { + if (i < VC4_LINEAR_PHASE_KERNEL_DWORDS) + writel(kernel[i], &dst_kernel[i]); + else { + writel(kernel[VC4_KERNEL_DWORDS - i - 1], + &dst_kernel[i]); + } + } + + return 0; +} + static int vc4_hvs_bind(struct device *dev, struct device *master, void *data) { struct platform_device *pdev = to_platform_device(dev); struct drm_device *drm = dev_get_drvdata(master); struct vc4_dev *vc4 = drm->dev_private; struct vc4_hvs *hvs = NULL; + int ret; hvs = devm_kzalloc(&pdev->dev, sizeof(*hvs), GFP_KERNEL); if (!hvs) @@ -119,6 +183,33 @@ static int vc4_hvs_bind(struct device *dev, struct device *master, void *data) hvs->dlist = hvs->regs + SCALER_DLIST_START; + spin_lock_init(&hvs->mm_lock); + + /* Set up the HVS display list memory manager. We never + * overwrite the setup from the bootloader (just 128b out of + * our 16K), since we don't want to scramble the screen when + * transitioning from the firmware's boot setup to runtime. + */ + drm_mm_init(&hvs->dlist_mm, + HVS_BOOTLOADER_DLIST_END, + (SCALER_DLIST_SIZE >> 2) - HVS_BOOTLOADER_DLIST_END); + + /* Set up the HVS LBM memory manager. We could have some more + * complicated data structure that allowed reuse of LBM areas + * between planes when they don't overlap on the screen, but + * for now we just allocate globally. + */ + drm_mm_init(&hvs->lbm_mm, 0, 96 * 1024); + + /* Upload filter kernels. We only have the one for now, so we + * keep it around for the lifetime of the driver. + */ + ret = vc4_hvs_upload_linear_kernel(hvs, + &hvs->mitchell_netravali_filter, + mitchell_netravali_1_3_1_3_kernel); + if (ret) + return ret; + vc4->hvs = hvs; return 0; } @@ -129,6 +220,12 @@ static void vc4_hvs_unbind(struct device *dev, struct device *master, struct drm_device *drm = dev_get_drvdata(master); struct vc4_dev *vc4 = drm->dev_private; + if (vc4->hvs->mitchell_netravali_filter.allocated) + drm_mm_remove_node(&vc4->hvs->mitchell_netravali_filter); + + drm_mm_takedown(&vc4->hvs->dlist_mm); + drm_mm_takedown(&vc4->hvs->lbm_mm); + vc4->hvs = NULL; } diff --git a/drivers/gpu/drm/vc4/vc4_kms.c b/drivers/gpu/drm/vc4/vc4_kms.c index f95f2df5f8d1..4718ae5176cc 100644 --- a/drivers/gpu/drm/vc4/vc4_kms.c +++ b/drivers/gpu/drm/vc4/vc4_kms.c @@ -49,6 +49,15 @@ vc4_atomic_complete_commit(struct vc4_commit *c) drm_atomic_helper_commit_modeset_enables(dev, state); + /* Make sure that drm_atomic_helper_wait_for_vblanks() + * actually waits for vblank. If we're doing a full atomic + * modeset (as opposed to a vc4_update_plane() short circuit), + * then we need to wait for scanout to be done with our + * display lists before we free it and potentially reallocate + * and overwrite the dlist memory with a new modeset. + */ + state->legacy_cursor_update = false; + drm_atomic_helper_wait_for_vblanks(dev, state); drm_atomic_helper_cleanup_planes(dev, state); diff --git a/drivers/gpu/drm/vc4/vc4_plane.c b/drivers/gpu/drm/vc4/vc4_plane.c index 0addbad15832..7b0c72ae02a0 100644 --- a/drivers/gpu/drm/vc4/vc4_plane.c +++ b/drivers/gpu/drm/vc4/vc4_plane.c @@ -24,19 +24,52 @@ #include "drm_fb_cma_helper.h" #include "drm_plane_helper.h" +enum vc4_scaling_mode { + VC4_SCALING_NONE, + VC4_SCALING_TPZ, + VC4_SCALING_PPF, +}; + struct vc4_plane_state { struct drm_plane_state base; + /* System memory copy of the display list for this element, computed + * at atomic_check time. + */ u32 *dlist; - u32 dlist_size; /* Number of dwords in allocated for the display list */ + u32 dlist_size; /* Number of dwords allocated for the display list */ u32 dlist_count; /* Number of used dwords in the display list. */ - /* Offset in the dlist to pointer word 0. */ - u32 pw0_offset; + /* Offset in the dlist to various words, for pageflip or + * cursor updates. + */ + u32 pos0_offset; + u32 pos2_offset; + u32 ptr0_offset; /* Offset where the plane's dlist was last stored in the - hardware at vc4_crtc_atomic_flush() time. - */ - u32 *hw_dlist; + * hardware at vc4_crtc_atomic_flush() time. + */ + u32 __iomem *hw_dlist; + + /* Clipped coordinates of the plane on the display. */ + int crtc_x, crtc_y, crtc_w, crtc_h; + /* Clipped area being scanned from in the FB. */ + u32 src_x, src_y; + + u32 src_w[2], src_h[2]; + + /* Scaling selection for the RGB/Y plane and the Cb/Cr planes. */ + enum vc4_scaling_mode x_scaling[2], y_scaling[2]; + bool is_unity; + bool is_yuv; + + /* Offset to start scanning out from the start of the plane's + * BO. + */ + u32 offsets[3]; + + /* Our allocation in LBM for temporary storage during scaling. */ + struct drm_mm_node lbm; }; static inline struct vc4_plane_state * @@ -50,6 +83,7 @@ static const struct hvs_format { u32 hvs; /* HVS_FORMAT_* */ u32 pixel_order; bool has_alpha; + bool flip_cbcr; } hvs_formats[] = { { .drm = DRM_FORMAT_XRGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888, @@ -59,6 +93,48 @@ static const struct hvs_format { .drm = DRM_FORMAT_ARGB8888, .hvs = HVS_PIXEL_FORMAT_RGBA8888, .pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true, }, + { + .drm = DRM_FORMAT_RGB565, .hvs = HVS_PIXEL_FORMAT_RGB565, + .pixel_order = HVS_PIXEL_ORDER_XRGB, .has_alpha = false, + }, + { + .drm = DRM_FORMAT_BGR565, .hvs = HVS_PIXEL_FORMAT_RGB565, + .pixel_order = HVS_PIXEL_ORDER_XBGR, .has_alpha = false, + }, + { + .drm = DRM_FORMAT_ARGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551, + .pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = true, + }, + { + .drm = DRM_FORMAT_XRGB1555, .hvs = HVS_PIXEL_FORMAT_RGBA5551, + .pixel_order = HVS_PIXEL_ORDER_ABGR, .has_alpha = false, + }, + { + .drm = DRM_FORMAT_YUV422, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE, + }, + { + .drm = DRM_FORMAT_YVU422, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE, + .flip_cbcr = true, + }, + { + .drm = DRM_FORMAT_YUV420, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE, + }, + { + .drm = DRM_FORMAT_YVU420, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE, + .flip_cbcr = true, + }, + { + .drm = DRM_FORMAT_NV12, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE, + }, + { + .drm = DRM_FORMAT_NV16, + .hvs = HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE, + }, }; static const struct hvs_format *vc4_get_hvs_format(u32 drm_format) @@ -73,6 +149,16 @@ static const struct hvs_format *vc4_get_hvs_format(u32 drm_format) return NULL; } +static enum vc4_scaling_mode vc4_get_scaling_mode(u32 src, u32 dst) +{ + if (dst > src) + return VC4_SCALING_PPF; + else if (dst < src) + return VC4_SCALING_TPZ; + else + return VC4_SCALING_NONE; +} + static bool plane_enabled(struct drm_plane_state *state) { return state->fb && state->crtc; @@ -89,6 +175,8 @@ static struct drm_plane_state *vc4_plane_duplicate_state(struct drm_plane *plane if (!vc4_state) return NULL; + memset(&vc4_state->lbm, 0, sizeof(vc4_state->lbm)); + __drm_atomic_helper_plane_duplicate_state(plane, &vc4_state->base); if (vc4_state->dlist) { @@ -108,8 +196,17 @@ static struct drm_plane_state *vc4_plane_duplicate_state(struct drm_plane *plane static void vc4_plane_destroy_state(struct drm_plane *plane, struct drm_plane_state *state) { + struct vc4_dev *vc4 = to_vc4_dev(plane->dev); struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); + if (vc4_state->lbm.allocated) { + unsigned long irqflags; + + spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags); + drm_mm_remove_node(&vc4_state->lbm); + spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags); + } + kfree(vc4_state->dlist); __drm_atomic_helper_plane_destroy_state(plane, &vc4_state->base); kfree(state); @@ -148,84 +245,400 @@ static void vc4_dlist_write(struct vc4_plane_state *vc4_state, u32 val) vc4_state->dlist[vc4_state->dlist_count++] = val; } +/* Returns the scl0/scl1 field based on whether the dimensions need to + * be up/down/non-scaled. + * + * This is a replication of a table from the spec. + */ +static u32 vc4_get_scl_field(struct drm_plane_state *state, int plane) +{ + struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); + + switch (vc4_state->x_scaling[plane] << 2 | vc4_state->y_scaling[plane]) { + case VC4_SCALING_PPF << 2 | VC4_SCALING_PPF: + return SCALER_CTL0_SCL_H_PPF_V_PPF; + case VC4_SCALING_TPZ << 2 | VC4_SCALING_PPF: + return SCALER_CTL0_SCL_H_TPZ_V_PPF; + case VC4_SCALING_PPF << 2 | VC4_SCALING_TPZ: + return SCALER_CTL0_SCL_H_PPF_V_TPZ; + case VC4_SCALING_TPZ << 2 | VC4_SCALING_TPZ: + return SCALER_CTL0_SCL_H_TPZ_V_TPZ; + case VC4_SCALING_PPF << 2 | VC4_SCALING_NONE: + return SCALER_CTL0_SCL_H_PPF_V_NONE; + case VC4_SCALING_NONE << 2 | VC4_SCALING_PPF: + return SCALER_CTL0_SCL_H_NONE_V_PPF; + case VC4_SCALING_NONE << 2 | VC4_SCALING_TPZ: + return SCALER_CTL0_SCL_H_NONE_V_TPZ; + case VC4_SCALING_TPZ << 2 | VC4_SCALING_NONE: + return SCALER_CTL0_SCL_H_TPZ_V_NONE; + default: + case VC4_SCALING_NONE << 2 | VC4_SCALING_NONE: + /* The unity case is independently handled by + * SCALER_CTL0_UNITY. + */ + return 0; + } +} + +static int vc4_plane_setup_clipping_and_scaling(struct drm_plane_state *state) +{ + struct drm_plane *plane = state->plane; + struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); + struct drm_framebuffer *fb = state->fb; + struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0); + u32 subpixel_src_mask = (1 << 16) - 1; + u32 format = fb->pixel_format; + int num_planes = drm_format_num_planes(format); + u32 h_subsample = 1; + u32 v_subsample = 1; + int i; + + for (i = 0; i < num_planes; i++) + vc4_state->offsets[i] = bo->paddr + fb->offsets[i]; + + /* We don't support subpixel source positioning for scaling. */ + if ((state->src_x & subpixel_src_mask) || + (state->src_y & subpixel_src_mask) || + (state->src_w & subpixel_src_mask) || + (state->src_h & subpixel_src_mask)) { + return -EINVAL; + } + + vc4_state->src_x = state->src_x >> 16; + vc4_state->src_y = state->src_y >> 16; + vc4_state->src_w[0] = state->src_w >> 16; + vc4_state->src_h[0] = state->src_h >> 16; + + vc4_state->crtc_x = state->crtc_x; + vc4_state->crtc_y = state->crtc_y; + vc4_state->crtc_w = state->crtc_w; + vc4_state->crtc_h = state->crtc_h; + + vc4_state->x_scaling[0] = vc4_get_scaling_mode(vc4_state->src_w[0], + vc4_state->crtc_w); + vc4_state->y_scaling[0] = vc4_get_scaling_mode(vc4_state->src_h[0], + vc4_state->crtc_h); + + if (num_planes > 1) { + vc4_state->is_yuv = true; + + h_subsample = drm_format_horz_chroma_subsampling(format); + v_subsample = drm_format_vert_chroma_subsampling(format); + vc4_state->src_w[1] = vc4_state->src_w[0] / h_subsample; + vc4_state->src_h[1] = vc4_state->src_h[0] / v_subsample; + + vc4_state->x_scaling[1] = + vc4_get_scaling_mode(vc4_state->src_w[1], + vc4_state->crtc_w); + vc4_state->y_scaling[1] = + vc4_get_scaling_mode(vc4_state->src_h[1], + vc4_state->crtc_h); + + /* YUV conversion requires that scaling be enabled, + * even on a plane that's otherwise 1:1. Choose TPZ + * for simplicity. + */ + if (vc4_state->x_scaling[0] == VC4_SCALING_NONE) + vc4_state->x_scaling[0] = VC4_SCALING_TPZ; + if (vc4_state->y_scaling[0] == VC4_SCALING_NONE) + vc4_state->y_scaling[0] = VC4_SCALING_TPZ; + } + + vc4_state->is_unity = (vc4_state->x_scaling[0] == VC4_SCALING_NONE && + vc4_state->y_scaling[0] == VC4_SCALING_NONE && + vc4_state->x_scaling[1] == VC4_SCALING_NONE && + vc4_state->y_scaling[1] == VC4_SCALING_NONE); + + /* No configuring scaling on the cursor plane, since it gets + non-vblank-synced updates, and scaling requires requires + LBM changes which have to be vblank-synced. + */ + if (plane->type == DRM_PLANE_TYPE_CURSOR && !vc4_state->is_unity) + return -EINVAL; + + /* Clamp the on-screen start x/y to 0. The hardware doesn't + * support negative y, and negative x wastes bandwidth. + */ + if (vc4_state->crtc_x < 0) { + for (i = 0; i < num_planes; i++) { + u32 cpp = drm_format_plane_cpp(fb->pixel_format, i); + u32 subs = ((i == 0) ? 1 : h_subsample); + + vc4_state->offsets[i] += (cpp * + (-vc4_state->crtc_x) / subs); + } + vc4_state->src_w[0] += vc4_state->crtc_x; + vc4_state->src_w[1] += vc4_state->crtc_x / h_subsample; + vc4_state->crtc_x = 0; + } + + if (vc4_state->crtc_y < 0) { + for (i = 0; i < num_planes; i++) { + u32 subs = ((i == 0) ? 1 : v_subsample); + + vc4_state->offsets[i] += (fb->pitches[i] * + (-vc4_state->crtc_y) / subs); + } + vc4_state->src_h[0] += vc4_state->crtc_y; + vc4_state->src_h[1] += vc4_state->crtc_y / v_subsample; + vc4_state->crtc_y = 0; + } + + return 0; +} + +static void vc4_write_tpz(struct vc4_plane_state *vc4_state, u32 src, u32 dst) +{ + u32 scale, recip; + + scale = (1 << 16) * src / dst; + + /* The specs note that while the reciprocal would be defined + * as (1<<32)/scale, ~0 is close enough. + */ + recip = ~0 / scale; + + vc4_dlist_write(vc4_state, + VC4_SET_FIELD(scale, SCALER_TPZ0_SCALE) | + VC4_SET_FIELD(0, SCALER_TPZ0_IPHASE)); + vc4_dlist_write(vc4_state, + VC4_SET_FIELD(recip, SCALER_TPZ1_RECIP)); +} + +static void vc4_write_ppf(struct vc4_plane_state *vc4_state, u32 src, u32 dst) +{ + u32 scale = (1 << 16) * src / dst; + + vc4_dlist_write(vc4_state, + SCALER_PPF_AGC | + VC4_SET_FIELD(scale, SCALER_PPF_SCALE) | + VC4_SET_FIELD(0, SCALER_PPF_IPHASE)); +} + +static u32 vc4_lbm_size(struct drm_plane_state *state) +{ + struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); + /* This is the worst case number. One of the two sizes will + * be used depending on the scaling configuration. + */ + u32 pix_per_line = max(vc4_state->src_w[0], (u32)vc4_state->crtc_w); + u32 lbm; + + if (!vc4_state->is_yuv) { + if (vc4_state->is_unity) + return 0; + else if (vc4_state->y_scaling[0] == VC4_SCALING_TPZ) + lbm = pix_per_line * 8; + else { + /* In special cases, this multiplier might be 12. */ + lbm = pix_per_line * 16; + } + } else { + /* There are cases for this going down to a multiplier + * of 2, but according to the firmware source, the + * table in the docs is somewhat wrong. + */ + lbm = pix_per_line * 16; + } + + lbm = roundup(lbm, 32); + + return lbm; +} + +static void vc4_write_scaling_parameters(struct drm_plane_state *state, + int channel) +{ + struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); + + /* Ch0 H-PPF Word 0: Scaling Parameters */ + if (vc4_state->x_scaling[channel] == VC4_SCALING_PPF) { + vc4_write_ppf(vc4_state, + vc4_state->src_w[channel], vc4_state->crtc_w); + } + + /* Ch0 V-PPF Words 0-1: Scaling Parameters, Context */ + if (vc4_state->y_scaling[channel] == VC4_SCALING_PPF) { + vc4_write_ppf(vc4_state, + vc4_state->src_h[channel], vc4_state->crtc_h); + vc4_dlist_write(vc4_state, 0xc0c0c0c0); + } + + /* Ch0 H-TPZ Words 0-1: Scaling Parameters, Recip */ + if (vc4_state->x_scaling[channel] == VC4_SCALING_TPZ) { + vc4_write_tpz(vc4_state, + vc4_state->src_w[channel], vc4_state->crtc_w); + } + + /* Ch0 V-TPZ Words 0-2: Scaling Parameters, Recip, Context */ + if (vc4_state->y_scaling[channel] == VC4_SCALING_TPZ) { + vc4_write_tpz(vc4_state, + vc4_state->src_h[channel], vc4_state->crtc_h); + vc4_dlist_write(vc4_state, 0xc0c0c0c0); + } +} + /* Writes out a full display list for an active plane to the plane's * private dlist state. */ static int vc4_plane_mode_set(struct drm_plane *plane, struct drm_plane_state *state) { + struct vc4_dev *vc4 = to_vc4_dev(plane->dev); struct vc4_plane_state *vc4_state = to_vc4_plane_state(state); struct drm_framebuffer *fb = state->fb; - struct drm_gem_cma_object *bo = drm_fb_cma_get_gem_obj(fb, 0); u32 ctl0_offset = vc4_state->dlist_count; const struct hvs_format *format = vc4_get_hvs_format(fb->pixel_format); - uint32_t offset = fb->offsets[0]; - int crtc_x = state->crtc_x; - int crtc_y = state->crtc_y; - int crtc_w = state->crtc_w; - int crtc_h = state->crtc_h; - - if (state->crtc_w << 16 != state->src_w || - state->crtc_h << 16 != state->src_h) { - /* We don't support scaling yet, which involves - * allocating the LBM memory for scaling temporary - * storage, and putting filter kernels in the HVS - * context. - */ - return -EINVAL; + int num_planes = drm_format_num_planes(format->drm); + u32 scl0, scl1; + u32 lbm_size; + unsigned long irqflags; + int ret, i; + + ret = vc4_plane_setup_clipping_and_scaling(state); + if (ret) + return ret; + + /* Allocate the LBM memory that the HVS will use for temporary + * storage due to our scaling/format conversion. + */ + lbm_size = vc4_lbm_size(state); + if (lbm_size) { + if (!vc4_state->lbm.allocated) { + spin_lock_irqsave(&vc4->hvs->mm_lock, irqflags); + ret = drm_mm_insert_node(&vc4->hvs->lbm_mm, + &vc4_state->lbm, + lbm_size, 32, 0); + spin_unlock_irqrestore(&vc4->hvs->mm_lock, irqflags); + } else { + WARN_ON_ONCE(lbm_size != vc4_state->lbm.size); + } } - if (crtc_x < 0) { - offset += drm_format_plane_cpp(fb->pixel_format, 0) * -crtc_x; - crtc_w += crtc_x; - crtc_x = 0; - } + if (ret) + return ret; - if (crtc_y < 0) { - offset += fb->pitches[0] * -crtc_y; - crtc_h += crtc_y; - crtc_y = 0; + /* SCL1 is used for Cb/Cr scaling of planar formats. For RGB + * and 4:4:4, scl1 should be set to scl0 so both channels of + * the scaler do the same thing. For YUV, the Y plane needs + * to be put in channel 1 and Cb/Cr in channel 0, so we swap + * the scl fields here. + */ + if (num_planes == 1) { + scl0 = vc4_get_scl_field(state, 1); + scl1 = scl0; + } else { + scl0 = vc4_get_scl_field(state, 1); + scl1 = vc4_get_scl_field(state, 0); } + /* Control word */ vc4_dlist_write(vc4_state, SCALER_CTL0_VALID | (format->pixel_order << SCALER_CTL0_ORDER_SHIFT) | (format->hvs << SCALER_CTL0_PIXEL_FORMAT_SHIFT) | - SCALER_CTL0_UNITY); + (vc4_state->is_unity ? SCALER_CTL0_UNITY : 0) | + VC4_SET_FIELD(scl0, SCALER_CTL0_SCL0) | + VC4_SET_FIELD(scl1, SCALER_CTL0_SCL1)); /* Position Word 0: Image Positions and Alpha Value */ + vc4_state->pos0_offset = vc4_state->dlist_count; vc4_dlist_write(vc4_state, VC4_SET_FIELD(0xff, SCALER_POS0_FIXED_ALPHA) | - VC4_SET_FIELD(crtc_x, SCALER_POS0_START_X) | - VC4_SET_FIELD(crtc_y, SCALER_POS0_START_Y)); - - /* Position Word 1: Scaled Image Dimensions. - * Skipped due to SCALER_CTL0_UNITY scaling. - */ + VC4_SET_FIELD(vc4_state->crtc_x, SCALER_POS0_START_X) | + VC4_SET_FIELD(vc4_state->crtc_y, SCALER_POS0_START_Y)); + + /* Position Word 1: Scaled Image Dimensions. */ + if (!vc4_state->is_unity) { + vc4_dlist_write(vc4_state, + VC4_SET_FIELD(vc4_state->crtc_w, + SCALER_POS1_SCL_WIDTH) | + VC4_SET_FIELD(vc4_state->crtc_h, + SCALER_POS1_SCL_HEIGHT)); + } /* Position Word 2: Source Image Size, Alpha Mode */ + vc4_state->pos2_offset = vc4_state->dlist_count; vc4_dlist_write(vc4_state, VC4_SET_FIELD(format->has_alpha ? SCALER_POS2_ALPHA_MODE_PIPELINE : SCALER_POS2_ALPHA_MODE_FIXED, SCALER_POS2_ALPHA_MODE) | - VC4_SET_FIELD(crtc_w, SCALER_POS2_WIDTH) | - VC4_SET_FIELD(crtc_h, SCALER_POS2_HEIGHT)); + VC4_SET_FIELD(vc4_state->src_w[0], SCALER_POS2_WIDTH) | + VC4_SET_FIELD(vc4_state->src_h[0], SCALER_POS2_HEIGHT)); /* Position Word 3: Context. Written by the HVS. */ vc4_dlist_write(vc4_state, 0xc0c0c0c0); - vc4_state->pw0_offset = vc4_state->dlist_count; - /* Pointer Word 0: RGB / Y Pointer */ - vc4_dlist_write(vc4_state, bo->paddr + offset); + /* Pointer Word 0/1/2: RGB / Y / Cb / Cr Pointers + * + * The pointers may be any byte address. + */ + vc4_state->ptr0_offset = vc4_state->dlist_count; + if (!format->flip_cbcr) { + for (i = 0; i < num_planes; i++) + vc4_dlist_write(vc4_state, vc4_state->offsets[i]); + } else { + WARN_ON_ONCE(num_planes != 3); + vc4_dlist_write(vc4_state, vc4_state->offsets[0]); + vc4_dlist_write(vc4_state, vc4_state->offsets[2]); + vc4_dlist_write(vc4_state, vc4_state->offsets[1]); + } - /* Pointer Context Word 0: Written by the HVS */ - vc4_dlist_write(vc4_state, 0xc0c0c0c0); + /* Pointer Context Word 0/1/2: Written by the HVS */ + for (i = 0; i < num_planes; i++) + vc4_dlist_write(vc4_state, 0xc0c0c0c0); - /* Pitch word 0: Pointer 0 Pitch */ - vc4_dlist_write(vc4_state, - VC4_SET_FIELD(fb->pitches[0], SCALER_SRC_PITCH)); + /* Pitch word 0/1/2 */ + for (i = 0; i < num_planes; i++) { + vc4_dlist_write(vc4_state, + VC4_SET_FIELD(fb->pitches[i], SCALER_SRC_PITCH)); + } + + /* Colorspace conversion words */ + if (vc4_state->is_yuv) { + vc4_dlist_write(vc4_state, SCALER_CSC0_ITR_R_601_5); + vc4_dlist_write(vc4_state, SCALER_CSC1_ITR_R_601_5); + vc4_dlist_write(vc4_state, SCALER_CSC2_ITR_R_601_5); + } + + if (!vc4_state->is_unity) { + /* LBM Base Address. */ + if (vc4_state->y_scaling[0] != VC4_SCALING_NONE || + vc4_state->y_scaling[1] != VC4_SCALING_NONE) { + vc4_dlist_write(vc4_state, vc4_state->lbm.start); + } + + if (num_planes > 1) { + /* Emit Cb/Cr as channel 0 and Y as channel + * 1. This matches how we set up scl0/scl1 + * above. + */ + vc4_write_scaling_parameters(state, 1); + } + vc4_write_scaling_parameters(state, 0); + + /* If any PPF setup was done, then all the kernel + * pointers get uploaded. + */ + if (vc4_state->x_scaling[0] == VC4_SCALING_PPF || + vc4_state->y_scaling[0] == VC4_SCALING_PPF || + vc4_state->x_scaling[1] == VC4_SCALING_PPF || + vc4_state->y_scaling[1] == VC4_SCALING_PPF) { + u32 kernel = VC4_SET_FIELD(vc4->hvs->mitchell_netravali_filter.start, + SCALER_PPF_KERNEL_OFFSET); + + /* HPPF plane 0 */ + vc4_dlist_write(vc4_state, kernel); + /* VPPF plane 0 */ + vc4_dlist_write(vc4_state, kernel); + /* HPPF plane 1 */ + vc4_dlist_write(vc4_state, kernel); + /* VPPF plane 1 */ + vc4_dlist_write(vc4_state, kernel); + } + } vc4_state->dlist[ctl0_offset] |= VC4_SET_FIELD(vc4_state->dlist_count, SCALER_CTL0_SIZE); @@ -303,13 +716,13 @@ void vc4_plane_async_set_fb(struct drm_plane *plane, struct drm_framebuffer *fb) * scanout will start from this address as soon as the FIFO * needs to refill with pixels. */ - writel(addr, &vc4_state->hw_dlist[vc4_state->pw0_offset]); + writel(addr, &vc4_state->hw_dlist[vc4_state->ptr0_offset]); /* Also update the CPU-side dlist copy, so that any later * atomic updates that don't do a new modeset on our plane * also use our updated address. */ - vc4_state->dlist[vc4_state->pw0_offset] = addr; + vc4_state->dlist[vc4_state->ptr0_offset] = addr; } static const struct drm_plane_helper_funcs vc4_plane_helper_funcs = { @@ -325,8 +738,83 @@ static void vc4_plane_destroy(struct drm_plane *plane) drm_plane_cleanup(plane); } +/* Implements immediate (non-vblank-synced) updates of the cursor + * position, or falls back to the atomic helper otherwise. + */ +static int +vc4_update_plane(struct drm_plane *plane, + struct drm_crtc *crtc, + struct drm_framebuffer *fb, + int crtc_x, int crtc_y, + unsigned int crtc_w, unsigned int crtc_h, + uint32_t src_x, uint32_t src_y, + uint32_t src_w, uint32_t src_h) +{ + struct drm_plane_state *plane_state; + struct vc4_plane_state *vc4_state; + + if (plane != crtc->cursor) + goto out; + + plane_state = plane->state; + vc4_state = to_vc4_plane_state(plane_state); + + if (!plane_state) + goto out; + + /* If we're changing the cursor contents, do that in the + * normal vblank-synced atomic path. + */ + if (fb != plane_state->fb) + goto out; + + /* No configuring new scaling in the fast path. */ + if (crtc_w != plane_state->crtc_w || + crtc_h != plane_state->crtc_h || + src_w != plane_state->src_w || + src_h != plane_state->src_h) { + goto out; + } + + /* Set the cursor's position on the screen. This is the + * expected change from the drm_mode_cursor_universal() + * helper. + */ + plane_state->crtc_x = crtc_x; + plane_state->crtc_y = crtc_y; + + /* Allow changing the start position within the cursor BO, if + * that matters. + */ + plane_state->src_x = src_x; + plane_state->src_y = src_y; + + /* Update the display list based on the new crtc_x/y. */ + vc4_plane_atomic_check(plane, plane_state); + + /* Note that we can't just call vc4_plane_write_dlist() + * because that would smash the context data that the HVS is + * currently using. + */ + writel(vc4_state->dlist[vc4_state->pos0_offset], + &vc4_state->hw_dlist[vc4_state->pos0_offset]); + writel(vc4_state->dlist[vc4_state->pos2_offset], + &vc4_state->hw_dlist[vc4_state->pos2_offset]); + writel(vc4_state->dlist[vc4_state->ptr0_offset], + &vc4_state->hw_dlist[vc4_state->ptr0_offset]); + + return 0; + +out: + return drm_atomic_helper_update_plane(plane, crtc, fb, + crtc_x, crtc_y, + crtc_w, crtc_h, + src_x, src_y, + src_w, src_h); +} + static const struct drm_plane_funcs vc4_plane_funcs = { - .update_plane = drm_atomic_helper_update_plane, + .update_plane = vc4_update_plane, .disable_plane = drm_atomic_helper_disable_plane, .destroy = vc4_plane_destroy, .set_property = NULL, @@ -341,6 +829,7 @@ struct drm_plane *vc4_plane_init(struct drm_device *dev, struct drm_plane *plane = NULL; struct vc4_plane *vc4_plane; u32 formats[ARRAY_SIZE(hvs_formats)]; + u32 num_formats = 0; int ret = 0; unsigned i; @@ -351,12 +840,20 @@ struct drm_plane *vc4_plane_init(struct drm_device *dev, goto fail; } - for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) - formats[i] = hvs_formats[i].drm; + for (i = 0; i < ARRAY_SIZE(hvs_formats); i++) { + /* Don't allow YUV in cursor planes, since that means + * tuning on the scaler, which we don't allow for the + * cursor. + */ + if (type != DRM_PLANE_TYPE_CURSOR || + hvs_formats[i].hvs < HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE) { + formats[num_formats++] = hvs_formats[i].drm; + } + } plane = &vc4_plane->base; ret = drm_universal_plane_init(dev, plane, 0xff, &vc4_plane_funcs, - formats, ARRAY_SIZE(formats), + formats, num_formats, type, NULL); drm_plane_helper_add(plane, &vc4_plane_helper_funcs); diff --git a/drivers/gpu/drm/vc4/vc4_regs.h b/drivers/gpu/drm/vc4/vc4_regs.h index 4e52a0a88551..25df20ef939c 100644 --- a/drivers/gpu/drm/vc4/vc4_regs.h +++ b/drivers/gpu/drm/vc4/vc4_regs.h @@ -503,7 +503,12 @@ enum hvs_pixel_format { HVS_PIXEL_FORMAT_RGB888 = 5, HVS_PIXEL_FORMAT_RGBA6666 = 6, /* 32bpp */ - HVS_PIXEL_FORMAT_RGBA8888 = 7 + HVS_PIXEL_FORMAT_RGBA8888 = 7, + + HVS_PIXEL_FORMAT_YCBCR_YUV420_3PLANE = 8, + HVS_PIXEL_FORMAT_YCBCR_YUV420_2PLANE = 9, + HVS_PIXEL_FORMAT_YCBCR_YUV422_3PLANE = 10, + HVS_PIXEL_FORMAT_YCBCR_YUV422_2PLANE = 11, }; /* Note: the LSB is the rightmost character shown. Only valid for @@ -536,6 +541,21 @@ enum hvs_pixel_format { #define SCALER_CTL0_ORDER_MASK VC4_MASK(14, 13) #define SCALER_CTL0_ORDER_SHIFT 13 +#define SCALER_CTL0_SCL1_MASK VC4_MASK(10, 8) +#define SCALER_CTL0_SCL1_SHIFT 8 + +#define SCALER_CTL0_SCL0_MASK VC4_MASK(7, 5) +#define SCALER_CTL0_SCL0_SHIFT 5 + +#define SCALER_CTL0_SCL_H_PPF_V_PPF 0 +#define SCALER_CTL0_SCL_H_TPZ_V_PPF 1 +#define SCALER_CTL0_SCL_H_PPF_V_TPZ 2 +#define SCALER_CTL0_SCL_H_TPZ_V_TPZ 3 +#define SCALER_CTL0_SCL_H_PPF_V_NONE 4 +#define SCALER_CTL0_SCL_H_NONE_V_PPF 5 +#define SCALER_CTL0_SCL_H_NONE_V_TPZ 6 +#define SCALER_CTL0_SCL_H_TPZ_V_NONE 7 + /* Set to indicate no scaling. */ #define SCALER_CTL0_UNITY BIT(4) @@ -551,6 +571,12 @@ enum hvs_pixel_format { #define SCALER_POS0_START_X_MASK VC4_MASK(11, 0) #define SCALER_POS0_START_X_SHIFT 0 +#define SCALER_POS1_SCL_HEIGHT_MASK VC4_MASK(27, 16) +#define SCALER_POS1_SCL_HEIGHT_SHIFT 16 + +#define SCALER_POS1_SCL_WIDTH_MASK VC4_MASK(11, 0) +#define SCALER_POS1_SCL_WIDTH_SHIFT 0 + #define SCALER_POS2_ALPHA_MODE_MASK VC4_MASK(31, 30) #define SCALER_POS2_ALPHA_MODE_SHIFT 30 #define SCALER_POS2_ALPHA_MODE_PIPELINE 0 @@ -564,6 +590,80 @@ enum hvs_pixel_format { #define SCALER_POS2_WIDTH_MASK VC4_MASK(11, 0) #define SCALER_POS2_WIDTH_SHIFT 0 +/* Color Space Conversion words. Some values are S2.8 signed + * integers, except that the 2 integer bits map as {0x0: 0, 0x1: 1, + * 0x2: 2, 0x3: -1} + */ +/* bottom 8 bits of S2.8 contribution of Cr to Blue */ +#define SCALER_CSC0_COEF_CR_BLU_MASK VC4_MASK(31, 24) +#define SCALER_CSC0_COEF_CR_BLU_SHIFT 24 +/* Signed offset to apply to Y before CSC. (Y' = Y + YY_OFS) */ +#define SCALER_CSC0_COEF_YY_OFS_MASK VC4_MASK(23, 16) +#define SCALER_CSC0_COEF_YY_OFS_SHIFT 16 +/* Signed offset to apply to CB before CSC (Cb' = Cb - 128 + CB_OFS). */ +#define SCALER_CSC0_COEF_CB_OFS_MASK VC4_MASK(15, 8) +#define SCALER_CSC0_COEF_CB_OFS_SHIFT 8 +/* Signed offset to apply to CB before CSC (Cr' = Cr - 128 + CR_OFS). */ +#define SCALER_CSC0_COEF_CR_OFS_MASK VC4_MASK(7, 0) +#define SCALER_CSC0_COEF_CR_OFS_SHIFT 0 +#define SCALER_CSC0_ITR_R_601_5 0x00f00000 +#define SCALER_CSC0_ITR_R_709_3 0x00f00000 +#define SCALER_CSC0_JPEG_JFIF 0x00000000 + +/* S2.8 contribution of Cb to Green */ +#define SCALER_CSC1_COEF_CB_GRN_MASK VC4_MASK(31, 22) +#define SCALER_CSC1_COEF_CB_GRN_SHIFT 22 +/* S2.8 contribution of Cr to Green */ +#define SCALER_CSC1_COEF_CR_GRN_MASK VC4_MASK(21, 12) +#define SCALER_CSC1_COEF_CR_GRN_SHIFT 12 +/* S2.8 contribution of Y to all of RGB */ +#define SCALER_CSC1_COEF_YY_ALL_MASK VC4_MASK(11, 2) +#define SCALER_CSC1_COEF_YY_ALL_SHIFT 2 +/* top 2 bits of S2.8 contribution of Cr to Blue */ +#define SCALER_CSC1_COEF_CR_BLU_MASK VC4_MASK(1, 0) +#define SCALER_CSC1_COEF_CR_BLU_SHIFT 0 +#define SCALER_CSC1_ITR_R_601_5 0xe73304a8 +#define SCALER_CSC1_ITR_R_709_3 0xf2b784a8 +#define SCALER_CSC1_JPEG_JFIF 0xea34a400 + +/* S2.8 contribution of Cb to Red */ +#define SCALER_CSC2_COEF_CB_RED_MASK VC4_MASK(29, 20) +#define SCALER_CSC2_COEF_CB_RED_SHIFT 20 +/* S2.8 contribution of Cr to Red */ +#define SCALER_CSC2_COEF_CR_RED_MASK VC4_MASK(19, 10) +#define SCALER_CSC2_COEF_CR_RED_SHIFT 10 +/* S2.8 contribution of Cb to Blue */ +#define SCALER_CSC2_COEF_CB_BLU_MASK VC4_MASK(19, 10) +#define SCALER_CSC2_COEF_CB_BLU_SHIFT 10 +#define SCALER_CSC2_ITR_R_601_5 0x00066204 +#define SCALER_CSC2_ITR_R_709_3 0x00072a1c +#define SCALER_CSC2_JPEG_JFIF 0x000599c5 + +#define SCALER_TPZ0_VERT_RECALC BIT(31) +#define SCALER_TPZ0_SCALE_MASK VC4_MASK(28, 8) +#define SCALER_TPZ0_SCALE_SHIFT 8 +#define SCALER_TPZ0_IPHASE_MASK VC4_MASK(7, 0) +#define SCALER_TPZ0_IPHASE_SHIFT 0 +#define SCALER_TPZ1_RECIP_MASK VC4_MASK(15, 0) +#define SCALER_TPZ1_RECIP_SHIFT 0 + +/* Skips interpolating coefficients to 64 phases, so just 8 are used. + * Required for nearest neighbor. + */ +#define SCALER_PPF_NOINTERP BIT(31) +/* Replaes the highest valued coefficient with one that makes all 4 + * sum to unity. + */ +#define SCALER_PPF_AGC BIT(30) +#define SCALER_PPF_SCALE_MASK VC4_MASK(24, 8) +#define SCALER_PPF_SCALE_SHIFT 8 +#define SCALER_PPF_IPHASE_MASK VC4_MASK(6, 0) +#define SCALER_PPF_IPHASE_SHIFT 0 + +#define SCALER_PPF_KERNEL_OFFSET_MASK VC4_MASK(13, 0) +#define SCALER_PPF_KERNEL_OFFSET_SHIFT 0 +#define SCALER_PPF_KERNEL_UNCACHED BIT(31) + #define SCALER_SRC_PITCH_MASK VC4_MASK(15, 0) #define SCALER_SRC_PITCH_SHIFT 0 diff --git a/drivers/gpu/drm/virtio/virtgpu_display.c b/drivers/gpu/drm/virtio/virtgpu_display.c index a165f03eaa79..429aa311685a 100644 --- a/drivers/gpu/drm/virtio/virtgpu_display.c +++ b/drivers/gpu/drm/virtio/virtgpu_display.c @@ -282,13 +282,6 @@ static const struct drm_crtc_helper_funcs virtio_gpu_crtc_helper_funcs = { .atomic_check = virtio_gpu_crtc_atomic_check, }; -static bool virtio_gpu_enc_mode_fixup(struct drm_encoder *encoder, - const struct drm_display_mode *mode, - struct drm_display_mode *adjusted_mode) -{ - return true; -} - static void virtio_gpu_enc_mode_set(struct drm_encoder *encoder, struct drm_display_mode *mode, struct drm_display_mode *adjusted_mode) @@ -362,7 +355,6 @@ virtio_gpu_best_encoder(struct drm_connector *connector) } static const struct drm_encoder_helper_funcs virtio_gpu_enc_helper_funcs = { - .mode_fixup = virtio_gpu_enc_mode_fixup, .mode_set = virtio_gpu_enc_mode_set, .enable = virtio_gpu_enc_enable, .disable = virtio_gpu_enc_disable, diff --git a/drivers/gpu/drm/virtio/virtgpu_drv.c b/drivers/gpu/drm/virtio/virtgpu_drv.c index b40ed6061f05..7f898cfdc746 100644 --- a/drivers/gpu/drm/virtio/virtgpu_drv.c +++ b/drivers/gpu/drm/virtio/virtgpu_drv.c @@ -118,7 +118,7 @@ static const struct file_operations virtio_gpu_driver_fops = { static struct drm_driver driver = { - .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | DRIVER_RENDER, + .driver_features = DRIVER_MODESET | DRIVER_GEM | DRIVER_PRIME | DRIVER_RENDER | DRIVER_ATOMIC, .set_busid = drm_virtio_set_busid, .load = virtio_gpu_driver_load, .unload = virtio_gpu_driver_unload, diff --git a/drivers/gpu/drm/virtio/virtgpu_plane.c b/drivers/gpu/drm/virtio/virtgpu_plane.c index 572fb351feab..70b44a2345ab 100644 --- a/drivers/gpu/drm/virtio/virtgpu_plane.c +++ b/drivers/gpu/drm/virtio/virtgpu_plane.c @@ -68,10 +68,17 @@ static void virtio_gpu_plane_atomic_update(struct drm_plane *plane, struct virtio_gpu_object *bo; uint32_t handle; - if (plane->fb) { - vgfb = to_virtio_gpu_framebuffer(plane->fb); + if (plane->state->fb) { + vgfb = to_virtio_gpu_framebuffer(plane->state->fb); bo = gem_to_virtio_gpu_obj(vgfb->obj); handle = bo->hw_res_handle; + if (bo->dumb) { + virtio_gpu_cmd_transfer_to_host_2d + (vgdev, handle, 0, + cpu_to_le32(plane->state->crtc_w), + cpu_to_le32(plane->state->crtc_h), + plane->state->crtc_x, plane->state->crtc_y, NULL); + } } else { handle = 0; } @@ -84,6 +91,11 @@ static void virtio_gpu_plane_atomic_update(struct drm_plane *plane, plane->state->crtc_h, plane->state->crtc_x, plane->state->crtc_y); + virtio_gpu_cmd_resource_flush(vgdev, handle, + plane->state->crtc_x, + plane->state->crtc_y, + plane->state->crtc_w, + plane->state->crtc_h); } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c index 24fb348a44e1..0ee76e523a90 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c @@ -972,15 +972,6 @@ static int vmw_driver_unload(struct drm_device *dev) return 0; } -static void vmw_preclose(struct drm_device *dev, - struct drm_file *file_priv) -{ - struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv); - struct vmw_private *dev_priv = vmw_priv(dev); - - vmw_event_fence_fpriv_gone(dev_priv->fman, &vmw_fp->fence_events); -} - static void vmw_postclose(struct drm_device *dev, struct drm_file *file_priv) { @@ -1011,7 +1002,6 @@ static int vmw_driver_open(struct drm_device *dev, struct drm_file *file_priv) if (unlikely(vmw_fp == NULL)) return ret; - INIT_LIST_HEAD(&vmw_fp->fence_events); vmw_fp->tfile = ttm_object_file_init(dev_priv->tdev, 10); if (unlikely(vmw_fp->tfile == NULL)) goto out_no_tfile; @@ -1501,7 +1491,6 @@ static struct drm_driver driver = { .master_set = vmw_master_set, .master_drop = vmw_master_drop, .open = vmw_driver_open, - .preclose = vmw_preclose, .postclose = vmw_postclose, .set_busid = drm_pci_set_busid, diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h index 469cdd520615..5cb1b1687cd4 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.h @@ -80,7 +80,6 @@ struct vmw_fpriv { struct drm_master *locked_master; struct ttm_object_file *tfile; - struct list_head fence_events; bool gb_aware; }; diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c index 8e689b439890..e959df6ede83 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c @@ -71,7 +71,6 @@ struct vmw_user_fence { */ struct vmw_event_fence_action { struct vmw_fence_action action; - struct list_head fpriv_head; struct drm_pending_event *event; struct vmw_fence_obj *fence; @@ -808,44 +807,6 @@ int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data, } /** - * vmw_event_fence_fpriv_gone - Remove references to struct drm_file objects - * - * @fman: Pointer to a struct vmw_fence_manager - * @event_list: Pointer to linked list of struct vmw_event_fence_action objects - * with pointers to a struct drm_file object about to be closed. - * - * This function removes all pending fence events with references to a - * specific struct drm_file object about to be closed. The caller is required - * to pass a list of all struct vmw_event_fence_action objects with such - * events attached. This function is typically called before the - * struct drm_file object's event management is taken down. - */ -void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman, - struct list_head *event_list) -{ - struct vmw_event_fence_action *eaction; - struct drm_pending_event *event; - unsigned long irq_flags; - - while (1) { - spin_lock_irqsave(&fman->lock, irq_flags); - if (list_empty(event_list)) - goto out_unlock; - eaction = list_first_entry(event_list, - struct vmw_event_fence_action, - fpriv_head); - list_del_init(&eaction->fpriv_head); - event = eaction->event; - eaction->event = NULL; - spin_unlock_irqrestore(&fman->lock, irq_flags); - event->destroy(event); - } -out_unlock: - spin_unlock_irqrestore(&fman->lock, irq_flags); -} - - -/** * vmw_event_fence_action_seq_passed * * @action: The struct vmw_fence_action embedded in a struct @@ -879,10 +840,8 @@ static void vmw_event_fence_action_seq_passed(struct vmw_fence_action *action) *eaction->tv_usec = tv.tv_usec; } - list_del_init(&eaction->fpriv_head); - list_add_tail(&eaction->event->link, &file_priv->event_list); + drm_send_event_locked(dev, eaction->event); eaction->event = NULL; - wake_up_all(&file_priv->event_wait); spin_unlock_irqrestore(&dev->event_lock, irq_flags); } @@ -899,12 +858,6 @@ static void vmw_event_fence_action_cleanup(struct vmw_fence_action *action) { struct vmw_event_fence_action *eaction = container_of(action, struct vmw_event_fence_action, action); - struct vmw_fence_manager *fman = fman_from_fence(eaction->fence); - unsigned long irq_flags; - - spin_lock_irqsave(&fman->lock, irq_flags); - list_del(&eaction->fpriv_head); - spin_unlock_irqrestore(&fman->lock, irq_flags); vmw_fence_obj_unreference(&eaction->fence); kfree(eaction); @@ -984,8 +937,6 @@ int vmw_event_fence_action_queue(struct drm_file *file_priv, { struct vmw_event_fence_action *eaction; struct vmw_fence_manager *fman = fman_from_fence(fence); - struct vmw_fpriv *vmw_fp = vmw_fpriv(file_priv); - unsigned long irq_flags; eaction = kzalloc(sizeof(*eaction), GFP_KERNEL); if (unlikely(eaction == NULL)) @@ -1002,10 +953,6 @@ int vmw_event_fence_action_queue(struct drm_file *file_priv, eaction->tv_sec = tv_sec; eaction->tv_usec = tv_usec; - spin_lock_irqsave(&fman->lock, irq_flags); - list_add_tail(&eaction->fpriv_head, &vmw_fp->fence_events); - spin_unlock_irqrestore(&fman->lock, irq_flags); - vmw_fence_obj_add_action(fence, &eaction->action); return 0; @@ -1025,38 +972,26 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv, struct vmw_event_fence_pending *event; struct vmw_fence_manager *fman = fman_from_fence(fence); struct drm_device *dev = fman->dev_priv->dev; - unsigned long irq_flags; int ret; - spin_lock_irqsave(&dev->event_lock, irq_flags); - - ret = (file_priv->event_space < sizeof(event->event)) ? -EBUSY : 0; - if (likely(ret == 0)) - file_priv->event_space -= sizeof(event->event); - - spin_unlock_irqrestore(&dev->event_lock, irq_flags); - - if (unlikely(ret != 0)) { - DRM_ERROR("Failed to allocate event space for this file.\n"); - goto out_no_space; - } - - event = kzalloc(sizeof(*event), GFP_KERNEL); if (unlikely(event == NULL)) { DRM_ERROR("Failed to allocate an event.\n"); ret = -ENOMEM; - goto out_no_event; + goto out_no_space; } event->event.base.type = DRM_VMW_EVENT_FENCE_SIGNALED; event->event.base.length = sizeof(*event); event->event.user_data = user_data; - event->base.event = &event->event.base; - event->base.file_priv = file_priv; - event->base.destroy = (void (*) (struct drm_pending_event *)) kfree; + ret = drm_event_reserve_init(dev, file_priv, &event->base, &event->event.base); + if (unlikely(ret != 0)) { + DRM_ERROR("Failed to allocate event space for this file.\n"); + kfree(event); + goto out_no_space; + } if (flags & DRM_VMW_FE_FLAG_REQ_TIME) ret = vmw_event_fence_action_queue(file_priv, fence, @@ -1076,11 +1011,7 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv, return 0; out_no_queue: - event->base.destroy(&event->base); -out_no_event: - spin_lock_irqsave(&dev->event_lock, irq_flags); - file_priv->event_space += sizeof(*event); - spin_unlock_irqrestore(&dev->event_lock, irq_flags); + drm_event_cancel_free(dev, &event->base); out_no_space: return ret; } diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h index 8be6c29f5eb5..83ae301ee141 100644 --- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h +++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.h @@ -116,8 +116,6 @@ extern int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); extern int vmw_fence_event_ioctl(struct drm_device *dev, void *data, struct drm_file *file_priv); -extern void vmw_event_fence_fpriv_gone(struct vmw_fence_manager *fman, - struct list_head *event_list); extern int vmw_event_fence_action_queue(struct drm_file *filee_priv, struct vmw_fence_obj *fence, struct drm_pending_event *event, diff --git a/drivers/gpu/vga/vga_switcheroo.c b/drivers/gpu/vga/vga_switcheroo.c index 665ab9fd0e01..cbd7c986d926 100644 --- a/drivers/gpu/vga/vga_switcheroo.c +++ b/drivers/gpu/vga/vga_switcheroo.c @@ -74,9 +74,17 @@ * there can thus be up to three clients: Two vga clients (GPUs) and one audio * client (on the discrete GPU). The code is mostly prepared to support * machines with more than two GPUs should they become available. + * * The GPU to which the outputs are currently switched is called the * active client in vga_switcheroo parlance. The GPU not in use is the - * inactive client. + * inactive client. When the inactive client's DRM driver is loaded, + * it will be unable to probe the panel's EDID and hence depends on + * VBIOS to provide its display modes. If the VBIOS modes are bogus or + * if there is no VBIOS at all (which is common on the MacBook Pro), + * a client may alternatively request that the DDC lines are temporarily + * switched to it, provided that the handler supports this. Switching + * only the DDC lines and not the entire output avoids unnecessary + * flickering. */ /** @@ -126,6 +134,10 @@ static DEFINE_MUTEX(vgasr_mutex); * (counting only vga clients, not audio clients) * @clients: list of registered clients * @handler: registered handler + * @handler_flags: flags of registered handler + * @mux_hw_lock: protects mux state + * (in particular while DDC lines are temporarily switched) + * @old_ddc_owner: client to which DDC lines will be switched back on unlock * * vga_switcheroo private data. Currently only one vga_switcheroo instance * per system is supported. @@ -142,6 +154,9 @@ struct vgasr_priv { struct list_head clients; const struct vga_switcheroo_handler *handler; + enum vga_switcheroo_handler_flags_t handler_flags; + struct mutex mux_hw_lock; + int old_ddc_owner; }; #define ID_BIT_AUDIO 0x100 @@ -156,6 +171,7 @@ static void vga_switcheroo_debugfs_fini(struct vgasr_priv *priv); /* only one switcheroo per system */ static struct vgasr_priv vgasr_priv = { .clients = LIST_HEAD_INIT(vgasr_priv.clients), + .mux_hw_lock = __MUTEX_INITIALIZER(vgasr_priv.mux_hw_lock), }; static bool vga_switcheroo_ready(void) @@ -190,13 +206,15 @@ static void vga_switcheroo_enable(void) /** * vga_switcheroo_register_handler() - register handler * @handler: handler callbacks + * @handler_flags: handler flags * * Register handler. Enable vga_switcheroo if two vga clients have already * registered. * * Return: 0 on success, -EINVAL if a handler was already registered. */ -int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler) +int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler, + enum vga_switcheroo_handler_flags_t handler_flags) { mutex_lock(&vgasr_mutex); if (vgasr_priv.handler) { @@ -205,6 +223,7 @@ int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler } vgasr_priv.handler = handler; + vgasr_priv.handler_flags = handler_flags; if (vga_switcheroo_ready()) { pr_info("enabled\n"); vga_switcheroo_enable(); @@ -222,16 +241,33 @@ EXPORT_SYMBOL(vga_switcheroo_register_handler); void vga_switcheroo_unregister_handler(void) { mutex_lock(&vgasr_mutex); + mutex_lock(&vgasr_priv.mux_hw_lock); + vgasr_priv.handler_flags = 0; vgasr_priv.handler = NULL; if (vgasr_priv.active) { pr_info("disabled\n"); vga_switcheroo_debugfs_fini(&vgasr_priv); vgasr_priv.active = false; } + mutex_unlock(&vgasr_priv.mux_hw_lock); mutex_unlock(&vgasr_mutex); } EXPORT_SYMBOL(vga_switcheroo_unregister_handler); +/** + * vga_switcheroo_handler_flags() - obtain handler flags + * + * Helper for clients to obtain the handler flags bitmask. + * + * Return: Handler flags. A value of 0 means that no handler is registered + * or that the handler has no special capabilities. + */ +enum vga_switcheroo_handler_flags_t vga_switcheroo_handler_flags(void) +{ + return vgasr_priv.handler_flags; +} +EXPORT_SYMBOL(vga_switcheroo_handler_flags); + static int register_client(struct pci_dev *pdev, const struct vga_switcheroo_client_ops *ops, enum vga_switcheroo_client_id id, bool active, @@ -413,6 +449,76 @@ void vga_switcheroo_client_fb_set(struct pci_dev *pdev, EXPORT_SYMBOL(vga_switcheroo_client_fb_set); /** + * vga_switcheroo_lock_ddc() - temporarily switch DDC lines to a given client + * @pdev: client pci device + * + * Temporarily switch DDC lines to the client identified by @pdev + * (but leave the outputs otherwise switched to where they are). + * This allows the inactive client to probe EDID. The DDC lines must + * afterwards be switched back by calling vga_switcheroo_unlock_ddc(), + * even if this function returns an error. + * + * Return: Previous DDC owner on success or a negative int on error. + * Specifically, %-ENODEV if no handler has registered or if the handler + * does not support switching the DDC lines. Also, a negative value + * returned by the handler is propagated back to the caller. + * The return value has merely an informational purpose for any caller + * which might be interested in it. It is acceptable to ignore the return + * value and simply rely on the result of the subsequent EDID probe, + * which will be %NULL if DDC switching failed. + */ +int vga_switcheroo_lock_ddc(struct pci_dev *pdev) +{ + enum vga_switcheroo_client_id id; + + mutex_lock(&vgasr_priv.mux_hw_lock); + if (!vgasr_priv.handler || !vgasr_priv.handler->switch_ddc) { + vgasr_priv.old_ddc_owner = -ENODEV; + return -ENODEV; + } + + id = vgasr_priv.handler->get_client_id(pdev); + vgasr_priv.old_ddc_owner = vgasr_priv.handler->switch_ddc(id); + return vgasr_priv.old_ddc_owner; +} +EXPORT_SYMBOL(vga_switcheroo_lock_ddc); + +/** + * vga_switcheroo_unlock_ddc() - switch DDC lines back to previous owner + * @pdev: client pci device + * + * Switch DDC lines back to the previous owner after calling + * vga_switcheroo_lock_ddc(). This must be called even if + * vga_switcheroo_lock_ddc() returned an error. + * + * Return: Previous DDC owner on success (i.e. the client identifier of @pdev) + * or a negative int on error. + * Specifically, %-ENODEV if no handler has registered or if the handler + * does not support switching the DDC lines. Also, a negative value + * returned by the handler is propagated back to the caller. + * Finally, invoking this function without calling vga_switcheroo_lock_ddc() + * first is not allowed and will result in %-EINVAL. + */ +int vga_switcheroo_unlock_ddc(struct pci_dev *pdev) +{ + enum vga_switcheroo_client_id id; + int ret = vgasr_priv.old_ddc_owner; + + if (WARN_ON_ONCE(!mutex_is_locked(&vgasr_priv.mux_hw_lock))) + return -EINVAL; + + if (vgasr_priv.old_ddc_owner >= 0) { + id = vgasr_priv.handler->get_client_id(pdev); + if (vgasr_priv.old_ddc_owner != id) + ret = vgasr_priv.handler->switch_ddc( + vgasr_priv.old_ddc_owner); + } + mutex_unlock(&vgasr_priv.mux_hw_lock); + return ret; +} +EXPORT_SYMBOL(vga_switcheroo_unlock_ddc); + +/** * DOC: Manual switching and manual power control * * In this mode of use, the file /sys/kernel/debug/vgaswitcheroo/switch @@ -549,7 +655,9 @@ static int vga_switchto_stage2(struct vga_switcheroo_client *new_client) console_unlock(); } + mutex_lock(&vgasr_priv.mux_hw_lock); ret = vgasr_priv.handler->switchto(new_client->id); + mutex_unlock(&vgasr_priv.mux_hw_lock); if (ret) return ret; @@ -664,7 +772,9 @@ vga_switcheroo_debugfs_write(struct file *filp, const char __user *ubuf, vgasr_priv.delayed_switch_active = false; if (just_mux) { + mutex_lock(&vgasr_priv.mux_hw_lock); ret = vgasr_priv.handler->switchto(client_id); + mutex_unlock(&vgasr_priv.mux_hw_lock); goto out; } @@ -876,8 +986,11 @@ static int vga_switcheroo_runtime_suspend(struct device *dev) if (ret) return ret; mutex_lock(&vgasr_mutex); - if (vgasr_priv.handler->switchto) + if (vgasr_priv.handler->switchto) { + mutex_lock(&vgasr_priv.mux_hw_lock); vgasr_priv.handler->switchto(VGA_SWITCHEROO_IGD); + mutex_unlock(&vgasr_priv.mux_hw_lock); + } vga_switcheroo_power_switch(pdev, VGA_SWITCHEROO_OFF); mutex_unlock(&vgasr_mutex); return 0; diff --git a/drivers/platform/x86/apple-gmux.c b/drivers/platform/x86/apple-gmux.c index f236250ac106..4034d2d4c507 100644 --- a/drivers/platform/x86/apple-gmux.c +++ b/drivers/platform/x86/apple-gmux.c @@ -19,6 +19,7 @@ #include <linux/acpi.h> #include <linux/pnp.h> #include <linux/apple_bl.h> +#include <linux/apple-gmux.h> #include <linux/slab.h> #include <linux/delay.h> #include <linux/pci.h> @@ -57,7 +58,9 @@ struct apple_gmux_data { /* switcheroo data */ acpi_handle dhandle; int gpe; - enum vga_switcheroo_client_id resume_client_id; + enum vga_switcheroo_client_id switch_state_display; + enum vga_switcheroo_client_id switch_state_ddc; + enum vga_switcheroo_client_id switch_state_external; enum vga_switcheroo_state power_state; struct completion powerchange_done; }; @@ -368,19 +371,70 @@ static const struct backlight_ops gmux_bl_ops = { * for the selected GPU. */ +static void gmux_read_switch_state(struct apple_gmux_data *gmux_data) +{ + if (gmux_read8(gmux_data, GMUX_PORT_SWITCH_DDC) == 1) + gmux_data->switch_state_ddc = VGA_SWITCHEROO_IGD; + else + gmux_data->switch_state_ddc = VGA_SWITCHEROO_DIS; + + if (gmux_read8(gmux_data, GMUX_PORT_SWITCH_DISPLAY) == 2) + gmux_data->switch_state_display = VGA_SWITCHEROO_IGD; + else + gmux_data->switch_state_display = VGA_SWITCHEROO_DIS; + + if (gmux_read8(gmux_data, GMUX_PORT_SWITCH_EXTERNAL) == 2) + gmux_data->switch_state_external = VGA_SWITCHEROO_IGD; + else + gmux_data->switch_state_external = VGA_SWITCHEROO_DIS; +} + +static void gmux_write_switch_state(struct apple_gmux_data *gmux_data) +{ + if (gmux_data->switch_state_ddc == VGA_SWITCHEROO_IGD) + gmux_write8(gmux_data, GMUX_PORT_SWITCH_DDC, 1); + else + gmux_write8(gmux_data, GMUX_PORT_SWITCH_DDC, 2); + + if (gmux_data->switch_state_display == VGA_SWITCHEROO_IGD) + gmux_write8(gmux_data, GMUX_PORT_SWITCH_DISPLAY, 2); + else + gmux_write8(gmux_data, GMUX_PORT_SWITCH_DISPLAY, 3); + + if (gmux_data->switch_state_external == VGA_SWITCHEROO_IGD) + gmux_write8(gmux_data, GMUX_PORT_SWITCH_EXTERNAL, 2); + else + gmux_write8(gmux_data, GMUX_PORT_SWITCH_EXTERNAL, 3); +} + static int gmux_switchto(enum vga_switcheroo_client_id id) { - if (id == VGA_SWITCHEROO_IGD) { + apple_gmux_data->switch_state_ddc = id; + apple_gmux_data->switch_state_display = id; + apple_gmux_data->switch_state_external = id; + + gmux_write_switch_state(apple_gmux_data); + + return 0; +} + +static int gmux_switch_ddc(enum vga_switcheroo_client_id id) +{ + enum vga_switcheroo_client_id old_ddc_owner = + apple_gmux_data->switch_state_ddc; + + if (id == old_ddc_owner) + return id; + + pr_debug("Switching DDC from %d to %d\n", old_ddc_owner, id); + apple_gmux_data->switch_state_ddc = id; + + if (id == VGA_SWITCHEROO_IGD) gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_DDC, 1); - gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_DISPLAY, 2); - gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_EXTERNAL, 2); - } else { + else gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_DDC, 2); - gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_DISPLAY, 3); - gmux_write8(apple_gmux_data, GMUX_PORT_SWITCH_EXTERNAL, 3); - } - return 0; + return old_ddc_owner; } /** @@ -440,17 +494,15 @@ static int gmux_get_client_id(struct pci_dev *pdev) return VGA_SWITCHEROO_DIS; } -static enum vga_switcheroo_client_id -gmux_active_client(struct apple_gmux_data *gmux_data) -{ - if (gmux_read8(gmux_data, GMUX_PORT_SWITCH_DISPLAY) == 2) - return VGA_SWITCHEROO_IGD; - - return VGA_SWITCHEROO_DIS; -} +static const struct vga_switcheroo_handler gmux_handler_indexed = { + .switchto = gmux_switchto, + .power_state = gmux_set_power_state, + .get_client_id = gmux_get_client_id, +}; -static const struct vga_switcheroo_handler gmux_handler = { +static const struct vga_switcheroo_handler gmux_handler_classic = { .switchto = gmux_switchto, + .switch_ddc = gmux_switch_ddc, .power_state = gmux_set_power_state, .get_client_id = gmux_get_client_id, }; @@ -513,7 +565,6 @@ static int gmux_suspend(struct device *dev) struct pnp_dev *pnp = to_pnp_dev(dev); struct apple_gmux_data *gmux_data = pnp_get_drvdata(pnp); - gmux_data->resume_client_id = gmux_active_client(gmux_data); gmux_disable_interrupts(gmux_data); return 0; } @@ -524,7 +575,7 @@ static int gmux_resume(struct device *dev) struct apple_gmux_data *gmux_data = pnp_get_drvdata(pnp); gmux_enable_interrupts(gmux_data); - gmux_switchto(gmux_data->resume_client_id); + gmux_write_switch_state(gmux_data); if (gmux_data->power_state == VGA_SWITCHEROO_OFF) gmux_set_discrete_state(gmux_data, gmux_data->power_state); return 0; @@ -704,9 +755,23 @@ static int gmux_probe(struct pnp_dev *pnp, const struct pnp_device_id *id) apple_gmux_data = gmux_data; init_completion(&gmux_data->powerchange_done); gmux_enable_interrupts(gmux_data); + gmux_read_switch_state(gmux_data); - if (vga_switcheroo_register_handler(&gmux_handler)) { - ret = -ENODEV; + /* + * Retina MacBook Pros cannot switch the panel's AUX separately + * and need eDP pre-calibration. They are distinguishable from + * pre-retinas by having an "indexed" gmux. + * + * Pre-retina MacBook Pros can switch the panel's DDC separately. + */ + if (gmux_data->indexed) + ret = vga_switcheroo_register_handler(&gmux_handler_indexed, + VGA_SWITCHEROO_NEEDS_EDP_CONFIG); + else + ret = vga_switcheroo_register_handler(&gmux_handler_classic, + VGA_SWITCHEROO_CAN_SWITCH_DDC); + if (ret) { + pr_err("Failed to register vga_switcheroo handler\n"); goto err_register_handler; } @@ -764,7 +829,7 @@ static void gmux_remove(struct pnp_dev *pnp) } static const struct pnp_device_id gmux_device_ids[] = { - {"APP000B", 0}, + {GMUX_ACPI_HID, 0}, {"", 0} }; diff --git a/drivers/staging/android/ion/ion.c b/drivers/staging/android/ion/ion.c index e237e9f3312d..0754a37c9674 100644 --- a/drivers/staging/android/ion/ion.c +++ b/drivers/staging/android/ion/ion.c @@ -1057,8 +1057,7 @@ static void ion_dma_buf_kunmap(struct dma_buf *dmabuf, unsigned long offset, { } -static int ion_dma_buf_begin_cpu_access(struct dma_buf *dmabuf, size_t start, - size_t len, +static int ion_dma_buf_begin_cpu_access(struct dma_buf *dmabuf, enum dma_data_direction direction) { struct ion_buffer *buffer = dmabuf->priv; @@ -1076,8 +1075,7 @@ static int ion_dma_buf_begin_cpu_access(struct dma_buf *dmabuf, size_t start, return PTR_ERR_OR_ZERO(vaddr); } -static void ion_dma_buf_end_cpu_access(struct dma_buf *dmabuf, size_t start, - size_t len, +static void ion_dma_buf_end_cpu_access(struct dma_buf *dmabuf, enum dma_data_direction direction) { struct ion_buffer *buffer = dmabuf->priv; diff --git a/drivers/staging/android/ion/ion_test.c b/drivers/staging/android/ion/ion_test.c index b8dcf5a26cc4..da34bc12cd7c 100644 --- a/drivers/staging/android/ion/ion_test.c +++ b/drivers/staging/android/ion/ion_test.c @@ -109,7 +109,7 @@ static int ion_handle_test_kernel(struct dma_buf *dma_buf, void __user *ptr, if (offset > dma_buf->size || size > dma_buf->size - offset) return -EINVAL; - ret = dma_buf_begin_cpu_access(dma_buf, offset, size, dir); + ret = dma_buf_begin_cpu_access(dma_buf, dir); if (ret) return ret; @@ -139,7 +139,7 @@ static int ion_handle_test_kernel(struct dma_buf *dma_buf, void __user *ptr, copy_offset = 0; } err: - dma_buf_end_cpu_access(dma_buf, offset, size, dir); + dma_buf_end_cpu_access(dma_buf, dir); return ret; } diff --git a/include/drm/drmP.h b/include/drm/drmP.h index d7162cf1c3e1..3c8422c69572 100644 --- a/include/drm/drmP.h +++ b/include/drm/drmP.h @@ -283,6 +283,7 @@ struct drm_ioctl_desc { struct drm_pending_event { struct drm_event *event; struct list_head link; + struct list_head pending_link; struct drm_file *file_priv; pid_t pid; /* pid of requester, no guarantee it's valid by the time we deliver the event, for tracing only */ @@ -346,6 +347,7 @@ struct drm_file { struct list_head blobs; wait_queue_head_t event_wait; + struct list_head pending_event_list; struct list_head event_list; int event_space; @@ -919,15 +921,25 @@ extern long drm_compat_ioctl(struct file *filp, unsigned int cmd, unsigned long arg); extern bool drm_ioctl_flags(unsigned int nr, unsigned int *flags); - /* Device support (drm_fops.h) */ -extern int drm_open(struct inode *inode, struct file *filp); -extern ssize_t drm_read(struct file *filp, char __user *buffer, - size_t count, loff_t *offset); -extern int drm_release(struct inode *inode, struct file *filp); -extern int drm_new_set_master(struct drm_device *dev, struct drm_file *fpriv); - - /* Mapping support (drm_vm.h) */ -extern unsigned int drm_poll(struct file *filp, struct poll_table_struct *wait); +/* File Operations (drm_fops.c) */ +int drm_open(struct inode *inode, struct file *filp); +ssize_t drm_read(struct file *filp, char __user *buffer, + size_t count, loff_t *offset); +int drm_release(struct inode *inode, struct file *filp); +int drm_new_set_master(struct drm_device *dev, struct drm_file *fpriv); +unsigned int drm_poll(struct file *filp, struct poll_table_struct *wait); +int drm_event_reserve_init_locked(struct drm_device *dev, + struct drm_file *file_priv, + struct drm_pending_event *p, + struct drm_event *e); +int drm_event_reserve_init(struct drm_device *dev, + struct drm_file *file_priv, + struct drm_pending_event *p, + struct drm_event *e); +void drm_event_cancel_free(struct drm_device *dev, + struct drm_pending_event *p); +void drm_send_event_locked(struct drm_device *dev, struct drm_pending_event *e); +void drm_send_event(struct drm_device *dev, struct drm_pending_event *e); /* Misc. IOCTL support (drm_ioctl.c) */ int drm_noop(struct drm_device *dev, void *data, diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h index c65a212db77e..8c7fb3d0f9d0 100644 --- a/include/drm/drm_crtc.h +++ b/include/drm/drm_crtc.h @@ -307,6 +307,7 @@ struct drm_plane_helper_funcs; * @connectors_changed: connectors to this crtc have been updated * @plane_mask: bitmask of (1 << drm_plane_index(plane)) of attached planes * @connector_mask: bitmask of (1 << drm_connector_index(connector)) of attached connectors + * @encoder_mask: bitmask of (1 << drm_encoder_index(encoder)) of attached encoders * @last_vblank_count: for helpers and drivers to capture the vblank of the * update to ensure framebuffer cleanup isn't done too early * @adjusted_mode: for use by helpers and drivers to compute adjusted mode timings @@ -341,6 +342,7 @@ struct drm_crtc_state { u32 plane_mask; u32 connector_mask; + u32 encoder_mask; /* last_vblank_count: for vblank waits before cleanup */ u32 last_vblank_count; @@ -2153,6 +2155,17 @@ struct drm_mode_config { list_for_each_entry((plane), &(dev)->mode_config.plane_list, head) \ for_each_if ((plane_mask) & (1 << drm_plane_index(plane))) +/** + * drm_for_each_encoder_mask - iterate over encoders specified by bitmask + * @encoder: the loop cursor + * @dev: the DRM device + * @encoder_mask: bitmask of encoder indices + * + * Iterate over all encoders specified by bitmask. + */ +#define drm_for_each_encoder_mask(encoder, dev, encoder_mask) \ + list_for_each_entry((encoder), &(dev)->mode_config.encoder_list, head) \ + for_each_if ((encoder_mask) & (1 << drm_encoder_index(encoder))) #define obj_to_crtc(x) container_of(x, struct drm_crtc, base) #define obj_to_connector(x) container_of(x, struct drm_connector, base) @@ -2225,6 +2238,7 @@ int drm_encoder_init(struct drm_device *dev, struct drm_encoder *encoder, const struct drm_encoder_funcs *funcs, int encoder_type, const char *name, ...); +extern unsigned int drm_encoder_index(struct drm_encoder *encoder); /** * drm_encoder_crtc_ok - can a given crtc drive a given encoder? @@ -2282,6 +2296,8 @@ extern void drm_property_destroy_user_blobs(struct drm_device *dev, extern bool drm_probe_ddc(struct i2c_adapter *adapter); extern struct edid *drm_get_edid(struct drm_connector *connector, struct i2c_adapter *adapter); +extern struct edid *drm_get_edid_switcheroo(struct drm_connector *connector, + struct i2c_adapter *adapter); extern struct edid *drm_edid_duplicate(const struct edid *edid); extern int drm_add_edid_modes(struct drm_connector *connector, struct edid *edid); extern void drm_mode_config_init(struct drm_device *dev); @@ -2482,6 +2498,8 @@ extern int drm_format_num_planes(uint32_t format); extern int drm_format_plane_cpp(uint32_t format, int plane); extern int drm_format_horz_chroma_subsampling(uint32_t format); extern int drm_format_vert_chroma_subsampling(uint32_t format); +extern int drm_format_plane_width(int width, uint32_t format, int plane); +extern int drm_format_plane_height(int height, uint32_t format, int plane); extern const char *drm_get_format_name(uint32_t format); extern struct drm_property *drm_mode_create_rotation_property(struct drm_device *dev, unsigned int supported_rotations); diff --git a/include/drm/drm_dp_aux_dev.h b/include/drm/drm_dp_aux_dev.h new file mode 100644 index 000000000000..1b76d990d8ab --- /dev/null +++ b/include/drm/drm_dp_aux_dev.h @@ -0,0 +1,62 @@ +/* + * Copyright © 2015 Intel Corporation + * + * Permission is hereby granted, free of charge, to any person obtaining a + * copy of this software and associated documentation files (the "Software"), + * to deal in the Software without restriction, including without limitation + * the rights to use, copy, modify, merge, publish, distribute, sublicense, + * and/or sell copies of the Software, and to permit persons to whom the + * Software is furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice (including the next + * paragraph) shall be included in all copies or substantial portions of the + * Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL + * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + * + * Authors: + * Rafael Antognolli <[email protected]> + * + */ + +#ifndef DRM_DP_AUX_DEV +#define DRM_DP_AUX_DEV + +#include <drm/drm_dp_helper.h> + +#ifdef CONFIG_DRM_DP_AUX_CHARDEV + +int drm_dp_aux_dev_init(void); +void drm_dp_aux_dev_exit(void); +int drm_dp_aux_register_devnode(struct drm_dp_aux *aux); +void drm_dp_aux_unregister_devnode(struct drm_dp_aux *aux); + +#else + +static inline int drm_dp_aux_dev_init(void) +{ + return 0; +} + +static inline void drm_dp_aux_dev_exit(void) +{ +} + +static inline int drm_dp_aux_register_devnode(struct drm_dp_aux *aux) +{ + return 0; +} + +static inline void drm_dp_aux_unregister_devnode(struct drm_dp_aux *aux) +{ +} + +#endif + +#endif diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h index d8a40dff0d1d..062723bdcabe 100644 --- a/include/drm/drm_fb_helper.h +++ b/include/drm/drm_fb_helper.h @@ -219,6 +219,7 @@ struct drm_fb_helper { }; #ifdef CONFIG_DRM_FBDEV_EMULATION +int drm_fb_helper_modinit(void); void drm_fb_helper_prepare(struct drm_device *dev, struct drm_fb_helper *helper, const struct drm_fb_helper_funcs *funcs); int drm_fb_helper_init(struct drm_device *dev, @@ -283,6 +284,11 @@ int drm_fb_helper_add_one_connector(struct drm_fb_helper *fb_helper, struct drm_ int drm_fb_helper_remove_one_connector(struct drm_fb_helper *fb_helper, struct drm_connector *connector); #else +static inline int drm_fb_helper_modinit(void) +{ + return 0; +} + static inline void drm_fb_helper_prepare(struct drm_device *dev, struct drm_fb_helper *helper, const struct drm_fb_helper_funcs *funcs) diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h index a126a0d7aed4..b61c2d45192e 100644 --- a/include/drm/drm_modeset_helper_vtables.h +++ b/include/drm/drm_modeset_helper_vtables.h @@ -439,7 +439,7 @@ struct drm_encoder_helper_funcs { * can be modified by this callback and does not need to match mode. * * This function is used by both legacy CRTC helpers and atomic helpers. - * With atomic helpers it is optional. + * This hook is optional. * * NOTE: * diff --git a/include/linux/apple-gmux.h b/include/linux/apple-gmux.h new file mode 100644 index 000000000000..b2d32e01dfe4 --- /dev/null +++ b/include/linux/apple-gmux.h @@ -0,0 +1,50 @@ +/* + * apple-gmux.h - microcontroller built into dual GPU MacBook Pro & Mac Pro + * Copyright (C) 2015 Lukas Wunner <[email protected]> + * + * This program is free software; you can redistribute it and/or modify + * it under the terms of the GNU General Public License (version 2) as + * published by the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, + * but WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef LINUX_APPLE_GMUX_H +#define LINUX_APPLE_GMUX_H + +#include <linux/acpi.h> + +#define GMUX_ACPI_HID "APP000B" + +#if IS_ENABLED(CONFIG_APPLE_GMUX) + +/** + * apple_gmux_present() - detect if gmux is built into the machine + * + * Drivers may use this to activate quirks specific to dual GPU MacBook Pros + * and Mac Pros, e.g. for deferred probing, runtime pm and backlight. + * + * Return: %true if gmux is present and the kernel was configured + * with CONFIG_APPLE_GMUX, %false otherwise. + */ +static inline bool apple_gmux_present(void) +{ + return acpi_dev_present(GMUX_ACPI_HID); +} + +#else /* !CONFIG_APPLE_GMUX */ + +static inline bool apple_gmux_present(void) +{ + return false; +} + +#endif /* !CONFIG_APPLE_GMUX */ + +#endif /* LINUX_APPLE_GMUX_H */ diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h index f98bd7068d55..532108ea0c1c 100644 --- a/include/linux/dma-buf.h +++ b/include/linux/dma-buf.h @@ -54,7 +54,7 @@ struct dma_buf_attachment; * @release: release this buffer; to be called after the last dma_buf_put. * @begin_cpu_access: [optional] called before cpu access to invalidate cpu * caches and allocate backing storage (if not yet done) - * respectively pin the objet into memory. + * respectively pin the object into memory. * @end_cpu_access: [optional] called after cpu access to flush caches. * @kmap_atomic: maps a page from the buffer into kernel address * space, users may not block until the subsequent unmap call. @@ -93,10 +93,8 @@ struct dma_buf_ops { /* after final dma_buf_put() */ void (*release)(struct dma_buf *); - int (*begin_cpu_access)(struct dma_buf *, size_t, size_t, - enum dma_data_direction); - void (*end_cpu_access)(struct dma_buf *, size_t, size_t, - enum dma_data_direction); + int (*begin_cpu_access)(struct dma_buf *, enum dma_data_direction); + void (*end_cpu_access)(struct dma_buf *, enum dma_data_direction); void *(*kmap_atomic)(struct dma_buf *, unsigned long); void (*kunmap_atomic)(struct dma_buf *, unsigned long, void *); void *(*kmap)(struct dma_buf *, unsigned long); @@ -224,9 +222,9 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *, enum dma_data_direction); void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *, enum dma_data_direction); -int dma_buf_begin_cpu_access(struct dma_buf *dma_buf, size_t start, size_t len, +int dma_buf_begin_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction dir); -void dma_buf_end_cpu_access(struct dma_buf *dma_buf, size_t start, size_t len, +void dma_buf_end_cpu_access(struct dma_buf *dma_buf, enum dma_data_direction dir); void *dma_buf_kmap_atomic(struct dma_buf *, unsigned long); void dma_buf_kunmap_atomic(struct dma_buf *, unsigned long, void *); diff --git a/include/linux/vga_switcheroo.h b/include/linux/vga_switcheroo.h index 69e1d4a1f1b3..b39a5f3153bd 100644 --- a/include/linux/vga_switcheroo.h +++ b/include/linux/vga_switcheroo.h @@ -36,6 +36,26 @@ struct pci_dev; /** + * enum vga_switcheroo_handler_flags_t - handler flags bitmask + * @VGA_SWITCHEROO_CAN_SWITCH_DDC: whether the handler is able to switch the + * DDC lines separately. This signals to clients that they should call + * drm_get_edid_switcheroo() to probe the EDID + * @VGA_SWITCHEROO_NEEDS_EDP_CONFIG: whether the handler is unable to switch + * the AUX channel separately. This signals to clients that the active + * GPU needs to train the link and communicate the link parameters to the + * inactive GPU (mediated by vga_switcheroo). The inactive GPU may then + * skip the AUX handshake and set up its output with these pre-calibrated + * values (DisplayPort specification v1.1a, section 2.5.3.3) + * + * Handler flags bitmask. Used by handlers to declare their capabilities upon + * registering with vga_switcheroo. + */ +enum vga_switcheroo_handler_flags_t { + VGA_SWITCHEROO_CAN_SWITCH_DDC = (1 << 0), + VGA_SWITCHEROO_NEEDS_EDP_CONFIG = (1 << 1), +}; + +/** * enum vga_switcheroo_state - client power state * @VGA_SWITCHEROO_OFF: off * @VGA_SWITCHEROO_ON: on @@ -82,6 +102,9 @@ enum vga_switcheroo_client_id { * Mandatory. For muxless machines this should be a no-op. Returning 0 * denotes success, anything else failure (in which case the switch is * aborted) + * @switch_ddc: switch DDC lines to given client. + * Optional. Should return the previous DDC owner on success or a + * negative int on failure * @power_state: cut or reinstate power of given client. * Optional. The return value is ignored * @get_client_id: determine if given pci device is integrated or discrete GPU. @@ -93,6 +116,7 @@ enum vga_switcheroo_client_id { struct vga_switcheroo_handler { int (*init)(void); int (*switchto)(enum vga_switcheroo_client_id id); + int (*switch_ddc)(enum vga_switcheroo_client_id id); int (*power_state)(enum vga_switcheroo_client_id id, enum vga_switcheroo_state state); enum vga_switcheroo_client_id (*get_client_id)(struct pci_dev *pdev); @@ -132,8 +156,12 @@ int vga_switcheroo_register_audio_client(struct pci_dev *pdev, void vga_switcheroo_client_fb_set(struct pci_dev *dev, struct fb_info *info); -int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler); +int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler, + enum vga_switcheroo_handler_flags_t handler_flags); void vga_switcheroo_unregister_handler(void); +enum vga_switcheroo_handler_flags_t vga_switcheroo_handler_flags(void); +int vga_switcheroo_lock_ddc(struct pci_dev *pdev); +int vga_switcheroo_unlock_ddc(struct pci_dev *pdev); int vga_switcheroo_process_delayed_switch(void); @@ -150,11 +178,15 @@ static inline void vga_switcheroo_unregister_client(struct pci_dev *dev) {} static inline int vga_switcheroo_register_client(struct pci_dev *dev, const struct vga_switcheroo_client_ops *ops, bool driver_power_control) { return 0; } static inline void vga_switcheroo_client_fb_set(struct pci_dev *dev, struct fb_info *info) {} -static inline int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler) { return 0; } +static inline int vga_switcheroo_register_handler(const struct vga_switcheroo_handler *handler, + enum vga_switcheroo_handler_flags_t handler_flags) { return 0; } static inline int vga_switcheroo_register_audio_client(struct pci_dev *pdev, const struct vga_switcheroo_client_ops *ops, enum vga_switcheroo_client_id id) { return 0; } static inline void vga_switcheroo_unregister_handler(void) {} +static inline enum vga_switcheroo_handler_flags_t vga_switcheroo_handler_flags(void) { return 0; } +static inline int vga_switcheroo_lock_ddc(struct pci_dev *pdev) { return -ENODEV; } +static inline int vga_switcheroo_unlock_ddc(struct pci_dev *pdev) { return -ENODEV; } static inline int vga_switcheroo_process_delayed_switch(void) { return 0; } static inline enum vga_switcheroo_state vga_switcheroo_get_client_state(struct pci_dev *dev) { return VGA_SWITCHEROO_ON; } diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h index b4e92eb12044..a0ebfe7c9a28 100644 --- a/include/uapi/drm/drm.h +++ b/include/uapi/drm/drm.h @@ -669,6 +669,7 @@ struct drm_set_client_cap { __u64 value; }; +#define DRM_RDWR O_RDWR #define DRM_CLOEXEC O_CLOEXEC struct drm_prime_handle { __u32 handle; diff --git a/include/uapi/drm/i915_drm.h b/include/uapi/drm/i915_drm.h index acf21026c78a..6a19371391fa 100644 --- a/include/uapi/drm/i915_drm.h +++ b/include/uapi/drm/i915_drm.h @@ -812,10 +812,35 @@ struct drm_i915_gem_busy { /** Handle of the buffer to check for busy */ __u32 handle; - /** Return busy status (1 if busy, 0 if idle). - * The high word is used to indicate on which rings the object - * currently resides: - * 16:31 - busy (r or r/w) rings (16 render, 17 bsd, 18 blt, etc) + /** Return busy status + * + * A return of 0 implies that the object is idle (after + * having flushed any pending activity), and a non-zero return that + * the object is still in-flight on the GPU. (The GPU has not yet + * signaled completion for all pending requests that reference the + * object.) + * + * The returned dword is split into two fields to indicate both + * the engines on which the object is being read, and the + * engine on which it is currently being written (if any). + * + * The low word (bits 0:15) indicate if the object is being written + * to by any engine (there can only be one, as the GEM implicit + * synchronisation rules force writes to be serialised). Only the + * engine for the last write is reported. + * + * The high word (bits 16:31) are a bitmask of which engines are + * currently reading from the object. Multiple engines may be + * reading from the object simultaneously. + * + * The value of each engine is the same as specified in the + * EXECBUFFER2 ioctl, i.e. I915_EXEC_RENDER, I915_EXEC_BSD etc. + * Note I915_EXEC_DEFAULT is a symbolic value and is mapped to + * the I915_EXEC_RENDER engine for execution, and so it is never + * reported as active itself. Some hardware may have parallel + * execution engines, e.g. multiple media engines, which are + * mapped to the same identifier in the EXECBUFFER2 ioctl and + * so are not separately reported for busyness. */ __u32 busy; }; diff --git a/include/uapi/drm/msm_drm.h b/include/uapi/drm/msm_drm.h index 81e6e0d1d360..0e3c4c48b326 100644 --- a/include/uapi/drm/msm_drm.h +++ b/include/uapi/drm/msm_drm.h @@ -50,6 +50,7 @@ struct drm_msm_timespec { #define MSM_PARAM_GPU_ID 0x01 #define MSM_PARAM_GMEM_SIZE 0x02 #define MSM_PARAM_CHIP_ID 0x03 +#define MSM_PARAM_MAX_FREQ 0x04 struct drm_msm_param { __u32 pipe; /* in, MSM_PIPE_x */ diff --git a/include/uapi/linux/dma-buf.h b/include/uapi/linux/dma-buf.h new file mode 100644 index 000000000000..fb0dedb7c121 --- /dev/null +++ b/include/uapi/linux/dma-buf.h @@ -0,0 +1,40 @@ +/* + * Framework for buffer objects that can be shared across devices/subsystems. + * + * Copyright(C) 2015 Intel Ltd + * + * This program is free software; you can redistribute it and/or modify it + * under the terms of the GNU General Public License version 2 as published by + * the Free Software Foundation. + * + * This program is distributed in the hope that it will be useful, but WITHOUT + * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or + * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for + * more details. + * + * You should have received a copy of the GNU General Public License along with + * this program. If not, see <http://www.gnu.org/licenses/>. + */ + +#ifndef _DMA_BUF_UAPI_H_ +#define _DMA_BUF_UAPI_H_ + +#include <linux/types.h> + +/* begin/end dma-buf functions used for userspace mmap. */ +struct dma_buf_sync { + __u64 flags; +}; + +#define DMA_BUF_SYNC_READ (1 << 0) +#define DMA_BUF_SYNC_WRITE (2 << 0) +#define DMA_BUF_SYNC_RW (DMA_BUF_SYNC_READ | DMA_BUF_SYNC_WRITE) +#define DMA_BUF_SYNC_START (0 << 2) +#define DMA_BUF_SYNC_END (1 << 2) +#define DMA_BUF_SYNC_VALID_FLAGS_MASK \ + (DMA_BUF_SYNC_RW | DMA_BUF_SYNC_END) + +#define DMA_BUF_BASE 'b' +#define DMA_BUF_IOCTL_SYNC _IOW(DMA_BUF_BASE, 0, struct dma_buf_sync) + +#endif diff --git a/mm/swapfile.c b/mm/swapfile.c index d2c37365e2d6..38884383ebac 100644 --- a/mm/swapfile.c +++ b/mm/swapfile.c @@ -48,6 +48,12 @@ static sector_t map_swap_entry(swp_entry_t, struct block_device**); DEFINE_SPINLOCK(swap_lock); static unsigned int nr_swapfiles; atomic_long_t nr_swap_pages; +/* + * Some modules use swappable objects and may try to swap them out under + * memory pressure (via the shrinker). Before doing so, they may wish to + * check to see if any swap space is available. + */ +EXPORT_SYMBOL_GPL(nr_swap_pages); /* protected with swap_lock. reading in vm_swap_full() doesn't need lock */ long total_swap_pages; static int least_priority; |