diff options
Diffstat (limited to 'drivers/gpu/drm/amd')
93 files changed, 1292 insertions, 484 deletions
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h index afec09930efa..9d92ca157677 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h @@ -248,6 +248,7 @@ extern int amdgpu_umsch_mm; extern int amdgpu_seamless; extern int amdgpu_user_partt_mode; +extern int amdgpu_agp; #define AMDGPU_VM_MAX_NUM_CTX 4096 #define AMDGPU_SG_THRESHOLD (256*1024*1024) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c index b8412202a1b0..75dc58470393 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c @@ -547,7 +547,7 @@ int amdgpu_amdkfd_get_xgmi_bandwidth_mbytes(struct amdgpu_device *dst, struct amdgpu_device *adev = dst, *peer_adev; int num_links; - if (adev->asic_type != CHIP_ALDEBARAN) + if (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(9, 4, 2)) return 0; if (src) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c index df3ecfa9e13f..e50be6500030 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_cs.c @@ -207,7 +207,7 @@ static int amdgpu_cs_pass1(struct amdgpu_cs_parser *p, } for (i = 0; i < p->nchunks; i++) { - struct drm_amdgpu_cs_chunk __user **chunk_ptr = NULL; + struct drm_amdgpu_cs_chunk __user *chunk_ptr = NULL; struct drm_amdgpu_cs_chunk user_chunk; uint32_t __user *cdata; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c index a53f436fa9f1..0e61ebdb3f3e 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_debugfs.c @@ -638,6 +638,9 @@ static ssize_t amdgpu_debugfs_regs_didt_read(struct file *f, char __user *buf, if (size & 0x3 || *pos & 0x3) return -EINVAL; + if (!adev->didt_rreg) + return -EOPNOTSUPP; + r = pm_runtime_get_sync(adev_to_drm(adev)->dev); if (r < 0) { pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); @@ -694,6 +697,9 @@ static ssize_t amdgpu_debugfs_regs_didt_write(struct file *f, const char __user if (size & 0x3 || *pos & 0x3) return -EINVAL; + if (!adev->didt_wreg) + return -EOPNOTSUPP; + r = pm_runtime_get_sync(adev_to_drm(adev)->dev); if (r < 0) { pm_runtime_put_autosuspend(adev_to_drm(adev)->dev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c index 7eeaf0aa7f81..93cf73d6fa11 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c @@ -2188,15 +2188,8 @@ static int amdgpu_device_parse_gpu_info_fw(struct amdgpu_device *adev) adev->firmware.gpu_info_fw = NULL; - if (adev->mman.discovery_bin) { - /* - * FIXME: The bounding box is still needed by Navi12, so - * temporarily read it from gpu_info firmware. Should be dropped - * when DAL no longer needs it. - */ - if (adev->asic_type != CHIP_NAVI12) - return 0; - } + if (adev->mman.discovery_bin) + return 0; switch (adev->asic_type) { default: @@ -3791,10 +3784,6 @@ static void amdgpu_device_set_mcbp(struct amdgpu_device *adev) adev->gfx.mcbp = true; else if (amdgpu_mcbp == 0) adev->gfx.mcbp = false; - else if ((amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 0, 0)) && - (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(10, 0, 0)) && - adev->gfx.num_gfx_rings) - adev->gfx.mcbp = true; if (amdgpu_sriov_vf(adev)) adev->gfx.mcbp = true; @@ -4520,8 +4509,6 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) amdgpu_ras_suspend(adev); - amdgpu_ttm_set_buffer_funcs_status(adev, false); - amdgpu_device_ip_suspend_phase1(adev); if (!adev->in_s0ix) @@ -4531,6 +4518,8 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) if (r) return r; + amdgpu_ttm_set_buffer_funcs_status(adev, false); + amdgpu_fence_driver_hw_fini(adev); amdgpu_device_ip_suspend_phase2(adev); @@ -4538,6 +4527,10 @@ int amdgpu_device_suspend(struct drm_device *dev, bool fbcon) if (amdgpu_sriov_vf(adev)) amdgpu_virt_release_full_gpu(adev, false); + r = amdgpu_dpm_notify_rlc_state(adev, false); + if (r) + return r; + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c index 0cacd0b9f8be..b8fbe97efe1d 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_display.c @@ -340,14 +340,11 @@ int amdgpu_display_crtc_set_config(struct drm_mode_set *set, adev->have_disp_power_ref = true; return ret; } - /* if we have no active crtcs, then drop the power ref - * we got before + /* if we have no active crtcs, then go to + * drop the power ref we got before */ - if (!active && adev->have_disp_power_ref) { - pm_runtime_put_autosuspend(dev->dev); + if (!active && adev->have_disp_power_ref) adev->have_disp_power_ref = false; - } - out: /* drop the power reference we got coming in here */ pm_runtime_put_autosuspend(dev->dev); diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c index 3095a3a864af..8b33b130ea36 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c @@ -207,6 +207,7 @@ int amdgpu_user_partt_mode = AMDGPU_AUTO_COMPUTE_PARTITION_MODE; int amdgpu_umsch_mm; int amdgpu_seamless = -1; /* auto */ uint amdgpu_debug_mask; +int amdgpu_agp = -1; /* auto */ static void amdgpu_drv_delayed_reset_work_handler(struct work_struct *work); @@ -961,6 +962,15 @@ module_param_named(seamless, amdgpu_seamless, int, 0444); MODULE_PARM_DESC(debug_mask, "debug options for amdgpu, disabled by default"); module_param_named(debug_mask, amdgpu_debug_mask, uint, 0444); +/** + * DOC: agp (int) + * Enable the AGP aperture. This provides an aperture in the GPU's internal + * address space for direct access to system memory. Note that these accesses + * are non-snooped, so they are only used for access to uncached memory. + */ +MODULE_PARM_DESC(agp, "AGP (-1 = auto (default), 0 = disable, 1 = enable)"); +module_param_named(agp, amdgpu_agp, int, 0444); + /* These devices are not supported by amdgpu. * They are supported by the mach64, r128, radeon drivers */ @@ -2253,6 +2263,8 @@ retry_init: pm_runtime_mark_last_busy(ddev->dev); pm_runtime_put_autosuspend(ddev->dev); + pci_wake_from_d3(pdev, TRUE); + /* * For runpm implemented via BACO, PMFW will handle the * timing for BACO in and out: diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c index 5f71414190e9..d2f273d77e59 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gmc.c @@ -181,6 +181,9 @@ uint64_t amdgpu_gmc_agp_addr(struct ttm_buffer_object *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->bdev); + if (!bo->ttm) + return AMDGPU_BO_INVALID_OFFSET; + if (bo->ttm->num_pages != 1 || bo->ttm->caching == ttm_cached) return AMDGPU_BO_INVALID_OFFSET; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h index 2b488fcf2f95..e51e8918e667 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_mca.h @@ -46,6 +46,8 @@ #define MCA_REG__STATUS__ERRORCODEEXT(x) MCA_REG_FIELD(x, 21, 16) #define MCA_REG__STATUS__ERRORCODE(x) MCA_REG_FIELD(x, 15, 0) +#define MCA_REG__SYND__ERRORINFORMATION(x) MCA_REG_FIELD(x, 17, 0) + enum amdgpu_mca_ip { AMDGPU_MCA_IP_UNKNOW = -1, AMDGPU_MCA_IP_PSP = 0, diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c index cef920a93924..5ad03f2afdb4 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_object.c @@ -1343,6 +1343,8 @@ void amdgpu_bo_release_notify(struct ttm_buffer_object *bo) abo = ttm_to_amdgpu_bo(bo); + WARN_ON(abo->vm_bo); + if (abo->kfd_bo) amdgpu_amdkfd_release_notify(abo); @@ -1527,10 +1529,14 @@ u64 amdgpu_bo_gpu_offset(struct amdgpu_bo *bo) u64 amdgpu_bo_gpu_offset_no_check(struct amdgpu_bo *bo) { struct amdgpu_device *adev = amdgpu_ttm_adev(bo->tbo.bdev); - uint64_t offset; + uint64_t offset = AMDGPU_BO_INVALID_OFFSET; + + if (bo->tbo.resource->mem_type == TTM_PL_TT) + offset = amdgpu_gmc_agp_addr(&bo->tbo); - offset = (bo->tbo.resource->start << PAGE_SHIFT) + - amdgpu_ttm_domain_start(adev, bo->tbo.resource->mem_type); + if (offset == AMDGPU_BO_INVALID_OFFSET) + offset = (bo->tbo.resource->start << PAGE_SHIFT) + + amdgpu_ttm_domain_start(adev, bo->tbo.resource->mem_type); return amdgpu_gmc_sign_extend(offset); } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c index 32b701cc0376..a21045d018f2 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c @@ -1473,6 +1473,11 @@ int psp_xgmi_get_topology_info(struct psp_context *psp, topology->nodes[i].num_links = (requires_reflection && topology->nodes[i].num_links) ? topology->nodes[i].num_links : node_num_links; } + /* popluate the connected port num info if supported and available */ + if (ta_port_num_support && topology->nodes[i].num_links) { + memcpy(topology->nodes[i].port_num, link_extend_info_output->nodes[i].port_num, + sizeof(struct xgmi_connected_port_num) * TA_XGMI__MAX_PORT_NUM); + } /* reflect the topology information for bi-directionality */ if (requires_reflection && topology->nodes[i].num_hops) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h index 5d36ad3f48c7..c4d9cbde55b9 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_psp.h @@ -150,6 +150,7 @@ struct psp_xgmi_node_info { uint8_t is_sharing_enabled; enum ta_xgmi_assigned_sdma_engine sdma_engine; uint8_t num_links; + struct xgmi_connected_port_num port_num[TA_XGMI__MAX_PORT_NUM]; }; struct psp_xgmi_topology_info { diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c index 84e5987b14e0..63fb4cd85e53 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras.c @@ -28,6 +28,7 @@ #include <linux/reboot.h> #include <linux/syscalls.h> #include <linux/pm_runtime.h> +#include <linux/list_sort.h> #include "amdgpu.h" #include "amdgpu_ras.h" @@ -1188,7 +1189,7 @@ static int amdgpu_ras_query_error_status_helper(struct amdgpu_device *adev, } if (block_obj->hw_ops->query_ras_error_count) - block_obj->hw_ops->query_ras_error_count(adev, &err_data); + block_obj->hw_ops->query_ras_error_count(adev, err_data); if ((info->head.block == AMDGPU_RAS_BLOCK__SDMA) || (info->head.block == AMDGPU_RAS_BLOCK__GFX) || @@ -3665,6 +3666,21 @@ static struct ras_err_node *amdgpu_ras_error_node_new(void) return err_node; } +static int ras_err_info_cmp(void *priv, const struct list_head *a, const struct list_head *b) +{ + struct ras_err_node *nodea = container_of(a, struct ras_err_node, node); + struct ras_err_node *nodeb = container_of(b, struct ras_err_node, node); + struct amdgpu_smuio_mcm_config_info *infoa = &nodea->err_info.mcm_info; + struct amdgpu_smuio_mcm_config_info *infob = &nodeb->err_info.mcm_info; + + if (unlikely(infoa->socket_id != infob->socket_id)) + return infoa->socket_id - infob->socket_id; + else + return infoa->die_id - infob->die_id; + + return 0; +} + static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_data, struct amdgpu_smuio_mcm_config_info *mcm_info) { @@ -3682,6 +3698,7 @@ static struct ras_err_info *amdgpu_ras_error_get_info(struct ras_err_data *err_d err_data->err_list_count++; list_add_tail(&err_node->node, &err_data->err_node_list); + list_sort(NULL, &err_data->err_node_list, ras_err_info_cmp); return &err_node->err_info; } diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c index 65aa218380be..2fde93b00cab 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ras_eeprom.c @@ -214,6 +214,12 @@ static bool __get_eeprom_i2c_addr(struct amdgpu_device *adev, control->i2c_address = EEPROM_I2C_MADDR_0; return true; case IP_VERSION(13, 0, 0): + if (strnstr(atom_ctx->vbios_pn, "D707", + sizeof(atom_ctx->vbios_pn))) + control->i2c_address = EEPROM_I2C_MADDR_0; + else + control->i2c_address = EEPROM_I2C_MADDR_4; + return true; case IP_VERSION(13, 0, 6): case IP_VERSION(13, 0, 10): control->i2c_address = EEPROM_I2C_MADDR_4; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c index 05991c5c8ddb..ab4a762aed5b 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c @@ -959,10 +959,8 @@ int amdgpu_ttm_alloc_gart(struct ttm_buffer_object *bo) return 0; addr = amdgpu_gmc_agp_addr(bo); - if (addr != AMDGPU_BO_INVALID_OFFSET) { - bo->resource->start = addr >> PAGE_SHIFT; + if (addr != AMDGPU_BO_INVALID_OFFSET) return 0; - } /* allocate GART space */ placement.num_placement = 1; diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c index 65949cc7abb9..07d930339b07 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_uvd.c @@ -398,6 +398,7 @@ int amdgpu_uvd_sw_fini(struct amdgpu_device *adev) * amdgpu_uvd_entity_init - init entity * * @adev: amdgpu_device pointer + * @ring: amdgpu_ring pointer to check * * Initialize the entity used for handle management in the kernel driver. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c index 0954447f689d..59acf424a078 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vce.c @@ -230,6 +230,7 @@ int amdgpu_vce_sw_fini(struct amdgpu_device *adev) * amdgpu_vce_entity_init - init entity * * @adev: amdgpu_device pointer + * @ring: amdgpu_ring pointer to check * * Initialize the entity used for handle management in the kernel driver. */ diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c index d1b8afd105c9..5baefb548a29 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm.c @@ -285,6 +285,7 @@ static void amdgpu_vm_bo_reset_state_machine(struct amdgpu_vm *vm) list_for_each_entry_safe(vm_bo, tmp, &vm->idle, vm_status) { struct amdgpu_bo *bo = vm_bo->bo; + vm_bo->moved = true; if (!bo || bo->tbo.type != ttm_bo_type_kernel) list_move(&vm_bo->vm_status, &vm_bo->vm->moved); else if (bo->parent) diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c index a2287bb25223..a160265ddc07 100644 --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_vm_pt.c @@ -642,13 +642,14 @@ static void amdgpu_vm_pt_free(struct amdgpu_vm_bo_base *entry) if (!entry->bo) return; + + entry->bo->vm_bo = NULL; shadow = amdgpu_bo_shadowed(entry->bo); if (shadow) { ttm_bo_set_bulk_move(&shadow->tbo, NULL); amdgpu_bo_unref(&shadow); } ttm_bo_set_bulk_move(&entry->bo->tbo, NULL); - entry->bo->vm_bo = NULL; spin_lock(&entry->vm->status_lock); list_del(&entry->vm_status); diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c index 0c6133cc5e57..8ed4a6fb147a 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v11_0.c @@ -89,6 +89,10 @@ MODULE_FIRMWARE("amdgpu/gc_11_5_0_me.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_0_mec.bin"); MODULE_FIRMWARE("amdgpu/gc_11_5_0_rlc.bin"); +static const struct soc15_reg_golden golden_settings_gc_11_0[] = { + SOC15_REG_GOLDEN_VALUE(GC, 0, regTCP_CNTL, 0x20000000, 0x20000000) +}; + static const struct soc15_reg_golden golden_settings_gc_11_0_1[] = { SOC15_REG_GOLDEN_VALUE(GC, 0, regCGTT_GS_NGG_CLK_CTRL, 0x9fff8fff, 0x00000010), @@ -304,6 +308,10 @@ static void gfx_v11_0_init_golden_registers(struct amdgpu_device *adev) default: break; } + soc15_program_register_sequence(adev, + golden_settings_gc_11_0, + (const u32)ARRAY_SIZE(golden_settings_gc_11_0)); + } static void gfx_v11_0_write_data_to_reg(struct amdgpu_ring *ring, int eng_sel, @@ -419,7 +427,7 @@ static int gfx_v11_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); cpu_ptr = &adev->wb.wb[index]; - r = amdgpu_ib_get(adev, NULL, 16, AMDGPU_IB_POOL_DIRECT, &ib); + r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); if (r) { DRM_ERROR("amdgpu: failed to get ib (%ld).\n", r); goto err1; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c index 885ebd703260..1943beb135c4 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v8_0.c @@ -883,8 +883,8 @@ static int gfx_v8_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) gpu_addr = adev->wb.gpu_addr + (index * 4); adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(adev, NULL, 16, - AMDGPU_IB_POOL_DIRECT, &ib); + + r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); if (r) goto err1; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c index e3ff6e46f3f7..69c500910746 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_0.c @@ -1039,8 +1039,8 @@ static int gfx_v9_0_ring_test_ib(struct amdgpu_ring *ring, long timeout) gpu_addr = adev->wb.gpu_addr + (index * 4); adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(adev, NULL, 16, - AMDGPU_IB_POOL_DIRECT, &ib); + + r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); if (r) goto err1; diff --git a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c index 40d06d32bb74..4a09cc0d8ce0 100644 --- a/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c +++ b/drivers/gpu/drm/amd/amdgpu/gfx_v9_4_3.c @@ -297,8 +297,8 @@ static int gfx_v9_4_3_ring_test_ib(struct amdgpu_ring *ring, long timeout) gpu_addr = adev->wb.gpu_addr + (index * 4); adev->wb.wb[index] = cpu_to_le32(0xCAFEDEAD); memset(&ib, 0, sizeof(ib)); - r = amdgpu_ib_get(adev, NULL, 16, - AMDGPU_IB_POOL_DIRECT, &ib); + + r = amdgpu_ib_get(adev, NULL, 20, AMDGPU_IB_POOL_DIRECT, &ib); if (r) goto err1; diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c index 0ec7b061d7c2..a5a05c16c10d 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v10_0.c @@ -675,7 +675,7 @@ static void gmc_v10_0_vram_gtt_location(struct amdgpu_device *adev, amdgpu_gmc_set_agp_default(adev, mc); amdgpu_gmc_vram_location(adev, &adev->gmc, base); amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT); - if (!amdgpu_sriov_vf(adev)) + if (!amdgpu_sriov_vf(adev) && (amdgpu_agp == 1)) amdgpu_gmc_agp_location(adev, mc); /* base offset of vram pages */ diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c index 6dce9b29f675..23d7b548d13f 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v11_0.c @@ -640,8 +640,9 @@ static void gmc_v11_0_vram_gtt_location(struct amdgpu_device *adev, amdgpu_gmc_set_agp_default(adev, mc); amdgpu_gmc_vram_location(adev, &adev->gmc, base); amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_HIGH); - if (!amdgpu_sriov_vf(adev) || - (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(11, 5, 0))) + if (!amdgpu_sriov_vf(adev) && + (amdgpu_ip_version(adev, GC_HWIP, 0) < IP_VERSION(11, 5, 0)) && + (amdgpu_agp == 1)) amdgpu_gmc_agp_location(adev, mc); /* base offset of vram pages */ diff --git a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c index bde25eb4ed8e..2ac5820e9c92 100644 --- a/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c +++ b/drivers/gpu/drm/amd/amdgpu/gmc_v9_0.c @@ -1630,7 +1630,7 @@ static void gmc_v9_0_vram_gtt_location(struct amdgpu_device *adev, } else { amdgpu_gmc_vram_location(adev, mc, base); amdgpu_gmc_gart_location(adev, mc, AMDGPU_GART_PLACEMENT_BEST_FIT); - if (!amdgpu_sriov_vf(adev)) + if (!amdgpu_sriov_vf(adev) && (amdgpu_agp == 1)) amdgpu_gmc_agp_location(adev, mc); } /* base offset of vram pages */ @@ -2170,8 +2170,6 @@ static int gmc_v9_0_sw_fini(void *handle) if (amdgpu_ip_version(adev, GC_HWIP, 0) == IP_VERSION(9, 4, 3)) amdgpu_gmc_sysfs_fini(adev); - adev->gmc.num_mem_partitions = 0; - kfree(adev->gmc.mem_partitions); amdgpu_gmc_ras_fini(adev); amdgpu_gem_force_release(adev); @@ -2185,6 +2183,9 @@ static int gmc_v9_0_sw_fini(void *handle) amdgpu_bo_free_kernel(&adev->gmc.pdb0_bo, NULL, &adev->gmc.ptr_pdb0); amdgpu_bo_fini(adev); + adev->gmc.num_mem_partitions = 0; + kfree(adev->gmc.mem_partitions); + return 0; } diff --git a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c index 49e934975719..4db6bb73ead4 100644 --- a/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c +++ b/drivers/gpu/drm/amd/amdgpu/hdp_v4_0.c @@ -129,6 +129,11 @@ static void hdp_v4_0_get_clockgating_state(struct amdgpu_device *adev, { int data; + if (amdgpu_ip_version(adev, HDP_HWIP, 0) == IP_VERSION(4, 4, 2)) { + /* Default enabled */ + *flags |= AMD_CG_SUPPORT_HDP_MGCG; + return; + } /* AMD_CG_SUPPORT_HDP_LS */ data = RREG32(SOC15_REG_OFFSET(HDP, 0, mmHDP_MEM_POWER_LS)); if (data & HDP_MEM_POWER_LS__LS_ENABLE_MASK) diff --git a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c index 9df011323d4b..6ede85b28cc8 100644 --- a/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c +++ b/drivers/gpu/drm/amd/amdgpu/jpeg_v4_0_5.c @@ -155,13 +155,6 @@ static int jpeg_v4_0_5_hw_init(void *handle) struct amdgpu_ring *ring = adev->jpeg.inst->ring_dec; int r; - adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, - (adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0); - - WREG32_SOC15(VCN, 0, regVCN_JPEG_DB_CTRL, - ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT | - VCN_JPEG_DB_CTRL__EN_MASK); - r = amdgpu_ring_test_helper(ring); if (r) return r; @@ -336,6 +329,14 @@ static int jpeg_v4_0_5_start(struct amdgpu_device *adev) if (adev->pm.dpm_enabled) amdgpu_dpm_enable_jpeg(adev, true); + /* doorbell programming is done for every playback */ + adev->nbio.funcs->vcn_doorbell_range(adev, ring->use_doorbell, + (adev->doorbell_index.vcn.vcn_ring0_1 << 1), 0); + + WREG32_SOC15(VCN, 0, regVCN_JPEG_DB_CTRL, + ring->doorbell_index << VCN_JPEG_DB_CTRL__OFFSET__SHIFT | + VCN_JPEG_DB_CTRL__EN_MASK); + /* disable power gating */ r = jpeg_v4_0_5_disable_static_power_gating(adev); if (r) diff --git a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c index ea142611be1c..9b0146732e13 100644 --- a/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c +++ b/drivers/gpu/drm/amd/amdgpu/mmhub_v1_8.c @@ -130,6 +130,9 @@ static void mmhub_v1_8_init_system_aperture_regs(struct amdgpu_device *adev) uint64_t value; int i; + if (amdgpu_sriov_vf(adev)) + return; + inst_mask = adev->aid_mask; for_each_inst(i, inst_mask) { /* Program the AGP BAR */ @@ -139,9 +142,6 @@ static void mmhub_v1_8_init_system_aperture_regs(struct amdgpu_device *adev) WREG32_SOC15(MMHUB, i, regMC_VM_AGP_TOP, adev->gmc.agp_end >> 24); - if (amdgpu_sriov_vf(adev)) - return; - /* Program the system aperture low logical page number. */ WREG32_SOC15(MMHUB, i, regMC_VM_SYSTEM_APERTURE_LOW_ADDR, min(adev->gmc.fb_start, adev->gmc.agp_start) >> 18); diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c index 676ab1d20d2f..1f52b4b1db03 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_11.c @@ -259,17 +259,17 @@ const struct nbio_hdp_flush_reg nbio_v7_11_hdp_flush_reg = { static void nbio_v7_11_init_registers(struct amdgpu_device *adev) { -/* uint32_t def, data; + uint32_t def, data; + + def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3); + data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, + CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1); + data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, + CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV, 1); - def = data = RREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3); - data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, - CI_SWUS_MAX_READ_REQUEST_SIZE_MODE, 1); - data = REG_SET_FIELD(data, BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, - CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV, 1); + if (def != data) + WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, data); - if (def != data) - WREG32_SOC15(NBIO, 0, regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3, data); -*/ } static void nbio_v7_11_update_medium_grain_clock_gating(struct amdgpu_device *adev, diff --git a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c index 23f26f8caad4..25a3da83e0fb 100644 --- a/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c +++ b/drivers/gpu/drm/amd/amdgpu/nbio_v7_9.c @@ -611,11 +611,6 @@ static void nbio_v7_9_handle_ras_controller_intr_no_bifring(struct amdgpu_device dev_info(adev->dev, "RAS controller interrupt triggered " "by NBIF error\n"); - - /* ras_controller_int is dedicated for nbif ras error, - * not the global interrupt for sync flood - */ - amdgpu_ras_reset_gpu(adev); } amdgpu_ras_error_data_fini(&err_data); diff --git a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c index 3cf4684d0d3f..df1844d0800f 100644 --- a/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c +++ b/drivers/gpu/drm/amd/amdgpu/psp_v13_0.c @@ -60,7 +60,7 @@ MODULE_FIRMWARE("amdgpu/psp_14_0_0_ta.bin"); #define GFX_CMD_USB_PD_USE_LFB 0x480 /* Retry times for vmbx ready wait */ -#define PSP_VMBX_POLLING_LIMIT 20000 +#define PSP_VMBX_POLLING_LIMIT 3000 /* VBIOS gfl defines */ #define MBOX_READY_MASK 0x80000000 @@ -161,14 +161,18 @@ static int psp_v13_0_wait_for_vmbx_ready(struct psp_context *psp) static int psp_v13_0_wait_for_bootloader(struct psp_context *psp) { struct amdgpu_device *adev = psp->adev; - int retry_loop, ret; + int retry_loop, retry_cnt, ret; + retry_cnt = + (amdgpu_ip_version(adev, MP0_HWIP, 0) == IP_VERSION(13, 0, 6)) ? + PSP_VMBX_POLLING_LIMIT : + 10; /* Wait for bootloader to signify that it is ready having bit 31 of * C2PMSG_35 set to 1. All other bits are expected to be cleared. * If there is an error in processing command, bits[7:0] will be set. * This is applicable for PSP v13.0.6 and newer. */ - for (retry_loop = 0; retry_loop < PSP_VMBX_POLLING_LIMIT; retry_loop++) { + for (retry_loop = 0; retry_loop < retry_cnt; retry_loop++) { ret = psp_wait_for( psp, SOC15_REG_OFFSET(MP0, 0, regMP0_SMN_C2PMSG_35), 0x80000000, 0xffffffff, false); @@ -821,7 +825,7 @@ static int psp_v13_0_query_boot_status(struct psp_context *psp) if (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6)) return 0; - if (RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_59) < 0x00a10007) + if (RREG32_SOC15(MP0, 0, regMP0_SMN_C2PMSG_59) < 0x00a10109) return 0; for_each_inst(i, inst_mask) { diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c index 45377a175250..8d5d86675a7f 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v2_4.c @@ -813,12 +813,12 @@ static int sdma_v2_4_early_init(void *handle) struct amdgpu_device *adev = (struct amdgpu_device *)handle; int r; + adev->sdma.num_instances = SDMA_MAX_INSTANCE; + r = sdma_v2_4_init_microcode(adev); if (r) return r; - adev->sdma.num_instances = SDMA_MAX_INSTANCE; - sdma_v2_4_set_ring_funcs(adev); sdma_v2_4_set_buffer_funcs(adev); sdma_v2_4_set_vm_pte_funcs(adev); diff --git a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c index 83c240f741b5..0058f3f7cf6e 100644 --- a/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c +++ b/drivers/gpu/drm/amd/amdgpu/sdma_v5_2.c @@ -1643,6 +1643,32 @@ static void sdma_v5_2_get_clockgating_state(void *handle, u64 *flags) *flags |= AMD_CG_SUPPORT_SDMA_LS; } +static void sdma_v5_2_ring_begin_use(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + /* SDMA 5.2.3 (RMB) FW doesn't seem to properly + * disallow GFXOFF in some cases leading to + * hangs in SDMA. Disallow GFXOFF while SDMA is active. + * We can probably just limit this to 5.2.3, + * but it shouldn't hurt for other parts since + * this GFXOFF will be disallowed anyway when SDMA is + * active, this just makes it explicit. + */ + amdgpu_gfx_off_ctrl(adev, false); +} + +static void sdma_v5_2_ring_end_use(struct amdgpu_ring *ring) +{ + struct amdgpu_device *adev = ring->adev; + + /* SDMA 5.2.3 (RMB) FW doesn't seem to properly + * disallow GFXOFF in some cases leading to + * hangs in SDMA. Allow GFXOFF when SDMA is complete. + */ + amdgpu_gfx_off_ctrl(adev, true); +} + const struct amd_ip_funcs sdma_v5_2_ip_funcs = { .name = "sdma_v5_2", .early_init = sdma_v5_2_early_init, @@ -1690,6 +1716,8 @@ static const struct amdgpu_ring_funcs sdma_v5_2_ring_funcs = { .test_ib = sdma_v5_2_ring_test_ib, .insert_nop = sdma_v5_2_ring_insert_nop, .pad_ib = sdma_v5_2_ring_pad_ib, + .begin_use = sdma_v5_2_ring_begin_use, + .end_use = sdma_v5_2_ring_end_use, .emit_wreg = sdma_v5_2_ring_emit_wreg, .emit_reg_wait = sdma_v5_2_ring_emit_reg_wait, .emit_reg_write_reg_wait = sdma_v5_2_ring_emit_reg_write_reg_wait, diff --git a/drivers/gpu/drm/amd/amdgpu/soc15.c b/drivers/gpu/drm/amd/amdgpu/soc15.c index d4b8d62f4294..51342809af03 100644 --- a/drivers/gpu/drm/amd/amdgpu/soc15.c +++ b/drivers/gpu/drm/amd/amdgpu/soc15.c @@ -1161,6 +1161,11 @@ static int soc15_common_early_init(void *handle) AMD_PG_SUPPORT_VCN_DPG | AMD_PG_SUPPORT_JPEG; adev->external_rev_id = adev->rev_id + 0x46; + /* GC 9.4.3 uses MMIO register region hole at a different offset */ + if (!amdgpu_sriov_vf(adev)) { + adev->rmmio_remap.reg_offset = 0x1A000; + adev->rmmio_remap.bus_addr = adev->rmmio_base + 0x1A000; + } break; default: /* FIXME: not supported yet */ @@ -1418,11 +1423,14 @@ static void soc15_common_get_clockgating_state(void *handle, u64 *flags) if (amdgpu_sriov_vf(adev)) *flags = 0; - adev->nbio.funcs->get_clockgating_state(adev, flags); + if (adev->nbio.funcs && adev->nbio.funcs->get_clockgating_state) + adev->nbio.funcs->get_clockgating_state(adev, flags); - adev->hdp.funcs->get_clock_gating_state(adev, flags); + if (adev->hdp.funcs && adev->hdp.funcs->get_clock_gating_state) + adev->hdp.funcs->get_clock_gating_state(adev, flags); - if (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 2)) { + if ((amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 2)) && + (amdgpu_ip_version(adev, MP0_HWIP, 0) != IP_VERSION(13, 0, 6))) { /* AMD_CG_SUPPORT_DRM_MGCG */ data = RREG32(SOC15_REG_OFFSET(MP0, 0, mmMP0_MISC_CGTT_CTRL0)); if (!(data & 0x01000000)) @@ -1435,9 +1443,11 @@ static void soc15_common_get_clockgating_state(void *handle, u64 *flags) } /* AMD_CG_SUPPORT_ROM_MGCG */ - adev->smuio.funcs->get_clock_gating_state(adev, flags); + if (adev->smuio.funcs && adev->smuio.funcs->get_clock_gating_state) + adev->smuio.funcs->get_clock_gating_state(adev, flags); - adev->df.funcs->get_clockgating_state(adev, flags); + if (adev->df.funcs && adev->df.funcs->get_clockgating_state) + adev->df.funcs->get_clockgating_state(adev, flags); } static int soc15_common_set_powergating_state(void *handle, diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h index 9cc32f577e38..4c8e278a0d0c 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_priv.h +++ b/drivers/gpu/drm/amd/amdkfd/kfd_priv.h @@ -1128,7 +1128,7 @@ static inline struct kfd_node *kfd_node_by_irq_ids(struct amdgpu_device *adev, struct kfd_dev *dev = adev->kfd.dev; uint32_t i; - if (adev->ip_versions[GC_HWIP][0] != IP_VERSION(9, 4, 3)) + if (KFD_GC_VERSION(dev) != IP_VERSION(9, 4, 3)) return dev->nodes[0]; for (i = 0; i < dev->num_nodes; i++) diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c index 77649392e233..77f493262e05 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_process_queue_manager.c @@ -169,16 +169,43 @@ int pqm_init(struct process_queue_manager *pqm, struct kfd_process *p) return 0; } +static void pqm_clean_queue_resource(struct process_queue_manager *pqm, + struct process_queue_node *pqn) +{ + struct kfd_node *dev; + struct kfd_process_device *pdd; + + dev = pqn->q->device; + + pdd = kfd_get_process_device_data(dev, pqm->process); + if (!pdd) { + pr_err("Process device data doesn't exist\n"); + return; + } + + if (pqn->q->gws) { + if (KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 4, 3) && + !dev->kfd->shared_resources.enable_mes) + amdgpu_amdkfd_remove_gws_from_process( + pqm->process->kgd_process_info, pqn->q->gws); + pdd->qpd.num_gws = 0; + } + + if (dev->kfd->shared_resources.enable_mes) { + amdgpu_amdkfd_free_gtt_mem(dev->adev, pqn->q->gang_ctx_bo); + if (pqn->q->wptr_bo) + amdgpu_amdkfd_free_gtt_mem(dev->adev, pqn->q->wptr_bo); + } +} + void pqm_uninit(struct process_queue_manager *pqm) { struct process_queue_node *pqn, *next; list_for_each_entry_safe(pqn, next, &pqm->queues, process_queue_list) { - if (pqn->q && pqn->q->gws && - KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 4, 3) && - !pqn->q->device->kfd->shared_resources.enable_mes) - amdgpu_amdkfd_remove_gws_from_process(pqm->process->kgd_process_info, - pqn->q->gws); + if (pqn->q) + pqm_clean_queue_resource(pqm, pqn); + kfd_procfs_del_queue(pqn->q); uninit_queue(pqn->q); list_del(&pqn->process_queue_list); @@ -461,22 +488,7 @@ int pqm_destroy_queue(struct process_queue_manager *pqm, unsigned int qid) goto err_destroy_queue; } - if (pqn->q->gws) { - if (KFD_GC_VERSION(pqn->q->device) != IP_VERSION(9, 4, 3) && - !dev->kfd->shared_resources.enable_mes) - amdgpu_amdkfd_remove_gws_from_process( - pqm->process->kgd_process_info, - pqn->q->gws); - pdd->qpd.num_gws = 0; - } - - if (dev->kfd->shared_resources.enable_mes) { - amdgpu_amdkfd_free_gtt_mem(dev->adev, - pqn->q->gang_ctx_bo); - if (pqn->q->wptr_bo) - amdgpu_amdkfd_free_gtt_mem(dev->adev, pqn->q->wptr_bo); - - } + pqm_clean_queue_resource(pqm, pqn); uninit_queue(pqn->q); } diff --git a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c index f2f3c338fd94..a15bfb5223e8 100644 --- a/drivers/gpu/drm/amd/amdkfd/kfd_svm.c +++ b/drivers/gpu/drm/amd/amdkfd/kfd_svm.c @@ -1653,18 +1653,24 @@ static int svm_range_validate_and_map(struct mm_struct *mm, if (test_bit(gpuidx, prange->bitmap_access)) bitmap_set(ctx->bitmap, gpuidx, 1); } + + /* + * If prange is already mapped or with always mapped flag, + * update mapping on GPUs with ACCESS attribute + */ + if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) { + if (prange->mapped_to_gpu || + prange->flags & KFD_IOCTL_SVM_FLAG_GPU_ALWAYS_MAPPED) + bitmap_copy(ctx->bitmap, prange->bitmap_access, MAX_GPU_INSTANCE); + } } else { bitmap_or(ctx->bitmap, prange->bitmap_access, prange->bitmap_aip, MAX_GPU_INSTANCE); } if (bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) { - bitmap_copy(ctx->bitmap, prange->bitmap_access, MAX_GPU_INSTANCE); - if (!prange->mapped_to_gpu || - bitmap_empty(ctx->bitmap, MAX_GPU_INSTANCE)) { - r = 0; - goto free_ctx; - } + r = 0; + goto free_ctx; } if (prange->actual_loc && !prange->ttm_res) { diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c index 6f99f6754c11..4e82ee4d74ac 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c @@ -2079,7 +2079,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) struct dmub_srv_create_params create_params; struct dmub_srv_region_params region_params; struct dmub_srv_region_info region_info; - struct dmub_srv_fb_params fb_params; + struct dmub_srv_memory_params memory_params; struct dmub_srv_fb_info *fb_info; struct dmub_srv *dmub_srv; const struct dmcub_firmware_header_v1_0 *hdr; @@ -2182,6 +2182,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) adev->dm.dmub_fw->data + le32_to_cpu(hdr->header.ucode_array_offset_bytes) + PSP_HEADER_BYTES; + region_params.is_mailbox_in_inbox = false; status = dmub_srv_calc_region_info(dmub_srv, ®ion_params, ®ion_info); @@ -2205,10 +2206,10 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) return r; /* Rebase the regions on the framebuffer address. */ - memset(&fb_params, 0, sizeof(fb_params)); - fb_params.cpu_addr = adev->dm.dmub_bo_cpu_addr; - fb_params.gpu_addr = adev->dm.dmub_bo_gpu_addr; - fb_params.region_info = ®ion_info; + memset(&memory_params, 0, sizeof(memory_params)); + memory_params.cpu_fb_addr = adev->dm.dmub_bo_cpu_addr; + memory_params.gpu_fb_addr = adev->dm.dmub_bo_gpu_addr; + memory_params.region_info = ®ion_info; adev->dm.dmub_fb_info = kzalloc(sizeof(*adev->dm.dmub_fb_info), GFP_KERNEL); @@ -2220,7 +2221,7 @@ static int dm_dmub_sw_init(struct amdgpu_device *adev) return -ENOMEM; } - status = dmub_srv_calc_fb_info(dmub_srv, &fb_params, fb_info); + status = dmub_srv_calc_mem_info(dmub_srv, &memory_params, fb_info); if (status != DMUB_STATUS_OK) { DRM_ERROR("Error calculating DMUB FB info: %d\n", status); return -EINVAL; @@ -5181,6 +5182,9 @@ static void fill_dc_dirty_rects(struct drm_plane *plane, if (plane->type == DRM_PLANE_TYPE_CURSOR) return; + if (new_plane_state->rotation != DRM_MODE_ROTATE_0) + goto ffu; + num_clips = drm_plane_get_damage_clips_count(new_plane_state); clips = drm_plane_get_damage_clips(new_plane_state); @@ -6166,8 +6170,9 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, if (stream->signal == SIGNAL_TYPE_HDMI_TYPE_A) mod_build_hf_vsif_infopacket(stream, &stream->vsp_infopacket); - - if (stream->link->psr_settings.psr_feature_enabled || stream->link->replay_settings.replay_feature_enabled) { + else if (stream->signal == SIGNAL_TYPE_DISPLAY_PORT || + stream->signal == SIGNAL_TYPE_DISPLAY_PORT_MST || + stream->signal == SIGNAL_TYPE_EDP) { // // should decide stream support vsc sdp colorimetry capability // before building vsc info packet @@ -6183,8 +6188,9 @@ create_stream_for_sink(struct amdgpu_dm_connector *aconnector, if (stream->out_transfer_func->tf == TRANSFER_FUNCTION_GAMMA22) tf = TRANSFER_FUNC_GAMMA_22; mod_build_vsc_infopacket(stream, &stream->vsc_infopacket, stream->output_color_space, tf); - aconnector->psr_skip_count = AMDGPU_DM_PSR_ENTRY_DELAY; + if (stream->link->psr_settings.psr_feature_enabled) + aconnector->psr_skip_count = AMDGPU_DM_PSR_ENTRY_DELAY; } finish: dc_sink_release(sink); @@ -6266,7 +6272,7 @@ int amdgpu_dm_connector_atomic_set_property(struct drm_connector *connector, dm_new_state->underscan_enable = val; ret = 0; } else if (property == adev->mode_info.abm_level_property) { - dm_new_state->abm_level = val; + dm_new_state->abm_level = val ?: ABM_LEVEL_IMMEDIATE_DISABLE; ret = 0; } @@ -6311,7 +6317,8 @@ int amdgpu_dm_connector_atomic_get_property(struct drm_connector *connector, *val = dm_state->underscan_enable; ret = 0; } else if (property == adev->mode_info.abm_level_property) { - *val = dm_state->abm_level; + *val = (dm_state->abm_level != ABM_LEVEL_IMMEDIATE_DISABLE) ? + dm_state->abm_level : 0; ret = 0; } @@ -6384,7 +6391,8 @@ void amdgpu_dm_connector_funcs_reset(struct drm_connector *connector) state->pbn = 0; if (connector->connector_type == DRM_MODE_CONNECTOR_eDP) - state->abm_level = amdgpu_dm_abm_level; + state->abm_level = amdgpu_dm_abm_level ?: + ABM_LEVEL_IMMEDIATE_DISABLE; __drm_atomic_helper_connector_reset(connector, &state->base); } @@ -6908,8 +6916,7 @@ static int dm_encoder_helper_atomic_check(struct drm_encoder *encoder, if (IS_ERR(mst_state)) return PTR_ERR(mst_state); - if (!mst_state->pbn_div) - mst_state->pbn_div = dm_mst_get_pbn_divider(aconnector->mst_root->dc_link); + mst_state->pbn_div = dm_mst_get_pbn_divider(aconnector->mst_root->dc_link); if (!state->duplicated) { int max_bpc = conn_state->max_requested_bpc; @@ -7481,6 +7488,9 @@ static int amdgpu_dm_i2c_xfer(struct i2c_adapter *i2c_adap, int i; int result = -EIO; + if (!ddc_service->ddc_pin || !ddc_service->ddc_pin->hw_info.hw_supported) + return result; + cmd.payloads = kcalloc(num, sizeof(struct i2c_payload), GFP_KERNEL); if (!cmd.payloads) @@ -9603,14 +9613,14 @@ static bool should_reset_plane(struct drm_atomic_state *state, struct drm_plane *other; struct drm_plane_state *old_other_state, *new_other_state; struct drm_crtc_state *new_crtc_state; + struct amdgpu_device *adev = drm_to_adev(plane->dev); int i; /* - * TODO: Remove this hack once the checks below are sufficient - * enough to determine when we need to reset all the planes on - * the stream. + * TODO: Remove this hack for all asics once it proves that the + * fast updates works fine on DCN3.2+. */ - if (state->allow_modeset) + if (adev->ip_versions[DCE_HWIP][0] < IP_VERSION(3, 2, 0) && state->allow_modeset) return true; /* Exit early if we know that we're adding or removing the plane. */ diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c index ed784cf27d39..aac98f93545a 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_helpers.c @@ -63,6 +63,12 @@ static void apply_edid_quirks(struct edid *edid, struct dc_edid_caps *edid_caps) DRM_DEBUG_DRIVER("Disabling FAMS on monitor with panel id %X\n", panel_id); edid_caps->panel_patch.disable_fams = true; break; + /* Workaround for some monitors that do not clear DPCD 0x317 if FreeSync is unsupported */ + case drm_edid_encode_panel_id('A', 'U', 'O', 0xA7AB): + case drm_edid_encode_panel_id('A', 'U', 'O', 0xE69B): + DRM_DEBUG_DRIVER("Clearing DPCD 0x317 on monitor with panel id %X\n", panel_id); + edid_caps->panel_patch.remove_sink_ext_caps = true; + break; default: return; } @@ -536,11 +542,8 @@ bool dm_helpers_dp_read_dpcd( struct amdgpu_dm_connector *aconnector = link->priv; - if (!aconnector) { - drm_dbg_dp(aconnector->base.dev, - "Failed to find connector for link!\n"); + if (!aconnector) return false; - } return drm_dp_dpcd_read(&aconnector->dm_dp_aux.aux, address, data, size) == size; diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c index d3b13d362eda..11da0eebee6c 100644 --- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c +++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_mst_types.c @@ -1604,31 +1604,31 @@ enum dc_status dm_dp_mst_is_port_support_mode( unsigned int upper_link_bw_in_kbps = 0, down_link_bw_in_kbps = 0; unsigned int max_compressed_bw_in_kbps = 0; struct dc_dsc_bw_range bw_range = {0}; - struct drm_dp_mst_topology_mgr *mst_mgr; + uint16_t full_pbn = aconnector->mst_output_port->full_pbn; /* - * check if the mode could be supported if DSC pass-through is supported - * AND check if there enough bandwidth available to support the mode - * with DSC enabled. + * Consider the case with the depth of the mst topology tree is equal or less than 2 + * A. When dsc bitstream can be transmitted along the entire path + * 1. dsc is possible between source and branch/leaf device (common dsc params is possible), AND + * 2. dsc passthrough supported at MST branch, or + * 3. dsc decoding supported at leaf MST device + * Use maximum dsc compression as bw constraint + * B. When dsc bitstream cannot be transmitted along the entire path + * Use native bw as bw constraint */ if (is_dsc_common_config_possible(stream, &bw_range) && - aconnector->mst_output_port->passthrough_aux) { - mst_mgr = aconnector->mst_output_port->mgr; - mutex_lock(&mst_mgr->lock); - + (aconnector->mst_output_port->passthrough_aux || + aconnector->dsc_aux == &aconnector->mst_output_port->aux)) { cur_link_settings = stream->link->verified_link_cap; upper_link_bw_in_kbps = dc_link_bandwidth_kbps(aconnector->dc_link, - &cur_link_settings - ); - down_link_bw_in_kbps = kbps_from_pbn(aconnector->mst_output_port->full_pbn); + &cur_link_settings); + down_link_bw_in_kbps = kbps_from_pbn(full_pbn); /* pick the bottleneck */ end_to_end_bw_in_kbps = min(upper_link_bw_in_kbps, down_link_bw_in_kbps); - mutex_unlock(&mst_mgr->lock); - /* * use the maximum dsc compression bandwidth as the required * bandwidth for the mode @@ -1643,8 +1643,7 @@ enum dc_status dm_dp_mst_is_port_support_mode( /* check if mode could be supported within full_pbn */ bpp = convert_dc_color_depth_into_bpc(stream->timing.display_color_depth) * 3; pbn = drm_dp_calc_pbn_mode(stream->timing.pix_clk_100hz / 10, bpp, false); - - if (pbn > aconnector->mst_output_port->full_pbn) + if (pbn > full_pbn) return DC_FAIL_BANDWIDTH_VALIDATE; } diff --git a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c index 7cdb1a8a0ba0..2d1f5efa9091 100644 --- a/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c +++ b/drivers/gpu/drm/amd/display/dc/bios/bios_parser2.c @@ -1014,13 +1014,20 @@ static enum bp_result get_ss_info_v4_5( DC_LOG_BIOS("AS_SIGNAL_TYPE_HDMI ss_percentage: %d\n", ss_info->spread_spectrum_percentage); break; case AS_SIGNAL_TYPE_DISPLAY_PORT: - ss_info->spread_spectrum_percentage = + if (bp->base.integrated_info) { + DC_LOG_BIOS("gpuclk_ss_percentage (unit of 0.001 percent): %d\n", bp->base.integrated_info->gpuclk_ss_percentage); + ss_info->spread_spectrum_percentage = + bp->base.integrated_info->gpuclk_ss_percentage; + ss_info->type.CENTER_MODE = + bp->base.integrated_info->gpuclk_ss_type; + } else { + ss_info->spread_spectrum_percentage = disp_cntl_tbl->dp_ss_percentage; - ss_info->spread_spectrum_range = + ss_info->spread_spectrum_range = disp_cntl_tbl->dp_ss_rate_10hz * 10; - if (disp_cntl_tbl->dp_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE) - ss_info->type.CENTER_MODE = true; - + if (disp_cntl_tbl->dp_ss_mode & ATOM_SS_CENTRE_SPREAD_MODE) + ss_info->type.CENTER_MODE = true; + } DC_LOG_BIOS("AS_SIGNAL_TYPE_DISPLAY_PORT ss_percentage: %d\n", ss_info->spread_spectrum_percentage); break; case AS_SIGNAL_TYPE_GPU_PLL: @@ -2814,6 +2821,8 @@ static enum bp_result get_integrated_info_v2_2( info->ma_channel_number = info_v2_2->umachannelnumber; info->dp_ss_control = le16_to_cpu(info_v2_2->reserved1); + info->gpuclk_ss_percentage = info_v2_2->gpuclk_ss_percentage; + info->gpuclk_ss_type = info_v2_2->gpuclk_ss_type; for (i = 0; i < NUMBER_OF_UCHAR_FOR_GUID; ++i) { info->ext_disp_conn_info.gu_id[i] = diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c index b2c4f97afc8b..8776055bbeaa 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn315/dcn315_clk_mgr.c @@ -334,7 +334,7 @@ static struct wm_table lpddr5_wm_table = { { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, - .pstate_latency_us = 11.65333, + .pstate_latency_us = 129.0, .sr_exit_time_us = 11.5, .sr_enter_plus_exit_time_us = 14.5, .valid = true, @@ -342,7 +342,7 @@ static struct wm_table lpddr5_wm_table = { { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, - .pstate_latency_us = 11.65333, + .pstate_latency_us = 129.0, .sr_exit_time_us = 11.5, .sr_enter_plus_exit_time_us = 14.5, .valid = true, @@ -350,7 +350,7 @@ static struct wm_table lpddr5_wm_table = { { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, - .pstate_latency_us = 11.65333, + .pstate_latency_us = 129.0, .sr_exit_time_us = 11.5, .sr_enter_plus_exit_time_us = 14.5, .valid = true, @@ -358,7 +358,7 @@ static struct wm_table lpddr5_wm_table = { { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, - .pstate_latency_us = 11.65333, + .pstate_latency_us = 129.0, .sr_exit_time_us = 11.5, .sr_enter_plus_exit_time_us = 14.5, .valid = true, diff --git a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c index 0fa4fcd00de2..d5fde7d23fbf 100644 --- a/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c +++ b/drivers/gpu/drm/amd/display/dc/clk_mgr/dcn35/dcn35_clk_mgr.c @@ -232,6 +232,10 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, if (dc->work_arounds.skip_clock_update) return; + /* DTBCLK is fixed, so set a default if unspecified. */ + if (new_clocks->dtbclk_en && !new_clocks->ref_dtbclk_khz) + new_clocks->ref_dtbclk_khz = 600000; + /* * if it is safe to lower, but we are already in the lower state, we don't have to do anything * also if safe to lower is false, we just go in the higher state @@ -265,8 +269,10 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, if (!clk_mgr_base->clks.dtbclk_en && new_clocks->dtbclk_en) { dcn35_smu_set_dtbclk(clk_mgr, true); - dcn35_update_clocks_update_dtb_dto(clk_mgr, context, clk_mgr_base->clks.ref_dtbclk_khz); clk_mgr_base->clks.dtbclk_en = new_clocks->dtbclk_en; + + dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz); + clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz; } /* check that we're not already in D0 */ @@ -314,17 +320,12 @@ void dcn35_update_clocks(struct clk_mgr *clk_mgr_base, update_dispclk = true; } - if (!new_clocks->dtbclk_en) { - new_clocks->ref_dtbclk_khz = 600000; - } - /* clock limits are received with MHz precision, divide by 1000 to prevent setting clocks at every call */ if (!dc->debug.disable_dtb_ref_clk_switch && - should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000, clk_mgr_base->clks.ref_dtbclk_khz / 1000)) { - /* DCCG requires KHz precision for DTBCLK */ - dcn35_smu_set_dtbclk(clk_mgr, true); - - dcn35_update_clocks_update_dtb_dto(clk_mgr, context, clk_mgr_base->clks.ref_dtbclk_khz); + should_set_clock(safe_to_lower, new_clocks->ref_dtbclk_khz / 1000, + clk_mgr_base->clks.ref_dtbclk_khz / 1000)) { + dcn35_update_clocks_update_dtb_dto(clk_mgr, context, new_clocks->ref_dtbclk_khz); + clk_mgr_base->clks.ref_dtbclk_khz = new_clocks->ref_dtbclk_khz; } if (dpp_clock_lowered) { @@ -443,32 +444,32 @@ static struct wm_table ddr5_wm_table = { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 9, - .sr_enter_plus_exit_time_us = 11, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 9, - .sr_enter_plus_exit_time_us = 11, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 9, - .sr_enter_plus_exit_time_us = 11, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.72, - .sr_exit_time_us = 9, - .sr_enter_plus_exit_time_us = 11, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, } @@ -480,32 +481,32 @@ static struct wm_table lpddr5_wm_table = { .wm_inst = WM_A, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 11.5, - .sr_enter_plus_exit_time_us = 14.5, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_B, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 11.5, - .sr_enter_plus_exit_time_us = 14.5, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_C, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 11.5, - .sr_enter_plus_exit_time_us = 14.5, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, { .wm_inst = WM_D, .wm_type = WM_TYPE_PSTATE_CHG, .pstate_latency_us = 11.65333, - .sr_exit_time_us = 11.5, - .sr_enter_plus_exit_time_us = 14.5, + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, .valid = true, }, } @@ -515,11 +516,6 @@ static DpmClocks_t_dcn35 dummy_clocks; static struct dcn35_watermarks dummy_wms = { 0 }; -static struct dcn35_ss_info_table ss_info_table = { - .ss_divider = 1000, - .ss_percentage = {0, 0, 375, 375, 375} -}; - static void dcn35_build_watermark_ranges(struct clk_bw_params *bw_params, struct dcn35_watermarks *table) { int i, num_valid_sets; @@ -653,27 +649,47 @@ static unsigned int convert_wck_ratio(uint8_t wck_ratio) return 1; } +static inline uint32_t calc_dram_speed_mts(const MemPstateTable_t *entry) +{ + return entry->UClk * convert_wck_ratio(entry->WckRatio) * 2; +} + static void dcn35_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk_mgr, struct integrated_info *bios_info, DpmClocks_t_dcn35 *clock_table) { struct clk_bw_params *bw_params = clk_mgr->base.bw_params; struct clk_limit_table_entry def_max = bw_params->clk_table.entries[bw_params->clk_table.num_entries - 1]; - uint32_t max_pstate = 0, max_uclk = 0, max_fclk = 0; - uint32_t min_pstate = 0, max_dispclk = 0, max_dppclk = 0; + uint32_t max_fclk = 0, min_pstate = 0, max_dispclk = 0, max_dppclk = 0; + uint32_t max_pstate = 0, max_dram_speed_mts = 0, min_dram_speed_mts = 0; int i; + /* Determine min/max p-state values. */ for (i = 0; i < clock_table->NumMemPstatesEnabled; i++) { - if (is_valid_clock_value(clock_table->MemPstateTable[i].UClk) && - clock_table->MemPstateTable[i].UClk > max_uclk) { - max_uclk = clock_table->MemPstateTable[i].UClk; + uint32_t dram_speed_mts = calc_dram_speed_mts(&clock_table->MemPstateTable[i]); + + if (is_valid_clock_value(dram_speed_mts) && dram_speed_mts > max_dram_speed_mts) { + max_dram_speed_mts = dram_speed_mts; max_pstate = i; } } - /* We expect the table to contain at least one valid Uclk entry. */ - ASSERT(is_valid_clock_value(max_uclk)); + min_dram_speed_mts = max_dram_speed_mts; + min_pstate = max_pstate; + + for (i = 0; i < clock_table->NumMemPstatesEnabled; i++) { + uint32_t dram_speed_mts = calc_dram_speed_mts(&clock_table->MemPstateTable[i]); + + if (is_valid_clock_value(dram_speed_mts) && dram_speed_mts < min_dram_speed_mts) { + min_dram_speed_mts = dram_speed_mts; + min_pstate = i; + } + } + /* We expect the table to contain at least one valid P-state entry. */ + ASSERT(clock_table->NumMemPstatesEnabled && + is_valid_clock_value(max_dram_speed_mts) && + is_valid_clock_value(min_dram_speed_mts)); /* dispclk and dppclk can be max at any voltage, same number of levels for both */ if (clock_table->NumDispClkLevelsEnabled <= NUM_DISPCLK_DPM_LEVELS && @@ -683,47 +699,46 @@ static void dcn35_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk max_dppclk = find_max_clk_value(clock_table->DppClocks, clock_table->NumDispClkLevelsEnabled); } else { + /* Invalid number of entries in the table from PMFW. */ ASSERT(0); } - if (clock_table->NumFclkLevelsEnabled <= NUM_FCLK_DPM_LEVELS) - max_fclk = find_max_clk_value(clock_table->FclkClocks_Freq, - clock_table->NumFclkLevelsEnabled); - for (i = 0; i < clock_table->NumMemPstatesEnabled; i++) { - uint32_t min_uclk = clock_table->MemPstateTable[0].UClk; - int j; + /* Base the clock table on dcfclk, need at least one entry regardless of pmfw table */ + ASSERT(clock_table->NumDcfClkLevelsEnabled > 0); - for (j = 1; j < clock_table->NumMemPstatesEnabled; j++) { - if (is_valid_clock_value(clock_table->MemPstateTable[j].UClk) && - clock_table->MemPstateTable[j].UClk < min_uclk && - clock_table->MemPstateTable[j].Voltage <= clock_table->SocVoltage[i]) { - min_uclk = clock_table->MemPstateTable[j].UClk; - min_pstate = j; - } - } + max_fclk = find_max_clk_value(clock_table->FclkClocks_Freq, clock_table->NumFclkLevelsEnabled); + + for (i = 0; i < clock_table->NumDcfClkLevelsEnabled; i++) { + int j; + /* First search defaults for the clocks we don't read using closest lower or equal default dcfclk */ for (j = bw_params->clk_table.num_entries - 1; j > 0; j--) if (bw_params->clk_table.entries[j].dcfclk_mhz <= clock_table->DcfClocks[i]) - break; + break; bw_params->clk_table.entries[i].phyclk_mhz = bw_params->clk_table.entries[j].phyclk_mhz; bw_params->clk_table.entries[i].phyclk_d18_mhz = bw_params->clk_table.entries[j].phyclk_d18_mhz; bw_params->clk_table.entries[i].dtbclk_mhz = bw_params->clk_table.entries[j].dtbclk_mhz; - bw_params->clk_table.entries[i].fclk_mhz = max_fclk; + + /* Now update clocks we do read */ bw_params->clk_table.entries[i].memclk_mhz = clock_table->MemPstateTable[min_pstate].MemClk; bw_params->clk_table.entries[i].voltage = clock_table->MemPstateTable[min_pstate].Voltage; bw_params->clk_table.entries[i].dcfclk_mhz = clock_table->DcfClocks[i]; bw_params->clk_table.entries[i].socclk_mhz = clock_table->SocClocks[i]; bw_params->clk_table.entries[i].dispclk_mhz = max_dispclk; bw_params->clk_table.entries[i].dppclk_mhz = max_dppclk; - bw_params->clk_table.entries[i].wck_ratio = convert_wck_ratio( - clock_table->MemPstateTable[min_pstate].WckRatio); - } + bw_params->clk_table.entries[i].wck_ratio = + convert_wck_ratio(clock_table->MemPstateTable[min_pstate].WckRatio); + + /* Dcfclk and Fclk are tied, but at a different ratio */ + bw_params->clk_table.entries[i].fclk_mhz = min(max_fclk, 2 * clock_table->DcfClocks[i]); + } /* Make sure to include at least one entry at highest pstate */ if (max_pstate != min_pstate || i == 0) { if (i > MAX_NUM_DPM_LVL - 1) i = MAX_NUM_DPM_LVL - 1; + bw_params->clk_table.entries[i].fclk_mhz = max_fclk; bw_params->clk_table.entries[i].memclk_mhz = clock_table->MemPstateTable[max_pstate].MemClk; bw_params->clk_table.entries[i].voltage = clock_table->MemPstateTable[max_pstate].Voltage; @@ -739,6 +754,7 @@ static void dcn35_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk } bw_params->clk_table.num_entries = i--; + /* Make sure all highest clocks are included*/ bw_params->clk_table.entries[i].socclk_mhz = find_max_clk_value(clock_table->SocClocks, NUM_SOCCLK_DPM_LEVELS); bw_params->clk_table.entries[i].dispclk_mhz = @@ -757,6 +773,11 @@ static void dcn35_clk_mgr_helper_populate_bw_params(struct clk_mgr_internal *clk bw_params->clk_table.num_entries_per_clk.num_fclk_levels = clock_table->NumFclkLevelsEnabled; bw_params->clk_table.num_entries_per_clk.num_memclk_levels = clock_table->NumMemPstatesEnabled; bw_params->clk_table.num_entries_per_clk.num_socclk_levels = clock_table->NumSocClkLevelsEnabled; + + /* + * Set any 0 clocks to max default setting. Not an issue for + * power since we aren't doing switching in such case anyway + */ for (i = 0; i < bw_params->clk_table.num_entries; i++) { if (!bw_params->clk_table.entries[i].fclk_mhz) { bw_params->clk_table.entries[i].fclk_mhz = def_max.fclk_mhz; @@ -820,22 +841,22 @@ static void dcn35_set_idle_state(struct clk_mgr *clk_mgr_base, bool allow_idle) if (dc->config.disable_ips == DMUB_IPS_ENABLE || dc->config.disable_ips == DMUB_IPS_DISABLE_DYNAMIC) { - val |= DMUB_IPS1_ALLOW_MASK; - val |= DMUB_IPS2_ALLOW_MASK; - } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS1) { val = val & ~DMUB_IPS1_ALLOW_MASK; val = val & ~DMUB_IPS2_ALLOW_MASK; - } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2) { - val |= DMUB_IPS1_ALLOW_MASK; - val = val & ~DMUB_IPS2_ALLOW_MASK; - } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2_Z10) { + } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS1) { val |= DMUB_IPS1_ALLOW_MASK; val |= DMUB_IPS2_ALLOW_MASK; + } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2) { + val = val & ~DMUB_IPS1_ALLOW_MASK; + val |= DMUB_IPS2_ALLOW_MASK; + } else if (dc->config.disable_ips == DMUB_IPS_DISABLE_IPS2_Z10) { + val = val & ~DMUB_IPS1_ALLOW_MASK; + val = val & ~DMUB_IPS2_ALLOW_MASK; } if (!allow_idle) { - val = val & ~DMUB_IPS1_ALLOW_MASK; - val = val & ~DMUB_IPS2_ALLOW_MASK; + val |= DMUB_IPS1_ALLOW_MASK; + val |= DMUB_IPS2_ALLOW_MASK; } dcn35_smu_write_ips_scratch(clk_mgr, val); @@ -965,21 +986,6 @@ struct clk_mgr_funcs dcn35_fpga_funcs = { .get_dtb_ref_clk_frequency = dcn31_get_dtb_ref_freq_khz, }; -static void dcn35_read_ss_info_from_lut(struct clk_mgr_internal *clk_mgr) -{ - uint32_t clock_source; - struct dc_context *ctx = clk_mgr->base.ctx; - - REG_GET(CLK1_CLK2_BYPASS_CNTL, CLK2_BYPASS_SEL, &clock_source); - - clk_mgr->dprefclk_ss_percentage = ss_info_table.ss_percentage[clock_source]; - - if (clk_mgr->dprefclk_ss_percentage != 0) { - clk_mgr->ss_on_dprefclk = true; - clk_mgr->dprefclk_ss_divider = ss_info_table.ss_divider; - } -} - void dcn35_clk_mgr_construct( struct dc_context *ctx, struct clk_mgr_dcn35 *clk_mgr, @@ -1043,17 +1049,11 @@ void dcn35_clk_mgr_construct( dcn35_dump_clk_registers(&clk_mgr->base.base.boot_snapshot, &clk_mgr->base.base, &log_info); clk_mgr->base.base.dprefclk_khz = dcn35_smu_get_dprefclk(&clk_mgr->base); - clk_mgr->base.base.clks.ref_dtbclk_khz = dcn35_smu_get_dtbclk(&clk_mgr->base); - - if (!clk_mgr->base.base.clks.ref_dtbclk_khz) - dcn35_smu_set_dtbclk(&clk_mgr->base, true); + clk_mgr->base.base.clks.ref_dtbclk_khz = 600000; - clk_mgr->base.base.clks.dtbclk_en = true; dce_clock_read_ss_info(&clk_mgr->base); /*when clk src is from FCH, it could have ss, same clock src as DPREF clk*/ - dcn35_read_ss_info_from_lut(&clk_mgr->base); - clk_mgr->base.base.bw_params = &dcn35_bw_params; if (clk_mgr->base.base.ctx->dc->debug.pstate_enabled) { @@ -1129,7 +1129,6 @@ void dcn35_clk_mgr_construct( ctx->dc->debug.disable_dpp_power_gate = false; ctx->dc->debug.disable_hubp_power_gate = false; ctx->dc->debug.disable_dsc_power_gate = false; - ctx->dc->debug.disable_hpo_power_gate = false; } else { /*let's reset the config control flag*/ ctx->dc->config.disable_ips = DMUB_IPS_DISABLE_ALL; /*pmfw not support it, disable it all*/ diff --git a/drivers/gpu/drm/amd/display/dc/core/dc.c b/drivers/gpu/drm/amd/display/dc/core/dc.c index 7b9bf5cb4529..5c1185206645 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc.c @@ -3178,7 +3178,7 @@ static bool update_planes_and_stream_state(struct dc *dc, struct pipe_ctx *otg_master = resource_get_otg_master_for_stream(&context->res_ctx, context->streams[i]); - if (otg_master->stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE) + if (otg_master && otg_master->stream->test_pattern.type != DP_TEST_PATTERN_VIDEO_MODE) resource_build_test_pattern_params(&context->res_ctx, otg_master); } } @@ -4934,8 +4934,8 @@ bool dc_dmub_is_ips_idle_state(struct dc *dc) if (dc->hwss.get_idle_state) idle_state = dc->hwss.get_idle_state(dc); - if ((idle_state & DMUB_IPS1_ALLOW_MASK) || - (idle_state & DMUB_IPS2_ALLOW_MASK)) + if (!(idle_state & DMUB_IPS1_ALLOW_MASK) || + !(idle_state & DMUB_IPS2_ALLOW_MASK)) return true; return false; @@ -5095,18 +5095,28 @@ void dc_mclk_switch_using_fw_based_vblank_stretch_shut_down(struct dc *dc) */ bool dc_is_dmub_outbox_supported(struct dc *dc) { - /* DCN31 B0 USB4 DPIA needs dmub notifications for interrupts */ - if (dc->ctx->asic_id.chip_family == FAMILY_YELLOW_CARP && - dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 && - !dc->debug.dpia_debug.bits.disable_dpia) - return true; + switch (dc->ctx->asic_id.chip_family) { - if (dc->ctx->asic_id.chip_family == AMDGPU_FAMILY_GC_11_0_1 && - !dc->debug.dpia_debug.bits.disable_dpia) - return true; + case FAMILY_YELLOW_CARP: + /* DCN31 B0 USB4 DPIA needs dmub notifications for interrupts */ + if (dc->ctx->asic_id.hw_internal_rev == YELLOW_CARP_B0 && + !dc->debug.dpia_debug.bits.disable_dpia) + return true; + break; + + case AMDGPU_FAMILY_GC_11_0_1: + case AMDGPU_FAMILY_GC_11_5_0: + if (!dc->debug.dpia_debug.bits.disable_dpia) + return true; + break; + + default: + break; + } /* dmub aux needs dmub notifications to be enabled */ return dc->debug.enable_dmub_aux_for_legacy_ddc; + } /** diff --git a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c index 1d48278cba96..a1f1d1003992 100644 --- a/drivers/gpu/drm/amd/display/dc/core/dc_resource.c +++ b/drivers/gpu/drm/amd/display/dc/core/dc_resource.c @@ -5190,6 +5190,9 @@ bool dc_resource_acquire_secondary_pipe_for_mpc_odm_legacy( sec_next = sec_pipe->next_odm_pipe; sec_prev = sec_pipe->prev_odm_pipe; + if (pri_pipe == NULL) + return false; + *sec_pipe = *pri_pipe; sec_pipe->top_pipe = sec_top; diff --git a/drivers/gpu/drm/amd/display/dc/dc.h b/drivers/gpu/drm/amd/display/dc/dc.h index 9316b737a8ba..2cafd644baff 100644 --- a/drivers/gpu/drm/amd/display/dc/dc.h +++ b/drivers/gpu/drm/amd/display/dc/dc.h @@ -874,6 +874,7 @@ struct dc_debug_options { unsigned int seamless_boot_odm_combine; unsigned int force_odm_combine_4to1; //bit vector based on otg inst int minimum_z8_residency_time; + int minimum_z10_residency_time; bool disable_z9_mpc; unsigned int force_fclk_khz; bool enable_tri_buf; @@ -1608,7 +1609,6 @@ struct dc_link { enum edp_revision edp_revision; union dpcd_sink_ext_caps dpcd_sink_ext_caps; - struct backlight_settings backlight_settings; struct psr_settings psr_settings; struct replay_settings replay_settings; diff --git a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c index e4c007203318..0e07699c1e83 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dc/dc_dmub_srv.c @@ -1202,11 +1202,11 @@ void dc_dmub_srv_exit_low_power_state(const struct dc *dc) allow_state = dc->hwss.get_idle_state(dc); dc->hwss.set_idle_state(dc, false); - if (allow_state & DMUB_IPS2_ALLOW_MASK) { + if (!(allow_state & DMUB_IPS2_ALLOW_MASK)) { // Wait for evaluation time udelay(dc->debug.ips2_eval_delay_us); commit_state = dc->hwss.get_idle_state(dc); - if (commit_state & DMUB_IPS2_COMMIT_MASK) { + if (!(commit_state & DMUB_IPS2_COMMIT_MASK)) { // Tell PMFW to exit low power state dc->clk_mgr->funcs->exit_low_power_state(dc->clk_mgr); @@ -1216,7 +1216,7 @@ void dc_dmub_srv_exit_low_power_state(const struct dc *dc) for (i = 0; i < max_num_polls; ++i) { commit_state = dc->hwss.get_idle_state(dc); - if (!(commit_state & DMUB_IPS2_COMMIT_MASK)) + if (commit_state & DMUB_IPS2_COMMIT_MASK) break; udelay(1); @@ -1235,10 +1235,10 @@ void dc_dmub_srv_exit_low_power_state(const struct dc *dc) } dc_dmub_srv_notify_idle(dc, false); - if (allow_state & DMUB_IPS1_ALLOW_MASK) { + if (!(allow_state & DMUB_IPS1_ALLOW_MASK)) { for (i = 0; i < max_num_polls; ++i) { commit_state = dc->hwss.get_idle_state(dc); - if (!(commit_state & DMUB_IPS1_COMMIT_MASK)) + if (commit_state & DMUB_IPS1_COMMIT_MASK) break; udelay(1); diff --git a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h index 9649934ea186..e2a3aa8812df 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_hw_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_hw_types.h @@ -465,6 +465,7 @@ struct dc_cursor_mi_param { struct fixed31_32 v_scale_ratio; enum dc_rotation_angle rotation; bool mirror; + struct dc_stream_state *stream; }; /* IPP related types */ diff --git a/drivers/gpu/drm/amd/display/dc/dc_types.h b/drivers/gpu/drm/amd/display/dc/dc_types.h index cea666ea66c6..35d146217aef 100644 --- a/drivers/gpu/drm/amd/display/dc/dc_types.h +++ b/drivers/gpu/drm/amd/display/dc/dc_types.h @@ -177,6 +177,7 @@ struct dc_panel_patch { unsigned int disable_fams; unsigned int skip_avmute; unsigned int mst_start_top_delay; + unsigned int remove_sink_ext_caps; }; struct dc_edid_caps { @@ -990,10 +991,6 @@ struct link_mst_stream_allocation_table { struct link_mst_stream_allocation stream_allocations[MAX_CONTROLLER_NUM]; }; -struct backlight_settings { - uint32_t backlight_millinits; -}; - /* PSR feature flags */ struct psr_settings { bool psr_feature_enabled; // PSR is supported by sink diff --git a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c index 139cf31d2e45..89c3bf0fe0c9 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c +++ b/drivers/gpu/drm/amd/display/dc/dcn20/dcn20_hubp.c @@ -1077,8 +1077,16 @@ void hubp2_cursor_set_position( if (src_y_offset < 0) src_y_offset = 0; /* Save necessary cursor info x, y position. w, h is saved in attribute func. */ - hubp->cur_rect.x = src_x_offset + param->viewport.x; - hubp->cur_rect.y = src_y_offset + param->viewport.y; + if (param->stream->link->psr_settings.psr_version >= DC_PSR_VERSION_SU_1 && + param->rotation != ROTATION_ANGLE_0) { + hubp->cur_rect.x = 0; + hubp->cur_rect.y = 0; + hubp->cur_rect.w = param->stream->timing.h_addressable; + hubp->cur_rect.h = param->stream->timing.v_addressable; + } else { + hubp->cur_rect.x = src_x_offset + param->viewport.x; + hubp->cur_rect.y = src_y_offset + param->viewport.y; + } } void hubp2_clk_cntl(struct hubp *hubp, bool enable) diff --git a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c index 677361d74a4e..c97391edb5ff 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c @@ -871,7 +871,7 @@ static const struct dc_plane_cap plane_cap = { static const struct dc_debug_options debug_defaults_drv = { .disable_z10 = false, .enable_z9_disable_interface = true, - .minimum_z8_residency_time = 2000, + .minimum_z8_residency_time = 2100, .psr_skip_crtc_disable = true, .replay_skip_crtc_disabled = true, .disable_dmcu = true, diff --git a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_stream_encoder.c b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_stream_encoder.c index 001f9eb66920..62a8f0b56006 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_stream_encoder.c +++ b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_dio_stream_encoder.c @@ -261,12 +261,6 @@ static void enc35_stream_encoder_enable( /* invalid mode ! */ ASSERT_CRITICAL(false); } - - REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 1); - REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 1); - } else { - REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 0); - REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 0); } } @@ -436,6 +430,8 @@ static void enc35_disable_fifo(struct stream_encoder *enc) struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc); REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_ENABLE, 0); + REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 0); + REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 0); } static void enc35_enable_fifo(struct stream_encoder *enc) @@ -443,6 +439,8 @@ static void enc35_enable_fifo(struct stream_encoder *enc) struct dcn10_stream_encoder *enc1 = DCN10STRENC_FROM_STRENC(enc); REG_UPDATE(DIG_FIFO_CTRL0, DIG_FIFO_READ_START_LEVEL, 0x7); + REG_UPDATE(DIG_FE_CLK_CNTL, DIG_FE_CLK_EN, 1); + REG_UPDATE(DIG_FE_EN_CNTL, DIG_FE_ENABLE, 1); enc35_reset_fifo(enc, true); enc35_reset_fifo(enc, false); diff --git a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_pg_cntl.c b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_pg_cntl.c index 46f71ff08fd1..d19db8e9b8a5 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_pg_cntl.c +++ b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_pg_cntl.c @@ -261,6 +261,7 @@ void pg_cntl35_hpo_pg_control(struct pg_cntl *pg_cntl, bool power_on) uint32_t power_gate = power_on ? 0 : 1; uint32_t pwr_status = power_on ? 0 : 2; uint32_t org_ip_request_cntl; + uint32_t power_forceon; bool block_enabled; if (pg_cntl->ctx->dc->debug.ignore_pg || @@ -277,6 +278,10 @@ void pg_cntl35_hpo_pg_control(struct pg_cntl *pg_cntl, bool power_on) return; } + REG_GET(DOMAIN25_PG_CONFIG, DOMAIN_POWER_FORCEON, &power_forceon); + if (power_forceon) + return; + REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl); if (org_ip_request_cntl == 0) REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1); @@ -304,6 +309,7 @@ void pg_cntl35_io_clk_pg_control(struct pg_cntl *pg_cntl, bool power_on) uint32_t power_gate = power_on ? 0 : 1; uint32_t pwr_status = power_on ? 0 : 2; uint32_t org_ip_request_cntl; + uint32_t power_forceon; bool block_enabled; if (pg_cntl->ctx->dc->debug.ignore_pg || @@ -319,6 +325,10 @@ void pg_cntl35_io_clk_pg_control(struct pg_cntl *pg_cntl, bool power_on) return; } + REG_GET(DOMAIN22_PG_CONFIG, DOMAIN_POWER_FORCEON, &power_forceon); + if (power_forceon) + return; + REG_GET(DC_IP_REQUEST_CNTL, IP_REQUEST_EN, &org_ip_request_cntl); if (org_ip_request_cntl == 0) REG_SET(DC_IP_REQUEST_CNTL, 0, IP_REQUEST_EN, 1); diff --git a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_resource.c b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_resource.c index c7e011d26d41..70ef1e7ff841 100644 --- a/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_resource.c +++ b/drivers/gpu/drm/amd/display/dc/dcn35/dcn35_resource.c @@ -1712,6 +1712,13 @@ static bool dcn35_validate_bandwidth(struct dc *dc, out = dml2_validate(dc, context, fast_validate); + if (fast_validate) + return out; + + DC_FP_START(); + dcn35_decide_zstate_support(dc, context); + DC_FP_END(); + return out; } @@ -1857,7 +1864,7 @@ static bool dcn35_resource_construct( /* Use pipe context based otg sync logic */ dc->config.use_pipe_ctx_sync_logic = true; - dc->config.use_default_clock_table = false; + /* read VBIOS LTTPR caps */ { if (ctx->dc_bios->funcs->get_lttpr_caps) { diff --git a/drivers/gpu/drm/amd/display/dc/dml/Makefile b/drivers/gpu/drm/amd/display/dc/dml/Makefile index ea7d60f9a9b4..6042a5a6a44f 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/Makefile +++ b/drivers/gpu/drm/amd/display/dc/dml/Makefile @@ -61,8 +61,12 @@ endif endif ifneq ($(CONFIG_FRAME_WARN),0) +ifeq ($(filter y,$(CONFIG_KASAN)$(CONFIG_KCSAN)),y) +frame_warn_flag := -Wframe-larger-than=3072 +else frame_warn_flag := -Wframe-larger-than=2048 endif +endif CFLAGS_$(AMDDALPATH)/dc/dml/display_mode_lib.o := $(dml_ccflags) CFLAGS_$(AMDDALPATH)/dc/dml/display_mode_vba.o := $(dml_ccflags) diff --git a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h index 2cbdd75429ff..6e669a2c5b2d 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dc_features.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dc_features.h @@ -36,7 +36,7 @@ * Define the maximum amount of states supported by the ASIC. Every ASIC has a * specific number of states; this macro defines the maximum number of states. */ -#define DC__VOLTAGE_STATES 20 +#define DC__VOLTAGE_STATES 40 #define DC__NUM_DPP__4 1 #define DC__NUM_DPP__0_PRESENT 1 #define DC__NUM_DPP__1_PRESENT 1 diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c index 7fc8b18096ba..d2271e308fa0 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn20/dcn20_fpu.c @@ -440,7 +440,115 @@ struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv14_soc = { .use_urgent_burst_bw = 0 }; -struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv12_soc = { 0 }; +struct _vcs_dpi_soc_bounding_box_st dcn2_0_nv12_soc = { + .clock_limits = { + { + .state = 0, + .dcfclk_mhz = 560.0, + .fabricclk_mhz = 560.0, + .dispclk_mhz = 513.0, + .dppclk_mhz = 513.0, + .phyclk_mhz = 540.0, + .socclk_mhz = 560.0, + .dscclk_mhz = 171.0, + .dram_speed_mts = 1069.0, + }, + { + .state = 1, + .dcfclk_mhz = 694.0, + .fabricclk_mhz = 694.0, + .dispclk_mhz = 642.0, + .dppclk_mhz = 642.0, + .phyclk_mhz = 600.0, + .socclk_mhz = 694.0, + .dscclk_mhz = 214.0, + .dram_speed_mts = 1324.0, + }, + { + .state = 2, + .dcfclk_mhz = 875.0, + .fabricclk_mhz = 875.0, + .dispclk_mhz = 734.0, + .dppclk_mhz = 734.0, + .phyclk_mhz = 810.0, + .socclk_mhz = 875.0, + .dscclk_mhz = 245.0, + .dram_speed_mts = 1670.0, + }, + { + .state = 3, + .dcfclk_mhz = 1000.0, + .fabricclk_mhz = 1000.0, + .dispclk_mhz = 1100.0, + .dppclk_mhz = 1100.0, + .phyclk_mhz = 810.0, + .socclk_mhz = 1000.0, + .dscclk_mhz = 367.0, + .dram_speed_mts = 2000.0, + }, + { + .state = 4, + .dcfclk_mhz = 1200.0, + .fabricclk_mhz = 1200.0, + .dispclk_mhz = 1284.0, + .dppclk_mhz = 1284.0, + .phyclk_mhz = 810.0, + .socclk_mhz = 1200.0, + .dscclk_mhz = 428.0, + .dram_speed_mts = 2000.0, + }, + { + .state = 5, + .dcfclk_mhz = 1200.0, + .fabricclk_mhz = 1200.0, + .dispclk_mhz = 1284.0, + .dppclk_mhz = 1284.0, + .phyclk_mhz = 810.0, + .socclk_mhz = 1200.0, + .dscclk_mhz = 428.0, + .dram_speed_mts = 2000.0, + }, + }, + + .num_states = 5, + .sr_exit_time_us = 1.9, + .sr_enter_plus_exit_time_us = 4.4, + .urgent_latency_us = 3.0, + .urgent_latency_pixel_data_only_us = 4.0, + .urgent_latency_pixel_mixed_with_vm_data_us = 4.0, + .urgent_latency_vm_data_only_us = 4.0, + .urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096, + .urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096, + .urgent_out_of_order_return_per_channel_vm_only_bytes = 4096, + .pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 40.0, + .pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 40.0, + .pct_ideal_dram_sdp_bw_after_urgent_vm_only = 40.0, + .max_avg_sdp_bw_use_normal_percent = 40.0, + .max_avg_dram_bw_use_normal_percent = 40.0, + .writeback_latency_us = 12.0, + .ideal_dram_bw_after_urgent_percent = 40.0, + .max_request_size_bytes = 256, + .dram_channel_width_bytes = 16, + .fabric_datapath_to_dcn_data_return_bytes = 64, + .dcn_downspread_percent = 0.5, + .downspread_percent = 0.5, + .dram_page_open_time_ns = 50.0, + .dram_rw_turnaround_time_ns = 17.5, + .dram_return_buffer_per_channel_bytes = 8192, + .round_trip_ping_latency_dcfclk_cycles = 131, + .urgent_out_of_order_return_per_channel_bytes = 4096, + .channel_interleave_bytes = 256, + .num_banks = 8, + .num_chans = 16, + .vmm_page_size_bytes = 4096, + .dram_clock_change_latency_us = 45.0, + .writeback_dram_clock_change_latency_us = 23.0, + .return_bus_width_bytes = 64, + .dispclk_dppclk_vco_speed_mhz = 3850, + .xfc_bus_transport_time_us = 20, + .xfc_xbuf_latency_tolerance_us = 50, + .use_urgent_burst_bw = 0, +}; struct _vcs_dpi_ip_params_st dcn2_1_ip = { .odm_capable = 1, @@ -950,10 +1058,8 @@ static enum dcn_zstate_support_state decide_zstate_support(struct dc *dc, struc { int plane_count; int i; - unsigned int min_dst_y_next_start_us; plane_count = 0; - min_dst_y_next_start_us = 0; for (i = 0; i < dc->res_pool->pipe_count; i++) { if (context->res_ctx.pipe_ctx[i].plane_state) plane_count++; @@ -975,26 +1081,15 @@ static enum dcn_zstate_support_state decide_zstate_support(struct dc *dc, struc else if (context->stream_count == 1 && context->streams[0]->signal == SIGNAL_TYPE_EDP) { struct dc_link *link = context->streams[0]->sink->link; struct dc_stream_status *stream_status = &context->stream_status[0]; - struct dc_stream_state *current_stream = context->streams[0]; int minmum_z8_residency = dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000; bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency; bool is_pwrseq0 = link->link_index == 0; - bool isFreesyncVideo; - - isFreesyncVideo = current_stream->adjust.v_total_min == current_stream->adjust.v_total_max; - isFreesyncVideo = isFreesyncVideo && current_stream->timing.v_total < current_stream->adjust.v_total_min; - for (i = 0; i < dc->res_pool->pipe_count; i++) { - if (context->res_ctx.pipe_ctx[i].stream == current_stream && isFreesyncVideo) { - min_dst_y_next_start_us = context->res_ctx.pipe_ctx[i].dlg_regs.min_dst_y_next_start_us; - break; - } - } /* Don't support multi-plane configurations */ if (stream_status->plane_count > 1) return DCN_ZSTATE_SUPPORT_DISALLOW; - if (is_pwrseq0 && (context->bw_ctx.dml.vba.StutterPeriod > 5000.0 || min_dst_y_next_start_us > 5000)) + if (is_pwrseq0 && context->bw_ctx.dml.vba.StutterPeriod > 5000.0) return DCN_ZSTATE_SUPPORT_ALLOW; else if (is_pwrseq0 && link->psr_settings.psr_version == DC_PSR_VERSION_1 && !link->panel_config.psr.disable_psr) return allow_z8 ? DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY : DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY; diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c index 9ec4172d1c2d..b46cde525066 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c @@ -1192,13 +1192,16 @@ static bool update_pipe_slice_table_with_split_flags( */ struct pipe_ctx *pipe; bool odm; - int i; + int dc_pipe_idx, dml_pipe_idx = 0; bool updated = false; - for (i = 0; i < dc->res_pool->pipe_count; i++) { - pipe = &context->res_ctx.pipe_ctx[i]; + for (dc_pipe_idx = 0; + dc_pipe_idx < dc->res_pool->pipe_count; dc_pipe_idx++) { + pipe = &context->res_ctx.pipe_ctx[dc_pipe_idx]; + if (resource_is_pipe_type(pipe, FREE_PIPE)) + continue; - if (merge[i]) { + if (merge[dc_pipe_idx]) { if (resource_is_pipe_type(pipe, OPP_HEAD)) /* merging OPP head means reducing ODM slice * count by 1 @@ -1213,17 +1216,18 @@ static bool update_pipe_slice_table_with_split_flags( updated = true; } - if (split[i]) { - odm = vba->ODMCombineEnabled[vba->pipe_plane[i]] != + if (split[dc_pipe_idx]) { + odm = vba->ODMCombineEnabled[vba->pipe_plane[dml_pipe_idx]] != dm_odm_combine_mode_disabled; if (odm && resource_is_pipe_type(pipe, OPP_HEAD)) update_slice_table_for_stream( - table, pipe->stream, split[i] - 1); + table, pipe->stream, split[dc_pipe_idx] - 1); else if (!odm && resource_is_pipe_type(pipe, DPP_PIPE)) update_slice_table_for_plane(table, pipe, - pipe->plane_state, split[i] - 1); + pipe->plane_state, split[dc_pipe_idx] - 1); updated = true; } + dml_pipe_idx++; } return updated; } @@ -2231,6 +2235,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, int i, pipe_idx, vlevel_temp = 0; double dcfclk = dcn3_2_soc.clock_limits[0].dcfclk_mhz; double dcfclk_from_validation = context->bw_ctx.dml.vba.DCFCLKState[vlevel][context->bw_ctx.dml.vba.maxMpcComb]; + double dram_speed_from_validation = context->bw_ctx.dml.vba.DRAMSpeed; double dcfclk_from_fw_based_mclk_switching = dcfclk_from_validation; bool pstate_en = context->bw_ctx.dml.vba.DRAMClockChangeSupport[vlevel][context->bw_ctx.dml.vba.maxMpcComb] != dm_dram_clock_change_unsupported; @@ -2418,7 +2423,7 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context, } if (dc->clk_mgr->bw_params->wm_table.nv_entries[WM_C].valid) { - min_dram_speed_mts = context->bw_ctx.dml.vba.DRAMSpeed; + min_dram_speed_mts = dram_speed_from_validation; min_dram_speed_mts_margin = 160; context->bw_ctx.dml.soc.dram_clock_change_latency_us = diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c index a5fe523668e9..f154a3eb1d1a 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.c @@ -124,7 +124,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .phyclk_mhz = 600.0, .phyclk_d18_mhz = 667.0, .dscclk_mhz = 186.0, - .dtbclk_mhz = 625.0, + .dtbclk_mhz = 600.0, }, { .state = 1, @@ -133,7 +133,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .phyclk_mhz = 810.0, .phyclk_d18_mhz = 667.0, .dscclk_mhz = 209.0, - .dtbclk_mhz = 625.0, + .dtbclk_mhz = 600.0, }, { .state = 2, @@ -142,7 +142,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .phyclk_mhz = 810.0, .phyclk_d18_mhz = 667.0, .dscclk_mhz = 209.0, - .dtbclk_mhz = 625.0, + .dtbclk_mhz = 600.0, }, { .state = 3, @@ -151,7 +151,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .phyclk_mhz = 810.0, .phyclk_d18_mhz = 667.0, .dscclk_mhz = 371.0, - .dtbclk_mhz = 625.0, + .dtbclk_mhz = 600.0, }, { .state = 4, @@ -160,14 +160,14 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_5_soc = { .phyclk_mhz = 810.0, .phyclk_d18_mhz = 667.0, .dscclk_mhz = 417.0, - .dtbclk_mhz = 625.0, + .dtbclk_mhz = 600.0, }, }, .num_states = 5, - .sr_exit_time_us = 9.0, - .sr_enter_plus_exit_time_us = 11.0, - .sr_exit_z8_time_us = 50.0, /*changed from 442.0*/ - .sr_enter_plus_exit_z8_time_us = 50.0,/*changed from 560.0*/ + .sr_exit_time_us = 14.0, + .sr_enter_plus_exit_time_us = 16.0, + .sr_exit_z8_time_us = 525.0, + .sr_enter_plus_exit_z8_time_us = 715.0, .fclk_change_latency_us = 20.0, .usr_retraining_latency_us = 2, .writeback_latency_us = 12.0, @@ -329,6 +329,52 @@ void dcn35_update_bw_bounding_box_fpu(struct dc *dc, /*temp till dml2 fully work without dml1*/ dml_init_instance(&dc->dml, &dcn3_5_soc, &dcn3_5_ip, DML_PROJECT_DCN31); + + /*copy to dml2, before dml2_create*/ + if (clk_table->num_entries > 2) { + + for (i = 0; i < clk_table->num_entries; i++) { + dc->dml2_options.bbox_overrides.clks_table.num_states = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dcfclk_mhz = + clock_limits[i].dcfclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].fclk_mhz = + clock_limits[i].fabricclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dispclk_mhz = + clock_limits[i].dispclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dppclk_mhz = + clock_limits[i].dppclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].socclk_mhz = + clock_limits[i].socclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].memclk_mhz = + clk_table->entries[i].memclk_mhz * clk_table->entries[i].wck_ratio; + dc->dml2_options.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz = + clock_limits[i].dtbclk_mhz; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dcfclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_fclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dppclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_socclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_memclk_levels = + clk_table->num_entries; + dc->dml2_options.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels = + clk_table->num_entries; + } + } + + /* Update latency values */ + dc->dml2_options.bbox_overrides.dram_clock_change_latency_us = dcn3_5_soc.dram_clock_change_latency_us; + + dc->dml2_options.bbox_overrides.sr_exit_latency_us = dcn3_5_soc.sr_exit_time_us; + dc->dml2_options.bbox_overrides.sr_enter_plus_exit_latency_us = dcn3_5_soc.sr_enter_plus_exit_time_us; + + dc->dml2_options.bbox_overrides.sr_exit_z8_time_us = dcn3_5_soc.sr_exit_z8_time_us; + dc->dml2_options.bbox_overrides.sr_enter_plus_exit_z8_time_us = dcn3_5_soc.sr_enter_plus_exit_z8_time_us; } static bool is_dual_plane(enum surface_pixel_format format) @@ -507,3 +553,37 @@ int dcn35_populate_dml_pipes_from_context_fpu(struct dc *dc, return pipe_cnt; } + +void dcn35_decide_zstate_support(struct dc *dc, struct dc_state *context) +{ + enum dcn_zstate_support_state support = DCN_ZSTATE_SUPPORT_DISALLOW; + unsigned int i, plane_count = 0; + + for (i = 0; i < dc->res_pool->pipe_count; i++) { + if (context->res_ctx.pipe_ctx[i].plane_state) + plane_count++; + } + + if (plane_count == 0) { + support = DCN_ZSTATE_SUPPORT_ALLOW; + } else if (plane_count == 1 && context->stream_count == 1 && context->streams[0]->signal == SIGNAL_TYPE_EDP) { + struct dc_link *link = context->streams[0]->sink->link; + bool is_pwrseq0 = link && link->link_index == 0; + bool is_psr1 = link && link->psr_settings.psr_version == DC_PSR_VERSION_1 && !link->panel_config.psr.disable_psr; + int minmum_z8_residency = + dc->debug.minimum_z8_residency_time > 0 ? dc->debug.minimum_z8_residency_time : 1000; + bool allow_z8 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z8_residency; + int minmum_z10_residency = + dc->debug.minimum_z10_residency_time > 0 ? dc->debug.minimum_z10_residency_time : 5000; + bool allow_z10 = context->bw_ctx.dml.vba.StutterPeriod > (double)minmum_z10_residency; + + if (is_pwrseq0 && allow_z10) + support = DCN_ZSTATE_SUPPORT_ALLOW; + else if (is_pwrseq0 && is_psr1) + support = allow_z8 ? DCN_ZSTATE_SUPPORT_ALLOW_Z8_Z10_ONLY : DCN_ZSTATE_SUPPORT_ALLOW_Z10_ONLY; + else if (allow_z8) + support = DCN_ZSTATE_SUPPORT_ALLOW_Z8_ONLY; + } + + context->bw_ctx.bw.dcn.clk.zstate_support = support; +} diff --git a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h index e8d5a170893e..067480fc3691 100644 --- a/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h +++ b/drivers/gpu/drm/amd/display/dc/dml/dcn35/dcn35_fpu.h @@ -39,4 +39,6 @@ int dcn35_populate_dml_pipes_from_context_fpu(struct dc *dc, display_e2e_pipe_params_st *pipes, bool fast_validate); +void dcn35_decide_zstate_support(struct dc *dc, struct dc_state *context); + #endif diff --git a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c index 510be909cd75..b95bf27f2fe2 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/display_mode_core.c @@ -5420,7 +5420,7 @@ static void CalculateOutputLink( *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, OutputFormat, DSCInputBitPerComponent, NumberOfDSCSlices, (dml_uint_t)AudioSampleRate, AudioSampleLayout, ODMModeNoDSC, ODMModeDSC, RequiredSlots); - if (OutBpp == 0 && PHYCLKD32PerState < 20000 / 32 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) { + if (*OutBpp == 0 && PHYCLKD32PerState < 20000 / 32 && DSCEnable == dml_dsc_enable_if_necessary && ForcedOutputLinkBPP == 0) { *RequiresDSC = true; LinkDSCEnable = true; *OutBpp = TruncToValidBPP((1 - Downspreading / 100) * 13500, OutputLinkDPLanes, HTotal, HActive, PixelClockBackEnd, ForcedOutputLinkBPP, LinkDSCEnable, Output, @@ -9447,12 +9447,12 @@ void dml_core_mode_programming(struct display_mode_lib_st *mode_lib, const struc // Output CalculateWatermarks_params->Watermark = &s->dummy_watermark; // Watermarks *Watermark - CalculateWatermarks_params->DRAMClockChangeSupport = &mode_lib->ms.support.DRAMClockChangeSupport[j]; + CalculateWatermarks_params->DRAMClockChangeSupport = &mode_lib->ms.support.DRAMClockChangeSupport[0]; CalculateWatermarks_params->MaxActiveDRAMClockChangeLatencySupported = &s->dummy_single_array[0][0]; // dml_float_t *MaxActiveDRAMClockChangeLatencySupported[] CalculateWatermarks_params->SubViewportLinesNeededInMALL = &mode_lib->ms.SubViewportLinesNeededInMALL[j]; // dml_uint_t SubViewportLinesNeededInMALL[] - CalculateWatermarks_params->FCLKChangeSupport = &mode_lib->ms.support.FCLKChangeSupport[j]; + CalculateWatermarks_params->FCLKChangeSupport = &mode_lib->ms.support.FCLKChangeSupport[0]; CalculateWatermarks_params->MaxActiveFCLKChangeLatencySupported = &s->dummy_single[0]; // dml_float_t *MaxActiveFCLKChangeLatencySupported - CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport[j]; + CalculateWatermarks_params->USRRetrainingSupport = &mode_lib->ms.support.USRRetrainingSupport[0]; CalculateWatermarksMALLUseAndDRAMSpeedChangeSupport( &mode_lib->scratch, diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c index 75171bee6f71..db06a5b749b4 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_translation_helper.c @@ -341,25 +341,42 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc, break; } - /* Override from passed values, mainly for debugging purposes, if available */ - if (dml2->config.bbox_overrides.sr_exit_latency_us) { - p->in_states->state_array[0].sr_exit_time_us = dml2->config.bbox_overrides.sr_exit_latency_us; - } + /* Override from passed values, if available */ + for (i = 0; i < p->in_states->num_states; i++) { + if (dml2->config.bbox_overrides.sr_exit_latency_us) { + p->in_states->state_array[i].sr_exit_time_us = + dml2->config.bbox_overrides.sr_exit_latency_us; + } - if (dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us) { - p->in_states->state_array[0].sr_enter_plus_exit_time_us = dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us; - } + if (dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us) { + p->in_states->state_array[i].sr_enter_plus_exit_time_us = + dml2->config.bbox_overrides.sr_enter_plus_exit_latency_us; + } - if (dml2->config.bbox_overrides.urgent_latency_us) { - p->in_states->state_array[0].urgent_latency_pixel_data_only_us = dml2->config.bbox_overrides.urgent_latency_us; - } + if (dml2->config.bbox_overrides.sr_exit_z8_time_us) { + p->in_states->state_array[i].sr_exit_z8_time_us = + dml2->config.bbox_overrides.sr_exit_z8_time_us; + } - if (dml2->config.bbox_overrides.dram_clock_change_latency_us) { - p->in_states->state_array[0].dram_clock_change_latency_us = dml2->config.bbox_overrides.dram_clock_change_latency_us; - } + if (dml2->config.bbox_overrides.sr_enter_plus_exit_z8_time_us) { + p->in_states->state_array[i].sr_enter_plus_exit_z8_time_us = + dml2->config.bbox_overrides.sr_enter_plus_exit_z8_time_us; + } + + if (dml2->config.bbox_overrides.urgent_latency_us) { + p->in_states->state_array[i].urgent_latency_pixel_data_only_us = + dml2->config.bbox_overrides.urgent_latency_us; + } - if (dml2->config.bbox_overrides.fclk_change_latency_us) { - p->in_states->state_array[0].fclk_change_latency_us = dml2->config.bbox_overrides.fclk_change_latency_us; + if (dml2->config.bbox_overrides.dram_clock_change_latency_us) { + p->in_states->state_array[i].dram_clock_change_latency_us = + dml2->config.bbox_overrides.dram_clock_change_latency_us; + } + + if (dml2->config.bbox_overrides.fclk_change_latency_us) { + p->in_states->state_array[i].fclk_change_latency_us = + dml2->config.bbox_overrides.fclk_change_latency_us; + } } /* DCFCLK stas values are project specific */ @@ -406,8 +423,9 @@ void dml2_init_soc_states(struct dml2_context *dml2, const struct dc *in_dc, } for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dtbclk_levels; i++) { - p->in_states->state_array[i].dtbclk_mhz = - dml2->config.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz; + if (dml2->config.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz > 0) + p->in_states->state_array[i].dtbclk_mhz = + dml2->config.bbox_overrides.clks_table.clk_entries[i].dtbclk_mhz; } for (i = 0; i < dml2->config.bbox_overrides.clks_table.num_entries_per_clk.num_dispclk_levels; i++) { @@ -498,8 +516,8 @@ void dml2_translate_socbb_params(const struct dc *in, struct soc_bounding_box_st out->do_urgent_latency_adjustment = in_soc_params->do_urgent_latency_adjustment; out->dram_channel_width_bytes = (dml_uint_t)in_soc_params->dram_channel_width_bytes; out->fabric_datapath_to_dcn_data_return_bytes = (dml_uint_t)in_soc_params->fabric_datapath_to_dcn_data_return_bytes; - out->gpuvm_min_page_size_kbytes = in_soc_params->gpuvm_min_page_size_bytes * 1024; - out->hostvm_min_page_size_kbytes = in_soc_params->hostvm_min_page_size_bytes * 1024; + out->gpuvm_min_page_size_kbytes = in_soc_params->gpuvm_min_page_size_bytes / 1024; + out->hostvm_min_page_size_kbytes = in_soc_params->hostvm_min_page_size_bytes / 1024; out->mall_allocated_for_dcn_mbytes = (dml_uint_t)in_soc_params->mall_allocated_for_dcn_mbytes; out->max_avg_dram_bw_use_normal_percent = in_soc_params->max_avg_dram_bw_use_normal_percent; out->max_avg_fabric_bw_use_normal_percent = in_soc_params->max_avg_fabric_bw_use_normal_percent; @@ -1040,9 +1058,12 @@ void map_dc_state_into_dml_display_cfg(struct dml2_context *dml2, struct dc_stat } //Generally these are set by referencing our latest BB/IP params in dcn32_resource.c file - dml_dispcfg->plane.GPUVMEnable = true; - dml_dispcfg->plane.GPUVMMaxPageTableLevels = 4; - dml_dispcfg->plane.HostVMEnable = false; + dml_dispcfg->plane.GPUVMEnable = dml2->v20.dml_core_ctx.ip.gpuvm_enable; + dml_dispcfg->plane.GPUVMMaxPageTableLevels = dml2->v20.dml_core_ctx.ip.gpuvm_max_page_table_levels; + dml_dispcfg->plane.HostVMEnable = dml2->v20.dml_core_ctx.ip.hostvm_enable; + dml_dispcfg->plane.HostVMMaxPageTableLevels = dml2->v20.dml_core_ctx.ip.hostvm_max_page_table_levels; + if (dml2->v20.dml_core_ctx.ip.hostvm_enable) + dml2->v20.dml_core_ctx.policy.AllowForPStateChangeOrStutterInVBlankFinal = dml_prefetch_support_uclk_fclk_and_stutter; dml2_populate_pipe_to_plane_index_mapping(dml2, context); diff --git a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h index 317f90776d97..fe15baa4bf09 100644 --- a/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h +++ b/drivers/gpu/drm/amd/display/dc/dml2/dml2_wrapper.h @@ -139,6 +139,8 @@ struct dml2_soc_bbox_overrides { double urgent_latency_us; double sr_exit_latency_us; double sr_enter_plus_exit_latency_us; + double sr_exit_z8_time_us; + double sr_enter_plus_exit_z8_time_us; double dram_clock_change_latency_us; double fclk_change_latency_us; unsigned int dram_num_chan; diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c index 2b8b8366538e..cdb903116eb7 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn10/dcn10_hwseq.c @@ -3417,7 +3417,8 @@ void dcn10_set_cursor_position(struct pipe_ctx *pipe_ctx) .h_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.horz, .v_scale_ratio = pipe_ctx->plane_res.scl_data.ratios.vert, .rotation = pipe_ctx->plane_state->rotation, - .mirror = pipe_ctx->plane_state->horizontal_mirror + .mirror = pipe_ctx->plane_state->horizontal_mirror, + .stream = pipe_ctx->stream, }; bool pipe_split_on = false; bool odm_combine_on = (pipe_ctx->next_odm_pipe != NULL) || diff --git a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c index 6a65af8c36b9..c1a9b746c43f 100644 --- a/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c +++ b/drivers/gpu/drm/amd/display/dc/hwss/dcn32/dcn32_hwseq.c @@ -487,8 +487,7 @@ bool dcn32_set_mcm_luts( if (plane_state->blend_tf->type == TF_TYPE_HWPWL) lut_params = &plane_state->blend_tf->pwl; else if (plane_state->blend_tf->type == TF_TYPE_DISTRIBUTED_POINTS) { - cm_helper_translate_curve_to_hw_format(plane_state->ctx, - plane_state->blend_tf, + cm3_helper_translate_curve_to_hw_format(plane_state->blend_tf, &dpp_base->regamma_params, false); lut_params = &dpp_base->regamma_params; } @@ -503,8 +502,7 @@ bool dcn32_set_mcm_luts( else if (plane_state->in_shaper_func->type == TF_TYPE_DISTRIBUTED_POINTS) { // TODO: dpp_base replace ASSERT(false); - cm_helper_translate_curve_to_hw_format(plane_state->ctx, - plane_state->in_shaper_func, + cm3_helper_translate_curve_to_hw_format(plane_state->in_shaper_func, &dpp_base->shaper_params, true); lut_params = &dpp_base->shaper_params; } @@ -962,6 +960,12 @@ void dcn32_init_hw(struct dc *dc) dc->caps.dmub_caps.subvp_psr = dc->ctx->dmub_srv->dmub->feature_caps.subvp_psr_support; dc->caps.dmub_caps.gecc_enable = dc->ctx->dmub_srv->dmub->feature_caps.gecc_enable; dc->caps.dmub_caps.mclk_sw = dc->ctx->dmub_srv->dmub->feature_caps.fw_assisted_mclk_switch; + + if (dc->ctx->dmub_srv->dmub->fw_version < + DMUB_FW_VERSION(7, 0, 35)) { + dc->debug.force_disable_subvp = true; + dc->debug.disable_fpo_optimizations = true; + } } } diff --git a/drivers/gpu/drm/amd/display/dc/link/link_detection.c b/drivers/gpu/drm/amd/display/dc/link/link_detection.c index d6f0f857c05a..24153b0df503 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_detection.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_detection.c @@ -879,7 +879,7 @@ static bool detect_link_and_local_sink(struct dc_link *link, (link->dpcd_sink_ext_caps.bits.oled == 1)) { dpcd_set_source_specific_data(link); msleep(post_oui_delay); - set_cached_brightness_aux(link); + set_default_brightness_aux(link); } return true; @@ -1088,6 +1088,9 @@ static bool detect_link_and_local_sink(struct dc_link *link, if (sink->edid_caps.panel_patch.skip_scdc_overwrite) link->ctx->dc->debug.hdmi20_disable = true; + if (sink->edid_caps.panel_patch.remove_sink_ext_caps) + link->dpcd_sink_ext_caps.raw = 0; + if (dc_is_hdmi_signal(link->connector_signal)) read_scdc_caps(link->ddc, link->local_sink); diff --git a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c index 34a4a8c0e18c..f8e01ca09d96 100644 --- a/drivers/gpu/drm/amd/display/dc/link/link_dpms.c +++ b/drivers/gpu/drm/amd/display/dc/link/link_dpms.c @@ -2142,8 +2142,7 @@ static enum dc_status enable_link_dp(struct dc_state *state, if (link->dpcd_sink_ext_caps.bits.oled == 1 || link->dpcd_sink_ext_caps.bits.sdr_aux_backlight_control == 1 || link->dpcd_sink_ext_caps.bits.hdr_aux_backlight_control == 1) { - set_cached_brightness_aux(link); - + set_default_brightness_aux(link); if (link->dpcd_sink_ext_caps.bits.oled == 1) msleep(bl_oled_enable_delay); edp_backlight_enable_aux(link, true); diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c index fd8f6f198146..68096d12f52f 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_dp_training_fixed_vs_pe_retimer.c @@ -115,7 +115,7 @@ static enum link_training_result perform_fixed_vs_pe_nontransparent_training_seq lt_settings->cr_pattern_time = 16000; /* Fixed VS/PE specific: Toggle link rate */ - apply_toggle_rate_wa = (link->vendor_specific_lttpr_link_rate_wa == target_rate); + apply_toggle_rate_wa = ((link->vendor_specific_lttpr_link_rate_wa == target_rate) || (link->vendor_specific_lttpr_link_rate_wa == 0)); target_rate = get_dpcd_link_rate(<_settings->link_settings); toggle_rate = (target_rate == 0x6) ? 0xA : 0x6; @@ -271,7 +271,7 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence_legacy( /* Vendor specific: Toggle link rate */ toggle_rate = (rate == 0x6) ? 0xA : 0x6; - if (link->vendor_specific_lttpr_link_rate_wa == rate) { + if (link->vendor_specific_lttpr_link_rate_wa == rate || link->vendor_specific_lttpr_link_rate_wa == 0) { core_link_write_dpcd( link, DP_LINK_BW_SET, @@ -617,7 +617,7 @@ enum link_training_result dp_perform_fixed_vs_pe_training_sequence( /* Vendor specific: Toggle link rate */ toggle_rate = (rate == 0x6) ? 0xA : 0x6; - if (link->vendor_specific_lttpr_link_rate_wa == rate) { + if (link->vendor_specific_lttpr_link_rate_wa == rate || link->vendor_specific_lttpr_link_rate_wa == 0) { core_link_write_dpcd( link, DP_LINK_BW_SET, diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c index e32a7974a4bc..e5cfaaef70b3 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.c @@ -170,7 +170,6 @@ bool edp_set_backlight_level_nits(struct dc_link *link, *(uint32_t *)&dpcd_backlight_set.backlight_level_millinits = backlight_millinits; *(uint16_t *)&dpcd_backlight_set.backlight_transition_time_ms = (uint16_t)transition_time_in_ms; - link->backlight_settings.backlight_millinits = backlight_millinits; if (!link->dpcd_caps.panel_luminance_control) { if (core_link_write_dpcd(link, DP_SOURCE_BACKLIGHT_LEVEL, @@ -290,7 +289,7 @@ bool set_default_brightness_aux(struct dc_link *link) default_backlight = 150000; // if < 1 nits or > 5000, it might be wrong readback if (default_backlight < 1000 || default_backlight > 5000000) - default_backlight = 150000; // + default_backlight = 150000; return edp_set_backlight_level_nits(link, true, default_backlight, 0); @@ -298,15 +297,6 @@ bool set_default_brightness_aux(struct dc_link *link) return false; } -bool set_cached_brightness_aux(struct dc_link *link) -{ - if (link->backlight_settings.backlight_millinits) - return edp_set_backlight_level_nits(link, true, - link->backlight_settings.backlight_millinits, 0); - else - return set_default_brightness_aux(link); - return false; -} bool edp_is_ilr_optimization_enabled(struct dc_link *link) { if (link->dpcd_caps.edp_supported_link_rates_count == 0 || !link->panel_config.ilr.optimize_edp_link_rate) diff --git a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.h b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.h index ebf7deb63d13..a034288ad75d 100644 --- a/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.h +++ b/drivers/gpu/drm/amd/display/dc/link/protocols/link_edp_panel_control.h @@ -30,7 +30,6 @@ enum dp_panel_mode dp_get_panel_mode(struct dc_link *link); void dp_set_panel_mode(struct dc_link *link, enum dp_panel_mode panel_mode); bool set_default_brightness_aux(struct dc_link *link); -bool set_cached_brightness_aux(struct dc_link *link); void edp_panel_backlight_power_on(struct dc_link *link, bool wait_for_hpd); int edp_get_backlight_level(const struct dc_link *link); bool edp_get_backlight_level_nits(struct dc_link *link, diff --git a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h index 9665ada0f894..df63aa8f01e9 100644 --- a/drivers/gpu/drm/amd/display/dmub/dmub_srv.h +++ b/drivers/gpu/drm/amd/display/dmub/dmub_srv.h @@ -195,6 +195,7 @@ struct dmub_srv_region_params { uint32_t vbios_size; const uint8_t *fw_inst_const; const uint8_t *fw_bss_data; + bool is_mailbox_in_inbox; }; /** @@ -214,20 +215,25 @@ struct dmub_srv_region_params { */ struct dmub_srv_region_info { uint32_t fb_size; + uint32_t inbox_size; uint8_t num_regions; struct dmub_region regions[DMUB_WINDOW_TOTAL]; }; /** - * struct dmub_srv_fb_params - parameters used for driver fb setup + * struct dmub_srv_memory_params - parameters used for driver fb setup * @region_info: region info calculated by dmub service - * @cpu_addr: base cpu address for the framebuffer - * @gpu_addr: base gpu virtual address for the framebuffer + * @cpu_fb_addr: base cpu address for the framebuffer + * @cpu_inbox_addr: base cpu address for the gart + * @gpu_fb_addr: base gpu virtual address for the framebuffer + * @gpu_inbox_addr: base gpu virtual address for the gart */ -struct dmub_srv_fb_params { +struct dmub_srv_memory_params { const struct dmub_srv_region_info *region_info; - void *cpu_addr; - uint64_t gpu_addr; + void *cpu_fb_addr; + void *cpu_inbox_addr; + uint64_t gpu_fb_addr; + uint64_t gpu_inbox_addr; }; /** @@ -563,8 +569,8 @@ dmub_srv_calc_region_info(struct dmub_srv *dmub, * DMUB_STATUS_OK - success * DMUB_STATUS_INVALID - unspecified error */ -enum dmub_status dmub_srv_calc_fb_info(struct dmub_srv *dmub, - const struct dmub_srv_fb_params *params, +enum dmub_status dmub_srv_calc_mem_info(struct dmub_srv *dmub, + const struct dmub_srv_memory_params *params, struct dmub_srv_fb_info *out); /** diff --git a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c index e43e8d4bfe37..38360adc53d9 100644 --- a/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c +++ b/drivers/gpu/drm/amd/display/dmub/src/dmub_srv.c @@ -434,7 +434,7 @@ dmub_srv_calc_region_info(struct dmub_srv *dmub, uint32_t fw_state_size = DMUB_FW_STATE_SIZE; uint32_t trace_buffer_size = DMUB_TRACE_BUFFER_SIZE; uint32_t scratch_mem_size = DMUB_SCRATCH_MEM_SIZE; - + uint32_t previous_top = 0; if (!dmub->sw_init) return DMUB_STATUS_INVALID; @@ -459,8 +459,15 @@ dmub_srv_calc_region_info(struct dmub_srv *dmub, bios->base = dmub_align(stack->top, 256); bios->top = bios->base + params->vbios_size; - mail->base = dmub_align(bios->top, 256); - mail->top = mail->base + DMUB_MAILBOX_SIZE; + if (params->is_mailbox_in_inbox) { + mail->base = 0; + mail->top = mail->base + DMUB_MAILBOX_SIZE; + previous_top = bios->top; + } else { + mail->base = dmub_align(bios->top, 256); + mail->top = mail->base + DMUB_MAILBOX_SIZE; + previous_top = mail->top; + } fw_info = dmub_get_fw_meta_info(params); @@ -479,7 +486,7 @@ dmub_srv_calc_region_info(struct dmub_srv *dmub, dmub->fw_version = fw_info->fw_version; } - trace_buff->base = dmub_align(mail->top, 256); + trace_buff->base = dmub_align(previous_top, 256); trace_buff->top = trace_buff->base + dmub_align(trace_buffer_size, 64); fw_state->base = dmub_align(trace_buff->top, 256); @@ -490,11 +497,14 @@ dmub_srv_calc_region_info(struct dmub_srv *dmub, out->fb_size = dmub_align(scratch_mem->top, 4096); + if (params->is_mailbox_in_inbox) + out->inbox_size = dmub_align(mail->top, 4096); + return DMUB_STATUS_OK; } -enum dmub_status dmub_srv_calc_fb_info(struct dmub_srv *dmub, - const struct dmub_srv_fb_params *params, +enum dmub_status dmub_srv_calc_mem_info(struct dmub_srv *dmub, + const struct dmub_srv_memory_params *params, struct dmub_srv_fb_info *out) { uint8_t *cpu_base; @@ -509,8 +519,8 @@ enum dmub_status dmub_srv_calc_fb_info(struct dmub_srv *dmub, if (params->region_info->num_regions != DMUB_NUM_WINDOWS) return DMUB_STATUS_INVALID; - cpu_base = (uint8_t *)params->cpu_addr; - gpu_base = params->gpu_addr; + cpu_base = (uint8_t *)params->cpu_fb_addr; + gpu_base = params->gpu_fb_addr; for (i = 0; i < DMUB_NUM_WINDOWS; ++i) { const struct dmub_region *reg = @@ -518,6 +528,12 @@ enum dmub_status dmub_srv_calc_fb_info(struct dmub_srv *dmub, out->fb[i].cpu_addr = cpu_base + reg->base; out->fb[i].gpu_addr = gpu_base + reg->base; + + if (i == DMUB_WINDOW_4_MAILBOX && params->cpu_inbox_addr != 0) { + out->fb[i].cpu_addr = (uint8_t *)params->cpu_inbox_addr + reg->base; + out->fb[i].gpu_addr = params->gpu_inbox_addr + reg->base; + } + out->fb[i].size = reg->top - reg->base; } @@ -707,9 +723,16 @@ enum dmub_status dmub_srv_sync_inbox1(struct dmub_srv *dmub) return DMUB_STATUS_INVALID; if (dmub->hw_funcs.get_inbox1_rptr && dmub->hw_funcs.get_inbox1_wptr) { - dmub->inbox1_rb.rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); - dmub->inbox1_rb.wrpt = dmub->hw_funcs.get_inbox1_wptr(dmub); - dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt; + uint32_t rptr = dmub->hw_funcs.get_inbox1_rptr(dmub); + uint32_t wptr = dmub->hw_funcs.get_inbox1_wptr(dmub); + + if (rptr > dmub->inbox1_rb.capacity || wptr > dmub->inbox1_rb.capacity) { + return DMUB_STATUS_HW_FAILURE; + } else { + dmub->inbox1_rb.rptr = rptr; + dmub->inbox1_rb.wrpt = wptr; + dmub->inbox1_last_wptr = dmub->inbox1_rb.wrpt; + } } return DMUB_STATUS_OK; @@ -743,6 +766,11 @@ enum dmub_status dmub_srv_cmd_queue(struct dmub_srv *dmub, if (!dmub->hw_init) return DMUB_STATUS_INVALID; + if (dmub->inbox1_rb.rptr > dmub->inbox1_rb.capacity || + dmub->inbox1_rb.wrpt > dmub->inbox1_rb.capacity) { + return DMUB_STATUS_HW_FAILURE; + } + if (dmub_rb_push_front(&dmub->inbox1_rb, cmd)) return DMUB_STATUS_OK; @@ -1049,6 +1077,7 @@ enum dmub_status dmub_srv_wait_for_inbox0_ack(struct dmub_srv *dmub, uint32_t ti ack = dmub->hw_funcs.read_inbox0_ack_register(dmub); if (ack) return DMUB_STATUS_OK; + udelay(1); } return DMUB_STATUS_TIMEOUT; } diff --git a/drivers/gpu/drm/amd/display/include/grph_object_ctrl_defs.h b/drivers/gpu/drm/amd/display/include/grph_object_ctrl_defs.h index bc96d0211360..813463ffe15c 100644 --- a/drivers/gpu/drm/amd/display/include/grph_object_ctrl_defs.h +++ b/drivers/gpu/drm/amd/display/include/grph_object_ctrl_defs.h @@ -417,6 +417,8 @@ struct integrated_info { /* V2.1 */ struct edp_info edp1_info; struct edp_info edp2_info; + uint32_t gpuclk_ss_percentage; + uint32_t gpuclk_ss_type; }; /* diff --git a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c index 84f9b412a4f1..738ee763f24a 100644 --- a/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c +++ b/drivers/gpu/drm/amd/display/modules/info_packet/info_packet.c @@ -147,12 +147,15 @@ void mod_build_vsc_infopacket(const struct dc_stream_state *stream, } /* VSC packet set to 4 for PSR-SU, or 2 for PSR1 */ - if (stream->link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) - vsc_packet_revision = vsc_packet_rev4; - else if (stream->link->replay_settings.config.replay_supported) + if (stream->link->psr_settings.psr_feature_enabled) { + if (stream->link->psr_settings.psr_version == DC_PSR_VERSION_SU_1) + vsc_packet_revision = vsc_packet_rev4; + else if (stream->link->psr_settings.psr_version == DC_PSR_VERSION_1) + vsc_packet_revision = vsc_packet_rev2; + } + + if (stream->link->replay_settings.config.replay_supported) vsc_packet_revision = vsc_packet_rev4; - else if (stream->link->psr_settings.psr_version == DC_PSR_VERSION_1) - vsc_packet_revision = vsc_packet_rev2; /* Update to revision 5 for extended colorimetry support */ if (stream->use_vsc_sdp_for_colorimetry) diff --git a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c index a522a7c02911..1675314a3ff2 100644 --- a/drivers/gpu/drm/amd/display/modules/power/power_helpers.c +++ b/drivers/gpu/drm/amd/display/modules/power/power_helpers.c @@ -839,6 +839,8 @@ bool is_psr_su_specific_panel(struct dc_link *link) ((dpcd_caps->sink_dev_id_str[1] == 0x08 && dpcd_caps->sink_dev_id_str[0] == 0x08) || (dpcd_caps->sink_dev_id_str[1] == 0x08 && dpcd_caps->sink_dev_id_str[0] == 0x07))) isPSRSUSupported = false; + else if (dpcd_caps->sink_dev_id_str[1] == 0x08 && dpcd_caps->sink_dev_id_str[0] == 0x03) + isPSRSUSupported = false; else if (dpcd_caps->psr_info.force_psrsu_cap == 0x1) isPSRSUSupported = true; } diff --git a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_offset.h index c92c4b83253f..4bff1ef8a9a6 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/gc/gc_11_0_0_offset.h @@ -6369,6 +6369,8 @@ #define regTCP_INVALIDATE_BASE_IDX 1 #define regTCP_STATUS 0x19a1 #define regTCP_STATUS_BASE_IDX 1 +#define regTCP_CNTL 0x19a2 +#define regTCP_CNTL_BASE_IDX 1 #define regTCP_CNTL2 0x19a3 #define regTCP_CNTL2_BASE_IDX 1 #define regTCP_DEBUG_INDEX 0x19a5 diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_offset.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_offset.h index ff30f04be591..7ee3d291120d 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_offset.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_offset.h @@ -781,6 +781,8 @@ #define regBIF_BIF256_CI256_RC3X4_USB4_PCIE_CNTL2_BASE_IDX 5 #define regBIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1 0x420187 #define regBIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1_BASE_IDX 5 +#define regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3 0x4201c6 +#define regBIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3_BASE_IDX 5 // addressBlock: nbio_nbif0_bif_cfg_dev0_rc_bifcfgdecp diff --git a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_sh_mask.h b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_sh_mask.h index 7f131999a263..eb8c556d9c93 100644 --- a/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_sh_mask.h +++ b/drivers/gpu/drm/amd/include/asic_reg/nbio/nbio_7_11_0_sh_mask.h @@ -24646,6 +24646,35 @@ //BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1 #define BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__MST_MEM_LS_EN_MASK 0x00000001L #define BIF_BIF256_CI256_RC3X4_USB4_PCIE_TX_POWER_CTRL_1__REPLAY_MEM_LS_EN_MASK 0x00000008L +//BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_PAYLOAD_SIZE_MODE__SHIFT 0x8 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_PRIV_MAX_PAYLOAD_SIZE__SHIFT 0x9 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_10BIT_TAG_EN_OVERRIDE__SHIFT 0xb +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_10BIT_TAG_EN_OVERRIDE__SHIFT 0xd +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__MST_DROP_SYNC_FLOOD_EN__SHIFT 0xf +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_PAYLOAD_SIZE_MODE__SHIFT 0x10 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_PRIV_MAX_PAYLOAD_SIZE__SHIFT 0x11 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_READ_REQUEST_SIZE_MODE__SHIFT 0x14 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_PRIV_MAX_READ_REQUEST_SIZE__SHIFT 0x15 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_READ_SAFE_MODE__SHIFT 0x18 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_EXTENDED_TAG_EN_OVERRIDE__SHIFT 0x19 +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_READ_REQUEST_SIZE_MODE__SHIFT 0x1b +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV__SHIFT 0x1c +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_EXTENDED_TAG_EN_OVERRIDE__SHIFT 0x1e +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_PAYLOAD_SIZE_MODE_MASK 0x00000100L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_PRIV_MAX_PAYLOAD_SIZE_MASK 0x00000600L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_10BIT_TAG_EN_OVERRIDE_MASK 0x00001800L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_10BIT_TAG_EN_OVERRIDE_MASK 0x00006000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__MST_DROP_SYNC_FLOOD_EN_MASK 0x00008000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_PAYLOAD_SIZE_MODE_MASK 0x00010000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_PRIV_MAX_PAYLOAD_SIZE_MASK 0x000E0000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_READ_REQUEST_SIZE_MODE_MASK 0x00100000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_PRIV_MAX_READ_REQUEST_SIZE_MASK 0x00E00000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_MAX_READ_SAFE_MODE_MASK 0x01000000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_EXTENDED_TAG_EN_OVERRIDE_MASK 0x06000000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_READ_REQUEST_SIZE_MODE_MASK 0x08000000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_MAX_READ_REQUEST_SIZE_PRIV_MASK 0x30000000L +#define BIF_BIF256_CI256_RC3X4_USB4_PCIE_MST_CTRL_3__CI_SWUS_EXTENDED_TAG_EN_OVERRIDE_MASK 0xC0000000L // addressBlock: nbio_nbif0_bif_cfg_dev0_rc_bifcfgdecp //BIF_CFG_DEV0_RC0_VENDOR_ID diff --git a/drivers/gpu/drm/amd/include/kgd_pp_interface.h b/drivers/gpu/drm/amd/include/kgd_pp_interface.h index cd3c40a86029..c2ccf3724e37 100644 --- a/drivers/gpu/drm/amd/include/kgd_pp_interface.h +++ b/drivers/gpu/drm/amd/include/kgd_pp_interface.h @@ -318,6 +318,7 @@ enum pp_xgmi_plpd_mode { #define MAX_GFX_CLKS 8 #define MAX_CLKS 4 #define NUM_VCN 4 +#define NUM_JPEG_ENG 32 struct seq_file; enum amd_pp_clock_type; @@ -444,6 +445,7 @@ struct amd_pm_funcs { struct dpm_clocks *clock_table); int (*get_smu_prv_buf_details)(void *handle, void **addr, size_t *size); void (*pm_compute_clocks)(void *handle); + int (*notify_rlc_state)(void *handle, bool en); }; struct metrics_table_header { @@ -773,6 +775,85 @@ struct gpu_metrics_v1_4 { uint16_t padding; }; +struct gpu_metrics_v1_5 { + struct metrics_table_header common_header; + + /* Temperature (Celsius) */ + uint16_t temperature_hotspot; + uint16_t temperature_mem; + uint16_t temperature_vrsoc; + + /* Power (Watts) */ + uint16_t curr_socket_power; + + /* Utilization (%) */ + uint16_t average_gfx_activity; + uint16_t average_umc_activity; // memory controller + uint16_t vcn_activity[NUM_VCN]; + uint16_t jpeg_activity[NUM_JPEG_ENG]; + + /* Energy (15.259uJ (2^-16) units) */ + uint64_t energy_accumulator; + + /* Driver attached timestamp (in ns) */ + uint64_t system_clock_counter; + + /* Throttle status */ + uint32_t throttle_status; + + /* Clock Lock Status. Each bit corresponds to clock instance */ + uint32_t gfxclk_lock_status; + + /* Link width (number of lanes) and speed (in 0.1 GT/s) */ + uint16_t pcie_link_width; + uint16_t pcie_link_speed; + + /* XGMI bus width and bitrate (in Gbps) */ + uint16_t xgmi_link_width; + uint16_t xgmi_link_speed; + + /* Utilization Accumulated (%) */ + uint32_t gfx_activity_acc; + uint32_t mem_activity_acc; + + /*PCIE accumulated bandwidth (GB/sec) */ + uint64_t pcie_bandwidth_acc; + + /*PCIE instantaneous bandwidth (GB/sec) */ + uint64_t pcie_bandwidth_inst; + + /* PCIE L0 to recovery state transition accumulated count */ + uint64_t pcie_l0_to_recov_count_acc; + + /* PCIE replay accumulated count */ + uint64_t pcie_replay_count_acc; + + /* PCIE replay rollover accumulated count */ + uint64_t pcie_replay_rover_count_acc; + + /* PCIE NAK sent accumulated count */ + uint32_t pcie_nak_sent_count_acc; + + /* PCIE NAK received accumulated count */ + uint32_t pcie_nak_rcvd_count_acc; + + /* XGMI accumulated data transfer size(KiloBytes) */ + uint64_t xgmi_read_data_acc[NUM_XGMI_LINKS]; + uint64_t xgmi_write_data_acc[NUM_XGMI_LINKS]; + + /* PMFW attached timestamp (10ns resolution) */ + uint64_t firmware_timestamp; + + /* Current clocks (Mhz) */ + uint16_t current_gfxclk[MAX_GFX_CLKS]; + uint16_t current_socclk[MAX_CLKS]; + uint16_t current_vclk0[MAX_CLKS]; + uint16_t current_dclk0[MAX_CLKS]; + uint16_t current_uclk; + + uint16_t padding; +}; + /* * gpu_metrics_v2_0 is not recommended as it's not naturally aligned. * Use gpu_metrics_v2_1 or later instead. @@ -1084,6 +1165,10 @@ struct gpu_metrics_v3_0 { uint16_t average_dram_reads; /* time filtered DRAM write bandwidth [MB/sec] */ uint16_t average_dram_writes; + /* time filtered IPU read bandwidth [MB/sec] */ + uint16_t average_ipu_reads; + /* time filtered IPU write bandwidth [MB/sec] */ + uint16_t average_ipu_writes; /* Driver attached timestamp (in ns) */ uint64_t system_clock_counter; @@ -1103,6 +1188,8 @@ struct gpu_metrics_v3_0 { uint32_t average_all_core_power; /* calculated core power [mW] */ uint16_t average_core_power[16]; + /* time filtered total system power [mW] */ + uint16_t average_sys_power; /* maximum IRM defined STAPM power limit [mW] */ uint16_t stapm_power_limit; /* time filtered STAPM power limit [mW] */ @@ -1115,6 +1202,8 @@ struct gpu_metrics_v3_0 { uint16_t average_ipuclk_frequency; uint16_t average_fclk_frequency; uint16_t average_vclk_frequency; + uint16_t average_uclk_frequency; + uint16_t average_mpipu_frequency; /* Current clocks */ /* target core frequency [MHz] */ @@ -1124,6 +1213,15 @@ struct gpu_metrics_v3_0 { /* GFXCLK frequency limit enforced on GFX [MHz] */ uint16_t current_gfx_maxfreq; + /* Throttle Residency (ASIC dependent) */ + uint32_t throttle_residency_prochot; + uint32_t throttle_residency_spl; + uint32_t throttle_residency_fppt; + uint32_t throttle_residency_sppt; + uint32_t throttle_residency_thm_core; + uint32_t throttle_residency_thm_gfx; + uint32_t throttle_residency_thm_soc; + /* Metrics table alpha filter time constant [us] */ uint32_t time_filter_alphavalue; }; diff --git a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c index 08cb79401410..8ec11da0319f 100644 --- a/drivers/gpu/drm/amd/pm/amdgpu_dpm.c +++ b/drivers/gpu/drm/amd/pm/amdgpu_dpm.c @@ -181,6 +181,24 @@ int amdgpu_dpm_set_mp1_state(struct amdgpu_device *adev, return ret; } +int amdgpu_dpm_notify_rlc_state(struct amdgpu_device *adev, bool en) +{ + int ret = 0; + const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; + + if (pp_funcs && pp_funcs->notify_rlc_state) { + mutex_lock(&adev->pm.mutex); + + ret = pp_funcs->notify_rlc_state( + adev->powerplay.pp_handle, + en); + + mutex_unlock(&adev->pm.mutex); + } + + return ret; +} + bool amdgpu_dpm_is_baco_supported(struct amdgpu_device *adev) { const struct amd_pm_funcs *pp_funcs = adev->powerplay.pp_funcs; diff --git a/drivers/gpu/drm/amd/pm/amdgpu_pm.c b/drivers/gpu/drm/amd/pm/amdgpu_pm.c index ca2ece24e1e0..20c53eefd680 100644 --- a/drivers/gpu/drm/amd/pm/amdgpu_pm.c +++ b/drivers/gpu/drm/amd/pm/amdgpu_pm.c @@ -2128,7 +2128,9 @@ static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_ if (amdgpu_dpm_is_overdrive_supported(adev)) *states = ATTR_STATE_SUPPORTED; } else if (DEVICE_ATTR_IS(mem_busy_percent)) { - if (adev->flags & AMD_IS_APU || gc_ver == IP_VERSION(9, 0, 1)) + if ((adev->flags & AMD_IS_APU && + gc_ver != IP_VERSION(9, 4, 3)) || + gc_ver == IP_VERSION(9, 0, 1)) *states = ATTR_STATE_UNSUPPORTED; } else if (DEVICE_ATTR_IS(pcie_bw)) { /* PCIe Perf counters won't work on APU nodes */ @@ -2198,10 +2200,10 @@ static int default_attr_update(struct amdgpu_device *adev, struct amdgpu_device_ } else if (DEVICE_ATTR_IS(xgmi_plpd_policy)) { if (amdgpu_dpm_get_xgmi_plpd_mode(adev, NULL) == XGMI_PLPD_NONE) *states = ATTR_STATE_UNSUPPORTED; - } else if (DEVICE_ATTR_IS(pp_dpm_mclk_od)) { + } else if (DEVICE_ATTR_IS(pp_mclk_od)) { if (amdgpu_dpm_get_mclk_od(adev) == -EOPNOTSUPP) *states = ATTR_STATE_UNSUPPORTED; - } else if (DEVICE_ATTR_IS(pp_dpm_sclk_od)) { + } else if (DEVICE_ATTR_IS(pp_sclk_od)) { if (amdgpu_dpm_get_sclk_od(adev) == -EOPNOTSUPP) *states = ATTR_STATE_UNSUPPORTED; } else if (DEVICE_ATTR_IS(apu_thermal_cap)) { diff --git a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h index feccd2a7120d..482ea30147ab 100644 --- a/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h +++ b/drivers/gpu/drm/amd/pm/inc/amdgpu_dpm.h @@ -415,6 +415,8 @@ int amdgpu_dpm_mode1_reset(struct amdgpu_device *adev); int amdgpu_dpm_set_mp1_state(struct amdgpu_device *adev, enum pp_mp1_state mp1_state); +int amdgpu_dpm_notify_rlc_state(struct amdgpu_device *adev, bool en); + int amdgpu_dpm_set_gfx_power_up_by_imu(struct amdgpu_device *adev); int amdgpu_dpm_baco_exit(struct amdgpu_device *adev); diff --git a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c index 1ead323f1c78..e1a5ee911dbb 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c +++ b/drivers/gpu/drm/amd/pm/swsmu/amdgpu_smu.c @@ -1710,6 +1710,16 @@ static int smu_disable_dpms(struct smu_context *smu) } } + /* Notify SMU RLC is going to be off, stop RLC and SMU interaction. + * otherwise SMU will hang while interacting with RLC if RLC is halted + * this is a WA for Vangogh asic which fix the SMU hang issue. + */ + ret = smu_notify_rlc_state(smu, false); + if (ret) { + dev_err(adev->dev, "Fail to notify rlc status!\n"); + return ret; + } + if (amdgpu_ip_version(adev, GC_HWIP, 0) >= IP_VERSION(9, 4, 2) && !((adev->flags & AMD_IS_APU) && adev->gfx.imu.funcs) && !amdgpu_sriov_vf(adev) && adev->gfx.rlc.funcs->stop) diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h index 8def291b18bc..f8b2e6cc2568 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/amdgpu_smu.h @@ -1360,6 +1360,11 @@ struct pptable_funcs { * management. */ int (*dpm_set_umsch_mm_enable)(struct smu_context *smu, bool enable); + + /** + * @notify_rlc_state: Notify RLC power state to SMU. + */ + int (*notify_rlc_state)(struct smu_context *smu, bool en); }; typedef enum { @@ -1403,6 +1408,16 @@ typedef enum { METRICS_PCIE_WIDTH, METRICS_CURR_FANPWM, METRICS_CURR_SOCKETPOWER, + METRICS_AVERAGE_VPECLK, + METRICS_AVERAGE_IPUCLK, + METRICS_AVERAGE_MPIPUCLK, + METRICS_THROTTLER_RESIDENCY_PROCHOT, + METRICS_THROTTLER_RESIDENCY_SPL, + METRICS_THROTTLER_RESIDENCY_FPPT, + METRICS_THROTTLER_RESIDENCY_SPPT, + METRICS_THROTTLER_RESIDENCY_THM_CORE, + METRICS_THROTTLER_RESIDENCY_THM_GFX, + METRICS_THROTTLER_RESIDENCY_THM_SOC, } MetricsMember_t; enum smu_cmn2asic_mapping_type { diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu14_driver_if_v14_0_0.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu14_driver_if_v14_0_0.h index 22f88842a7fd..8f42771e1f0a 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu14_driver_if_v14_0_0.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu14_driver_if_v14_0_0.h @@ -27,7 +27,7 @@ // *** IMPORTANT *** // SMU TEAM: Always increment the interface version if // any structure is changed in this file -#define PMFW_DRIVER_IF_VERSION 6 +#define PMFW_DRIVER_IF_VERSION 7 typedef struct { int32_t value; @@ -150,37 +150,50 @@ typedef struct { } DpmClocks_t; typedef struct { - uint16_t CoreFrequency[16]; //Target core frequency [MHz] - uint16_t CorePower[16]; //CAC calculated core power [mW] - uint16_t CoreTemperature[16]; //TSEN measured core temperature [centi-C] - uint16_t GfxTemperature; //TSEN measured GFX temperature [centi-C] - uint16_t SocTemperature; //TSEN measured SOC temperature [centi-C] - uint16_t StapmOpnLimit; //Maximum IRM defined STAPM power limit [mW] - uint16_t StapmCurrentLimit; //Time filtered STAPM power limit [mW] - uint16_t InfrastructureCpuMaxFreq; //CCLK frequency limit enforced on classic cores [MHz] - uint16_t InfrastructureGfxMaxFreq; //GFXCLK frequency limit enforced on GFX [MHz] - uint16_t SkinTemp; //Maximum skin temperature reported by APU and HS2 chassis sensors [centi-C] - uint16_t GfxclkFrequency; //Time filtered target GFXCLK frequency [MHz] - uint16_t FclkFrequency; //Time filtered target FCLK frequency [MHz] - uint16_t GfxActivity; //Time filtered GFX busy % [0-100] - uint16_t SocclkFrequency; //Time filtered target SOCCLK frequency [MHz] - uint16_t VclkFrequency; //Time filtered target VCLK frequency [MHz] - uint16_t VcnActivity; //Time filtered VCN busy % [0-100] - uint16_t VpeclkFrequency; //Time filtered target VPECLK frequency [MHz] - uint16_t IpuclkFrequency; //Time filtered target IPUCLK frequency [MHz] - uint16_t IpuBusy[8]; //Time filtered IPU per-column busy % [0-100] - uint16_t DRAMReads; //Time filtered DRAM read bandwidth [MB/sec] - uint16_t DRAMWrites; //Time filtered DRAM write bandwidth [MB/sec] - uint16_t CoreC0Residency[16]; //Time filtered per-core C0 residency % [0-100] - uint16_t IpuPower; //Time filtered IPU power [mW] - uint32_t ApuPower; //Time filtered APU power [mW] - uint32_t GfxPower; //Time filtered GFX power [mW] - uint32_t dGpuPower; //Time filtered dGPU power [mW] - uint32_t SocketPower; //Time filtered power used for PPT/STAPM [APU+dGPU] [mW] - uint32_t AllCorePower; //Time filtered sum of core power across all cores in the socket [mW] - uint32_t FilterAlphaValue; //Metrics table alpha filter time constant [us] - uint32_t MetricsCounter; //Counter that is incremented on every metrics table update [PM_TIMER cycles] - uint32_t spare[16]; + uint16_t CoreFrequency[16]; //Target core frequency [MHz] + uint16_t CorePower[16]; //CAC calculated core power [mW] + uint16_t CoreTemperature[16]; //TSEN measured core temperature [centi-C] + uint16_t GfxTemperature; //TSEN measured GFX temperature [centi-C] + uint16_t SocTemperature; //TSEN measured SOC temperature [centi-C] + uint16_t StapmOpnLimit; //Maximum IRM defined STAPM power limit [mW] + uint16_t StapmCurrentLimit; //Time filtered STAPM power limit [mW] + uint16_t InfrastructureCpuMaxFreq; //CCLK frequency limit enforced on classic cores [MHz] + uint16_t InfrastructureGfxMaxFreq; //GFXCLK frequency limit enforced on GFX [MHz] + uint16_t SkinTemp; //Maximum skin temperature reported by APU and HS2 chassis sensors [centi-C] + uint16_t GfxclkFrequency; //Time filtered target GFXCLK frequency [MHz] + uint16_t FclkFrequency; //Time filtered target FCLK frequency [MHz] + uint16_t GfxActivity; //Time filtered GFX busy % [0-100] + uint16_t SocclkFrequency; //Time filtered target SOCCLK frequency [MHz] + uint16_t VclkFrequency; //Time filtered target VCLK frequency [MHz] + uint16_t VcnActivity; //Time filtered VCN busy % [0-100] + uint16_t VpeclkFrequency; //Time filtered target VPECLK frequency [MHz] + uint16_t IpuclkFrequency; //Time filtered target IPUCLK frequency [MHz] + uint16_t IpuBusy[8]; //Time filtered IPU per-column busy % [0-100] + uint16_t DRAMReads; //Time filtered DRAM read bandwidth [MB/sec] + uint16_t DRAMWrites; //Time filtered DRAM write bandwidth [MB/sec] + uint16_t CoreC0Residency[16]; //Time filtered per-core C0 residency % [0-100] + uint16_t IpuPower; //Time filtered IPU power [mW] + uint32_t ApuPower; //Time filtered APU power [mW] + uint32_t GfxPower; //Time filtered GFX power [mW] + uint32_t dGpuPower; //Time filtered dGPU power [mW] + uint32_t SocketPower; //Time filtered power used for PPT/STAPM [APU+dGPU] [mW] + uint32_t AllCorePower; //Time filtered sum of core power across all cores in the socket [mW] + uint32_t FilterAlphaValue; //Metrics table alpha filter time constant [us] + uint32_t MetricsCounter; //Counter that is incremented on every metrics table update [PM_TIMER cycles] + uint16_t MemclkFrequency; //Time filtered target MEMCLK frequency [MHz] + uint16_t MpipuclkFrequency; //Time filtered target MPIPUCLK frequency [MHz] + uint16_t IpuReads; //Time filtered IPU read bandwidth [MB/sec] + uint16_t IpuWrites; //Time filtered IPU write bandwidth [MB/sec] + uint32_t ThrottleResidency_PROCHOT; //Counter that is incremented on every metrics table update when PROCHOT was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_SPL; //Counter that is incremented on every metrics table update when SPL was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_FPPT; //Counter that is incremented on every metrics table update when fast PPT was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_SPPT; //Counter that is incremented on every metrics table update when slow PPT was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_THM_CORE; //Counter that is incremented on every metrics table update when CORE thermal throttling was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_THM_GFX; //Counter that is incremented on every metrics table update when GFX thermal throttling was engaged [PM_TIMER cycles] + uint32_t ThrottleResidency_THM_SOC; //Counter that is incremented on every metrics table update when SOC thermal throttling was engaged [PM_TIMER cycles] + uint16_t Psys; //Time filtered Psys power [mW] + uint16_t spare1; + uint32_t spare[6]; } SmuMetrics_t; //ISP tile definitions diff --git a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h index dab35d878a90..7b812b9994d7 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h +++ b/drivers/gpu/drm/amd/pm/swsmu/inc/pmfw_if/smu_v13_0_6_pmfw.h @@ -123,7 +123,7 @@ typedef enum { VOLTAGE_GUARDBAND_COUNT } GFX_GUARDBAND_e; -#define SMU_METRICS_TABLE_VERSION 0x8 +#define SMU_METRICS_TABLE_VERSION 0xB typedef struct __attribute__((packed, aligned(4))) { uint32_t AccumulationCounter; @@ -211,7 +211,111 @@ typedef struct __attribute__((packed, aligned(4))) { //XGMI Data tranfser size uint64_t XgmiReadDataSizeAcc[8];//in KByte uint64_t XgmiWriteDataSizeAcc[8];//in KByte -} MetricsTable_t; + + //PCIE BW Data and error count + uint32_t PcieBandwidth[4]; + uint32_t PCIeL0ToRecoveryCountAcc; // The Pcie counter itself is accumulated + uint32_t PCIenReplayAAcc; // The Pcie counter itself is accumulated + uint32_t PCIenReplayARolloverCountAcc; // The Pcie counter itself is accumulated + uint32_t PCIeNAKSentCountAcc; // The Pcie counter itself is accumulated + uint32_t PCIeNAKReceivedCountAcc; // The Pcie counter itself is accumulated + + // VCN/JPEG ACTIVITY + uint32_t VcnBusy[4]; + uint32_t JpegBusy[32]; +} MetricsTableX_t; + +typedef struct __attribute__((packed, aligned(4))) { + uint32_t AccumulationCounter; + + //TEMPERATURE + uint32_t MaxSocketTemperature; + uint32_t MaxVrTemperature; + uint32_t MaxHbmTemperature; + uint64_t MaxSocketTemperatureAcc; + uint64_t MaxVrTemperatureAcc; + uint64_t MaxHbmTemperatureAcc; + + //POWER + uint32_t SocketPowerLimit; + uint32_t MaxSocketPowerLimit; + uint32_t SocketPower; + + //ENERGY + uint64_t Timestamp; + uint64_t SocketEnergyAcc; + uint64_t CcdEnergyAcc; + uint64_t XcdEnergyAcc; + uint64_t AidEnergyAcc; + uint64_t HbmEnergyAcc; + + //FREQUENCY + uint32_t CclkFrequencyLimit; + uint32_t GfxclkFrequencyLimit; + uint32_t FclkFrequency; + uint32_t UclkFrequency; + uint32_t SocclkFrequency[4]; + uint32_t VclkFrequency[4]; + uint32_t DclkFrequency[4]; + uint32_t LclkFrequency[4]; + uint64_t GfxclkFrequencyAcc[8]; + uint64_t CclkFrequencyAcc[96]; + + //FREQUENCY RANGE + uint32_t MaxCclkFrequency; + uint32_t MinCclkFrequency; + uint32_t MaxGfxclkFrequency; + uint32_t MinGfxclkFrequency; + uint32_t FclkFrequencyTable[4]; + uint32_t UclkFrequencyTable[4]; + uint32_t SocclkFrequencyTable[4]; + uint32_t VclkFrequencyTable[4]; + uint32_t DclkFrequencyTable[4]; + uint32_t LclkFrequencyTable[4]; + uint32_t MaxLclkDpmRange; + uint32_t MinLclkDpmRange; + + //XGMI + uint32_t XgmiWidth; + uint32_t XgmiBitrate; + uint64_t XgmiReadBandwidthAcc[8]; + uint64_t XgmiWriteBandwidthAcc[8]; + + //ACTIVITY + uint32_t SocketC0Residency; + uint32_t SocketGfxBusy; + uint32_t DramBandwidthUtilization; + uint64_t SocketC0ResidencyAcc; + uint64_t SocketGfxBusyAcc; + uint64_t DramBandwidthAcc; + uint32_t MaxDramBandwidth; + uint64_t DramBandwidthUtilizationAcc; + uint64_t PcieBandwidthAcc[4]; + + //THROTTLERS + uint32_t ProchotResidencyAcc; + uint32_t PptResidencyAcc; + uint32_t SocketThmResidencyAcc; + uint32_t VrThmResidencyAcc; + uint32_t HbmThmResidencyAcc; + uint32_t GfxLockXCDMak; + + // New Items at end to maintain driver compatibility + uint32_t GfxclkFrequency[8]; + + //PSNs + uint64_t PublicSerialNumber_AID[4]; + uint64_t PublicSerialNumber_XCD[8]; + uint64_t PublicSerialNumber_CCD[12]; + + //XGMI Data tranfser size + uint64_t XgmiReadDataSizeAcc[8];//in KByte + uint64_t XgmiWriteDataSizeAcc[8];//in KByte + + // VCN/JPEG ACTIVITY + uint32_t VcnBusy[4]; + uint32_t JpegBusy[32]; +} MetricsTableA_t; #define SMU_VF_METRICS_TABLE_VERSION 0x3 diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c index 762b31455a0b..2ff6deedef95 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu11/vangogh_ppt.c @@ -2193,8 +2193,7 @@ static int vangogh_get_dpm_clock_table(struct smu_context *smu, struct dpm_clock return 0; } - -static int vangogh_system_features_control(struct smu_context *smu, bool en) +static int vangogh_notify_rlc_state(struct smu_context *smu, bool en) { struct amdgpu_device *adev = smu->adev; int ret = 0; @@ -2523,7 +2522,7 @@ static const struct pptable_funcs vangogh_ppt_funcs = { .print_clk_levels = vangogh_common_print_clk_levels, .set_default_dpm_table = vangogh_set_default_dpm_tables, .set_fine_grain_gfx_freq_parameters = vangogh_set_fine_grain_gfx_freq_parameters, - .system_features_control = vangogh_system_features_control, + .notify_rlc_state = vangogh_notify_rlc_state, .feature_is_enabled = smu_cmn_feature_is_enabled, .set_power_profile_mode = vangogh_set_power_profile_mode, .get_power_profile_mode = vangogh_get_power_profile_mode, diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c index 1a6675d70a4b..f1440869d1ce 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/aldebaran_ppt.c @@ -257,8 +257,11 @@ static int aldebaran_tables_init(struct smu_context *smu) } smu_table->ecc_table = kzalloc(tables[SMU_TABLE_ECCINFO].size, GFP_KERNEL); - if (!smu_table->ecc_table) + if (!smu_table->ecc_table) { + kfree(smu_table->metrics_table); + kfree(smu_table->gpu_metrics_table); return -ENOMEM; + } return 0; } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c index 891605d4975f..b64e07b75937 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu13/smu_v13_0_6_ppt.c @@ -245,6 +245,8 @@ struct PPTable_t { #define SMUQ10_TO_UINT(x) ((x) >> 10) #define SMUQ10_FRAC(x) ((x) & 0x3ff) #define SMUQ10_ROUND(x) ((SMUQ10_TO_UINT(x)) + ((SMUQ10_FRAC(x)) >= 0x200)) +#define GET_METRIC_FIELD(field) ((adev->flags & AMD_IS_APU) ?\ + (metrics_a->field) : (metrics_x->field)) struct smu_v13_0_6_dpm_map { enum smu_clk_type clk_type; @@ -327,7 +329,8 @@ static int smu_v13_0_6_tables_init(struct smu_context *smu) SMU_TABLE_INIT(tables, SMU_TABLE_PMSTATUSLOG, SMU13_TOOL_SIZE, PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM); - SMU_TABLE_INIT(tables, SMU_TABLE_SMU_METRICS, sizeof(MetricsTable_t), + SMU_TABLE_INIT(tables, SMU_TABLE_SMU_METRICS, + max(sizeof(MetricsTableX_t), sizeof(MetricsTableA_t)), PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT); @@ -335,12 +338,13 @@ static int smu_v13_0_6_tables_init(struct smu_context *smu) PAGE_SIZE, AMDGPU_GEM_DOMAIN_VRAM | AMDGPU_GEM_DOMAIN_GTT); - smu_table->metrics_table = kzalloc(sizeof(MetricsTable_t), GFP_KERNEL); + smu_table->metrics_table = kzalloc(max(sizeof(MetricsTableX_t), + sizeof(MetricsTableA_t)), GFP_KERNEL); if (!smu_table->metrics_table) return -ENOMEM; smu_table->metrics_time = 0; - smu_table->gpu_metrics_table_size = sizeof(struct gpu_metrics_v1_4); + smu_table->gpu_metrics_table_size = sizeof(struct gpu_metrics_v1_5); smu_table->gpu_metrics_table = kzalloc(smu_table->gpu_metrics_table_size, GFP_KERNEL); if (!smu_table->gpu_metrics_table) { @@ -431,9 +435,11 @@ static int smu_v13_0_6_get_metrics_table(struct smu_context *smu, static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) { struct smu_table_context *smu_table = &smu->smu_table; - MetricsTable_t *metrics = (MetricsTable_t *)smu_table->metrics_table; + MetricsTableX_t *metrics_x = (MetricsTableX_t *)smu_table->metrics_table; + MetricsTableA_t *metrics_a = (MetricsTableA_t *)smu_table->metrics_table; struct PPTable_t *pptable = (struct PPTable_t *)smu_table->driver_pptable; + struct amdgpu_device *adev = smu->adev; int ret, i, retry = 100; /* Store one-time values in driver PPTable */ @@ -444,7 +450,7 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) return ret; /* Ensure that metrics have been updated */ - if (metrics->AccumulationCounter) + if (GET_METRIC_FIELD(AccumulationCounter)) break; usleep_range(1000, 1100); @@ -454,29 +460,29 @@ static int smu_v13_0_6_setup_driver_pptable(struct smu_context *smu) return -ETIME; pptable->MaxSocketPowerLimit = - SMUQ10_ROUND(metrics->MaxSocketPowerLimit); + SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketPowerLimit)); pptable->MaxGfxclkFrequency = - SMUQ10_ROUND(metrics->MaxGfxclkFrequency); + SMUQ10_ROUND(GET_METRIC_FIELD(MaxGfxclkFrequency)); pptable->MinGfxclkFrequency = - SMUQ10_ROUND(metrics->MinGfxclkFrequency); + SMUQ10_ROUND(GET_METRIC_FIELD(MinGfxclkFrequency)); for (i = 0; i < 4; ++i) { pptable->FclkFrequencyTable[i] = - SMUQ10_ROUND(metrics->FclkFrequencyTable[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequencyTable)[i]); pptable->UclkFrequencyTable[i] = - SMUQ10_ROUND(metrics->UclkFrequencyTable[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequencyTable)[i]); pptable->SocclkFrequencyTable[i] = SMUQ10_ROUND( - metrics->SocclkFrequencyTable[i]); + GET_METRIC_FIELD(SocclkFrequencyTable)[i]); pptable->VclkFrequencyTable[i] = - SMUQ10_ROUND(metrics->VclkFrequencyTable[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequencyTable)[i]); pptable->DclkFrequencyTable[i] = - SMUQ10_ROUND(metrics->DclkFrequencyTable[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequencyTable)[i]); pptable->LclkFrequencyTable[i] = - SMUQ10_ROUND(metrics->LclkFrequencyTable[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(LclkFrequencyTable)[i]); } /* use AID0 serial number by default */ - pptable->PublicSerialNumber_AID = metrics->PublicSerialNumber_AID[0]; + pptable->PublicSerialNumber_AID = GET_METRIC_FIELD(PublicSerialNumber_AID)[0]; pptable->Init = true; } @@ -778,7 +784,8 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, uint32_t *value) { struct smu_table_context *smu_table = &smu->smu_table; - MetricsTable_t *metrics = (MetricsTable_t *)smu_table->metrics_table; + MetricsTableX_t *metrics_x = (MetricsTableX_t *)smu_table->metrics_table; + MetricsTableA_t *metrics_a = (MetricsTableA_t *)smu_table->metrics_table; struct amdgpu_device *adev = smu->adev; int ret = 0; int xcc_id; @@ -793,50 +800,50 @@ static int smu_v13_0_6_get_smu_metrics_data(struct smu_context *smu, case METRICS_AVERAGE_GFXCLK: if (smu->smc_fw_version >= 0x552F00) { xcc_id = GET_INST(GC, 0); - *value = SMUQ10_ROUND(metrics->GfxclkFrequency[xcc_id]); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency)[xcc_id]); } else { *value = 0; } break; case METRICS_CURR_SOCCLK: case METRICS_AVERAGE_SOCCLK: - *value = SMUQ10_ROUND(metrics->SocclkFrequency[0]); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency)[0]); break; case METRICS_CURR_UCLK: case METRICS_AVERAGE_UCLK: - *value = SMUQ10_ROUND(metrics->UclkFrequency); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency)); break; case METRICS_CURR_VCLK: - *value = SMUQ10_ROUND(metrics->VclkFrequency[0]); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency)[0]); break; case METRICS_CURR_DCLK: - *value = SMUQ10_ROUND(metrics->DclkFrequency[0]); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency)[0]); break; case METRICS_CURR_FCLK: - *value = SMUQ10_ROUND(metrics->FclkFrequency); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(FclkFrequency)); break; case METRICS_AVERAGE_GFXACTIVITY: - *value = SMUQ10_ROUND(metrics->SocketGfxBusy); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy)); break; case METRICS_AVERAGE_MEMACTIVITY: - *value = SMUQ10_ROUND(metrics->DramBandwidthUtilization); + *value = SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization)); break; case METRICS_CURR_SOCKETPOWER: - *value = SMUQ10_ROUND(metrics->SocketPower) << 8; + *value = SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower)) << 8; break; case METRICS_TEMPERATURE_HOTSPOT: - *value = SMUQ10_ROUND(metrics->MaxSocketTemperature) * + *value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature)) * SMU_TEMPERATURE_UNITS_PER_CENTIGRADES; break; case METRICS_TEMPERATURE_MEM: - *value = SMUQ10_ROUND(metrics->MaxHbmTemperature) * + *value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature)) * SMU_TEMPERATURE_UNITS_PER_CENTIGRADES; break; /* This is the max of all VRs and not just SOC VR. * No need to define another data type for the same. */ case METRICS_TEMPERATURE_VRSOC: - *value = SMUQ10_ROUND(metrics->MaxVrTemperature) * + *value = SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature)) * SMU_TEMPERATURE_UNITS_PER_CENTIGRADES; break; default: @@ -1454,7 +1461,7 @@ static int smu_v13_0_6_register_irq_handler(struct smu_context *smu) static int smu_v13_0_6_notify_unload(struct smu_context *smu) { - if (smu->smc_fw_version <= 0x553500) + if (amdgpu_in_reset(smu->adev)) return 0; dev_dbg(smu->adev->dev, "Notify PMFW about driver unload"); @@ -2022,67 +2029,70 @@ static int smu_v13_0_6_get_current_pcie_link_speed(struct smu_context *smu) static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table) { struct smu_table_context *smu_table = &smu->smu_table; - struct gpu_metrics_v1_4 *gpu_metrics = - (struct gpu_metrics_v1_4 *)smu_table->gpu_metrics_table; + struct gpu_metrics_v1_5 *gpu_metrics = + (struct gpu_metrics_v1_5 *)smu_table->gpu_metrics_table; struct amdgpu_device *adev = smu->adev; - int ret = 0, xcc_id, inst, i; - MetricsTable_t *metrics; + int ret = 0, xcc_id, inst, i, j; + MetricsTableX_t *metrics_x; + MetricsTableA_t *metrics_a; u16 link_width_level; - metrics = kzalloc(sizeof(MetricsTable_t), GFP_KERNEL); - ret = smu_v13_0_6_get_metrics_table(smu, metrics, true); + metrics_x = kzalloc(max(sizeof(MetricsTableX_t), sizeof(MetricsTableA_t)), GFP_KERNEL); + ret = smu_v13_0_6_get_metrics_table(smu, metrics_x, true); if (ret) { - kfree(metrics); + kfree(metrics_x); return ret; } - smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 4); + metrics_a = (MetricsTableA_t *)metrics_x; + + smu_cmn_init_soft_gpu_metrics(gpu_metrics, 1, 5); gpu_metrics->temperature_hotspot = - SMUQ10_ROUND(metrics->MaxSocketTemperature); + SMUQ10_ROUND(GET_METRIC_FIELD(MaxSocketTemperature)); /* Individual HBM stack temperature is not reported */ gpu_metrics->temperature_mem = - SMUQ10_ROUND(metrics->MaxHbmTemperature); + SMUQ10_ROUND(GET_METRIC_FIELD(MaxHbmTemperature)); /* Reports max temperature of all voltage rails */ gpu_metrics->temperature_vrsoc = - SMUQ10_ROUND(metrics->MaxVrTemperature); + SMUQ10_ROUND(GET_METRIC_FIELD(MaxVrTemperature)); gpu_metrics->average_gfx_activity = - SMUQ10_ROUND(metrics->SocketGfxBusy); + SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusy)); gpu_metrics->average_umc_activity = - SMUQ10_ROUND(metrics->DramBandwidthUtilization); + SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilization)); gpu_metrics->curr_socket_power = - SMUQ10_ROUND(metrics->SocketPower); + SMUQ10_ROUND(GET_METRIC_FIELD(SocketPower)); /* Energy counter reported in 15.259uJ (2^-16) units */ - gpu_metrics->energy_accumulator = metrics->SocketEnergyAcc; + gpu_metrics->energy_accumulator = GET_METRIC_FIELD(SocketEnergyAcc); for (i = 0; i < MAX_GFX_CLKS; i++) { xcc_id = GET_INST(GC, i); if (xcc_id >= 0) gpu_metrics->current_gfxclk[i] = - SMUQ10_ROUND(metrics->GfxclkFrequency[xcc_id]); + SMUQ10_ROUND(GET_METRIC_FIELD(GfxclkFrequency)[xcc_id]); if (i < MAX_CLKS) { gpu_metrics->current_socclk[i] = - SMUQ10_ROUND(metrics->SocclkFrequency[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(SocclkFrequency)[i]); inst = GET_INST(VCN, i); if (inst >= 0) { gpu_metrics->current_vclk0[i] = - SMUQ10_ROUND(metrics->VclkFrequency[inst]); + SMUQ10_ROUND(GET_METRIC_FIELD(VclkFrequency)[inst]); gpu_metrics->current_dclk0[i] = - SMUQ10_ROUND(metrics->DclkFrequency[inst]); + SMUQ10_ROUND(GET_METRIC_FIELD(DclkFrequency)[inst]); } } } - gpu_metrics->current_uclk = SMUQ10_ROUND(metrics->UclkFrequency); + gpu_metrics->current_uclk = SMUQ10_ROUND(GET_METRIC_FIELD(UclkFrequency)); /* Throttle status is not reported through metrics now */ gpu_metrics->throttle_status = 0; /* Clock Lock Status. Each bit corresponds to each GFXCLK instance */ - gpu_metrics->gfxclk_lock_status = metrics->GfxLockXCDMak >> GET_INST(GC, 0); + gpu_metrics->gfxclk_lock_status = GET_METRIC_FIELD(GfxLockXCDMak) >> GET_INST(GC, 0); if (!(adev->flags & AMD_IS_APU)) { link_width_level = smu_v13_0_6_get_current_pcie_link_width_level(smu); @@ -2094,30 +2104,57 @@ static ssize_t smu_v13_0_6_get_gpu_metrics(struct smu_context *smu, void **table gpu_metrics->pcie_link_speed = smu_v13_0_6_get_current_pcie_link_speed(smu); gpu_metrics->pcie_bandwidth_acc = - SMUQ10_ROUND(metrics->PcieBandwidthAcc[0]); + SMUQ10_ROUND(metrics_x->PcieBandwidthAcc[0]); + gpu_metrics->pcie_bandwidth_inst = + SMUQ10_ROUND(metrics_x->PcieBandwidth[0]); + gpu_metrics->pcie_l0_to_recov_count_acc = + metrics_x->PCIeL0ToRecoveryCountAcc; + gpu_metrics->pcie_replay_count_acc = + metrics_x->PCIenReplayAAcc; + gpu_metrics->pcie_replay_rover_count_acc = + metrics_x->PCIenReplayARolloverCountAcc; + gpu_metrics->pcie_nak_sent_count_acc = + metrics_x->PCIeNAKSentCountAcc; + gpu_metrics->pcie_nak_rcvd_count_acc = + metrics_x->PCIeNAKReceivedCountAcc; } gpu_metrics->system_clock_counter = ktime_get_boottime_ns(); gpu_metrics->gfx_activity_acc = - SMUQ10_ROUND(metrics->SocketGfxBusyAcc); + SMUQ10_ROUND(GET_METRIC_FIELD(SocketGfxBusyAcc)); gpu_metrics->mem_activity_acc = - SMUQ10_ROUND(metrics->DramBandwidthUtilizationAcc); + SMUQ10_ROUND(GET_METRIC_FIELD(DramBandwidthUtilizationAcc)); for (i = 0; i < NUM_XGMI_LINKS; i++) { gpu_metrics->xgmi_read_data_acc[i] = - SMUQ10_ROUND(metrics->XgmiReadDataSizeAcc[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(XgmiReadDataSizeAcc)[i]); gpu_metrics->xgmi_write_data_acc[i] = - SMUQ10_ROUND(metrics->XgmiWriteDataSizeAcc[i]); + SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWriteDataSizeAcc)[i]); + } + + for (i = 0; i < adev->jpeg.num_jpeg_inst; ++i) { + inst = GET_INST(JPEG, i); + for (j = 0; j < adev->jpeg.num_jpeg_rings; ++j) { + gpu_metrics->jpeg_activity[(i * adev->jpeg.num_jpeg_rings) + j] = + SMUQ10_ROUND(GET_METRIC_FIELD(JpegBusy) + [(inst * adev->jpeg.num_jpeg_rings) + j]); + } } - gpu_metrics->xgmi_link_width = SMUQ10_ROUND(metrics->XgmiWidth); - gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(metrics->XgmiBitrate); + for (i = 0; i < adev->vcn.num_vcn_inst; ++i) { + inst = GET_INST(VCN, i); + gpu_metrics->vcn_activity[i] = + SMUQ10_ROUND(GET_METRIC_FIELD(VcnBusy)[inst]); + } - gpu_metrics->firmware_timestamp = metrics->Timestamp; + gpu_metrics->xgmi_link_width = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiWidth)); + gpu_metrics->xgmi_link_speed = SMUQ10_ROUND(GET_METRIC_FIELD(XgmiBitrate)); + + gpu_metrics->firmware_timestamp = GET_METRIC_FIELD(Timestamp); *table = (void *)gpu_metrics; - kfree(metrics); + kfree(metrics_x); return sizeof(*gpu_metrics); } @@ -2585,13 +2622,20 @@ static bool mca_gfx_smu_bank_is_valid(const struct mca_ras_info *mca_ras, struct static bool mca_smu_bank_is_valid(const struct mca_ras_info *mca_ras, struct amdgpu_device *adev, enum amdgpu_mca_error_type type, struct mca_bank_entry *entry) { + struct smu_context *smu = adev->powerplay.pp_handle; uint32_t errcode, instlo; instlo = REG_GET_FIELD(entry->regs[MCA_REG_IDX_IPID], MCMP1_IPIDT0, InstanceIdLo); if (instlo != 0x03b30400) return false; - errcode = REG_GET_FIELD(entry->regs[MCA_REG_IDX_STATUS], MCMP1_STATUST0, ErrorCode); + if (!(adev->flags & AMD_IS_APU) && smu->smc_fw_version >= 0x00555600) { + errcode = MCA_REG__SYND__ERRORINFORMATION(entry->regs[MCA_REG_IDX_SYND]); + errcode &= 0xff; + } else { + errcode = REG_GET_FIELD(entry->regs[MCA_REG_IDX_STATUS], MCMP1_STATUST0, ErrorCode); + } + return mca_smu_check_error_code(adev, mca_ras, errcode); } diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_0_ppt.c b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_0_ppt.c index 03b38c3a9968..94ccdbfd7090 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_0_ppt.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu14/smu_v14_0_0_ppt.c @@ -246,11 +246,20 @@ static int smu_v14_0_0_get_smu_metrics_data(struct smu_context *smu, *value = 0; break; case METRICS_AVERAGE_UCLK: - *value = 0; + *value = metrics->MemclkFrequency; break; case METRICS_AVERAGE_FCLK: *value = metrics->FclkFrequency; break; + case METRICS_AVERAGE_VPECLK: + *value = metrics->VpeclkFrequency; + break; + case METRICS_AVERAGE_IPUCLK: + *value = metrics->IpuclkFrequency; + break; + case METRICS_AVERAGE_MPIPUCLK: + *value = metrics->MpipuclkFrequency; + break; case METRICS_AVERAGE_GFXACTIVITY: *value = metrics->GfxActivity / 100; break; @@ -270,8 +279,26 @@ static int smu_v14_0_0_get_smu_metrics_data(struct smu_context *smu, *value = metrics->SocTemperature / 100 * SMU_TEMPERATURE_UNITS_PER_CENTIGRADES; break; - case METRICS_THROTTLER_STATUS: - *value = 0; + case METRICS_THROTTLER_RESIDENCY_PROCHOT: + *value = metrics->ThrottleResidency_PROCHOT; + break; + case METRICS_THROTTLER_RESIDENCY_SPL: + *value = metrics->ThrottleResidency_SPL; + break; + case METRICS_THROTTLER_RESIDENCY_FPPT: + *value = metrics->ThrottleResidency_FPPT; + break; + case METRICS_THROTTLER_RESIDENCY_SPPT: + *value = metrics->ThrottleResidency_SPPT; + break; + case METRICS_THROTTLER_RESIDENCY_THM_CORE: + *value = metrics->ThrottleResidency_THM_CORE; + break; + case METRICS_THROTTLER_RESIDENCY_THM_GFX: + *value = metrics->ThrottleResidency_THM_GFX; + break; + case METRICS_THROTTLER_RESIDENCY_THM_SOC: + *value = metrics->ThrottleResidency_THM_SOC; break; case METRICS_VOLTAGE_VDDGFX: *value = 0; @@ -498,6 +525,8 @@ static ssize_t smu_v14_0_0_get_gpu_metrics(struct smu_context *smu, sizeof(uint16_t) * 16); gpu_metrics->average_dram_reads = metrics.DRAMReads; gpu_metrics->average_dram_writes = metrics.DRAMWrites; + gpu_metrics->average_ipu_reads = metrics.IpuReads; + gpu_metrics->average_ipu_writes = metrics.IpuWrites; gpu_metrics->average_socket_power = metrics.SocketPower; gpu_metrics->average_ipu_power = metrics.IpuPower; @@ -505,6 +534,7 @@ static ssize_t smu_v14_0_0_get_gpu_metrics(struct smu_context *smu, gpu_metrics->average_gfx_power = metrics.GfxPower; gpu_metrics->average_dgpu_power = metrics.dGpuPower; gpu_metrics->average_all_core_power = metrics.AllCorePower; + gpu_metrics->average_sys_power = metrics.Psys; memcpy(&gpu_metrics->average_core_power[0], &metrics.CorePower[0], sizeof(uint16_t) * 16); @@ -515,6 +545,8 @@ static ssize_t smu_v14_0_0_get_gpu_metrics(struct smu_context *smu, gpu_metrics->average_fclk_frequency = metrics.FclkFrequency; gpu_metrics->average_vclk_frequency = metrics.VclkFrequency; gpu_metrics->average_ipuclk_frequency = metrics.IpuclkFrequency; + gpu_metrics->average_uclk_frequency = metrics.MemclkFrequency; + gpu_metrics->average_mpipu_frequency = metrics.MpipuclkFrequency; memcpy(&gpu_metrics->current_coreclk[0], &metrics.CoreFrequency[0], @@ -522,6 +554,14 @@ static ssize_t smu_v14_0_0_get_gpu_metrics(struct smu_context *smu, gpu_metrics->current_core_maxfreq = metrics.InfrastructureCpuMaxFreq; gpu_metrics->current_gfx_maxfreq = metrics.InfrastructureGfxMaxFreq; + gpu_metrics->throttle_residency_prochot = metrics.ThrottleResidency_PROCHOT; + gpu_metrics->throttle_residency_spl = metrics.ThrottleResidency_SPL; + gpu_metrics->throttle_residency_fppt = metrics.ThrottleResidency_FPPT; + gpu_metrics->throttle_residency_sppt = metrics.ThrottleResidency_SPPT; + gpu_metrics->throttle_residency_thm_core = metrics.ThrottleResidency_THM_CORE; + gpu_metrics->throttle_residency_thm_gfx = metrics.ThrottleResidency_THM_GFX; + gpu_metrics->throttle_residency_thm_soc = metrics.ThrottleResidency_THM_SOC; + gpu_metrics->time_filter_alphavalue = metrics.FilterAlphaValue; gpu_metrics->system_clock_counter = ktime_get_boottime_ns(); diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c index 001a5cf09657..00cd615bbcdc 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c +++ b/drivers/gpu/drm/amd/pm/swsmu/smu_cmn.c @@ -989,6 +989,9 @@ void smu_cmn_init_soft_gpu_metrics(void *table, uint8_t frev, uint8_t crev) case METRICS_VERSION(1, 4): structure_size = sizeof(struct gpu_metrics_v1_4); break; + case METRICS_VERSION(1, 5): + structure_size = sizeof(struct gpu_metrics_v1_5); + break; case METRICS_VERSION(2, 0): structure_size = sizeof(struct gpu_metrics_v2_0); break; diff --git a/drivers/gpu/drm/amd/pm/swsmu/smu_internal.h b/drivers/gpu/drm/amd/pm/swsmu/smu_internal.h index 80b3c3efc006..64766ac69c53 100644 --- a/drivers/gpu/drm/amd/pm/swsmu/smu_internal.h +++ b/drivers/gpu/drm/amd/pm/swsmu/smu_internal.h @@ -97,6 +97,7 @@ #define smu_get_default_config_table_settings(smu, config_table) smu_ppt_funcs(get_default_config_table_settings, -EOPNOTSUPP, smu, config_table) #define smu_set_config_table(smu, config_table) smu_ppt_funcs(set_config_table, -EOPNOTSUPP, smu, config_table) #define smu_init_pptable_microcode(smu) smu_ppt_funcs(init_pptable_microcode, 0, smu) +#define smu_notify_rlc_state(smu, en) smu_ppt_funcs(notify_rlc_state, 0, smu, en) #endif #endif |