aboutsummaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
AgeCommit message (Collapse)AuthorFilesLines
2023-01-09drm/amdgpu: Fix potential NULL dereferenceLuben Tuikov1-2/+3
Fix potential NULL dereference, in the case when "man", the resource manager might be NULL, when/if we print debug information. Cc: Alex Deucher <[email protected]> Cc: Christian König <[email protected]> Cc: AMD Graphics <[email protected]> Cc: Dan Carpenter <[email protected]> Cc: kernel test robot <[email protected]> Fixes: 7554886daa31ea ("drm/amdgpu: Fix size validation for non-exclusive domains (v4)") Signed-off-by: Luben Tuikov <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-09drm/amd: Add a new helper for loading/validating microcodeMario Limonciello2-0/+39
All microcode runs a basic validation after it's been loaded. Each IP block as part of init will run both. Introduce a wrapper for request_firmware and amdgpu_ucode_validate. This wrapper will also remap any error codes from request_firmware to -ENODEV. This is so that early_init will fail if firmware couldn't be loaded instead of the IP block being disabled. Reviewed-by: Lijo Lazar <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-09drm/amd: Add a legacy mapping to "amdgpu_ucode_ip_version_decode"Mario Limonciello1-0/+221
This will allow other parts of the driver that currently special case firmware file names to before IP version style naming to just have a single call to `amdgpu_ucode_ip_version_decode`. Signed-off-by: Mario Limonciello <[email protected]> Acked-by: Christian König <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Reviewed-by: Lijo Lazar <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-09drm/amd: Delay removal of the firmware framebufferMario Limonciello2-6/+8
Removing the firmware framebuffer from the driver means that even if the driver doesn't support the IP blocks in a GPU it will no longer be functional after the driver fails to initialize. This change will ensure that unsupported IP blocks at least cause the driver to work with the EFI framebuffer. Cc: [email protected] Suggested-by: Alex Deucher <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Reviewed-by: Lijo Lazar <[email protected]> Signed-off-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-09drm/amdgpu: Fix potential NULL dereferenceLuben Tuikov1-2/+3
Fix potential NULL dereference, in the case when "man", the resource manager might be NULL, when/if we print debug information. Cc: Alex Deucher <[email protected]> Cc: Christian König <[email protected]> Cc: AMD Graphics <[email protected]> Cc: Dan Carpenter <[email protected]> Cc: kernel test robot <[email protected]> Fixes: 7554886daa31ea ("drm/amdgpu: Fix size validation for non-exclusive domains (v4)") Signed-off-by: Luben Tuikov <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-05drm/amdgpu: fix missing dma_fence_put in error pathChristian König1-1/+3
When the fence can't be added we need to drop the reference. Suggested-by: Bert Karwatzki <[email protected]> Signed-off-by: Christian König <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Reviewed-by: Alex Deucher <[email protected]> Reviewed-by: Luben Tuikov <[email protected]>
2023-01-05drm/amdgpu: fix another missing fence reference in the CS codeChristian König1-1/+4
drm_sched_job_add_dependency() consumes the references of the gang members. Only triggered by mesh shaders. Signed-off-by: Christian König <[email protected]> Fixes: 1728baa7e4e6 ("drm/amdgpu: use scheduler dependencies for CS") Tested-by: Mike Lothian <[email protected]> Tested-by: Bert Karwatzki <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/[email protected] Reviewed-by: Alex Deucher <[email protected]> Reviewed-by: Luben Tuikov <[email protected]>
2023-01-05drm/amdgpu: update ta_secureDisplay_if.h to v27.00.00.08Aaron Liu4-18/+18
1. Rename securedisplay_cmd to ta_securedisplay_cmd. 2. Rename ta_securedisplay_max_phy to ta_securedisplay_phy_ID. 3. update securedisplay_cmd to ta_securedisplay_cmd Signed-off-by: Aaron Liu <[email protected]> Signed-off-by: Shane Xiao <[email protected]> Reviewed-by: Alan Liu <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-05drm/amdgpu: allow query error counters for specific IP blockHawking Zhang2-21/+71
amdgpu_ras_block_late_init will be invoked in IP specific ras_late_init call as a common helper for all the IP blocks. However, when amdgpu_ras_block_late_init call amdgpu_ras_query_error_count to query ras error counters, amdgpu_ras_query_error_count queries all the IP blocks that support ras query interface. This results to wrong error counters cached in software copies when there are ras errors detected at time zero or warm reset procedure. i.e., in sdma_ras_late_init phase, it counts on sdma/mmhub errors, while, in mmhub_ras_late_init phase, it still counts on sdma/mmhub errors. The change updates amdgpu_ras_query_error_count interface to allow query specific ip error counter. It introduces a new input parameter: query_info. if query_info is NULL, it means query all the IP blocks, otherwise, only query the ip block specified by query_info. Signed-off-by: Hawking Zhang <[email protected]> Reviewed-by: Tao Zhou <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-05drm/amdgpu: Retry DDC probing on DVI on failure if we got an HPD interruptxurui8-15/+36
HPD signals on DVI ports can be fired off before the pins required for DDC probing actually make contact, due to the pins for HPD making contact first. This results in a HPD signal being asserted but DDC probing failing, resulting in hotplugging occasionally failing. Rescheduling the hotplug work for a second when we run into an HPD signal with a failing DDC probe usually gives enough time for the rest of the connector's pins to make contact, and fixes this issue. Signed-off-by: xurui <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-04Revert "drm/amd/display: Enable Freesync Video Mode by default"Michel Dänzer2-0/+28
This reverts commit de05abe6b9d0fe08f65d744f7f75a4cba4df27ad. The bug referenced below was bisected to this commit. There has been no activity toward fixing it in 3 months, so let's revert for now. Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2162 Signed-off-by: Michel Dänzer <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected]
2023-01-04Merge tag 'drm-misc-next-2023-01-03' of ↵Daniel Vetter14-15/+18
git://anongit.freedesktop.org/drm/drm-misc into drm-next drm-misc-next for v6.3: UAPI Changes: * connector: Support analog-TV mode property * media: Add MEDIA_BUS_FMT_RGB565_1X24_CPADHI, MEDIA_BUS_FMT_RGB666_1X18 and MEDIA_BUS_FMT_RGB666_1X24_CPADHI Cross-subsystem Changes: * dma-buf: Documentation fixes * i2c: Introduce i2c_client_get_device_id() helper Core Changes: * Improve support for analog TV output * bridge: Remove unused drm_bridge_chain functions * debugfs: Add per-device helpers and convert various DRM drivers * dp-mst: Various fixes * fbdev emulation: Always pick 32 bpp as default * KUnit: Add tests for managed helpers; Various cleanups * panel-orientation: Add quirks for Lenovo Yoga Tab 3 X90F and DynaBook K50 * TTM: Open-code ttm_bo_wait() and remove the helper Driver Changes: * Fix preferred depth and bpp values throughout DRM drivers * Remove #CONFIG_PM guards throughout DRM drivers * ast: Various fixes * bridge: Implement i2c's probe_new in various drivers; Fixes; ite-it6505: Locking fixes, Cache EDID data; ite-it66121: Support IT6610 chip, Cleanups; lontium-tl9611: Fix HDMI on DragonBoard 845c; parade-ps8640: Use atomic bridge functions * gud: Convert to DRM shadow-plane helpers; Perform flushing synchronously during atomic update * ili9486: Support 16-bit pixel data * imx: Split off IPUv3 driver; Various fixes * mipi-dbi: Convert to DRM shadow-plane helpers plus rsp driver changes;i Support separate I/O-voltage supply * mxsfb: Depend on ARCH_MXS or ARCH_MXC * omapdrm: Various fixes * panel: Use ktime_get_boottime() to measure power-down delay in various drivers; Fix auto-suspend delay in various drivers; orisetech-ota5601a: Add support * sprd: Cleanups * sun4i: Convert to new TV-mode property * tidss: Various fixes * v3d: Various fixes * vc4: Convert to new TV-mode property; Support Kunit tests; Cleanups; dpi: Support RGB565 and RGB666 formats; dsi: Convert DSI driver to bridge * virtio: Improve tracing * vkms: Support small cursors in IGT tests; Various fixes Signed-off-by: Daniel Vetter <[email protected]> From: Thomas Zimmermann <[email protected]> Link: https://patchwork.freedesktop.org/patch/msgid/Y7QIwlfElAYWxRcR@linux-uq9g
2023-01-03Revert "drm/amd/display: Enable Freesync Video Mode by default"Michel Dänzer2-0/+28
This reverts commit de05abe6b9d0fe08f65d744f7f75a4cba4df27ad. The bug referenced below was bisected to this commit. There has been no activity toward fixing it in 3 months, so let's revert for now. Bug: https://gitlab.freedesktop.org/drm/amd/-/issues/2162 Signed-off-by: Michel Dänzer <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: remove enable ras cmd call traceStanley.Yang1-3/+13
[Why] [ 41.285804] RIP: 0010:amdgpu_ras_feature_enable+0x15c/0x310 [amdgpu] [ 41.285945] Code: 48 89 c1 48 c7 c2 b9 f2 88 c1 48 c7 c0 c0 f2 88 c1 49 8b 3c 24 48 0f 44 d0 48 c7 c6 98 33 80 c1 e8 5f 52 75 d9 e9 fa fe ff ff <0f> 0b e9 66 ff ff ff 48 8b 3d 86 8c 0f da ba 00 04 00 00 be c0 0d [ 41.285946] RSP: 0018:ffffbccdc72efc90 EFLAGS: 00010246 [ 41.285948] RAX: 0000000000000004 RBX: ffff931897406980 RCX: 0000000000000002 [ 41.285949] RDX: 0000000000000dc0 RSI: 0000000000000002 RDI: ffff931500042b00 [ 41.285950] RBP: ffffbccdc72efcc0 R08: 0000000000000002 R09: ffff931885b87000 [ 41.285951] R10: 0000000000ffff10 R11: 0000000000000001 R12: ffff931893e20000 [ 41.285952] R13: 0000000000000001 R14: ffff931885b87000 R15: 0000000000000000 [ 41.285953] FS: 0000000000000000(0000) GS:ffff931c6f200000(0000) knlGS:0000000000000000 [ 41.285954] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 41.285955] CR2: 000055dd6f532008 CR3: 000000061b010006 CR4: 00000000003706e0 [ 41.285956] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 [ 41.285957] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 [ 41.285958] Call Trace: [ 41.285959] <TASK> [ 41.285963] ? gfx_v11_0_early_init+0x250/0x250 [amdgpu] [ 41.286117] gfx_v11_0_late_init+0x8c/0xb0 [amdgpu] [ 41.286271] amdgpu_device_ip_late_init+0x8d/0x3c0 [amdgpu] [ 41.286401] amdgpu_device_init.cold+0x1677/0x1fda [amdgpu] [ 41.286616] ? pci_bus_read_config_word+0x4a/0x70 [ 41.286621] ? do_pci_enable_device+0xdb/0x110 [ 41.286625] amdgpu_driver_load_kms+0x1a/0x160 [amdgpu] [ 41.286762] amdgpu_pci_probe+0x18d/0x3a0 [amdgpu] [ 41.286898] local_pci_probe+0x4b/0x90 [ 41.286901] work_for_cpu_fn+0x1a/0x30 [ 41.286903] process_one_work+0x22b/0x3d0 [ 41.286905] worker_thread+0x223/0x420 [ 41.286907] ? process_one_work+0x3d0/0x3d0 [ 41.286908] kthread+0x12a/0x150 [ 41.286911] ? set_kthread_struct+0x50/0x50 [ 41.286913] ret_from_fork+0x22/0x30 [How] For specific asic, only mem ecc is enabled, sram ecc is not enabled, but it still need to send ras enable cmd to gfx block to support poison mode, so add check posion mode. Signed-off-by: Stanley.Yang <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: correct umc poison mode set valueStanley.Yang1-20/+4
For GFX 11.0.3, Due to security policy, there is no way to check UcFatalEn field of UMCCH0_0_GeccCtrl to identify UMC poison mode. This is workaround force set umc poison mode as 1 for GFX 11.0.3 Signed-off-by: Stanley.Yang <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: allow zero as vram limitChristian König3-3/+3
This allows testing the driver without any VRAM. Signed-off-by: Christian König <[email protected]> Acked-by: Luben Tuikov <[email protected]> Acked-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: cleanup visible vram size handlingChristian König6-19/+7
Centralize the limit handling and validation in one place instead of spreading that around in different hw generations. Signed-off-by: Christian König <[email protected]> Acked-by: Luben Tuikov <[email protected]> Acked-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: rename vram_scratch into mem_scratchChristian König18-32/+32
Rename vram_scratch into mem_scratch and allow allocating it into GTT as well. The only problem with that is that we won't have a default page for the system aperture any more. Signed-off-by: Christian König <[email protected]> Signed-off-by: Luben Tuikov <[email protected]> Acked-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: use VRAM|GTT for a bunch of kernel allocationsChristian König16-49/+101
Technically all of those can use GTT as well, no need to force things into VRAM. Signed-off-by: Christian König <[email protected]> Signed-off-by: Luben Tuikov <[email protected]> Acked-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: enable VCN DPG for GC IP v11.0.4Saleemkhan Jamadar1-0/+1
Enable VCN Dynamic Power Gating control for GC IP v11.0.4. Signed-off-by: Saleemkhan Jamadar <[email protected]> Reviewed-by: Veerabadhran Gopalakrishnan <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2023-01-03drm/amdgpu: adjust the sequence to check soft resetLikun Gao1-6/+2
1.Drop soft reset check when do should recover gpu check. (As it will skip gpu reset operation if some ip is hang but not support soft reset) 2.Check soft reset status before do soft reset when pre asic reset. a. If check soft reset return true, it means: some ip is hang and it also support soft reset, will try soft reset first. b. If check soft reset return false, it means: I. All the ip are not hang, will skip gpu reset. II. Some ip is hang but not support soft reset, will skip soft reset and retry with full reset later. Signed-off-by: Likun Gao <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-23Merge tag 'drm-next-2022-12-23' of git://anongit.freedesktop.org/drm/drmLinus Torvalds22-234/+355
Pull drm fixes from Dave Airlie: "Holiday fixes! Two batches from amd, and one group of i915 changes. amdgpu: - Spelling fix - BO pin fix - Properly handle polaris 10/11 overlap asics - GMC9 fix - SR-IOV suspend fix - DCN 3.1.4 fix - KFD userptr locking fix - SMU13.x fixes - GDS/GWS/OA handling fix - Reserved VMID handling fixes - FRU EEPROM fix - BO validation fixes - Avoid large variable on the stack - S0ix fixes - SMU 13.x fixes - VCN fix - Add missing fence reference amdkfd: - Fix init vm error handling - Fix double release of compute pasid i915 - Documentation fixes - OA-perf related fix - VLV/CHV HDMI/DP audio fix - Display DDI/Transcoder fix - Migrate fixes" * tag 'drm-next-2022-12-23' of git://anongit.freedesktop.org/drm/drm: (39 commits) drm/amdgpu: grab extra fence reference for drm_sched_job_add_dependency drm/amdgpu: enable VCN DPG for GC IP v11.0.4 drm/amdgpu: skip mes self test after s0i3 resume for MES IP v11.0 drm/amd/pm: correct the fan speed retrieving in PWM for some SMU13 asics drm/amd/pm: bump SMU13.0.0 driver_if header to version 0x34 drm/amdgpu: skip MES for S0ix as well since it's part of GFX drm/amd/pm: avoid large variable on kernel stack drm/amdkfd: Fix double release compute pasid drm/amdkfd: Fix kfd_process_device_init_vm error handling drm/amd/pm: update SMU13.0.0 reported maximum shader clock drm/amd/pm: correct SMU13.0.0 pstate profiling clock settings drm/amd/pm: enable GPO dynamic control support for SMU13.0.7 drm/amd/pm: enable GPO dynamic control support for SMU13.0.0 drm/amdgpu: revert "generally allow over-commit during BO allocation" drm/amdgpu: Remove unnecessary domain argument drm/amdgpu: Fix size validation for non-exclusive domains (v4) drm/amdgpu: Check if fru_addr is not NULL (v2) drm/i915/ttm: consider CCS for backup objects drm/i915/migrate: fix corner case in CCS aux copying drm/amdgpu: rework reserved VMID handling ...
2022-12-21drm/amdgpu: grab extra fence reference for drm_sched_job_add_dependencyChristian König1-0/+2
That function consumes the reference. Reviewed-by: Luben Tuikov <[email protected]> Reported-by: Borislav Petkov (AMD) <[email protected]> Tested-by: Borislav Petkov (AMD) <[email protected]> Signed-off-by: Christian König <[email protected]> Fixes: aab9cf7b6954 ("drm/amdgpu: use scheduler dependencies for VM updates") Signed-off-by: Alex Deucher <[email protected]>
2022-12-21drm/amdgpu: enable VCN DPG for GC IP v11.0.4Saleemkhan Jamadar1-0/+1
Enable VCN Dynamic Power Gating control for GC IP v11.0.4. Signed-off-by: Saleemkhan Jamadar <[email protected]> Reviewed-by: Veerabadhran Gopalakrishnan <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected] # 6.0, 6.1
2022-12-20drm/amdgpu: skip mes self test after s0i3 resume for MES IP v11.0Tim Huang1-1/+2
MES is part of gfxoff and MES suspend and resume are skipped for S0i3. But the mes_self_test call path is still in the amdgpu_device_ip_late_init. it's should also be skipped for s0ix as no hardware re-initialization happened. Besides, mes_self_test will free the BO that triggers a lot of warning messages while in the suspend state. [ 81.656085] WARNING: CPU: 2 PID: 1550 at drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:425 amdgpu_bo_free_kernel+0xfc/0x110 [amdgpu] [ 81.679435] Call Trace: [ 81.679726] <TASK> [ 81.679981] amdgpu_mes_remove_hw_queue+0x17a/0x230 [amdgpu] [ 81.680857] amdgpu_mes_self_test+0x390/0x430 [amdgpu] [ 81.681665] mes_v11_0_late_init+0x37/0x50 [amdgpu] [ 81.682423] amdgpu_device_ip_late_init+0x53/0x280 [amdgpu] [ 81.683257] amdgpu_device_resume+0xae/0x2a0 [amdgpu] [ 81.684043] amdgpu_pmops_resume+0x37/0x70 [amdgpu] [ 81.684818] pci_pm_resume+0x5c/0xa0 [ 81.685247] ? pci_pm_thaw+0x90/0x90 [ 81.685658] dpm_run_callback+0x4e/0x160 [ 81.686110] device_resume+0xad/0x210 [ 81.686529] async_resume+0x1e/0x40 [ 81.686931] async_run_entry_fn+0x33/0x120 [ 81.687405] process_one_work+0x21d/0x3f0 [ 81.687869] worker_thread+0x4a/0x3c0 [ 81.688293] ? process_one_work+0x3f0/0x3f0 [ 81.688777] kthread+0xff/0x130 [ 81.689157] ? kthread_complete_and_exit+0x20/0x20 [ 81.689707] ret_from_fork+0x22/0x30 [ 81.690118] </TASK> [ 81.690380] ---[ end trace 0000000000000000 ]--- v2: make the comment clean and use adev->in_s0ix instead of adev->suspend Signed-off-by: Tim Huang <[email protected]> Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]> Cc: [email protected] # 6.0, 6.1
2022-12-20drm/amdgpu: skip MES for S0ix as well since it's part of GFXAlex Deucher1-2/+3
It's also part of gfxoff. Cc: [email protected] # 6.0, 6.1 Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdkfd: Fix double release compute pasidPhilip Yang2-12/+31
If kfd_process_device_init_vm returns failure after vm is converted to compute vm and vm->pasid set to compute pasid, KFD will not take pdd->drm_file reference. As a result, drm close file handler maybe called to release the compute pasid before KFD process destroy worker to release the same pasid and set vm->pasid to zero, this generates below WARNING backtrace and NULL pointer access. Add helper amdgpu_amdkfd_gpuvm_set_vm_pasid and call it at the last step of kfd_process_device_init_vm, to ensure vm pasid is the original pasid if acquiring vm failed or is the compute pasid with pdd->drm_file reference taken to avoid double release same pasid. amdgpu: Failed to create process VM object ida_free called for id=32770 which is not allocated. WARNING: CPU: 57 PID: 72542 at ../lib/idr.c:522 ida_free+0x96/0x140 RIP: 0010:ida_free+0x96/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 BUG: kernel NULL pointer dereference, address: 0000000000000000 RIP: 0010:ida_free+0x76/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 Signed-off-by: Philip Yang <[email protected]> Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu: skip mes self test after s0i3 resume for MES IP v11.0Tim Huang1-1/+2
MES is part of gfxoff and MES suspend and resume are skipped for S0i3. But the mes_self_test call path is still in the amdgpu_device_ip_late_init. it's should also be skipped for s0ix as no hardware re-initialization happened. Besides, mes_self_test will free the BO that triggers a lot of warning messages while in the suspend state. [ 81.656085] WARNING: CPU: 2 PID: 1550 at drivers/gpu/drm/amd/amdgpu/amdgpu_object.c:425 amdgpu_bo_free_kernel+0xfc/0x110 [amdgpu] [ 81.679435] Call Trace: [ 81.679726] <TASK> [ 81.679981] amdgpu_mes_remove_hw_queue+0x17a/0x230 [amdgpu] [ 81.680857] amdgpu_mes_self_test+0x390/0x430 [amdgpu] [ 81.681665] mes_v11_0_late_init+0x37/0x50 [amdgpu] [ 81.682423] amdgpu_device_ip_late_init+0x53/0x280 [amdgpu] [ 81.683257] amdgpu_device_resume+0xae/0x2a0 [amdgpu] [ 81.684043] amdgpu_pmops_resume+0x37/0x70 [amdgpu] [ 81.684818] pci_pm_resume+0x5c/0xa0 [ 81.685247] ? pci_pm_thaw+0x90/0x90 [ 81.685658] dpm_run_callback+0x4e/0x160 [ 81.686110] device_resume+0xad/0x210 [ 81.686529] async_resume+0x1e/0x40 [ 81.686931] async_run_entry_fn+0x33/0x120 [ 81.687405] process_one_work+0x21d/0x3f0 [ 81.687869] worker_thread+0x4a/0x3c0 [ 81.688293] ? process_one_work+0x3f0/0x3f0 [ 81.688777] kthread+0xff/0x130 [ 81.689157] ? kthread_complete_and_exit+0x20/0x20 [ 81.689707] ret_from_fork+0x22/0x30 [ 81.690118] </TASK> [ 81.690380] ---[ end trace 0000000000000000 ]--- v2: make the comment clean and use adev->in_s0ix instead of adev->suspend Signed-off-by: Tim Huang <[email protected]> Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu: skip MES for S0ix as well since it's part of GFXAlex Deucher1-2/+3
It's also part of gfxoff. Cc: [email protected] # 6.0, 6.1 Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20Revert "drm/amdgpu: force exit gfxoff on sdma resume for rmb s0ix"Alex Deucher1-8/+0
This reverts commit e5d59cfa330523e47cba62a496864acc3948fc27. This is no longer needed since we no longer suspend SDMA during S0ix. Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20Revert "drm/amdgpu: disallow gfxoff until GC IP blocks complete s2idle resume"Alex Deucher1-16/+0
This reverts commit f543d28687480fad06b708bc6e0b0b6ec953b078. This is no longer needed since we no longer touch SDMA 5.x for s0i3. Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu: for S0ix, skip SDMA 5.x+ suspend/resumeAlex Deucher1-0/+6
SDMA 5.x is part of the GFX block so it's controlled via GFXOFF. Skip suspend as it should be handled the same as GFX. v2: drop SDMA 4.x. That requires special handling. Reviewed-by: Mario Limonciello <[email protected]> Acked-by: Rajneesh Bhardwaj <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu: don't mess with SDMA clock or powergating in S0ixAlex Deucher1-4/+6
It's handled by GFXOFF for SDMA 5.x and SMU saves the state on SDMA 4.x. Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu/gmc11: don't touch gfxhub registers during S0ixAlex Deucher1-2/+14
gfxhub registers are part of gfx IP and should not need to be changed. Doing so without disabling gfxoff can hang the gfx IP. v2: add comments explaining why we can skip the interrupt control for S0i3 Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu/gmc10: don't touch gfxhub registers during S0ixAlex Deucher1-9/+27
gfxhub registers are part of gfx IP and should not need to be changed. Doing so without disabling gfxoff can hang the gfx IP. v2: add comments explaining why we can skip the interrupt control for S0i3 Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdgpu/gmc9: don't touch gfxhub registers during S0ixAlex Deucher1-6/+30
gfxhub registers are part of gfx IP and should not need to be changed. Doing so without disabling gfxoff can hang the gfx IP. v2: add comments explaining why we can skip the interrupt control for S0i3 Reviewed-by: Mario Limonciello <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-20drm/amdkfd: Fix double release compute pasidPhilip Yang2-12/+31
If kfd_process_device_init_vm returns failure after vm is converted to compute vm and vm->pasid set to compute pasid, KFD will not take pdd->drm_file reference. As a result, drm close file handler maybe called to release the compute pasid before KFD process destroy worker to release the same pasid and set vm->pasid to zero, this generates below WARNING backtrace and NULL pointer access. Add helper amdgpu_amdkfd_gpuvm_set_vm_pasid and call it at the last step of kfd_process_device_init_vm, to ensure vm pasid is the original pasid if acquiring vm failed or is the compute pasid with pdd->drm_file reference taken to avoid double release same pasid. amdgpu: Failed to create process VM object ida_free called for id=32770 which is not allocated. WARNING: CPU: 57 PID: 72542 at ../lib/idr.c:522 ida_free+0x96/0x140 RIP: 0010:ida_free+0x96/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 BUG: kernel NULL pointer dereference, address: 0000000000000000 RIP: 0010:ida_free+0x76/0x140 Call Trace: amdgpu_pasid_free_delayed+0xe1/0x2a0 [amdgpu] amdgpu_driver_postclose_kms+0x2d8/0x340 [amdgpu] drm_file_free.part.13+0x216/0x270 [drm] drm_close_helper.isra.14+0x60/0x70 [drm] drm_release+0x6e/0xf0 [drm] __fput+0xcc/0x280 ____fput+0xe/0x20 task_work_run+0x96/0xc0 do_exit+0x3d0/0xc10 Signed-off-by: Philip Yang <[email protected]> Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: Add poison mode query for df v4_3Candice Li4-1/+98
Add poison mode query support on df v4_3. Signed-off-by: Candice Li <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: bump minor version number for DEV_INFO and SENSOR IOCTLs updateEvan Quan1-2/+4
Update AMDGPU_INFO_DEV_INFO IOCTL for minimum engine and memory clock. And update AMDGPU_INFO_SENSOR IOCTL for PEAK_PSTATE engine and memory clock. User applications can better utilize these IOCTLs to get needed informations. Increase the minor version number to indicate that the new flags are available. Proposed mesa patch: https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/278 Signed-off-by: Evan Quan <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: expose the minimum shader/memory clock frequencyEvan Quan1-2/+8
Otherwise, some UMD tools will treate them as 0 at default while actually they are not. Signed-off-by: Evan Quan <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: expose peak profiling mode shader/memory clocksEvan Quan1-0/+18
Expose those informations to UMD who need them as for standard profiling mode. Signed-off-by: Evan Quan <[email protected]> Reviewed-by: Alex Deucher <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: define RAS query poison mode functionTao Zhou1-21/+33
1. no need to query poison mode on SRIOV guest side, host can handle it. 2. define the function to simplify code. v2: rename amdgpu_ras_poison_mode_query to amdgpu_ras_query_poison_mode. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: update VCN/JPEG RAS settingTao Zhou1-11/+13
Support VCN/JPEG RAS in both bare metal and SRIOV environment. v2: update commit description. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: skip RAS error injection in SRIOVTao Zhou1-0/+4
Injection on guest is not allowed. v2: return directly in SRIOV environment. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: add VCN poison consumption handler for SRIOVTao Zhou1-2/+10
Inform host and let host handle consumption interrupt. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: add RAS poison consumption handler for SRIOVTao Zhou1-18/+26
Send message to PF if VF receives RAS poison consumption interrupt. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: add RAS poison consumption handler for NV SRIOVTao Zhou2-0/+7
Send handling request to host. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-15drm/amdgpu: add RAS poison consumption handler for AI SRIOVTao Zhou3-0/+8
Send message to host and host will handle it. v2: split the patch into two parts, one is for mxgpu ai and another one is for common poison consumption handler. Signed-off-by: Tao Zhou <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-14drm/amdgpu: revert "generally allow over-commit during BO allocation"Christian König2-4/+18
This reverts commit f9d00a4a8dc8fff951c97b3213f90d6bc7a72175. This causes problem for KFD because when we overcommit we accidentially bind the BO to GTT for moving it into VRAM. We also need to make sure that this is done only as fallback after trying to evict first. Signed-off-by: Christian König <[email protected]> Reviewed-by: Felix Kuehling <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
2022-12-14drm/amdgpu: Remove unnecessary domain argumentLuben Tuikov4-14/+6
Remove the "domain" argument to amdgpu_bo_create_kernel_at() since this function takes an "offset" argument which is the offset off of VRAM, and as such allocation always takes place in VRAM. Thus, the "domain" argument is unnecessary. Cc: Alex Deucher <[email protected]> Cc: Christian König <[email protected]> Cc: AMD Graphics <[email protected]> Signed-off-by: Luben Tuikov <[email protected]> Reviewed-by: Christian König <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.