summaryrefslogtreecommitdiff
path: root/drivers/gpu/drm/amd/amdgpu
AgeCommit message (Collapse)AuthorFilesLines
14 daysdrm/amd: Fix unbind/rebind for VCN 4.0.5Mario Limonciello (AMD)1-0/+2
commit 93a01629c8bfd30906c76921ec986802d76920c6 upstream. Unbinding amdgpu has no problems, but binding it again leads to an error of sysfs file already existing. This is because it wasn't actually cleaned up on unbind. Add the missing cleanup step. Fixes: 547aad32edac ("drm/amdgpu: add VCN4 ip block support") Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit d717e62e9b6ccff0e3cec78a58dfbd00858448b3) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu/gmc11: add amdgpu_vm_handle_fault() handlingAlex Deucher1-0/+27
commit 3f2289b56cd98f5741056bdb6e521324eff07ce5 upstream. We need to call amdgpu_vm_handle_fault() on page fault on all gfx9 and newer parts to properly update the page tables, not just for recoverable page faults. Cc: stable@vger.kernel.org Reviewed-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu/sdma6: Update SDMA 6.0.3 FW version to include UMQ ↵Srinivasan Shanmugam1-1/+1
protected-fence fix commit c8e7e3c2215e286ebfe66fe828ed426546c519e6 upstream. On GFX11.0.3, earlier SDMA firmware versions issue the PROTECTED_FENCE write from the user VMID (e.g. VMID 8) instead of VMID 0. This causes a GPU VM protection fault when SDMA tries to write the secure fence location, as seen in the UMQ SDMA test (cs-sdma-with-IP-DMA-UMQ) Fixes the below GPU page fault: [ 514.037189] amdgpu 0000:0b:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:40 vmid:8 pasid:32770) [ 514.037199] amdgpu 0000:0b:00.0: amdgpu: Process pid 0 thread pid 0 [ 514.037205] amdgpu 0000:0b:00.0: amdgpu: in page starting at address 0x00007fff00409000 from client 10 [ 514.037212] amdgpu 0000:0b:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00841A51 [ 514.037217] amdgpu 0000:0b:00.0: amdgpu: Faulty UTCL2 client ID: SDMA0 (0xd) [ 514.037223] amdgpu 0000:0b:00.0: amdgpu: MORE_FAULTS: 0x1 [ 514.037227] amdgpu 0000:0b:00.0: amdgpu: WALKER_ERROR: 0x0 [ 514.037232] amdgpu 0000:0b:00.0: amdgpu: PERMISSION_FAULTS: 0x5 [ 514.037236] amdgpu 0000:0b:00.0: amdgpu: MAPPING_ERROR: 0x0 [ 514.037241] amdgpu 0000:0b:00.0: amdgpu: RW: 0x1 v2: Updated commit message v3: s/gfx11.0.3/sdma 6.0.3/ in patch title (Alex) Cc: Alex Deucher <alexander.deucher@amd.com> Cc: Christian König <christian.koenig@amd.com> Cc: stable@vger.kernel.org Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu: add missing lock to amdgpu_ttm_access_memory_sdmaPierre-Eric Pelloux-Prayer1-0/+2
commit 4fa944255be521b1bbd9780383f77206303a3a5c upstream. Users of ttm entities need to hold the gtt_window_lock before using them to guarantee proper ordering of jobs. Cc: stable@vger.kernel.org Fixes: cb5cc4f573e1 ("drm/amdgpu: improve debug VRAM access performance using sdma") Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu: Forward VMID reservation errorsNatalie Vock1-2/+1
commit 8defb4f081a5feccc3ea8372d0c7af3522124e1f upstream. Otherwise userspace may be fooled into believing it has a reserved VMID when in reality it doesn't, ultimately leading to GPU hangs when SPM is used. Fixes: 80e709ee6ecc ("drm/amdgpu: add option params to enforce process isolation between graphics and compute") Cc: stable@vger.kernel.org Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Natalie Vock <natalie.vock@gmx.de> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu/gmc12: add amdgpu_vm_handle_fault() handlingAlex Deucher1-0/+27
commit ff28ff98db6a8eeb469e02fb8bd1647b353232a9 upstream. We need to call amdgpu_vm_handle_fault() on page fault on all gfx9 and newer parts to properly update the page tables, not just for recoverable page faults. Cc: stable@vger.kernel.org Reviewed-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysRevert "drm/amd: Skip power ungate during suspend for VPE"Mario Limonciello (AMD)1-2/+1
commit 3925683515e93844be204381d2d5a1df5de34f31 upstream. Skipping power ungate exposed some scenarios that will fail like below: ``` amdgpu: Register(0) [regVPEC_QUEUE_RESET_REQ] failed to reach value 0x00000000 != 0x00000001n amdgpu 0000:c1:00.0: amdgpu: VPE queue reset failed ... amdgpu: [drm] *ERROR* wait_for_completion_timeout timeout! ``` The underlying s2idle issue that prompted this commit is going to be fixed in BIOS. This reverts commit 2a6c826cfeedd7714611ac115371a959ead55bda. Fixes: 2a6c826cfeed ("drm/amd: Skip power ungate during suspend for VPE") Cc: stable@vger.kernel.org Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reported-by: Konstantin <answer2019@yandex.ru> Closes: https://bugzilla.kernel.org/show_bug.cgi?id=220812 Reported-by: Matthew Schwartz <matthew.schwartz@linux.dev> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
14 daysdrm/amdgpu: don't attach the tlb fence for SIAlex Deucher1-1/+3
commit eb296c09805ee37dd4ea520a7fb3ec157c31090f upstream. SI hardware doesn't support pasids, user mode queues, or KIQ/MES so there is no need for this. Doing so results in a segfault as these callbacks are non-existent for SI. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4744 Fixes: f3854e04b708 ("drm/amdgpu: attach tlb fence to the PTs update") Reviewed-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 820b3d376e8a102c6aeab737ec6edebbbb710e04) Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2026-01-02drm/amdgpu: fix a job->pasid access race in gpu recoveryAlex Deucher1-2/+8
[ Upstream commit 77f73253015cbc7893fca1821ac3eae9eb4bc943 ] Avoid a possible UAF in GPU recovery due to a race between the sched timeout callback and the tdr work queue. The gpu recovery function calls drm_sched_stop() and later drm_sched_start(). drm_sched_start() restarts the tdr queue which will eventually free the job. If the tdr queue frees the job before time out callback completes, the job will be freed and we'll get a UAF when accessing the pasid. Cache it early to avoid the UAF. Example KASAN trace: [ 493.058141] BUG: KASAN: slab-use-after-free in amdgpu_device_gpu_recover+0x968/0x990 [amdgpu] [ 493.067530] Read of size 4 at addr ffff88b0ce3f794c by task kworker/u128:1/323 [ 493.074892] [ 493.076485] CPU: 9 UID: 0 PID: 323 Comm: kworker/u128:1 Tainted: G E 6.16.0-1289896.2.zuul.bf4f11df81c1410bbe901c4373305a31 #1 PREEMPT(voluntary) [ 493.076493] Tainted: [E]=UNSIGNED_MODULE [ 493.076495] Hardware name: TYAN B8021G88V2HR-2T/S8021GM2NR-2T, BIOS V1.03.B10 04/01/2019 [ 493.076500] Workqueue: amdgpu-reset-dev drm_sched_job_timedout [gpu_sched] [ 493.076512] Call Trace: [ 493.076515] <TASK> [ 493.076518] dump_stack_lvl+0x64/0x80 [ 493.076529] print_report+0xce/0x630 [ 493.076536] ? _raw_spin_lock_irqsave+0x86/0xd0 [ 493.076541] ? __pfx__raw_spin_lock_irqsave+0x10/0x10 [ 493.076545] ? amdgpu_device_gpu_recover+0x968/0x990 [amdgpu] [ 493.077253] kasan_report+0xb8/0xf0 [ 493.077258] ? amdgpu_device_gpu_recover+0x968/0x990 [amdgpu] [ 493.077965] amdgpu_device_gpu_recover+0x968/0x990 [amdgpu] [ 493.078672] ? __pfx_amdgpu_device_gpu_recover+0x10/0x10 [amdgpu] [ 493.079378] ? amdgpu_coredump+0x1fd/0x4c0 [amdgpu] [ 493.080111] amdgpu_job_timedout+0x642/0x1400 [amdgpu] [ 493.080903] ? pick_task_fair+0x24e/0x330 [ 493.080910] ? __pfx_amdgpu_job_timedout+0x10/0x10 [amdgpu] [ 493.081702] ? _raw_spin_lock+0x75/0xc0 [ 493.081708] ? __pfx__raw_spin_lock+0x10/0x10 [ 493.081712] drm_sched_job_timedout+0x1b0/0x4b0 [gpu_sched] [ 493.081721] ? __pfx__raw_spin_lock_irq+0x10/0x10 [ 493.081725] process_one_work+0x679/0xff0 [ 493.081732] worker_thread+0x6ce/0xfd0 [ 493.081736] ? __pfx_worker_thread+0x10/0x10 [ 493.081739] kthread+0x376/0x730 [ 493.081744] ? __pfx_kthread+0x10/0x10 [ 493.081748] ? __pfx__raw_spin_lock_irq+0x10/0x10 [ 493.081751] ? __pfx_kthread+0x10/0x10 [ 493.081755] ret_from_fork+0x247/0x330 [ 493.081761] ? __pfx_kthread+0x10/0x10 [ 493.081764] ret_from_fork_asm+0x1a/0x30 [ 493.081771] </TASK> Fixes: a72002cb181f ("drm/amdgpu: Make use of drm_wedge_task_info") Link: https://github.com/HansKristian-Work/vkd3d-proton/pull/2670 Cc: SRINIVASAN.SHANMUGAM@amd.com Cc: vitaly.prosyak@amd.com Cc: christian.koenig@amd.com Suggested-by: Matthew Brost <matthew.brost@intel.com> Reviewed-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 20880a3fd5dd7bca1a079534cf6596bda92e107d) Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-12-18drm/amdgpu/userq: fix SDMA and compute validationAlex Deucher1-5/+14
[ Upstream commit a0559012a18a5a6ad87516e982892765a403b8ab ] The CSA and EOP buffers have different alignement requirements. Hardcode them for now as a bug fix. A proper query will be added in a subsequent patch. v2: verify gfx shadow helper callback (Prike) Fixes: 9e46b8bb0539 ("drm/amdgpu: validate userq buffer virtual address and size") Reviewed-by: Prike Liang <Prike.Liang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-12-18drm/amdgpu: add userq object va track helpersPrike Liang4-21/+52
[ Upstream commit 5cfa33fabf01f2cc0af6b1feed6e65cb81806a37 ] Add the userq object virtual address list_add() helpers for tracking the userq obj va address usage. Signed-off-by: Prike Liang <Prike.Liang@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Stable-dep-of: a0559012a18a ("drm/amdgpu/userq: fix SDMA and compute validation") Signed-off-by: Sasha Levin <sashal@kernel.org>
2025-11-26drm/amdgpu: fix cyan_skillfish2 gpu info fw handlingAlex Deucher1-0/+2
If the board supports IP discovery, we don't need to parse the gpu info firmware. Backport to 6.18. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4721 Fixes: fa819e3a7c1e ("drm/amdgpu: add support for cyan skillfish gpu_info") Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 5427e32fa3a0ba9a016db83877851ed277b065fb)
2025-11-26drm/amdgpu: attach tlb fence to the PTs updatePrike Liang1-1/+1
Ensure the userq TLB flush is emitted only after the VM update finishes and the PT BOs have been annotated with bookkeeping fences. Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Prike Liang <Prike.Liang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit f3854e04b708d73276c4488231a8bd66d30b4671) Cc: stable@vger.kernel.org
2025-11-24drm/amd/amdgpu: reserve vm invalidation engine for uni_mesMichael Chen1-0/+3
Reserve vm invalidation engine 6 when uni_mes enabled. It is used in processing tlb flush request from host. Signed-off-by: Michael Chen <michael.chen@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Shaoyun liu <Shaoyun.liu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 873373739b9b150720ea2c5390b4e904a4d21505) Cc: stable@vger.kernel.org
2025-11-20drm/amdgpu: Add sriov vf check for VCN per queue reset support.Shikang Fan2-2/+4
Add SRIOV check when setting VCN ring's supported reset mask. Signed-off-by: Shikang Fan <shikang.fan@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit ee9b603ad43f9870eb75184f9fb0a84f8c3cc852) Cc: stable@vger.kernel.org
2025-11-20drm/amdgpu/ttm: Fix crash when handling MMIO_REMAP in PDE flagsSrinivasan Shanmugam1-1/+1
The MMIO_REMAP BO is a special 4K IO page that does not have a ttm_tt behind it. However, amdgpu_ttm_tt_pde_flags() was treating it like normal TT/doorbell/preempt memory and unconditionally accessed ttm->caching. For the MMIO_REMAP BO, ttm is NULL, so this leads to a NULL pointer dereference when computing PDE flags. Fix this by checking that ttm is non-NULL before reading ttm->caching. This prevents the crash for MMIO_REMAP and also makes the code more defensive if other BOs ever come through without a ttm_tt. Fixes: fb5a52dbe9fe ("drm/amdgpu: Implement TTM handling for MMIO_REMAP placement") Suggested-by: Jesse Zhang <Jesse.Zhang@amd.com> Suggested-by: Christian König <christian.koenig@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Srinivasan Shanmugam <srinivasan.shanmugam@amd.com> Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Tested-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 0db94da5a0a1cacda080b9ec8425fcbe4babc141)
2025-11-20drm/amdgpu/vm: Check PRT uAPI flag instead of PTE flagTimur Kristóf1-2/+2
This fixes sparse mappings (aka. partially resident textures). Check the correct flags. Since a recent refactor, the code works with uAPI flags (for mapping buffer objects), and not PTE (page table entry) flags. Fixes: 6716a823d18d ("drm/amdgpu: rework how PTE flags are generated v3") Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 8feeab26c80635b802f72b3ed986c693ff8f3212)
2025-11-20drm/amdgpu: Skip emit de meta data on gfx11 with rs64 enabledYifan Zha1-2/+2
[Why] Accoreding to CP updated to RS64 on gfx11, WRITE_DATA with PREEMPTION_META_MEMORY(dst_sel=8) is illegal for CP FW. That packet is used for MCBP on F32 based system. So it would lead to incorrect GRBM write and FW is not handling that extra case correctly. [How] With gfx11 rs64 enabled, skip emit de meta data. Signed-off-by: Yifan Zha <Yifan.Zha@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 8366cd442d226463e673bed5d199df916f4ecbcf) Cc: stable@vger.kernel.org
2025-11-20drm/amd: Skip power ungate during suspend for VPEMario Limonciello1-1/+2
During the suspend sequence VPE is already going to be power gated as part of vpe_suspend(). It's unnecessary to call during calls to amdgpu_device_set_pg_state(). It actually can expose a race condition with the firmware if s0i3 sequence starts as well. Drop these calls. Cc: Peyton.Lee@amd.com Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 2a6c826cfeedd7714611ac115371a959ead55bda) Cc: stable@vger.kernel.org
2025-11-12drm/amdgpu/jpeg: Add parse_cs for JPEG5_0_1Sathishkumar S1-0/+1
enable parse_cs callback for JPEG5_0_1. Signed-off-by: Sathishkumar S <sathishkumar.sundararaju@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 547985579932c1de13f57f8bcf62cd9361b9d3d3) Cc: stable@vger.kernel.org
2025-11-12drm/amd/amdgpu: Ensure isp_kernel_buffer_alloc() creates a new BOSultan Alsawaf1-0/+2
When the BO pointer provided to amdgpu_bo_create_kernel() points to non-NULL, amdgpu_bo_create_kernel() takes it as a hint to pin that address rather than allocate a new BO. This functionality is never desired for allocating ISP buffers. A new BO should always be created when isp_kernel_buffer_alloc() is called, per the description for isp_kernel_buffer_alloc(). Ensure this by zeroing *bo right before the amdgpu_bo_create_kernel() call. Fixes: 55d42f616976 ("drm/amd/amdgpu: Add helper functions for isp buffers") Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org> Reviewed-by: Pratap Nirujogi <pratap.nirujogi@amd.com> Signed-off-by: Sultan Alsawaf <sultan@kerneltoast.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 73c8c29baac7f0c7e703d92eba009008cbb5228e)
2025-11-12drm/amdgpu: fix lock warning in amdgpu_userq_fence_driver_processJesse.Zhang1-2/+3
Fix a potential deadlock caused by inconsistent spinlock usage between interrupt and process contexts in the userq fence driver. The issue occurs when amdgpu_userq_fence_driver_process() is called from both: - Interrupt context: gfx_v11_0_eop_irq() -> amdgpu_userq_fence_driver_process() - Process context: amdgpu_eviction_fence_suspend_worker() -> amdgpu_userq_fence_driver_force_completion() -> amdgpu_userq_fence_driver_process() In interrupt context, the spinlock was acquired without disabling interrupts, leaving it in {IN-HARDIRQ-W} state. When the same lock is acquired in process context, the kernel detects inconsistent locking since the process context acquisition would enable interrupts while holding a lock previously acquired in interrupt context. Kernel log shows: [ 4039.310790] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. [ 4039.310804] kworker/7:2/409 [HC0[0]:SC0[0]:HE1:SE1] takes: [ 4039.310818] ffff9284e1bed000 (&fence_drv->fence_list_lock){?...}-{3:3}, [ 4039.310993] {IN-HARDIRQ-W} state was registered at: [ 4039.311004] lock_acquire+0xc6/0x300 [ 4039.311018] _raw_spin_lock+0x39/0x80 [ 4039.311031] amdgpu_userq_fence_driver_process.part.0+0x30/0x180 [amdgpu] [ 4039.311146] amdgpu_userq_fence_driver_process+0x17/0x30 [amdgpu] [ 4039.311257] gfx_v11_0_eop_irq+0x132/0x170 [amdgpu] Fix by using spin_lock_irqsave()/spin_unlock_irqrestore() to properly manage interrupt state regardless of calling context. Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit ded3ad780cf97a04927773c4600823b84f7f3cc2) Cc: stable@vger.kernel.org
2025-11-12drm/amdgpu: jump to the correct label on failurePierre-Eric Pelloux-Prayer1-1/+1
drm_sched_entity_init wasn't called yet, so the only thing to do is to release allocated memory. This doesn't fix any bug since entity is zero allocated and drm_sched_entity_fini does nothing in this case. Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com> Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com> Acked-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit ec49374ccb8da86b465beaf09c367f3dfd648d8f)
2025-11-12drm/amdgpu: disable peer-to-peer access for DCC-enabled GC12 VRAM surfacesVitaly Prosyak1-0/+12
Certain multi-GPU configurations (especially GFX12) may hit data corruption when a DCC-compressed VRAM surface is shared across GPUs using peer-to-peer (P2P) DMA transfers. Such surfaces rely on device-local metadata and cannot be safely accessed through a remote GPU’s page tables. Attempting to import a DCC-enabled surface through P2P leads to incorrect rendering or GPU faults. This change disables P2P for DCC-enabled VRAM buffers that are contiguous and allocated on GFX12+ hardware. In these cases, the importer falls back to the standard system-memory path, avoiding invalid access to compressed surfaces. Future work could consider optional migration (VRAM→System→VRAM) if a performance regression is observed when `attach->peer2peer = false`. Tested on: - Dual RX 9700 XT (Navi4x) setup - GNOME and Wayland compositor scenarios - Confirmed no corruption after disabling P2P under these conditions v2: Remove check TTM_PL_VRAM & TTM_PL_FLAG_CONTIGUOUS. v3: simplify for upsteam and fix ip version check (Alex) Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Vitaly Prosyak <vitaly.prosyak@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 9dff2bb709e6fbd97e263fd12bf12802d2b5a0cf) Cc: stable@vger.kernel.org
2025-11-06drm/amdgpu: Fix wait after reset sequence in S3Lijo Lazar2-3/+32
For a mode-1 reset done at the end of S3 on PSPv11 dGPUs, only check if TOS is unloaded. Fixes: 32f73741d6ee ("drm/amdgpu: Wait for bootloader after PSPv11 reset") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4649 Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 1ad25fd272753db14c5d1cc8c68e20ce01f3f888)
2025-11-06drm/amd: Fix suspend failure with secure display TAMario Limonciello1-1/+4
commit c760bcda83571 ("drm/amd: Check whether secure display TA loaded successfully") attempted to fix extra messages, but failed to port the cleanup that was in commit 5c6d52ff4b61e ("drm/amd: Don't try to enable secure display TA multiple times") to prevent multiple tries. Add that to the failure handling path even on a quick failure. Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4679 Fixes: c760bcda8357 ("drm/amd: Check whether secure display TA loaded successfully") Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 4104c0a454f6a4d1e0d14895d03c0e7bdd0c8240)
2025-11-06drm/amdgpu: fix gpu page fault after hibernation on PF passthroughSamuel Zhang2-2/+5
On PF passthrough environment, after hibernate and then resume, coralgemm will cause gpu page fault. Mode1 reset happens during hibernate, but partition mode is not restored on resume, register mmCP_HYP_XCP_CTL and mmCP_PSP_XCP_CTL is not right after resume. When CP access the MQD BO, wrong stride size is used, this will cause out of bound access on the MQD BO, resulting page fault. The fix is to ensure gfx_v9_4_3_switch_compute_partition() is called when resume from a hibernation. KFD resume is called separately during a reset recovery or resume from suspend sequence. Hence it's not required to be called as part of partition switch. Signed-off-by: Samuel Zhang <guoqing.zhang@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 5d1b32cfe4a676fe552416cb5ae847b215463a1a)
2025-11-04drm/amdkfd: Don't clear PT after process killedPhilip Yang1-0/+4
If process is killed. the vm entity is stopped, submit pt update job will trigger the error message "*ERROR* Trying to push to a killed entity", job will not execute. Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Philip Yang <Philip.Yang@amd.com> Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 10c382ec6c6d1e11975a11962bec21cba6360391) Cc: stable@vger.kernel.org
2025-11-04drm/amdgpu: Drop PMFW RLC notifier from amdgpu_device_suspend()Alex Deucher1-4/+0
For S3 on vangogh, PMFW needs to be notified before the driver powers down RLC. This already happens in smu_disable_dpms() so drop the superfluous call in amdgpu_device_suspend(). Co-developed-by: Mario Limonciello (AMD) <superm1@kernel.org> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Mario Limonciello (AMD) <superm1@kernel.org> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 960e30a61e1a7ca5341a6cf9481e770e1cda24aa)
2025-11-04drm/amdgpu: set default gfx reset masks for gfx6-8Alex Deucher3-0/+15
These were not set so soft recovery was inadvertantly disabled. Fixes: 6ac55eab4fc4 ("drm/amdgpu: move reset support type checks into the caller") Reviewed-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 1972763505d728c604b537180727ec8132e619df)
2025-10-28drm/amdgpu: fix SPDX header on cyan_skillfish_reg_init.cAlex Deucher1-1/+1
This should be MIT. The driver in general is MIT and the license text at the top of the file is MIT so fix it. Fixes: e8529dbc75ca ("drm/amdgpu: add ip offset support for cyan skillfish") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654 Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 102c4f7c554ac5a5ecf0023fa0612beb58e3b0bd)
2025-10-28drm/amdgpu: fix SPDX headers on amdgpu_cper.c/hAlex Deucher2-2/+2
These should be MIT. The driver in general is MIT and the license text at the top of the files is MIT so fix it. Fixes: 92d5d2a09de1 ("drm/amdgpu: Introduce funcs for populating CPER") Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4654 Reviewed-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit abd3f876404cafb107cb34bacb74706bfee11cbe)
2025-10-28drm/amd: Check that VPE has reached DPM0 in idle handlerMario Limonciello1-4/+30
[Why] Newer VPE microcode has functionality that will decrease DPM level only when a workload has run for 2 or more seconds. If VPE is turned off before this DPM decrease and the PMFW doesn't reset it when power gating VPE, the SOC can get stuck with a higher DPM level. This can happen from amdgpu's ring buffer test because it's a short quick workload for VPE and VPE is turned off after 1s. [How] In idle handler besides checking fences are drained check PMFW version to determine if it will reset DPM when power gating VPE. If PMFW will not do this, then check VPE DPM level. If it is not DPM0 reschedule delayed work again until it is. v2: squash in return fix (Alex) Cc: Peyton.Lee@amd.com Reported-by: Sultan Alsawaf <sultan@kerneltoast.com> Reviewed-by: Sultan Alsawaf <sultan@kerneltoast.com> Tested-by: Sultan Alsawaf <sultan@kerneltoast.com> Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4615 Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Mario Limonciello <mario.limonciello@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> (cherry picked from commit 3ac635367eb589bee8edcc722f812a89970e14b7) Cc: stable@vger.kernel.org
2025-10-13drm/amdgpu: enable suspend/resume all for gfx 12Jonathan Kim1-7/+4
Suspend/resume all gangs has been available for GFX12 for a while now so enable it. Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: fix hung reset queue array memory allocationJonathan Kim5-10/+20
By design the MES will return an array result that is twice the number of hung doorbells it can report. i.e. if up k reported doorbells are supported, then the second half of the array, also of length k, holds the HQD information (type/queue/pipe) where queue 1 corresponds to index 0 and k, queue 2 corresponds to index 1 and k + 1 etc ... The driver will use the HDQ info to target queue/pipe reset for hardware scheduled user compute queues. Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: fix initialization of doorbell array for detect and hangJonathan Kim1-1/+1
Initialized doorbells should be set to invalid rather than 0 to prevent driver from over counting hung doorbells since it checks against the invalid value to begin with. Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: fix gfx12 mes packet status return checkJonathan Kim1-1/+6
GFX12 MES uses low 32 bits of status return for success (1 or 0) and high bits for debug information if low bits are 0. GFX11 MES doesn't do this so checking full 64-bit status return for 1 or 0 is still valid. Signed-off-by: Jonathan Kim <jonathan.kim@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com> Cc: stable@vger.kernel.org
2025-10-13drm/amdgpu: Fix NULL pointer dereference in VRAM logic for APU devicesJesse.Zhang3-6/+7
Previously, APU platforms (and other scenarios with uninitialized VRAM managers) triggered a NULL pointer dereference in `ttm_resource_manager_usage()`. The root cause is not that the `struct ttm_resource_manager *man` pointer itself is NULL, but that `man->bdev` (the backing device pointer within the manager) remains uninitialized (NULL) on APUs—since APUs lack dedicated VRAM and do not fully set up VRAM manager structures. When `ttm_resource_manager_usage()` attempts to acquire `man->bdev->lru_lock`, it dereferences the NULL `man->bdev`, leading to a kernel OOPS. 1. **amdgpu_cs.c**: Extend the existing bandwidth control check in `amdgpu_cs_get_threshold_for_moves()` to include a check for `ttm_resource_manager_used()`. If the manager is not used (uninitialized `bdev`), return 0 for migration thresholds immediately—skipping VRAM-specific logic that would trigger the NULL dereference. 2. **amdgpu_kms.c**: Update the `AMDGPU_INFO_VRAM_USAGE` ioctl and memory info reporting to use a conditional: if the manager is used, return the real VRAM usage; otherwise, return 0. This avoids accessing `man->bdev` when it is NULL. 3. **amdgpu_virt.c**: Modify the vf2pf (virtual function to physical function) data write path. Use `ttm_resource_manager_used()` to check validity: if the manager is usable, calculate `fb_usage` from VRAM usage; otherwise, set `fb_usage` to 0 (APUs have no discrete framebuffer to report). This approach is more robust than APU-specific checks because it: - Works for all scenarios where the VRAM manager is uninitialized (not just APUs), - Aligns with TTM's design by using its native helper function, - Preserves correct behavior for discrete GPUs (which have fully initialized `man->bdev` and pass the `ttm_resource_manager_used()` check). v4: use ttm_resource_manager_used(&adev->mman.vram_mgr.manager) instead of checking the adev->gmc.is_app_apu flag (Christian) Reviewed-by: Christian König <christian.koenig@amd.com> Suggested-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Jesse Zhang <Jesse.Zhang@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: hide VRAM sysfs attributes on GPUs without VRAMChristian König1-0/+3
Otherwise accessing them can cause a crash. Signed-off-by: Christian König <christian.koenig@amd.com> Tested-by: Mangesh Gadre <Mangesh.Gadre@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: fix bit shift logicSathishkumar S1-1/+1
BIT_ULL(n) sets nth bit, remove explicit shift and set the position Fixes: a7a411e24626 ("drm/amdgpu: fix shift-out-of-bounds in amdgpu_debugfs_jpeg_sched_mask_set") Signed-off-by: Sathishkumar S <sathishkumar.sundararaju@amd.com> Reviewed-by: Leo Liu <leo.liu@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: use atomic functions with memory barriers for vm fault infoGui-Dong Han3-11/+8
The atomic variable vm_fault_info_updated is used to synchronize access to adev->gmc.vm_fault_info between the interrupt handler and get_vm_fault_info(). The default atomic functions like atomic_set() and atomic_read() do not provide memory barriers. This allows for CPU instruction reordering, meaning the memory accesses to vm_fault_info and the vm_fault_info_updated flag are not guaranteed to occur in the intended order. This creates a race condition that can lead to inconsistent or stale data being used. The previous implementation, which used an explicit mb(), was incomplete and inefficient. It failed to account for all potential CPU reorderings, such as the access of vm_fault_info being reordered before the atomic_read of the flag. This approach is also more verbose and less performant than using the proper atomic functions with acquire/release semantics. Fix this by switching to atomic_set_release() and atomic_read_acquire(). These functions provide the necessary acquire and release semantics, which act as memory barriers to ensure the correct order of operations. It is also more efficient and idiomatic than using explicit full memory barriers. Fixes: b97dfa27ef3a ("drm/amdgpu: save vm fault information for amdkfd") Cc: stable@vger.kernel.org Signed-off-by: Gui-Dong Han <hanguidong02@gmail.com> Signed-off-by: Felix Kuehling <felix.kuehling@amd.com> Reviewed-by: Felix Kuehling <felix.kuehling@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: set an error on all fences from a bad contextAlex Deucher3-6/+37
When we backup ring contents to reemit after a queue reset, we don't backup ring contents from the bad context. When we signal the fences, we should set an error on those fences as well. v2: misc cleanups v3: add locking for fence error, fix comment (Christian) v4: fix wrap around, locking (Christian) Fixes: 77cc0da39c7c ("drm/amdgpu: track ring state associated with a fence") Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: handle wrap around in reemit handlingAlex Deucher1-5/+10
Compare the sequence numbers directly. Fixes: 77cc0da39c7c ("drm/amdgpu: track ring state associated with a fence") Reviewed-by: Christian König <christian.koenig@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: fix handling of harvesting for ip_discovery firmwareAlex Deucher1-1/+17
Chips which use the IP discovery firmware loaded by the driver reported incorrect harvesting information in the ip discovery table in sysfs because the driver only uses the ip discovery firmware for populating sysfs and not for direct parsing for the driver itself as such, the fields that are used to print the harvesting info in sysfs report incorrect data for some IPs. Populate the relevant fields for this case as well. Fixes: 514678da56da ("drm/amdgpu/discovery: fix fw based ip discovery") Acked-by: Tom St Denis <tom.stdenis@amd.com> Reviewed-by: Lijo Lazar <lijo.lazar@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: block CE CS if not explicitely allowed by module optionChristian König3-1/+14
The Constant Engine found on gfx6-gfx10 HW has been a notorious source of problems. RADV never used it in the first place, radeonsi only used it for a few releases around 2017 for gfx6-gfx9 before dropping support for it as well. While investigating another problem I just recently found that submitting to the CE seems to be completely broken on gfx9 for quite a while. Since nobody complained about that problem it most likely means that nobody is using any of the affected radeonsi versions on current Linux kernels any more. So to potentially phase out the support for the CE and eliminate another source of problems block submitting CE IBs unless it is enabled again using a debug flag. Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amdgpu: remove two invalid BUG_ON()sChristian König2-4/+0
Those can be triggered trivially by userspace. Signed-off-by: Christian König <christian.koenig@amd.com> Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Acked-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-13drm/amd: Disable ASPM on SITimur Kristóf1-0/+7
Enabling ASPM causes randoms hangs on Tahiti and Oland on Zen4. It's unclear if this is a platform-specific or GPU-specific issue. Disable ASPM on SI for the time being. Reviewed-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Timur Kristóf <timur.kristof@gmail.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-11Merge tag 'drm-next-2025-10-11-1' of https://gitlab.freedesktop.org/drm/kernelLinus Torvalds10-146/+173
Pull more drm fixes from Dave Airlie: "Just the follow up fixes for rc1 from the next branch, amdgpu and xe mostly with a single v3d fix in there. amdgpu: - DC DCE6 fixes - GPU reset fixes - Secure diplay messaging cleanup - MES fix - GPUVM locking fixes - PMFW messaging cleanup - PCI US/DS switch handling fix - VCN queue reset fix - DC FPU handling fix - DCN 3.5 fix - DC mirroring fix amdkfd: - Fix kfd process ref leak - mmap write lock handling fix - Fix comments in IOCTL xe: - Fix build with clang 16 - Fix handling of invalid configfs syntax usage and spell out the expected syntax in the documentation - Do not try late bind firmware when running as VF since it shouldn't handle firmware loading - Fix idle assertion for local BOs - Fix uninitialized variable for late binding - Do not require perfmon_capable to expose free memory at page granularity. Handle it like other drm drivers do - Fix lock handling on suspend error path - Fix I2C controller resume after S3 v3d: - fix fence locking" * tag 'drm-next-2025-10-11-1' of https://gitlab.freedesktop.org/drm/kernel: (34 commits) drm/amd/display: Incorrect Mirror Cositing drm/amd/display: Enable Dynamic DTBCLK Switch drm/amdgpu: Report individual reset error drm/amdgpu: partially revert "revert to old status lock handling v3" drm/amd/display: Fix unsafe uses of kernel mode FPU drm/amd/pm: Disable VCN queue reset on SMU v13.0.6 due to regression drm/amdgpu: Fix general protection fault in amdgpu_vm_bo_reset_state_machine drm/amdgpu: Check swus/ds for switch state save drm/amdkfd: Fix two comments in kfd_ioctl.h drm/amd/pm: Avoid interface mismatch messaging drm/amdgpu: Merge amdgpu_vm_set_pasid into amdgpu_vm_init drm/amd/amdgpu: Fix the mes version that support inv_tlbs drm/amd: Check whether secure display TA loaded successfully drm/amdkfd: Fix mmap write lock not release drm/amdkfd: Fix kfd process ref leaking when userptr unmapping drm/amdgpu: Fix for GPU reset being blocked by KIQ I/O. drm/amd/display: Disable scaling on DCE6 for now drm/amd/display: Properly disable scaling on DCE6 drm/amd/display: Properly clear SCL_*_FILTER_CONTROL on DCE6 drm/amd/display: Add missing DCE6 SCL_HORZ_FILTER_INIT* SRIs ...
2025-10-07drm/amdgpu: Report individual reset errorLijo Lazar1-10/+15
If reinitialization of one of the GPUs fails after reset, it logs failure on all subsequent GPUs eventhough they have resumed successfully. A sample log where only device at 0000:95:00.0 had a failure - amdgpu 0000:15:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:65:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:75:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:85:00.0: amdgpu: GPU reset(19) succeeded! amdgpu 0000:95:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:e5:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:f5:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:05:00.0: amdgpu: GPU reset(19) failed amdgpu 0000:15:00.0: amdgpu: GPU reset end with ret = -5 To avoid confusion, report the error for each device separately and return the first error as the overall result. Signed-off-by: Lijo Lazar <lijo.lazar@amd.com> Reviewed-by: Asad Kamal <asad.kamal@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
2025-10-07drm/amdgpu: partially revert "revert to old status lock handling v3"Christian König4-68/+105
The CI systems are pointing out list corruptions, so we still need to fix something here. Keep the asserts, but revert the lock changes for now. Fixes: 59e4405e9ee2 ("drm/amdgpu: revert to old status lock handling v3") Signed-off-by: Christian König <christian.koenig@amd.com> Acked-by: Alex Deucher <alexander.deucher@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
generated by cgit v1.2.3 (git 2.39.0) at 2026-01-22 02:36:08 +0300