kernel/linux.git/drivers/gpu/drm/amd/amdgpu, branch v7.0-rc7

drm/amdgpu: Fix wait after reset sequence in S4

2026-03-30T20:20:43+00:00

For a mode-1 reset done at the end of S4 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/-/work_items/4853 Signed-off-by: Lijo Lazar Reviewed-by: Alex Deucher Signed-off-by: Alex Deucher (cherry picked from commit 2fb4883b884a437d760bd7bdf7695a7e5a60bba3) Cc: stable@vger.kernel.org

drm/amdgpu: Change AMDGPU_VA_RESERVED_TRAP_SIZE to 64KB

2026-03-30T20:14:11+00:00

Currently, AMDGPU_VA_RESERVED_TRAP_SIZE is hardcoded to 8KB, while KFD_CWSR_TBA_TMA_SIZE is defined as 2 * PAGE_SIZE. On systems with 4K pages, both values match (8KB), so allocation and reserved space are consistent. However, on 64K page-size systems, KFD_CWSR_TBA_TMA_SIZE becomes 128KB, while the reserved trap area remains 8KB. This mismatch causes the kernel to crash when running rocminfo or rccl unit tests. Kernel attempted to read user page (2) - exploit attempt? (uid: 1001) BUG: Kernel NULL pointer dereference on read at 0x00000002 Faulting instruction address: 0xc0000000002c8a64 Oops: Kernel access of bad area, sig: 11 [#1] LE PAGE_SIZE=64K MMU=Radix SMP NR_CPUS=2048 NUMA pSeries CPU: 34 UID: 1001 PID: 9379 Comm: rocminfo Tainted: G E 6.19.0-rc4-amdgpu-00320-gf23176405700 #56 VOLUNTARY Tainted: [E]=UNSIGNED_MODULE Hardware name: IBM,9105-42A POWER10 (architected) 0x800200 0xf000006 of:IBM,FW1060.30 (ML1060_896) hv:phyp pSeries NIP: c0000000002c8a64 LR: c00000000125dbc8 CTR: c00000000125e730 REGS: c0000001e0957580 TRAP: 0300 Tainted: G E MSR: 8000000000009033 CR: 24008268 XER: 00000036 CFAR: c00000000125dbc4 DAR: 0000000000000002 DSISR: 40000000 IRQMASK: 1 GPR00: c00000000125d908 c0000001e0957820 c0000000016e8100 c00000013d814540 GPR04: 0000000000000002 c00000013d814550 0000000000000045 0000000000000000 GPR08: c00000013444d000 c00000013d814538 c00000013d814538 0000000084002268 GPR12: c00000000125e730 c000007e2ffd5f00 ffffffffffffffff 0000000000020000 GPR16: 0000000000000000 0000000000000002 c00000015f653000 0000000000000000 GPR20: c000000138662400 c00000013d814540 0000000000000000 c00000013d814500 GPR24: 0000000000000000 0000000000000002 c0000001e0957888 c0000001e0957878 GPR28: c00000013d814548 0000000000000000 c00000013d814540 c0000001e0957888 NIP [c0000000002c8a64] __mutex_add_waiter+0x24/0xc0 LR [c00000000125dbc8] __mutex_lock.constprop.0+0x318/0xd00 Call Trace: 0xc0000001e0957890 (unreliable) __mutex_lock.constprop.0+0x58/0xd00 amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu+0x6fc/0xb60 [amdgpu] kfd_process_alloc_gpuvm+0x54/0x1f0 [amdgpu] kfd_process_device_init_cwsr_dgpu+0xa4/0x1a0 [amdgpu] kfd_process_device_init_vm+0xd8/0x2e0 [amdgpu] kfd_ioctl_acquire_vm+0xd0/0x130 [amdgpu] kfd_ioctl+0x514/0x670 [amdgpu] sys_ioctl+0x134/0x180 system_call_exception+0x114/0x300 system_call_vectored_common+0x15c/0x2ec This patch changes AMDGPU_VA_RESERVED_TRAP_SIZE to 64 KB and KFD_CWSR_TBA_TMA_SIZE to the AMD GPU page size. This means we reserve 64 KB for the trap in the address space, but only allocate 8 KB within it. With this approach, the allocation size never exceeds the reserved area. Fixes: 34a1de0f7935 ("drm/amdkfd: Relocate TBA/TMA to opposite side of VM hole") Reviewed-by: Christian König Suggested-by: Felix Kuehling Suggested-by: Christian König Signed-off-by: Donet Tom Signed-off-by: Alex Deucher (cherry picked from commit 31b8de5e55666f26ea7ece5f412b83eab3f56dbb) Cc: stable@vger.kernel.org

drm/amdgpu/userq: fix memory leak in MQD creation error paths

2026-03-30T20:13:47+00:00

In mes_userq_mqd_create(), the memdup_user() allocations for IP-specific MQD structs are not freed when subsequent VA validation fails. The goto free_mqd label only cleans up the MQD BO object and userq_props. Fix by adding kfree() before each goto free_mqd on VA validation failure in the COMPUTE, GFX, and SDMA branches. Fixes: 9e46b8bb0539 ("drm/amdgpu: validate userq buffer virtual address and size") Reported-by: Yuhao Jiang Signed-off-by: Junrui Luo Reviewed-by: Prike Liang Signed-off-by: Alex Deucher (cherry picked from commit 27f5ff9e4a4150d7cf8b4085aedd3b77ddcc5d08) Cc: stable@vger.kernel.org

drm/amd: Fix MQD and control stack alignment for non-4K

2026-03-30T20:12:27+00:00

For gfxV9, due to a hardware bug ("based on the comments in the code here [1]"), the control stack of a user-mode compute queue must be allocated immediately after the page boundary of its regular MQD buffer. To handle this, we allocate an enlarged MQD buffer where the first page is used as the MQD and the remaining pages store the control stack. Although these regions share the same BO, they require different memory types: the MQD must be UC (uncached), while the control stack must be NC (non-coherent), matching the behavior when the control stack is allocated in user space. This logic works correctly on systems where the CPU page size matches the GPU page size (4K). However, the current implementation aligns both the MQD and the control stack to the CPU PAGE_SIZE. On systems with a larger CPU page size, the entire first CPU page is marked UC—even though that page may contain multiple GPU pages. The GPU treats the second 4K GPU page inside that CPU page as part of the control stack, but it is incorrectly mapped as UC. This patch fixes the issue by aligning both the MQD and control stack sizes to the GPU page size (4K). The first 4K page is correctly marked as UC for the MQD, and the remaining GPU pages are marked NC for the control stack. This ensures proper memory type assignment on systems with larger CPU page sizes. [1]: https://elixir.bootlin.com/linux/v6.18/source/drivers/gpu/drm/amd/amdkfd/kfd_mqd_manager_v9.c#L118 Acked-by: Felix Kuehling Signed-off-by: Donet Tom Signed-off-by: Alex Deucher (cherry picked from commit 998d6781410de1c4b787fdbf6c56e851ea7fa553)

drm/amdgpu: fix the idr allocation flags

2026-03-30T20:10:19+00:00

Fix the IDR allocation flags by using atomic GFP flags in non‑sleepable contexts to avoid the __might_sleep() complaint. 268.290239] [drm] Initialized amdgpu 3.64.0 for 0000:03:00.0 on minor 0 [ 268.294900] BUG: sleeping function called from invalid context at ./include/linux/sched/mm.h:323 [ 268.295355] in_atomic(): 1, irqs_disabled(): 0, non_block: 0, pid: 1744, name: modprobe [ 268.295705] preempt_count: 1, expected: 0 [ 268.295886] RCU nest depth: 0, expected: 0 [ 268.296072] 2 locks held by modprobe/1744: [ 268.296077] #0: ffff8c3a44abd1b8 (&dev->mutex){....}-{4:4}, at: __driver_attach+0xe4/0x210 [ 268.296100] #1: ffffffffc1a6ea78 (amdgpu_pasid_idr_lock){+.+.}-{3:3}, at: amdgpu_pasid_alloc+0x26/0xe0 [amdgpu] [ 268.296494] CPU: 12 UID: 0 PID: 1744 Comm: modprobe Tainted: G U OE 6.19.0-custom #16 PREEMPT(voluntary) [ 268.296498] Tainted: [U]=USER, [O]=OOT_MODULE, [E]=UNSIGNED_MODULE [ 268.296499] Hardware name: AMD Majolica-RN/Majolica-RN, BIOS RMJ1009A 06/13/2021 [ 268.296501] Call Trace: Fixes: 8f1de51f49be ("drm/amdgpu: prevent immediate PASID reuse case") Tested-by: Borislav Petkov (AMD) Signed-off-by: Prike Liang Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit ea56aa2625708eaf96f310032391ff37746310ef) Cc: stable@vger.kernel.org

drm/amdgpu: validate doorbell_offset in user queue creation

2026-03-30T20:08:17+00:00

amdgpu_userq_get_doorbell_index() passes the user-provided doorbell_offset to amdgpu_doorbell_index_on_bar() without bounds checking. An arbitrarily large doorbell_offset can cause the calculated doorbell index to fall outside the allocated doorbell BO, potentially corrupting kernel doorbell space. Validate that doorbell_offset falls within the doorbell BO before computing the BAR index, using u64 arithmetic to prevent overflow. Fixes: f09c1e6077ab ("drm/amdgpu: generate doorbell index for userqueue") Reported-by: Yuhao Jiang Signed-off-by: Junrui Luo Signed-off-by: Alex Deucher (cherry picked from commit de1ef4ffd70e1d15f0bf584fd22b1f28cbd5e2ec) Cc: stable@vger.kernel.org

drm/amdgpu: Handle GPU page faults correctly on non-4K page systems

2026-03-24T17:55:29+00:00

During a GPU page fault, the driver restores the SVM range and then maps it into the GPU page tables. The current implementation passes a GPU-page-size (4K-based) PFN to svm_range_restore_pages() to restore the range. SVM ranges are tracked using system-page-size PFNs. On systems where the system page size is larger than 4K, using GPU-page-size PFNs to restore the range causes two problems: Range lookup fails: Because the restore function receives PFNs in GPU (4K) units, the SVM range lookup does not find the existing range. This will result in a duplicate SVM range being created. VMA lookup failure: The restore function also tries to locate the VMA for the faulting address. It converts the GPU-page-size PFN into an address using the system page size, which results in an incorrect address on non-4K page-size systems. As a result, the VMA lookup fails with the message: "address 0xxxx VMA is removed". This patch passes the system-page-size PFN to svm_range_restore_pages() so that the SVM range is restored correctly on non-4K page systems. Acked-by: Christian König Signed-off-by: Donet Tom Signed-off-by: Alex Deucher (cherry picked from commit 074fe395fb13247b057f60004c7ebcca9f38ef46)

drm/amdgpu: Fix fence put before wait in amdgpu_amdkfd_submit_ib

2026-03-24T17:53:33+00:00

amdgpu_amdkfd_submit_ib() submits a GPU job and gets a fence from amdgpu_ib_schedule(). This fence is used to wait for job completion. Currently, the code drops the fence reference using dma_fence_put() before calling dma_fence_wait(). If dma_fence_put() releases the last reference, the fence may be freed before dma_fence_wait() is called. This can lead to a use-after-free. Fix this by waiting on the fence first and releasing the reference only after dma_fence_wait() completes. Fixes the below: drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.c:697 amdgpu_amdkfd_submit_ib() warn: passing freed memory 'f' (line 696) Fixes: 9ae55f030dc5 ("drm/amdgpu: Follow up change to previous drm scheduler change.") Cc: Felix Kuehling Cc: Dan Carpenter Cc: Christian König Cc: Alex Deucher Signed-off-by: Srinivasan Shanmugam Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit 8b9e5259adc385b61a6590a13b82ae0ac2bd3482)

drm/amdgpu: prevent immediate PASID reuse case

2026-03-23T18:48:06+00:00

PASID resue could cause interrupt issue when process immediately runs into hw state left by previous process exited with the same PASID, it's possible that page faults are still pending in the IH ring buffer when the process exits and frees up its PASID. To prevent the case, it uses idr cyclic allocator same as kernel pid's. Signed-off-by: Eric Huang Reviewed-by: Christian König Signed-off-by: Alex Deucher (cherry picked from commit 8f1de51f49be692de137c8525106e0fce2d1912d) Cc: stable@vger.kernel.org

drm/amdgpu: fix strsep() corrupting lockup_timeout on multi-GPU (v3)

2026-03-23T18:47:47+00:00

amdgpu_device_get_job_timeout_settings() passes a pointer directly to the global amdgpu_lockup_timeout[] buffer into strsep(). strsep() destructively replaces delimiter characters with '\0' in-place. On multi-GPU systems, this function is called once per device. When a multi-value setting like "0,0,0,-1" is used, the first GPU's call transforms the global buffer into "0\00\00\0-1". The second GPU then sees only "0" (terminated at the first '\0'), parses a single value, hits the single-value fallthrough (index == 1), and applies timeout=0 to all rings — causing immediate false job timeouts. Fix this by copying into a stack-local array before calling strsep(), so the global module parameter buffer remains intact across calls. The buffer is AMDGPU_MAX_TIMEOUT_PARAM_LENGTH (256) bytes, which is safe for the stack. v2: wrap commit message to 72 columns, add Assisted-by tag. v3: use stack array with strscpy() instead of kstrdup()/kfree() to avoid unnecessary heap allocation (Christian). This patch was developed with assistance from Claude (claude-opus-4-6). Assisted-by: Claude:claude-opus-4-6 Reviewed-by: Christian König Reviewed-by: Alex Deucher Signed-off-by: Ruijing Dong Signed-off-by: Alex Deucher (cherry picked from commit 94d79f51efecb74be1d88dde66bdc8bfcca17935) Cc: stable@vger.kernel.org