linux.git/drivers/gpu/drm/amd/amdgpu, branch master

drm/amdgpu/ttm: Consider concurrent VM flushes for buffer entities

2026-07-17T21:41:01+00:00

Allow using multiple SDMA schedulers only on GPUs where
we are allowed to do concurrent VM flushes.
This consideration is necessary because all GART windows
are mapped in VMID 0 (the kernel VMID) so each buffer
entity would flush VMID 0 concurrently.

Practically this means that we can't use multiple SDMA
engines for TTM on GFX6-8 and Navi 1x.

Fixes: 01c836788b37 ("drm/amdgpu: pass all the sdma scheds to amdgpu_mman")
Fixes: e4029f7a9474 ("drm/amdgpu: only use working sdma schedulers for ttm")
Cc: Pierre-Eric Pelloux-Prayer 
Signed-off-by: Timur Kristóf 
Reviewed-by: Christian König 
Signed-off-by: Alex Deucher 
(cherry picked from commit a8171229bc836607fbc225d323ebc4d14489cfbb)

drm/amdgpu: Disable PCIe dynamic speed switching on Ryzen Pinnacle Ridge

2026-07-17T21:40:31+00:00

AMD Ryzen Pinnacle Ridge (Zen+, family 0x17 model 0x08) CPUs have
PCI controllers that don't support PCIe dynamic speed switching,
causing system freezes during GPU initialization when enabled.

Disable dynamic speed switching when this CPU is detected.

Assisted-by: Claude:sonnet
Fixes: 466a7d115326 ("drm/amd: Use the first non-dGPU PCI device for BW limits")
Closes: https://gitlab.freedesktop.org/drm/amd/-/work_items/5436
Reviewed-by: Lijo Lazar 
Link: https://patch.msgid.link/20260709031520.841611-1-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello 
Signed-off-by: Alex Deucher 
(cherry picked from commit 9ceb4e034a327a04155f32f1cd1a5031dfa5fe02)
Cc: stable@vger.kernel.org

drm/amdgpu: always emit the job vm fence

2026-07-17T21:40:23+00:00

We need the fence to reemit the gds switch or spm update
after a queue reset.

Fixes: a17ef941212b ("drm/amdgpu: rework ring reset backup and reemit v9")
Cc: timur.kristof@gmail.com
Cc: christian.koenig@amd.com
Reviewed-by: Timur Kristóf 
Reviewed-by: Christian König 
Signed-off-by: Alex Deucher 
(cherry picked from commit bc639a9eadc75822f7f15a4315c198a4b5513bd2)
Cc: stable@vger.kernel.org

drm/amdgpu: Reserve space for IB contents in devcoredumps

2026-07-17T21:39:16+00:00

Currently the contents of IBs are abruptly cut off and don't
show the full contents. This patch makes sure to reserve
space for those contents too so they may be printed.

Signed-off-by: Timur Kristóf 
Acked-by: Christian König 
Signed-off-by: Alex Deucher 
(cherry picked from commit 4e2c0821509fed754e8c31d5053d152fbb3484a5)
Cc: stable@vger.kernel.org

drm/amdgpu: Print vmid, pasid and more task info in devcoredump

2026-07-17T21:39:08+00:00

These are in the dmesg logs but are missing from devcoredumps.

Signed-off-by: Timur Kristóf 
Reviewed-by: Christian König 
Signed-off-by: Alex Deucher 
(cherry picked from commit fed7aa36d79802c3e02acd05aeae8b0a877e47c2)
Cc: stable@vger.kernel.org

drm/amdgpu: Release VFCT ACPI table reference

2026-07-17T21:39:00+00:00

amdgpu_acpi_vfct_bios() fetches the VFCT table with acpi_get_table()
but never releases it. acpi_get_table() takes a reference on the
table (incrementing its validation_count and mapping it on the 0->1
transition); without a paired acpi_put_table() the mapping is leaked
on every call, whether or not a matching VBIOS image is found.

Route all exit paths after the table is acquired through a common
acpi_put_table(). The VBIOS image is copied out with kmemdup() before
the table is released, so it remains valid for the caller.

Reviewed-by: Alex Deucher 
Link: https://patch.msgid.link/20260708193518.702584-3-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello 
Signed-off-by: Alex Deucher 
(cherry picked from commit ca5988682b4cba4cd125a0fa99b2de1239164ae4)
Cc: stable@vger.kernel.org

drm/amdgpu: Fix VFCT bus number matching with soft filter

2026-07-17T21:38:48+00:00

On systems where PCI bus renumbering occurs (e.g. pci=realloc,
resource conflicts), the runtime bus number may differ from the
BIOS POST bus number recorded in the VFCT table. This causes
amdgpu_acpi_vfct_bios() to fail finding the VBIOS even though
the correct device entry exists.

Introduce amdgpu_acpi_vfct_match() which treats the bus number
as a soft filter: vendor/device/function identity is the hard
requirement, while exact bus match is the preferred path. When
bus numbers disagree but device identity matches, accept the
VFCT entry and log a dev_notice for diagnostics.

Reported-by: Oz Tiram 
Closes: https://lore.kernel.org/amd-gfx/20260621173211.28443-1-oz@shift-computing.de/
Reviewed-by: Alex Deucher 
Link: https://patch.msgid.link/20260708193518.702584-2-mario.limonciello@amd.com
Signed-off-by: Mario Limonciello 
Signed-off-by: Alex Deucher 
(cherry picked from commit 11c141672045ffc0187aa604f2c0f597bc334fb2)
Cc: stable@vger.kernel.org

drm/amdgpu: fix bo->pin leaking in amdgpu_bo_create_reserved

2026-07-17T21:38:33+00:00

amdgpu_bo_create_reserved() only allocates a new BO when
*bo_ptr (struct amdgpu_bo **bo_ptr as input parameter) is
NULL, it simply skips creation when *bo_ptr is non-NULL.
But it unconditionally reserves, pins, gart allocates
and maps the BO afterwards.

When the same non-NULL BO pointer is passed in again,
for example firmware buffers that live in adev and are
re-loaded on every resume / cp_resume / start
under AMDGPU_FW_LOAD_DIRECT, amdgpu_bo_pin() just increases
pin_count unconditionally, however the matching teardown only unpins
once, so pin_count never drops to zero, so TTM is not able
to move, swap or evict a BO, causing BO leaks.

This commit fixes this issue by only pinning the bo
once at creation, and repeated calls no longer
take additional pin references.

Signed-off-by: Zhu Lingshan 
Reviewed-by: Alex Deucher 
Reviewed-by: Christian König 
Signed-off-by: Alex Deucher 
(cherry picked from commit 3ddc0ae76202c447b6aec61e907b852bc94671cf)
Cc: stable@vger.kernel.org

drm/amdgpu/userq: fix indefinite fence wait during GPU reset

2026-07-17T21:38:16+00:00

pre_reset only force-completes fences of MAPPED queues. A queue in any
other state (e.g. mid-eviction) keeps its last_fence pending; after a
GPU reset that fence never signals, so the eviction/suspend worker and
process teardown (amdgpu_evf_mgr_flush_suspend) wait on it forever and
wedge the machine:

  INFO: task kworker/6:28 blocked for more than 120 seconds.
  Workqueue: events amdgpu_eviction_fence_suspend_worker [amdgpu]
  Call Trace:
   dma_fence_wait_timeout+0x7e/0x130
   amdgpu_userq_evict+0x67/0x140 [amdgpu]
   amdgpu_eviction_fence_suspend_worker+0xd8/0x160 [amdgpu]
   process_scheduled_works+0xa6/0x420

Force-complete every queue's fence regardless of state. The unmap and
mark-hung step stays gated on MAPPED, since unmapping a queue that is
not mapped is invalid.

Fixes: 290f46cf5726 ("drm/amdgpu: Implement user queue reset functionality")
Reviewed-by: Christian König 
Signed-off-by: Jesse Zhang 
Signed-off-by: Alex Deucher 
(cherry picked from commit 9102b39fa924dcc3dc75a3137bfa9633c40b88c0)
Cc: stable@vger.kernel.org

drm/amdgpu/discovery: Fix device family for DCN42

2026-07-17T21:36:57+00:00

GC 11.7.0 and 11.7.1 should map to AMDGPU_FAMILY_GC_11_5_4 for DCN42.

Fixes: cf591e67c095 ("drm/amdgpu: add support for GC IP version 11.7.0")
Fixes: a928d8d81ec5 ("drm/amdgpu: add support for GC IP version 11.7.1")
Signed-off-by: Roman Li 
Acked-by: Alex Deucher 
Signed-off-by: Alex Deucher 
(cherry picked from commit f8ee6447e7ec1d75d6663c817e45566dd01f440b)