linux.git/drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h, branch v7.3-rc2

drm/amdkfd: Remove svm_bo eviction fence

2026-07-29T21:52:20+00:00

SVM BOs are now migrated back to system memory synchronously from the TTM
eviction path (svm_range_evict_svm_bo), so the per-svm_bo eviction fence
is no longer used.

Remove the eviction fence from svm_range_bo, drop the
amdgpu_amdkfd_fence->svm_bo back pointer and the
amdgpu_amdkfd_evict_svm_bo() helper, and stop special-casing svm_bo
fences in the KFD fence enable_signaling and check_mm paths. Embed struct
amdgpu_bo directly in svm_range_bo with a dedicated svm_range_bo_destroy()
callback, and keep the owning mm via mmgrab()/mmdrop() instead of through
the fence.

Signed-off-by: Philip Yang 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher

drm/amdkfd: Evict SVM BOs synchronously from TTM eviction

2026-07-28T23:17:34+00:00

svm_range_evict_svm_bo_worker() migrated an SVM BO's pages back to system
memory from a work item that took mmap_read_lock. When an mmap writer was
pending, that read lock blocked behind the writer while the thread
allocating a new migration VRAM BO waited on this BO's eviction fence - a
circular wait that hung the SVM workers.

Evict the SVM BO synchronously from the TTM eviction path
(amdgpu_ttm_bo_eviction_valuable) instead of deferring to a work item.
The BO is already reserved and the lock order is mmap_lock -> BO
reservation, so only trylock the owning process's mmap lock; on
contention return -EBUSY so TTM skips this BO. This removes the eviction
work item and the enable_signaling path, so no worker can block on
mmap_read_lock.

The SVM BO uses AMDGPU_GEM_CREATE_DISCARDABLE, so ttm_bo_evict takes the
pipeline_gutting path and skips allocating a system memory placement.
That would be wasted work, since svm_migrate_vram_to_ram allocates the
system pages and copies the data back itself.

Eviction now migrates ranges directly, so it must serialize with the
owning process: it trylocks migrate_mutex under svm_bo->list_lock before
unlinking the range, and svm_range_free() unlinks the range then waits on
migrate_mutex, so a concurrent eviction cannot free a range under it.

Drop the mm reference with mmput_async so exit_mmap() does not run under
the BO reservation.

Signed-off-by: Philip Yang 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher

drm/amdkfd: Add domain parameter to kernel BO mapping function

2026-07-01T15:29:43+00:00

This change allows amdgpu_amdkfd_gpuvm_map_bo_to_kernel() to pin buffers
in either GTT or VRAM based on caller specification, providing flexibility
for different memory placement requirements across various kernel buffers.

The domain parameter accepts AMDGPU_GEM_DOMAIN_GTT, AMDGPU_GEM_DOMAIN_VRAM,
or their combination (GTT|VRAM) to let amdgpu_bo_pin() choose the optimal
placement via amdgpu_bo_get_preferred_domain(). This flexible validation
allows callers to specify their preference while delegating final placement
decisions to the driver when appropriate.

CPU visibility is automatically enforced by amdgpu_bo_pin() regardless of
the domain parameter (see amdgpu_bo_pin() line 975-976 which sets
AMDGPU_GEM_CREATE_CPU_ACCESS_REQUIRED for kernel mappings).

-v3: update amdgpu_amdkfd_gpuvm_map_bo_to_kernel description

Signed-off-by: James Zhu 
Reviewed-by: Vladimir Indic 
Reviewed-by: Philip Yang 
Signed-off-by: Alex Deucher

drm/amdkfd: Let driver decide buffer size at AMDKFD_IOC_GET_DMABUF_INFO ioctl

2026-06-17T20:30:25+00:00

amdkfd driver needs allocate buffer to return bo metadata to user space. The
buffer size is controlled by user currently. It is a potential security issue
that hostile value (e.g. 2 GiB) lets any render-group user trigger order-MAX
allocation/OOM in kernel context.

This patch first finds bo metadata size. If the size is smaller than user
provided value drive can safely allocate buffer in kernel space and copy to
user space buffer. If not, driver will let user know, not allocate and copy.
User will redo with new buffer in user space.

This patch lets driver decide buffer allocation size to avoid potential hostile
size from user space.

Signed-off-by: Xiaogang Chen 
Reviewed-by: Alex Deucher 
Signed-off-by: Alex Deucher

drm/amdgpu: add ioctl to handle RAS poison error

2026-06-17T19:51:36+00:00

Add a new DRM_IOCTL_AMDGPU_PROC_OPTIONS ioctl with the
AMDGPU_PROC_OPTIONS_OP_KFD_SIGBUS_DELAY option, allowing userspace (ROCr)
to control per-process SIGBUS delivery.

Userspace for this can be found at:
https://github.com/ROCm/rocm-systems/pull/6190

Reviewed-by: Lijo Lazar 
Reviewed-by: Alex Deucher 
Signed-off-by: Yifan Zhang 
Signed-off-by: Alex Deucher

drm/amdkfd: plumb a helper to reset a KFD user queue

2026-06-17T19:51:35+00:00

Can be called from KGD.

Reviewed-by: Amber Lin 
Signed-off-by: Alex Deucher

drm/amdgpu: unmap all user mappings of framebuffer and doorbell before mode1 reset

2026-05-19T15:45:41+00:00

During Mode 1 reset, the ASIC undergoes a reset cycle and becomes temporarily
inaccessible via PCIe. Any attempt to access framebuffer or MMIO registers during
this window can result in uncompleted PCIe transactions, leading to NMI panics or
system hangs.

To prevent this, Unmap all of the applications mappings of the framebuffer
and doorbell BARs before mode1 reset. Also prevent new mappings from coming in
during the reset process.

v2: remove inode in kfd_dev (Christian)
v3: correct unmap offset (Felix), remove prevent new mappings part
to avoid deadlock (Christian)

Reviewed-by: Felix Kuehling 
Signed-off-by: Yifan Zhang 
Signed-off-by: Alex Deucher

drm/amdgpu: Move KFD sched stop/start into PTL control path

2026-05-11T19:55:56+00:00

Move amdgpu_amdkfd_stop/start_sched calls from kfd_ptl_control()
into amdgpu_ptl_perf_monitor_ctrl() so all PTL callers (KFD ioctl,
sysfs, GFX init) get consistent scheduling management.

Add amdgpu_amdkfd_stop/start_sched_all() wrappers to stop and
restart KFD scheduling on all nodes without assuming node ID ordering.

v3:
 * call start/stop for PTL Set Only
v2:
 * move the stop/start sched function to
   amdgpu_ptl_perf_monitor_ctrl(Lijo)
 * add wrapper amdgpu_amdkfd_stop_sched_all and
   amdgpu_amdkfd_start_sched_all (Lijo)

Signed-off-by: Perry Yuan 
Reviewed-by: Yifan Zhang 
Signed-off-by: Alex Deucher

drm/amdgpu: add psp interfaces for peak tops limiter driver

2026-05-11T19:55:55+00:00

Introduce a Peak Tops Limiter (PTL) driver that dynamically caps
engine frequency to ensure delivered TOPS never exceeds a defined
TOPS_limit. This initial implementation provides core data structures
and kernel-space interfaces (set/get, enable/disable) to manage PTL state.

PTL performs a firmware handshake to initialize its state and update
predefined format types. It supports updating these format types at
runtime while user-space tools automatically switch PTL state, and
also allows explicitly switching PTL state via newly added commands.

Signed-off-by: Perry Yuan 
Reviewed-by: Lijo Lazar 
Acked-by: Alex Deucher 
Signed-off-by: Alex Deucher

drm/amdkfd: Add domain parameter to alloc kernel BO

2026-01-14T19:28:59+00:00

To allocate kernel BO from VRAM domain for MQD in the following patch.
No functional change because kernel BO allocate all from GTT domain.

Rename amdgpu_amdkfd_alloc_gtt_mem to amdgpu_amdkfd_alloc_kernel_mem
Rename amdgpu_amdkfd_free_gtt_mem to amdgpu_amdkfd_free_kernel_mem
Rename mem_kfd_mem_obj gtt_mem to mem

Signed-off-by: Philip Yang 
Reviewed-by: Kent Russell 
Reviewed-by: Felix Kuehling 
Signed-off-by: Alex Deucher