linux/include/uapi/drm
Oak Zeng ae28e34400 drm/xe: Allow scratch page under fault mode for certain platform
Normally scratch page is not allowed when a vm is operate under page
fault mode, i.e., in the existing codes, DRM_XE_VM_CREATE_FLAG_SCRATCH_PAGE
and DRM_XE_VM_CREATE_FLAG_FAULT_MODE are mutual exclusive. The reason
is fault mode relies on recoverable page to work, while scratch page
can mute recoverable page fault.

On xe2 and xe3, out of bound prefetch can cause page fault and further
system hang because xekmd can't resolve such page fault. SYCL and OCL
language runtime requires out of bound prefetch to be silently dropped
without causing any functional problem, thus the existing behavior
doesn't meet language runtime requirement.

At the same time, HW prefetching can cause page fault interrupt. Due to
page fault interrupt overhead (i.e., need Guc and KMD involved to fix
the page fault), HW prefetching can be slowed by many orders of magnitude.

Fix those problems by allowing scratch page under fault mode for xe2 and
xe3. With scratch page in place, HW prefetching could always hit scratch
page instead of causing interrupt.

A side effect is, scratch page could hide application program error.
Application out of bound accesses are hided by scratch page mapping,
instead of get reported to user.

v2: Refine commit message (Thomas)

v3: Move the scratch page flag check to after scratch page wa (Thomas)

v4: drop NEEDS_SCRATCH macro (matt)
    Add a comment to DRM_XE_VM_CREATE_FLAG_SCRATCH_PAGE

Signed-off-by: Oak Zeng <oak.zeng@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Link: https://lore.kernel.org/r/20250403165328.2438690-4-oak.zeng@intel.com
Signed-off-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
2025-04-07 11:17:30 +05:30
..
amdgpu_drm.h drm/amdgpu: Add flags to distinguish vf/pf/pt mode 2025-02-12 21:04:08 -05:00
amdxdna_accel.h accel/amdxdna: Remove DRM_AMDXDNA_HWCTX_CONFIG_NUM 2024-12-17 12:10:07 -06:00
armada_drm.h
drm_fourcc.h drm: add modifiers for MediaTek tiled formats 2025-01-24 11:13:52 +00:00
drm_mode.h drm: Add missing documentation for struct drm_plane_size_hint 2024-08-12 11:15:14 +02:00
drm_sarea.h
drm.h drm: add DRM_SET_CLIENT_NAME ioctl 2024-10-08 10:00:30 +02:00
etnaviv_drm.h Revert "drm/etnaviv: Expose a few more chipspecs to userspace" 2024-04-25 16:56:20 +02:00
exynos_drm.h
habanalabs_accel.h
i915_drm.h drm/i915: Support replaying GPU hangs with captured context image 2024-05-16 07:37:05 +00:00
ivpu_accel.h accel/ivpu: Add API for command queue create/destroy/submit 2025-01-09 09:35:44 +01:00
lima_drm.h
msm_drm.h drm/msm: Expose uche trap base via uapi 2025-01-03 07:20:27 -08:00
nouveau_drm.h drm/nouveau: use tile_mode and pte_kind for VM_BIND bo allocations 2024-05-13 22:27:33 +02:00
omap_drm.h
panfrost_drm.h drm/panfrost: Add cycle counter job requirement 2024-09-02 10:13:38 +01:00
panthor_drm.h drm/panthor: Convert IOCTL defines to an enum 2025-02-07 18:35:25 +01:00
pvr_drm.h
qaic_accel.h accel/qaic: Fix typo for struct qaic_manage_trans_passthrough 2024-12-13 08:55:09 -07:00
qxl_drm.h
radeon_drm.h
tegra_drm.h
v3d_drm.h drm/v3d: Add DRM_IOCTL_V3D_PERFMON_SET_GLOBAL 2024-12-04 08:44:27 -03:00
vc4_drm.h
vgem_drm.h
virtgpu_drm.h
vmwgfx_drm.h drm/vmwgfx: Add SPDX header to vmwgfx_drm.h 2024-01-30 14:18:19 -05:00
xe_drm.h drm/xe: Allow scratch page under fault mode for certain platform 2025-04-07 11:17:30 +05:30