diff options
| author | Arvind Yadav <arvind.yadav@intel.com> | 2026-05-26 19:24:47 +0530 |
|---|---|---|
| committer | Thomas Hellström <thomas.hellstrom@linux.intel.com> | 2026-07-20 18:18:41 +0200 |
| commit | 57441577bac3637473da2c9644336eaa0ac5732f (patch) | |
| tree | af87f356a7711654bdbe61c47f8706252a7862e0 /drivers/gpu/drm/xe | |
| parent | 130910bac905a42225f93841e338bbea4a431b1a (diff) | |
drm/xe/madvise: Skip invalidation for purgeable state updates
Purgeable state updates only change VMA/BO metadata. They do not zap
PTEs when switching between DONTNEED and WILLNEED. PTEs are zapped
later if the BO is actually purged.
xe_vm_invalidate_madvise_range() waits on the VM dma-resv before checking
vma->skip_invalidation. Since purgeable madvise marks all affected VMAs to
skip invalidation, this wait is unnecessary and can stall on unrelated
in-flight work.
Skip the invalidate path entirely for purgeable state updates.
v2:
- Replace inline 'args->type != DRM_XE_VMA_ATTR_PURGEABLE_STATE'
check with a small helper madvise_range_needs_invalidation().
(Himal)
Suggested-by: Matthew Brost <matthew.brost@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Signed-off-by: Arvind Yadav <arvind.yadav@intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Link: https://patch.msgid.link/20260526135447.2973029-1-arvind.yadav@intel.com
Signed-off-by: Tejas Upadhyay <tejas.upadhyay@intel.com>
Fixes: ada7486c5668 ("drm/xe: Implement madvise ioctl for xe")
Cc: <stable@vger.kernel.org> # v6.18+
(cherry picked from commit 134377098b9c14abd31c3bcac00c9653f0f0c4c3)
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Diffstat (limited to 'drivers/gpu/drm/xe')
| -rw-r--r-- | drivers/gpu/drm/xe/xe_vm_madvise.c | 19 |
1 files changed, 17 insertions, 2 deletions
diff --git a/drivers/gpu/drm/xe/xe_vm_madvise.c b/drivers/gpu/drm/xe/xe_vm_madvise.c index 246fe1843142..0474768a38aa 100644 --- a/drivers/gpu/drm/xe/xe_vm_madvise.c +++ b/drivers/gpu/drm/xe/xe_vm_madvise.c @@ -332,6 +332,20 @@ static int xe_vm_invalidate_madvise_range(struct xe_vm *vm, u64 start, u64 end) return err; } +/** + * madvise_range_needs_invalidation() - Check whether madvise needs invalidation + * @args: madvise ioctl arguments + * + * Purgeable state updates only touch VMA/BO metadata. PTEs stay valid and are + * zapped only if the BO is later purged. + * + * Return: true when the update needs PTE invalidation. + */ +static bool madvise_range_needs_invalidation(const struct drm_xe_madvise *args) +{ + return args->type != DRM_XE_VMA_ATTR_PURGEABLE_STATE; +} + static bool madvise_args_are_sane(struct xe_device *xe, const struct drm_xe_madvise *args) { if (XE_IOCTL_DBG(xe, !args)) @@ -708,8 +722,9 @@ int xe_vm_madvise_ioctl(struct drm_device *dev, void *data, struct drm_file *fil madvise_funcs[attr_type](xe, vm, madvise_range.vmas, madvise_range.num_vmas, args, &details); - err = xe_vm_invalidate_madvise_range(vm, madvise_range.addr, - madvise_range.addr + args->range); + if (madvise_range_needs_invalidation(args)) + err = xe_vm_invalidate_madvise_range(vm, madvise_range.addr, + madvise_range.addr + args->range); if (madvise_range.has_svm_userptr_vmas) xe_svm_notifier_unlock(vm); |