CPU mappings created before a device wedge can keep valid PTEs and
continue accessing VRAM. Mapping new faults to a dummy page does not
replace these existing mappings.

Use the common device I/O SRCU gate to wait for active CPU faults to
finish. Then invalidate all tracked VRAM mappings. Faults starting
after the wedge use the per-BO dummy page.

This ensures userspace is notified only after existing VRAM mappings
have been removed.

v2:
- Drain active faults before invalidating VRAM mappings

Cc: Matthew Brost <[email protected]>
Cc: Thomas Hellström <[email protected]>
Cc: Himal Prasad Ghimiray <[email protected]>
Cc: Rodrigo Vivi <[email protected]>
Assisted-by: Claude:claude-opus-4-8
Signed-off-by: Arvind Yadav <[email protected]>
---
 drivers/gpu/drm/xe/xe_bo.c     | 20 ++++++++++++++++++++
 drivers/gpu/drm/xe/xe_bo.h     |  1 +
 drivers/gpu/drm/xe/xe_device.c |  4 ++++
 3 files changed, 25 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index 7902ce3fe012..73dcd397dc13 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -4155,6 +4155,26 @@ void xe_bo_runtime_pm_release_mmap_offset(struct xe_bo 
*bo)
        list_del_init(&bo->vram_userfault_link);
 }
 
+/**
+ * xe_bo_wedged_invalidate_mmaps - Invalidate CPU mappings backed by VRAM
+ * @xe: xe device instance
+ *
+ * The caller must drain the common device I/O gate before calling this
+ * function. Remove all tracked VRAM mappings so later faults map the
+ * per-BO dummy page.
+ */
+void xe_bo_wedged_invalidate_mmaps(struct xe_device *xe)
+{
+       struct xe_bo *bo, *next;
+
+       mutex_lock(&xe->mem_access.vram_userfault.lock);
+       list_for_each_entry_safe(bo, next,
+                                &xe->mem_access.vram_userfault.list,
+                                vram_userfault_link)
+               xe_bo_runtime_pm_release_mmap_offset(bo);
+       mutex_unlock(&xe->mem_access.vram_userfault.lock);
+}
+
 #if IS_ENABLED(CONFIG_DRM_XE_KUNIT_TEST)
 #include "tests/xe_bo.c"
 #endif
diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h
index 290ca624e2a7..103eaacb56f2 100644
--- a/drivers/gpu/drm/xe/xe_bo.h
+++ b/drivers/gpu/drm/xe/xe_bo.h
@@ -449,6 +449,7 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
 int xe_gem_mmap_offset_ioctl(struct drm_device *dev, void *data,
                             struct drm_file *file);
 void xe_bo_runtime_pm_release_mmap_offset(struct xe_bo *bo);
+void xe_bo_wedged_invalidate_mmaps(struct xe_device *xe);
 
 int xe_bo_dumb_create(struct drm_file *file_priv,
                      struct drm_device *dev,
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 045b5844e5c3..3549cf89a353 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -896,6 +896,10 @@ static void xe_device_wedged_work(struct work_struct *work)
        unsigned long method;
        int err;
 
+       /* Drain active faults before invalidating VRAM mappings. */
+       xe_device_io_drain(xe);
+       xe_bo_wedged_invalidate_mmaps(xe);
+
        /* Report at most one recovery method per worker invocation. */
        method = READ_ONCE(xe->wedged.method);
        if (method != READ_ONCE(xe->wedged.reported_method)) {
-- 
2.43.0

Reply via email to