drm/xe: Strict migration policy for atomic SVM faults

Mixing GPU and CPU atomics does not work unless a strict migration
policy of GPU atomics must be device memory. Enforce a policy of must be
in VRAM with a retry loop of 3 attempts, if retry loop fails abort
fault.

Removing always_migrate_to_vram modparam as we now have real migration
policy.

v2:
 - Only retry migration on atomics
 - Drop alway migrate modparam
v3:
 - Only set vram_only on DGFX (Himal)
 - Bail on get_pages failure if vram_only and retry count exceeded (Himal)
 - s/vram_only/devmem_only
 - Update xe_svm_range_is_valid to accept devmem_only argument
v4:
 - Fix logic bug get_pages failure
v5:
 - Fix commit message (Himal)
 - Mention removing always_migrate_to_vram in commit message (Lucas)
 - Fix xe_svm_range_is_valid to check for devmem pages
 - Bail on devmem_only && !migrate_devmem (Thomas)
v6:
 - Add READ_ONCE barriers for opportunistic checks (Thomas)
 - Pair READ_ONCE with WRITE_ONCE (Thomas)
v7:
 - Adjust comments (Thomas)

Fixes: 2f118c949160 ("drm/xe: Add SVM VRAM migration")
Cc: stable@vger.kernel.org
Signed-off-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Acked-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Link: https://lore.kernel.org/r/20250512135500.1405019-3-matthew.brost@intel.com

1 files changed, 18 insertions, 5 deletions

diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
index a58d03e6cac2..41f6616bcf76 100644
--- a/drivers/gpu/drm/drm_gpusvm.c
+++ b/drivers/gpu/drm/drm_gpusvm.c
@@ -1118,6 +1118,10 @@ static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
 	lockdep_assert_held(&gpusvm->notifier_lock);
 
 	if (range->flags.has_dma_mapping) {
+		struct drm_gpusvm_range_flags flags = {
+			.__flags = range->flags.__flags,
+		};
+
 		for (i = 0, j = 0; i < npages; j++) {
 			struct drm_pagemap_device_addr *addr = &range->dma_addr[j];
 
@@ -1131,8 +1135,12 @@ static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
 							    dev, *addr);
 			i += 1 << addr->order;
 		}
-		range->flags.has_devmem_pages = false;
-		range->flags.has_dma_mapping = false;
+
+		/* WRITE_ONCE pairs with READ_ONCE for opportunistic checks */
+		flags.has_devmem_pages = false;
+		flags.has_dma_mapping = false;
+		WRITE_ONCE(range->flags.__flags, flags.__flags);
+
 		range->dpagemap = NULL;
 	}
 }
@@ -1334,6 +1342,7 @@ int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
 	int err = 0;
 	struct dev_pagemap *pagemap;
 	struct drm_pagemap *dpagemap;
+	struct drm_gpusvm_range_flags flags;
 
 retry:
 	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
@@ -1378,7 +1387,8 @@ map_pages:
 	 */
 	drm_gpusvm_notifier_lock(gpusvm);
 
-	if (range->flags.unmapped) {
+	flags.__flags = range->flags.__flags;
+	if (flags.unmapped) {
 		drm_gpusvm_notifier_unlock(gpusvm);
 		err = -EFAULT;
 		goto err_free;
@@ -1474,14 +1484,17 @@ map_pages:
 		}
 		i += 1 << order;
 		num_dma_mapped = i;
-		range->flags.has_dma_mapping = true;
+		flags.has_dma_mapping = true;
 	}
 
 	if (zdd) {
-		range->flags.has_devmem_pages = true;
+		flags.has_devmem_pages = true;
 		range->dpagemap = dpagemap;
 	}
 
+	/* WRITE_ONCE pairs with READ_ONCE for opportunistic checks */
+	WRITE_ONCE(range->flags.__flags, flags.__flags);
+
 	drm_gpusvm_notifier_unlock(gpusvm);
 	kvfree(pfns);
 set_seqno: