drm/xe/client: add missing bo locking in show_meminfo()

[Matthew Auld <matthew.auld@xxxxxxxxx>]

From: Matthew Auld <matthew.auld@xxxxxxxxx>

commit 94c4aa266111262c96c98f822d1bccc494786fee upstream.

bo_meminfo() wants to inspect bo state like tt and the ttm resource,
however this state can change at any point leading to stuff like NPD and
UAF, if the bo lock is not held. Grab the bo lock when calling
bo_meminfo(), ensuring we drop any spinlocks first. In the case of
object_idr we now also need to hold a ref.

v2 (MattB)
  - Also add xe_bo_assert_held()

Fixes: 0845233388f8 ("drm/xe: Implement fdinfo memory stats printing")
Signed-off-by: Matthew Auld <matthew.auld@xxxxxxxxx>
Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@xxxxxxxxx>
Cc: Tejas Upadhyay <tejas.upadhyay@xxxxxxxxx>
Cc: "Thomas Hellström" <thomas.hellstrom@xxxxxxxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx> # v6.8+
Reviewed-by: Matthew Brost <matthew.brost@xxxxxxxxx>
Reviewed-by: Tejas Upadhyay <tejas.upadhyay@xxxxxxxxx>
Link: https://patchwork.freedesktop.org/patch/msgid/20240911155527.178910-6-matthew.auld@xxxxxxxxx
(cherry picked from commit 4f63d712fa104c3ebefcb289d1e733e86d8698c7)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
 drivers/gpu/drm/xe/xe_drm_client.c |   39 ++++++++++++++++++++++++++++++++++---
 1 file changed, 36 insertions(+), 3 deletions(-)

--- a/drivers/gpu/drm/xe/xe_drm_client.c
+++ b/drivers/gpu/drm/xe/xe_drm_client.c
@@ -9,6 +9,7 @@
 #include <linux/slab.h>
 #include <linux/types.h>
 
+#include "xe_assert.h"
 #include "xe_bo.h"
 #include "xe_bo_types.h"
 #include "xe_device_types.h"
@@ -93,10 +94,13 @@ void xe_drm_client_add_bo(struct xe_drm_
  */
 void xe_drm_client_remove_bo(struct xe_bo *bo)
 {
+	struct xe_device *xe = ttm_to_xe_device(bo->ttm.bdev);
 	struct xe_drm_client *client = bo->client;
 
+	xe_assert(xe, !kref_read(&bo->ttm.base.refcount));
+
 	spin_lock(&client->bos_lock);
-	list_del(&bo->client_link);
+	list_del_init(&bo->client_link);
 	spin_unlock(&client->bos_lock);
 
 	xe_drm_client_put(client);
@@ -108,6 +112,8 @@ static void bo_meminfo(struct xe_bo *bo,
 	u64 sz = bo->size;
 	u32 mem_type;
 
+	xe_bo_assert_held(bo);
+
 	if (bo->placement.placement)
 		mem_type = bo->placement.placement->mem_type;
 	else
@@ -149,7 +155,20 @@ static void show_meminfo(struct drm_prin
 	idr_for_each_entry(&file->object_idr, obj, id) {
 		struct xe_bo *bo = gem_to_xe_bo(obj);
 
-		bo_meminfo(bo, stats);
+		if (dma_resv_trylock(bo->ttm.base.resv)) {
+			bo_meminfo(bo, stats);
+			xe_bo_unlock(bo);
+		} else {
+			xe_bo_get(bo);
+			spin_unlock(&file->table_lock);
+
+			xe_bo_lock(bo, false);
+			bo_meminfo(bo, stats);
+			xe_bo_unlock(bo);
+
+			xe_bo_put(bo);
+			spin_lock(&file->table_lock);
+		}
 	}
 	spin_unlock(&file->table_lock);
 
@@ -159,7 +178,21 @@ static void show_meminfo(struct drm_prin
 		if (!kref_get_unless_zero(&bo->ttm.base.refcount))
 			continue;
 
-		bo_meminfo(bo, stats);
+		if (dma_resv_trylock(bo->ttm.base.resv)) {
+			bo_meminfo(bo, stats);
+			xe_bo_unlock(bo);
+		} else {
+			spin_unlock(&client->bos_lock);
+
+			xe_bo_lock(bo, false);
+			bo_meminfo(bo, stats);
+			xe_bo_unlock(bo);
+
+			spin_lock(&client->bos_lock);
+			/* The bo ref will prevent this bo from being removed from the list */
+			xe_assert(xef->xe, !list_empty(&bo->client_link));
+		}
+
 		xe_bo_put_deferred(bo, &deferred);
 	}
 	spin_unlock(&client->bos_lock);


Patches currently in stable-queue which might be from matthew.auld@xxxxxxxxx are

queue-6.10/drm-xe-use-devm-instead-of-drmm-for-managed-bo.patch
queue-6.10/drm-xe-client-fix-deadlock-in-show_meminfo.patch
queue-6.10/drm-xe-client-add-missing-bo-locking-in-show_meminfo.patch