This is a note to let you know that I've just added the patch titled

    drm/ttm: Fix ttm_bo_swapout() infinite LRU walk on swapout failure

to the 7.0-stable tree which can be found at:
    
http://www.kernel.org/git/?p=linux/kernel/git/stable/stable-queue.git;a=summary

The filename of the patch is:
     drm-ttm-fix-ttm_bo_swapout-infinite-lru-walk-on-swapout-failure.patch
and it can be found in the queue-7.0 subdirectory.

If you, or anyone else, feels it should not be added to the stable tree,
please let <[email protected]> know about it.


>From b2ed01e7ad3de80333e9b962a44024b094bc0b2b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Thomas=20Hellstr=C3=B6m?= <[email protected]>
Date: Tue, 28 Apr 2026 11:44:42 +0200
Subject: drm/ttm: Fix ttm_bo_swapout() infinite LRU walk on swapout failure
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

From: Thomas Hellström <[email protected]>

commit b2ed01e7ad3de80333e9b962a44024b094bc0b2b upstream.

When ttm_tt_swapout() fails, the current code calls
ttm_resource_add_bulk_move() followed by ttm_resource_move_to_lru_tail()
to restore the resource's bulk_move membership.

However, ttm_resource_move_to_lru_tail() places the resource at the tail
of the LRU list which, relative to the walk cursor's hitch node (placed
immediately after the resource when it was yielded), puts the resource
*in front of the* the hitch. The next list_for_each_entry_continue() from
the hitch finds the same resource again, causing an infinite loop.

Fix by deferring del_bulk_move to the success path only.

On the success path, TTM_TT_FLAG_SWAPPED has just been set by
ttm_tt_swapout() but the resource is still tracked in the bulk_move range,
so ttm_resource_del_bulk_move()'s !ttm_resource_unevictable() guard would
incorrectly skip the removal. Introduce
ttm_resource_del_bulk_move_unevictable() which bypasses that guard.

Reported-by: Jatin Kataria <[email protected]>
Fixes: fc5d96670eb2 ("drm/ttm: Move swapped objects off the manager's LRU list")
Cc: Christian König <[email protected]>
Cc: Matthew Brost <[email protected]>
Cc: <[email protected]>
Cc: <[email protected]> # v6.13+
Assisted-by: GitHub_Copilot:claude-sonnet-4.6
Signed-off-by: Thomas Hellström <[email protected]>
Reviewed-by: Christian König <[email protected]>
Tested-by: Boqun Feng <[email protected]>
Link: 
https://patch.msgid.link/[email protected]
Signed-off-by: Greg Kroah-Hartman <[email protected]>
---
 drivers/gpu/drm/ttm/ttm_bo.c       |   16 ++++++----------
 drivers/gpu/drm/ttm/ttm_resource.c |   13 +++++++++++++
 include/drm/ttm/ttm_resource.h     |    2 ++
 3 files changed, 21 insertions(+), 10 deletions(-)

--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -1178,17 +1178,13 @@ ttm_bo_swapout_cb(struct ttm_lru_walk *w
                bdev->funcs->swap_notify(bo);
 
        if (ttm_tt_is_populated(tt)) {
-               spin_lock(&bdev->lru_lock);
-               ttm_resource_del_bulk_move(bo->resource, bo);
-               spin_unlock(&bdev->lru_lock);
-
                ret = ttm_tt_swapout(bdev, tt, swapout_walk->gfp_flags);
-
-               spin_lock(&bdev->lru_lock);
-               if (ret)
-                       ttm_resource_add_bulk_move(bo->resource, bo);
-               ttm_resource_move_to_lru_tail(bo->resource);
-               spin_unlock(&bdev->lru_lock);
+               if (!ret) {
+                       spin_lock(&bdev->lru_lock);
+                       ttm_resource_del_bulk_move_unevictable(bo->resource, 
bo);
+                       ttm_resource_move_to_lru_tail(bo->resource);
+                       spin_unlock(&bdev->lru_lock);
+               }
        }
 
 out:
--- a/drivers/gpu/drm/ttm/ttm_resource.c
+++ b/drivers/gpu/drm/ttm/ttm_resource.c
@@ -292,6 +292,19 @@ void ttm_resource_del_bulk_move(struct t
                ttm_lru_bulk_move_del(bo->bulk_move, res);
 }
 
+/*
+ * Remove a resource from its bulk_move, bypassing the unevictable check.
+ * Use only when the resource is known to still be tracked in the range despite
+ * the BO having just become unevictable; asserts that this is the case.
+ */
+void ttm_resource_del_bulk_move_unevictable(struct ttm_resource *res,
+                                           struct ttm_buffer_object *bo)
+{
+       WARN_ON_ONCE(!ttm_resource_unevictable(res, bo));
+       if (bo->bulk_move)
+               ttm_lru_bulk_move_del(bo->bulk_move, res);
+}
+
 /* Move a resource to the LRU or bulk tail */
 void ttm_resource_move_to_lru_tail(struct ttm_resource *res)
 {
--- a/include/drm/ttm/ttm_resource.h
+++ b/include/drm/ttm/ttm_resource.h
@@ -448,6 +448,8 @@ void ttm_resource_add_bulk_move(struct t
                                struct ttm_buffer_object *bo);
 void ttm_resource_del_bulk_move(struct ttm_resource *res,
                                struct ttm_buffer_object *bo);
+void ttm_resource_del_bulk_move_unevictable(struct ttm_resource *res,
+                                           struct ttm_buffer_object *bo);
 void ttm_resource_move_to_lru_tail(struct ttm_resource *res);
 
 void ttm_resource_init(struct ttm_buffer_object *bo,


Patches currently in stable-queue which might be from 
[email protected] are

queue-7.0/drm-xe-dma-buf-fix-uaf-with-retry-loop.patch
queue-7.0/drm-ttm-fix-ttm_bo_shrink-infinite-lru-walk-on-backup-failure.patch
queue-7.0/drm-ttm-fix-ttm_bo_swapout-infinite-lru-walk-on-swapout-failure.patch
queue-7.0/drm-gpusvm-fix-unbalanced-unlock-in-drm_gpusvm_scan_.patch
queue-7.0/drm-xe-xelp-fix-wa_18022495364.patch
queue-7.0/drm-xe-dma-buf-handle-empty-bo-and-uaf-races.patch
queue-7.0/drm-ttm-convert-eagain-from-dmem_cgroup_try_charge-to-enospc.patch

Reply via email to