drm/vmwgfx: take fman->lock around fence list mutation in fifo_down
authorZack Rusin <zack.rusin@broadcom.com>
Tue, 5 May 2026 22:22:25 +0000 (18:22 -0400)
committerZack Rusin <zack.rusin@broadcom.com>
Mon, 27 Jul 2026 15:29:24 +0000 (11:29 -0400)
commit250af2e8c3e90dc978e062a936b633870a22e660
treebaa9246434b4272c4875467e070fc0eb7b52489c
parentf47d542d5912f236273399d7139b522f6950e2e4
drm/vmwgfx: take fman->lock around fence list mutation in fifo_down

vmw_fence_fifo_down() drops fman->lock to wait on a fence and, on
timeout, mutates fman->fence_list via list_del_init() and signals
the fence without re-acquiring the lock.  __vmw_fences_update() walks
and removes entries from the same list under fman->lock from any
other waiter, the fence-IRQ thread, or vmw_fences_update(), so the
unlocked list_del_init() can corrupt the list head.

Re-take fman->lock before manipulating fence->head and use
dma_fence_signal_locked().  Wrap the locked signalling in
dma_fence_begin_signalling() / dma_fence_end_signalling() so the
lockdep annotation that dma_fence_signal() previously provided is
preserved (the same pattern as __vmw_fences_update()).

dma_fence_put() is moved outside the lock to avoid a recursive
acquire from vmw_fence_obj_destroy(), which also takes fman->lock.

Fixes: ae2a104058e2 ("vmwgfx: Implement fence objects")
Cc: stable@vger.kernel.org
Assisted-by: Claude:claude-opus-4.7
Signed-off-by: Zack Rusin <zack.rusin@broadcom.com>
Reviewed-by: Ian Forbes <ian.forbes@broadcom.com>
Link: https://patch.msgid.link/20260505222728.519626-5-zack.rusin@broadcom.com
drivers/gpu/drm/vmwgfx/vmwgfx_fence.c