[PATCH v2 1/2] drm/nouveau: fix stale locked mutex in nouveau_gem_ioctl_pushbuf

Karol Herbst kherbst at redhat.com
Tue Mar 5 13:38:52 UTC 2024


If VM_BIND is enabled on the client the legacy submission ioctl can't be
used, however if a client tries to do so regardless it will return an
error. In this case the clients mutex remained unlocked leading to a
deadlock inside nouveau_drm_postclose or any other nouveau ioctl call.

Fixes: b88baab82871 ("drm/nouveau: implement new VM_BIND uAPI")
Cc: Danilo Krummrich <dakr at redhat.com>
Cc: <stable at vger.kernel.org> # v6.6+
Signed-off-by: Karol Herbst <kherbst at redhat.com>
Reviewed-by: Lyude Paul <lyude at redhat.com>
Reviewed-by: Danilo Krummrich <dakr at redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240304183157.1587152-1-kherbst@redhat.com
---
 drivers/gpu/drm/nouveau/nouveau_gem.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 49c2bcbef1299..5a887d67dc0e8 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -764,7 +764,7 @@ nouveau_gem_ioctl_pushbuf(struct drm_device *dev, void *data,
 		return -ENOMEM;
 
 	if (unlikely(nouveau_cli_uvmm(cli)))
-		return -ENOSYS;
+		return nouveau_abi16_put(abi16, -ENOSYS);
 
 	list_for_each_entry(temp, &abi16->channels, head) {
 		if (temp->chan->chid == req->channel) {
-- 
2.44.0



More information about the dri-devel mailing list