On Sat, 2011-05-07 at 18:03 +0200, Daniel Vetter wrote:
Otherwise we have a use-after free.
Tested-and-Reported-by: Bruno Prémont bonbons@linux-vserver.org Signed-off-by: Daniel Vetter daniel.vetter@ffwll.ch
Ah, we actually have a patch in the nouveau git tree fixing this already.
I'll get this upstream ASAP.
Ben.
drivers/gpu/drm/nouveau/nouveau_mem.c | 2 -- drivers/gpu/drm/nouveau/nouveau_state.c | 2 ++ 2 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/nouveau/nouveau_mem.c b/drivers/gpu/drm/nouveau/nouveau_mem.c index 5045f8b..c3e953b 100644 --- a/drivers/gpu/drm/nouveau/nouveau_mem.c +++ b/drivers/gpu/drm/nouveau/nouveau_mem.c @@ -152,8 +152,6 @@ nouveau_mem_vram_fini(struct drm_device *dev) { struct drm_nouveau_private *dev_priv = dev->dev_private;
nouveau_bo_ref(NULL, &dev_priv->vga_ram);
ttm_bo_device_release(&dev_priv->ttm.bdev);
nouveau_ttm_global_release(dev_priv);
diff --git a/drivers/gpu/drm/nouveau/nouveau_state.c b/drivers/gpu/drm/nouveau/nouveau_state.c index a30adec..1fe6503 100644 --- a/drivers/gpu/drm/nouveau/nouveau_state.c +++ b/drivers/gpu/drm/nouveau/nouveau_state.c @@ -768,6 +768,8 @@ static void nouveau_card_takedown(struct drm_device *dev) engine->mc.takedown(dev); engine->display.late_takedown(dev);
- nouveau_bo_ref(NULL, &dev_priv->vga_ram);
- mutex_lock(&dev->struct_mutex); ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_VRAM); ttm_bo_clean_mm(&dev_priv->ttm.bdev, TTM_PL_TT);