It's unclear whether ->map is allowed to fail with -EINTR, but looking at current callers it's pretty clear that they don't expect this to happen. So use a blocking mutex_lock call. Since we don't wait for the gpu in our ->map callback the lack of the gpu hang checks doesn't matter.
Furthermore the goal is to eventually have per dma-buf locking done by callers with ww mutexes, so this will then be removed.
Cc: Chris Wilson chris@chris-wilson.co.uk Signed-off-by: Daniel Vetter daniel.vetter@ffwll.ch --- drivers/gpu/drm/i915/i915_gem_dmabuf.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_gem_dmabuf.c b/drivers/gpu/drm/i915/i915_gem_dmabuf.c index f7e1682..63c0818 100644 --- a/drivers/gpu/drm/i915/i915_gem_dmabuf.c +++ b/drivers/gpu/drm/i915/i915_gem_dmabuf.c @@ -35,9 +35,7 @@ static struct sg_table *i915_gem_map_dma_buf(struct dma_buf_attachment *attachme struct scatterlist *src, *dst; int ret, i;
- ret = i915_mutex_lock_interruptible(obj->base.dev); - if (ret) - return ERR_PTR(ret); + mutex_lock(&obj->base.dev->struct_mutex);
ret = i915_gem_object_get_pages(obj); if (ret) {