| Message ID | 20230903170736.513347-7-dmitry.osipenko@collabora.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | Add generic memory shrinker to VirtIO-GPU and Panfrost DRM drivers | expand |
On Sun, 3 Sep 2023 20:07:22 +0300 Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote: > Prepare virtio_gpu_object_create() to addition of memory shrinker support > by replacing open-coded drm_gem_shmem_free() with drm_gem_object_put() that > decrements GEM refcount to 0, which becomes important for drm-shmem because > it will start to use GEM's refcount during the shmem's BO freeing time in > order to prevent spurious lockdep warning about resv lock ordering vs > fs_reclaim code paths. I think I'm okay with the change (assuming virtio_gpu_free_object() can deal with partially initialized objects), not with the explanation :-). I don't really see why we need to take the resv lock in drm_gem_shmem_free(). As said in my v15 review, I think we should replace the drm_gem_shmem_put_pages() call we have in drm_gem_shmem_free() by a call to a new drm_gem_shmem_free_pages() helper that does exactly what drm_gem_shmem_put_pages() does without the refcounting/locking, because all that should remain at the time drm_gem_shmem_free() is called is the implicit pages ref owned by shmem->sgt, and there's no risk of other threads accessing the GEM object at that point. > > Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com> > --- > drivers/gpu/drm/virtio/virtgpu_object.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c > index c7e74cf13022..343b13428125 100644 > --- a/drivers/gpu/drm/virtio/virtgpu_object.c > +++ b/drivers/gpu/drm/virtio/virtgpu_object.c > @@ -244,6 +244,6 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, > err_put_id: > virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle); > err_free_gem: > - drm_gem_shmem_free(shmem_obj); > + drm_gem_object_put(&bo->base.base); > return ret; > }
On 9/5/23 10:20, Boris Brezillon wrote: > On Sun, 3 Sep 2023 20:07:22 +0300 > Dmitry Osipenko <dmitry.osipenko@collabora.com> wrote: > >> Prepare virtio_gpu_object_create() to addition of memory shrinker support >> by replacing open-coded drm_gem_shmem_free() with drm_gem_object_put() that >> decrements GEM refcount to 0, which becomes important for drm-shmem because >> it will start to use GEM's refcount during the shmem's BO freeing time in >> order to prevent spurious lockdep warning about resv lock ordering vs >> fs_reclaim code paths. > > I think I'm okay with the change (assuming virtio_gpu_free_object() > can deal with partially initialized objects), not with the explanation > :-). I don't really see why we need to take the resv lock in > drm_gem_shmem_free(). As said in my v15 review, I think we should > replace the drm_gem_shmem_put_pages() call we have in > drm_gem_shmem_free() by a call to a new drm_gem_shmem_free_pages() > helper that does exactly what drm_gem_shmem_put_pages() does without > the refcounting/locking, because all that should remain at the time > drm_gem_shmem_free() is called is the implicit pages ref owned by > shmem->sgt, and there's no risk of other threads accessing the GEM > object at that point. Apparently I forgot to drop these drm_gem_object_put() patches by accident in v16, the drm_gem_shmem_free() doesn't touch resv lock anymore. Will re-check for v17.
diff --git a/drivers/gpu/drm/virtio/virtgpu_object.c b/drivers/gpu/drm/virtio/virtgpu_object.c index c7e74cf13022..343b13428125 100644 --- a/drivers/gpu/drm/virtio/virtgpu_object.c +++ b/drivers/gpu/drm/virtio/virtgpu_object.c @@ -244,6 +244,6 @@ int virtio_gpu_object_create(struct virtio_gpu_device *vgdev, err_put_id: virtio_gpu_resource_id_put(vgdev, bo->hw_res_handle); err_free_gem: - drm_gem_shmem_free(shmem_obj); + drm_gem_object_put(&bo->base.base); return ret; }
Prepare virtio_gpu_object_create() to addition of memory shrinker support by replacing open-coded drm_gem_shmem_free() with drm_gem_object_put() that decrements GEM refcount to 0, which becomes important for drm-shmem because it will start to use GEM's refcount during the shmem's BO freeing time in order to prevent spurious lockdep warning about resv lock ordering vs fs_reclaim code paths. Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com> --- drivers/gpu/drm/virtio/virtgpu_object.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)