Maarten, for this and the other patches in this series,
I seem to recall we have this discussion before?
IIRC I stated that reservation was a too heavy-weight lock to hold to
determine whether a buffer was idle? It's a pretty nasty thing to build in.
/Thomas
On 01/21/2014 02:04 PM, Maarten Lankhorst wrote:
> Apart from some code inside ttm itself and nouveau_bo_vma_del,
> this is the only place where ttm_bo_wait is used without a reservation.
> Fix this so we can remove the fence_lock later on.
>
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst at canonical.com>
> ---
> drivers/gpu/drm/nouveau/nouveau_gem.c | 22 ++++++++++++++++++----
> 1 file changed, 18 insertions(+), 4 deletions(-)
>
> diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c
> b/drivers/gpu/drm/nouveau/nouveau_gem.c
> index 78a27f8ad7d9..24e9c58da8aa 100644
> --- a/drivers/gpu/drm/nouveau/nouveau_gem.c
> +++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
> @@ -894,17 +894,31 @@ nouveau_gem_ioctl_cpu_prep(struct drm_device *dev, void
> *data,
> struct drm_gem_object *gem;
> struct nouveau_bo *nvbo;
> bool no_wait = !!(req->flags & NOUVEAU_GEM_CPU_PREP_NOWAIT);
> - int ret = -EINVAL;
> + int ret;
> + struct nouveau_fence *fence = NULL;
>
> gem = drm_gem_object_lookup(dev, file_priv, req->handle);
> if (!gem)
> return -ENOENT;
> nvbo = nouveau_gem_object(gem);
>
> - spin_lock(&nvbo->bo.bdev->fence_lock);
> - ret = ttm_bo_wait(&nvbo->bo, true, true, no_wait);
> - spin_unlock(&nvbo->bo.bdev->fence_lock);
> + ret = ttm_bo_reserve(&nvbo->bo, true, false, false, 0);
> + if (!ret) {
> + spin_lock(&nvbo->bo.bdev->fence_lock);
> + ret = ttm_bo_wait(&nvbo->bo, true, true, true);
> + if (!no_wait && ret)
> + fence = nouveau_fence_ref(nvbo->bo.sync_obj);
> + spin_unlock(&nvbo->bo.bdev->fence_lock);
> +
> + ttm_bo_unreserve(&nvbo->bo);
> + }
> drm_gem_object_unreference_unlocked(gem);
> +
> + if (fence) {
> + ret = nouveau_fence_wait(fence, true, no_wait);
> + nouveau_fence_unref(&fence);
> + }
> +
> return ret;
> }
>
