drm/ttm: Improved fencing of buffer object lists

Drastically reduce the number of spin lock / unlock operations by performing
unreserving and fencing under global locks.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Jerome Glisse <j.glisse@redhat.com>
Signed-off-by: Dave Airlie <airlied@redhat.com>

2 files changed, 33 insertions, 14 deletions

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index d93c73b1c471..551a5d31cadf 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -299,14 +299,19 @@ int ttm_bo_reserve(struct ttm_buffer_object *bo,
         return ret;
 }
 
+void ttm_bo_unreserve_locked(struct ttm_buffer_object *bo)
+{
+        ttm_bo_add_to_lru(bo);
+        atomic_set(&bo->reserved, 0);
+        wake_up_all(&bo->event_queue);
+}
+
 void ttm_bo_unreserve(struct ttm_buffer_object *bo)
 {
         struct ttm_bo_global *glob = bo->glob;
 
         spin_lock(&glob->lru_lock);
-        ttm_bo_add_to_lru(bo);
-        atomic_set(&bo->reserved, 0);
-        wake_up_all(&bo->event_queue);
+        ttm_bo_unreserve_locked(bo);
         spin_unlock(&glob->lru_lock);
 }
 EXPORT_SYMBOL(ttm_bo_unreserve);
diff --git a/drivers/gpu/drm/ttm/ttm_execbuf_util.c b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
index c3a2100bace6..b6da65cc502a 100644
--- a/drivers/gpu/drm/ttm/ttm_execbuf_util.c
+++ b/drivers/gpu/drm/ttm/ttm_execbuf_util.c
@@ -200,22 +200,36 @@ EXPORT_SYMBOL(ttm_eu_reserve_buffers);
 void ttm_eu_fence_buffer_objects(struct list_head *list, void *sync_obj)
 {
         struct ttm_validate_buffer *entry;
+        struct ttm_buffer_object *bo;
+        struct ttm_bo_global *glob;
+        struct ttm_bo_device *bdev;
+        struct ttm_bo_driver *driver;
 
-        list_for_each_entry(entry, list, head) {
-                struct ttm_buffer_object *bo = entry->bo;
-                struct ttm_bo_device *bdev = bo->bdev;
-                struct ttm_bo_driver *driver = bdev->driver;
-                void *old_sync_obj;
+        if (list_empty(list))
+                return;
+
+        bo = list_first_entry(list, struct ttm_validate_buffer, head)->bo;
+        bdev = bo->bdev;
+        driver = bdev->driver;
+        glob = bo->glob;
 
-                spin_lock(&bdev->fence_lock);
-                old_sync_obj = bo->sync_obj;
+        spin_lock(&bdev->fence_lock);
+        spin_lock(&glob->lru_lock);
+
+        list_for_each_entry(entry, list, head) {
+                bo = entry->bo;
+                entry->old_sync_obj = bo->sync_obj;
                 bo->sync_obj = driver->sync_obj_ref(sync_obj);
                 bo->sync_obj_arg = entry->new_sync_obj_arg;
-                spin_unlock(&bdev->fence_lock);
-                ttm_bo_unreserve(bo);
+                ttm_bo_unreserve_locked(bo);
                 entry->reserved = false;
-                if (old_sync_obj)
-                        driver->sync_obj_unref(&old_sync_obj);
+        }
+        spin_unlock(&glob->lru_lock);
+        spin_unlock(&bdev->fence_lock);
+
+        list_for_each_entry(entry, list, head) {
+                if (entry->old_sync_obj)
+                        driver->sync_obj_unref(&entry->old_sync_obj);
         }
 }
 EXPORT_SYMBOL(ttm_eu_fence_buffer_objects);