[PATCH 2/3] drm/amd/display: Skip determining update type for async updates

[Alex Deucher <alexdeucher@gmail.com>]

From: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>

[Why]
By passing through the dm_determine_update_type_for_commit for atomic
commits that can be done asynchronously we are incurring a
performance penalty by locking access to the global private object
and holding that access until the end of the programming sequence.

This is also allocating a new large dc_state on every access in addition
to retaining all the references on each stream and plane until the end
of the programming sequence.

[How]
Shift the determination for async update before validation. Return early
if it's going to be an async update.

Bug: https://bugzilla.kernel.org/show_bug.cgi?id=204181
Signed-off-by: Nicholas Kazlauskas <nicholas.kazlauskas@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: David Francis <david.francis@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 43d10d30df156f7834fa91aecb69614fefc8bb0a)
Cc: stable@vger.kernel.org
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 27 ++++++++++++++-----
 1 file changed, 20 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index ab341fca9647..8df49740518e 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -7274,6 +7274,26 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 	if (ret)
 		goto fail;
 
+	if (state->legacy_cursor_update) {
+		/*
+		 * This is a fast cursor update coming from the plane update
+		 * helper, check if it can be done asynchronously for better
+		 * performance.
+		 */
+		state->async_update =
+			!drm_atomic_helper_async_check(dev, state);
+
+		/*
+		 * Skip the remaining global validation if this is an async
+		 * update. Cursor updates can be done without affecting
+		 * state or bandwidth calcs and this avoids the performance
+		 * penalty of locking the private state object and
+		 * allocating a new dc_state.
+		 */
+		if (state->async_update)
+			return 0;
+	}
+
 	/* Check scaling and underscan changes*/
 	/* TODO Removed scaling changes validation due to inability to commit
 	 * new stream into context w\o causing full reset. Need to
@@ -7326,13 +7346,6 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 			ret = -EINVAL;
 			goto fail;
 		}
-	} else if (state->legacy_cursor_update) {
-		/*
-		 * This is a fast cursor update coming from the plane update
-		 * helper, check if it can be done asynchronously for better
-		 * performance.
-		 */
-		state->async_update = !drm_atomic_helper_async_check(dev, state);
 	}
 
 	/* Must be success */
-- 
2.20.1