Skip to contentSkip to conversation

amdgpu DC fixes for stable

PATCH In review Alex Deucher 3 participants 7 messages Sep 20, 2019 – Sep 22, 2019

Subsystems (4)
lore
Share & tools
lore
#0 lore raw

cover amdgpu DC fixes for stable

This set of patches is cherry-picked from 5.4 to stable to fix:
https://bugzilla.kernel.org/show_bug.cgi?id=204181

Please apply!

Thanks,

Alex

Nicholas Kazlauskas (3):
drm/amd/display: Allow cursor async updates for framebuffer swaps drm/amd/display: Skip determining update type for async updates drm/amd/display: Don't replace the dc_state for fast updates

 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 56 ++++++++++++++-----
 1 file changed, 41 insertions(+), 15 deletions(-)
Signature · 1 line
2.20.1
#1 lore raw

PATCH 1/3 drm/amd/display: Allow cursor async updates for framebuffer swaps

From: Nicholas Kazlauskas <[email protected]>

[Why]
We previously allowed framebuffer swaps as async updates for cursor planes but had to disable them due to a bug in DRM with async update handling and incorrect ref counting. The check to block framebuffer swaps has been added to DRM for a while now, so this check is redundant.

The real fix that allows this to properly in DRM has also finally been merged and is getting backported into stable branches, so dropping this now seems to be the right time to do so.

[How]
Drop the redundant check for old_fb != new_fb.

With the proper fix in DRM, this should also fix some cursor stuttering issues with xf86-video-amdgpu since it double buffers the cursor.

IGT tests that swap framebuffers (-varying-size for example) should also pass again.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 10 ----------
 1 file changed, 10 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +0 -10
index 45be7a2132bb..ab341fca9647 100644
@@ -4548,20 +4548,10 @@ static int dm_plane_atomic_check(struct drm_plane *plane,
 static int dm_plane_atomic_async_check(struct drm_plane *plane,
 				       struct drm_plane_state *new_plane_state)
 {
-	struct drm_plane_state *old_plane_state =
-		drm_atomic_get_old_plane_state(new_plane_state->state, plane);
-
 	/* Only support async updates on cursor planes. */
 	if (plane->type != DRM_PLANE_TYPE_CURSOR)
 		return -EINVAL;
 
-	/*
-	 * DRM calls prepare_fb and cleanup_fb on new_plane_state for
-	 * async commits so don't allow fb changes.
-	 */
-	if (old_plane_state->fb != new_plane_state->fb)
-		return -EINVAL;
-
 	return 0;
 }
 
Signature · 1 line
2.20.1
#2 lore raw

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

In reply to Alex Deucher

From: Nicholas Kazlauskas <[email protected]>

[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.

---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 27 ++++++++++++++-----
 1 file changed, 20 insertions(+), 7 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +20 -7
index ab341fca9647..8df49740518e 100644
@@ -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 */
Signature · 1 line
2.20.1
#3 lore raw

PATCH 3/3 drm/amd/display: Don't replace the dc_state for fast updates

In reply to Alex Deucher

From: Nicholas Kazlauskas <[email protected]>

[Why]
DRM private objects have no hw_done/flip_done fencing mechanism on their own and cannot be used to sequence commits accordingly.

When issuing commits that don't touch the same set of hardware resources like page-flips on different CRTCs we can run into the issue below because of this:

1. Client requests non-blocking Commit #1, has a new dc_state #1, state is swapped, commit tail is deferred to work queue

2. Client requests non-blocking Commit #2, has a new dc_state #2, state is swapped, commit tail is deferred to work queue

3. Commit #2 work starts, commit tail finishes,
atomic state is cleared, dc_state #1 is freed

4. Commit #1 work starts,
commit tail encounters null pointer deref on dc_state #1

In order to change the DC state as in the private object we need to ensure that we wait for all outstanding commits to finish and that any other pending commits must wait for the current one to finish as well.

We do this for MEDIUM and FULL updates. But not for FAST updates, nor would we want to since it would cause stuttering from the delays.

FAST updates that go through dm_determine_update_type_for_commit always create a new dc_state and lock the DRM private object if there are any changed planes.

We need the old state to validate, but we don't actually need the new state here.

[How]
If the commit isn't a full update then the use after free can be resolved by simply discarding the new state entirely and retaining the existing one instead.

With this change the sequence above can be reexamined. Commit #2 will still free Commit #1's reference, but before this happens we actually added an additional reference as part of Commit #2.

If an update comes in during this that needs to change the dc_state it will need to wait on Commit #1 and Commit #2 to finish. Then it'll swap the state, finish the work in commit tail and drop the last reference on Commit #2's dc_state.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=204181

Fixes 004b3938e637 ("drm/amd/display: Check scaling info when determing update type")
---
 .../gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 23 +++++++++++++++++++
 1 file changed, 23 insertions(+)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +23 -0
index 8df49740518e..beb2d268d1ef 100644
@@ -7346,6 +7346,29 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 			ret = -EINVAL;
 			goto fail;
 		}
+	} else {
+		/*
+		 * The commit is a fast update. Fast updates shouldn't change
+		 * the DC context, affect global validation, and can have their
+		 * commit work done in parallel with other commits not touching
+		 * the same resource. If we have a new DC context as part of
+		 * the DM atomic state from validation we need to free it and
+		 * retain the existing one instead.
+		 */
+		struct dm_atomic_state *new_dm_state, *old_dm_state;
+
+		new_dm_state = dm_atomic_get_new_state(state);
+		old_dm_state = dm_atomic_get_old_state(state);
+
+		if (new_dm_state && old_dm_state) {
+			if (new_dm_state->context)
+				dc_release_state(new_dm_state->context);
+
+			new_dm_state->context = old_dm_state->context;
+
+			if (old_dm_state->context)
+				dc_retain_state(old_dm_state->context);
+		}
 	}
 
 	/* Must be success */
Signature · 1 line
2.20.1
#5 lore raw
Quoted 9 lines

-----Original Message-----
From: Greg KH <[email protected]>
Sent: Friday, September 20, 2019 10:12 AM
To: Alex Deucher <[email protected]>
Cc: [email protected]; Kazlauskas, Nicholas <[email protected]>; Deucher, Alexander <[email protected]>
Subject: Re: [PATCH 0/3] amdgpu DC fixes for stable

Quoted 7 lines
Quoted 4 lines · Alex Deucher

This set of patches is cherry-picked from 5.4 to stable to fix:
https://bugzilla.kernel.org/show_bug.cgi?id=204181

Please apply!

What stable tree(s) do you wish to see this applied to?

5.3 and 5.2 would be great. Thanks!

Alex

Quoted 4 lines

thanks,

greg k-h

maintainer
#6 lore raw
Quoted 10 lines · Alexander
Quoted 10 lines

-----Original Message-----
From: Greg KH <[email protected]>
Sent: Friday, September 20, 2019 10:12 AM
To: Alex Deucher <[email protected]>
Cc: [email protected]; Kazlauskas, Nicholas <[email protected]>; Deucher, Alexander <[email protected]>
Subject: Re: [PATCH 0/3] amdgpu DC fixes for stable

On Fri, Sep 20, 2019 at 09:03:35AM -0500, Alex Deucher wrote:

Quoted 8 lines
Quoted 6 lines
Quoted 4 lines

This set of patches is cherry-picked from 5.4 to stable to fix:
https://bugzilla.kernel.org/show_bug.cgi?id=204181

Please apply!

What stable tree(s) do you wish to see this applied to?

5.3 and 5.2 would be great. Thanks!

All queued up now, thanks.

greg k-h