Bug 7811 - Buffer corruption after resizing wayland client
Summary: Buffer corruption after resizing wayland client
Status: RESOLVED OBSOLETE
Alias: None
Product: BSPs
Classification: Build System, Metadata & Runtime
Component: bsps-meta-fsl-arm (show other bugs)
Version: 1.8
Hardware: Other arm
: Medium normal
Target Milestone: Future
Assignee: Thomas McGuire
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2015-05-27 13:15 UTC by Thomas McGuire
Modified: 2018-10-12 16:38 UTC (History)
0 users

See Also:
OS type for building Yocto: ---
Type of Regression: New (Never tested)
Verified:
Documentation change: No (bug/feature does not impact docs)


Attachments
Photo of the buffer corruption in action (1.25 MB, image/jpeg)
2015-05-27 13:15 UTC, Thomas McGuire
no flags Details
Minimal testcase to reproduce the problem (36.53 KB, application/x-gzip)
2015-05-27 13:15 UTC, Thomas McGuire
no flags Details

Note You need to log in before you can comment on or make changes to this bug.
Description Thomas McGuire 2015-05-27 13:15:05 UTC
Created attachment 2525 [details]
Photo of the buffer corruption in action

When resizing a wayland client using wl_egl_window_resize(), the buffer for that frame gets corrupted. This seems to be a bug in the Vivante GPU driver.

Tested on dizzy on on fido, reproducible in both.

A workaround of this issue would be to destroy and re-create the window with a new size, but then we run into a crash, see https://community.freescale.com/message/496179.

This is the second project I've been working on that had this problem - I assume resizing a window is quite a common operation, so this bug should be visible by many people using Wayland.

Since only the first one or two frames after a resize are affected, this mostly is a problem for applications that display mostly static content, such as user interfaces.
Comment 1 Thomas McGuire 2015-05-27 13:15:58 UTC
Created attachment 2526 [details]
Minimal testcase to reproduce the problem
Comment 2 Thomas McGuire 2015-05-27 13:18:13 UTC
The board I'm testing with is imx6qsabresd, using core-image-weston from fsl-community-bsp-platform (manually added "DISTRO_FEATURES_remove = "x11 directfb"" to local.conf to work around weston build failing, but otherwise pristine setup).

The attached testcase contains both a binary and the source code for it, including a README on how to build the binary from source.
Comment 3 Thomas McGuire 2016-01-13 15:19:10 UTC
According to https://community.freescale.com/message/603901, this is fixed in driver version 5.0.11.p7, which AFAIK is contained in Yocto jethro.
Comment 4 Thomas McGuire 2016-04-20 16:00:22 UTC
Just tested with Yocto Jethro, which has GPU driver version 5.0.11.p7.4. The problem persists.
Comment 5 Otavio Salvador 2018-01-26 11:44:53 UTC
Is it possible for you to try to reproduce in current release?
Comment 6 Thomas McGuire 2018-01-26 12:17:56 UTC
I can confirm the bug with Krogoth. I am not able to test any newer releases at the moment.
Comment 7 Otavio Salvador 2018-01-26 12:21:40 UTC
(In reply to comment #6)
> I can confirm the bug with Krogoth. I am not able to test any newer releases
> at the moment.

Great; it would be good if you could test with rocko and master.
Comment 8 Thomas McGuire 2018-02-19 12:33:17 UTC
I can't test with a newer Yocto version right now, so closing as RESOLVED/OBSOLETE as instructed by the bug triage email.