X11 crash on running application 2nd time on custom target bassed on i.MX6S

[vikaspatil]

Hi All,

I am encountering the following issue on both the release of graphics driver (i.e. 3.10.17-1.0.1 and 3.10.17-1.0.1) on i.MX6s based custom board. I have tried 3.10.17-1.0.0  and 3.10.17-1.0.1 version of recipe's for graphics driver and x ddx driver from daisy branch of meta-fsl-arm and also took the binaries of xorg 1.14.4 after building daisy branch using community-bsp to our custom rootfs.

Is this the known issue? Is there any plan to fix this? Could you give some inputs regarding how should I debug and fix if possible?

When I run some x11/gles2.0 example first time after starting the X11 (1.14.4), it works fine, however when I try second time I am seeing following errors. In short X11/GLES2.0 application on exit or killing makes X11 unusable and need to restart the X11 to run the application second time. Is this you also observed?

root@imx6solosabreauto:/opt/

viv_samples/vdk# ./tutorial2_es20 -f 100
systemd[1]: xorg.service: main process exited, code=killed, status=6/ABRT
XIO:  fatal IO error 11 (Resource temporarily unavailable) on X server ":0.0"
      after 28 requests (28 known processed) with 0 events remaining.
systemd[1]: Unit xorg.service entered failed state.

I also see following in xorg log

[    49.212] (**) Option "XkbModel" "pc105"
[    49.213] (**) Option "XkbLayout" "us"
[    49.349] (II) config/udev: Adding drm device (/dev/dri/card0)
[    54.636] (EE)
[    54.636] (EE) Backtrace:
[    54.636] (EE)
[    54.637] (EE) Segmentation fault at address 0x65d24000
[    54.637] (EE)
Fatal server error:
[    54.638] (EE) Caught signal 11 (Segmentation fault). Server aborting
[    54.638] (EE)
[    54.638] (EE)

Regards & thanks,
Vikash

[Bio_TICFSL · Accepted solution]

Hi Vikash,

I try to reproduce this issue without success, however I tried on Mx6Q, and tutorial2 runs well, So  Please try to use the 3.10.17-1.02 version that Includes a graphics driver update with memory, x11 support improvements, deadlock and bug hang fixes

http://www.freescale.com/webapp/sps/site/prod_summary.jsp?code=i.MX6Q&fpsp=1&tab=Design_Tools_Tab

If the problem persist please send us the step by step instruction to reproduce the issue on our side with Freescale boards.

Regards

[vikaspatil]

Thanks for your reply.

It means you can run tutorial2_es02 (or any other application based on x11/gles2.0 api) multiple times on release 3.10.17-1.0.2 however I am using 3.10.17-1.0.0 (only kernel, graphics driver, xddx and xorg). Is it possible to test this on 3.10.17-1.0.0 or  3.10.17-1.0.1 too?

I will try 3.10.17-1.0.2 when it is upstream in daisy or dizzy branch of meta-fsl-arm https://github.com/Freescale/meta-fsl-arm and update.

To reproduce the issue:

- Boot the target

- Start X11

- export DISPLAY=:0.0

- cd /opt/viv_samples/vdk

- ./tutorial2_es02 -f 100 (after this completes run it again)

- ./tutorial2_es02 -f 100 (can't run, seeing the errors I mentioned, to run it again I need to start x11 again)

Regards & Thanks,

Vikash

[vikaspatil]

Now it is working fine. I reinstalled rootfs.

Regards,

Vikash