Bug 110614

Summary: [Regression] Freeze at desktop manager startup
Product: Mesa Reporter: raffarti
Component: Drivers/Gallium/radeonsiAssignee: Default DRI bug account <dri-devel>
Status: RESOLVED MOVED QA Contact: Default DRI bug account <dri-devel>
Severity: normal    
Priority: medium CC: julien.isorce
Version: unspecified   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments: backtrace from journal

Description raffarti 2019-05-05 15:01:19 UTC 
Freeze at desktop manager startup (tested with sddm and lightdm).

Bisected as

commit 1cec049d4db1c4dcd121bad17df4a77273dd9bb1
Author: Julien Isorce <jisorce@oblong.com>
Date:   Tue Apr 23 14:28:48 2019 -0700

    radeonsi: implement resource_get_info
    
    Re-use existing si_texture_get_offset.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110443
    Signed-off-by: Julien Isorce <jisorce@oblong.com>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>

commit a3c202de0a963c0562796cf75e3a9b3eedf1afad
Author: Julien Isorce <jisorce@oblong.com>
Date:   Tue Apr 23 14:26:33 2019 -0700

    gallium: add resource_get_info to pipe_screen
    
    Generic plumbing.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=110443
    Signed-off-by: Julien Isorce <jisorce@oblong.com>
    Reviewed-by: Marek Olšák <marek.olsak@amd.com>

GPU: RX 480
OS: OpenSuse Tumbleweed
Comment 1 raffarti 2019-05-05 15:20:10 UTC 
Created attachment 144166 [details]
backtrace from journal

Debugoptimized build, trace from journalctl.
Comment 2 Julien Isorce 2019-05-06 14:35:31 UTC 
Can you try with latest master ? Should be fixed with:
https://cgit.freedesktop.org/mesa/mesa/commit/?id=8cd71f399e73c5d87e9162cc74da76e317a9f41f
Sorry for the inconvenience
Comment 3 raffarti 2019-05-06 15:13:23 UTC 
Tested both master and 8cd71f399e73c5d87e9162cc74da76e317a9f41f specifically, but it still hangs and the stacktrace looks the same.
Any way I can help debug this?
Comment 4 Julien Isorce 2019-05-06 15:22:50 UTC 
So somehow the factoring is incorrect https://cgit.freedesktop.org/mesa/mesa/commit/?id=1cec049d4db1c4dcd121bad17df4a77273dd9bb1

If you can log the values for offset and stride before and after this patch it might help to understand the issue. Thx
Comment 5 raffarti 2019-05-06 18:36:08 UTC 
Maybe I should have stated so explicitely, ec6c2297634eba77248a929048cf4201887a5f0a is the first commit to crash. It looks like it crashes even before calling si_texture_get_handle as there's nothing logged on bad builds. Possibly there is something wrong with my setup? I'm only compiling mesa from git, all other libs are from repos.
Comment 7 raffarti 2019-05-07 07:53:21 UTC 
While trying to use the coredump, I've noticed /usr/lib64/dri/radeonsi_dri.so was used, so I've tried checking Xorg process with lsof in the good build... turns out both /usr/lib64/dri/radeonsi_dri.so and /opt/mesa/lib64/dri/radeonsi_dri.so were listed. With bad builds, renaming /usr/lib64/dri results in kde working but glxgears and glxinfo failing, while symlinking it to opt makes everything fine.
So, looks like there's something wrong with my setup rather than mesa itself. I'll keep investigating why this happens but I guess this is invalid anyway. Sorry for the trouble.
Comment 8 GitLab Migration User 2019-09-25 18:49:44 UTC 
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1398.

Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.