Bug 93201

Summary:

When running a basic OpenGL program with DRI_PRIME=1, Xorg crashes

Product:

xorg

Reporter:

Marcos Souza <marcos.souza.org>

Component:

Driver/nouveau

Assignee:

Nouveau Project <nouveau>

Status:

RESOLVED MOVED

QA Contact:

Xorg Project Team <xorg-team>

Severity:

normal

Priority:

medium

CC:

amillhouse

Version:

unspecified

Hardware:

x86-64 (AMD64)

OS:

Linux (All)

Whiteboard:

i915 platform:

i915 features:

Attachments:

Description	Flags
dmesg	none
Xorg old, showing the crash	none
current Xorg.0.log	none

Description Marcos Souza 2015-12-01 23:23:22 UTC

Created attachment 120247 [details]
dmesg

The program app is here: https://github.com/marcosps/experimentations/blob/master/glsl/test_shader.cxx

I just load two shaders and the draw a triangle. The same program works without DRI_PRIME=1.

I already compiled the xf86-video-intel, but I don't know how to setup the new driver in Fedora 23... Can this be related to DRI2?

Please let me know if you you need any additional info.

Comment 1 Marcos Souza 2015-12-01 23:24:34 UTC

Created attachment 120248 [details]
Xorg old, showing the crash

Comment 2 Marcos Souza 2015-12-01 23:25:02 UTC

Created attachment 120249 [details]
current Xorg.0.log

Comment 3 Marcos Souza 2015-12-01 23:48:10 UTC

ABRT also created a bug in RedHat bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1286874

Comment 4 Marcos Souza 2015-12-02 00:26:43 UTC

When requesting the intel driver to use DRI3, I can't see the crashes:

[marcos@localhost glsl]$ cat /etc/X11/xorg.conf.d/30-dri3.conf 
Section "Device"
	Identifier "isrequired"
	Driver "intel"
	Option "AccelMethod" "sna"
	Option "DRI" "3"
#	Driver "modesetting"
#	Option "AccelMethod" "glamor"
#	Option "AccelMethod" "none"
#	Option "ShadowFB" "off"
EndSection


But instead, I get a black window, instead of my shaders...

Comment 5 Chris Wilson 2015-12-02 12:14:17 UTC

I'm going to assume the DRICloseScreen in the middle of the stack is just the unwinder getting confused and so this looks like a nouveau bug and not a core DRI2 bug.

Comment 6 Marcos Souza 2015-12-03 00:40:29 UTC

(In reply to Chris Wilson from comment #5)
> I'm going to assume the DRICloseScreen in the middle of the stack is just
> the unwinder getting confused and so this looks like a nouveau bug and not a
> core DRI2 bug.

So, in this case, there is a start point to look at to solve this problem? I would like ot help fixing this issue...

Thanks,

Comment 7 Martin Peres 2019-12-04 09:08:05 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/xorg/driver/xf86-video-nouveau/issues/242.

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.