Bug 71438

The relevant bit of the stacktrace:

#10 0x42834a87 in __assert_fail () from /lib/libc.so.6
#11 0xaf348070 in nouveau_scratch_init () from /usr/lib/dri/nouveau_vieux_dri.so
#12 0xaf3436ed in nouveau_context_init () from /usr/lib/dri/nouveau_vieux_dri.so
#13 0xaf34d811 in nv10_context_create () from /usr/lib/dri/nouveau_vieux_dri.so
#14 0xaf343abd in nouveau_context_create () from /usr/lib/dri/nouveau_vieux_dri.so
#15 0xaf36045f in dri2CreateContextAttribs () from /usr/lib/dri/nouveau_vieux_dri.so
#16 0xaf360638 in dri2CreateNewContext () from /usr/lib/dri/nouveau_vieux_dri.so
#17 0xb4032024 in dri2_create_context () from /lib/libGL.so.1
#18 0xb400824a in CreateContext () from /lib/libGL.so.1
#19 0xb4008541 in glXCreateContext () from /lib/libGL.so.1

And the code in nouveau_scratch_init:

	for (i = 0; i < NOUVEAU_SCRATCH_COUNT; i++) {
		ret = nouveau_bo_new(context_dev(ctx), NOUVEAU_BO_GART |
				     NOUVEAU_BO_MAP, 0, NOUVEAU_SCRATCH_SIZE,
				     NULL, &scratch->bo[i]);
		assert(!ret);

Looks like it's trying to get 2 3MB bo's, but it can't. Unfortunately error handling is often done rather poorly in the nouveau_vieux driver (via asserts rather than returning error codes).

What card are you using? How much VRAM does it have? What resolution are you running your screen at? What kernel are you on? What version of xf86-video-nouveau and mesa are you using?

Oh, and another relevant question -- what is the GART size? Actually, just attach your dmesg, that should answer a bunch of questions. (Would still want to know versions of the various things I mentioned in the previous comment.)

Created attachment 88949 [details]
attachment-29197-0.html

1. Dmesg output attached.

2. Kernel is :kernel-PAE.i686 3.11.6

3. Display resolution is 1920x1200

4. Video card : IBM Intellistation Graphics Card FRU IBM 25P6684 NVIDIA Quadro4 200NVS 64MB

5. xorg-x11-drv-nouveau.i686        1:1.0.9-1.fc19

6. mesa-dri-drivers.i686                 9.2-1.20130919.fc19
   mesa-filesystem.i686                  9.2-1.20130919.fc19
   mesa-libEGL.i686                       9.2-1.20130919.fc19
   mesa-libGL.i686                         9.2-1.20130919.fc19
   mesa-libGLU.i686                      9.0.0-2.fc19
   mesa-libgbm.i686                        9.2-1.20130919.fc19
   mesa-libglapi.i686                       9.2-1.20130919.fc19
   mesa-libwayland-egl.i686           9.2-1.20130919.fc19
   mesa-libxatracker.i686                9.2-1.20130919.fc19

7. OS  is Linux Fedora 19

8. System has 4GB memory.


________________________________
From: bugzilla-daemon@freedesktop.org <bugzilla-daemon@freedesktop.org>
Sent: November-09-13 17:40
To: Brian Kaye
Subject: [Bug 71438] [NV10/nouveau_vieux] Application Konqueror crashes when connecting to asus.com

Ilia Mirkin<mailto:imirkin@alum.mit.edu> changed bug 71438<https://bugs.freedesktop.org/show_bug.cgi?id=71438>
What    Removed Added
Summary Application Konqueror crashes when connecting to asus.com       [NV10/nouveau_vieux] Application Konqueror crashes when connecting to asus.com

Comment # 1<https://bugs.freedesktop.org/show_bug.cgi?id=71438#c1> on bug 71438<https://bugs.freedesktop.org/show_bug.cgi?id=71438> from Ilia Mirkin<mailto:imirkin@alum.mit.edu>

The relevant bit of the stacktrace:

#10 0x42834a87 in __assert_fail () from /lib/libc.so.6
#11 0xaf348070 in nouveau_scratch_init () from
/usr/lib/dri/nouveau_vieux_dri.so
#12 0xaf3436ed in nouveau_context_init () from
/usr/lib/dri/nouveau_vieux_dri.so
#13 0xaf34d811 in nv10_context_create () from /usr/lib/dri/nouveau_vieux_dri.so
#14 0xaf343abd in nouveau_context_create () from
/usr/lib/dri/nouveau_vieux_dri.so
#15 0xaf36045f in dri2CreateContextAttribs () from
/usr/lib/dri/nouveau_vieux_dri.so
#16 0xaf360638 in dri2CreateNewContext () from
/usr/lib/dri/nouveau_vieux_dri.so
#17 0xb4032024 in dri2_create_context () from /lib/libGL.so.1
#18 0xb400824a in CreateContext () from /lib/libGL.so.1
#19 0xb4008541 in glXCreateContext () from /lib/libGL.so.1

And the code in nouveau_scratch_init:

    for (i = 0; i < NOUVEAU_SCRATCH_COUNT; i++) {
        ret = nouveau_bo_new(context_dev(ctx), NOUVEAU_BO_GART |
                     NOUVEAU_BO_MAP, 0, NOUVEAU_SCRATCH_SIZE,
                     NULL, &scratch->bo[i]);
        assert(!ret);

Looks like it's trying to get 2 3MB bo's, but it can't. Unfortunately error
handling is often done rather poorly in the nouveau_vieux driver (via asserts
rather than returning error codes).

What card are you using? How much VRAM does it have? What resolution are you
running your screen at? What kernel are you on? What version of
xf86-video-nouveau and mesa are you using?

________________________________
You are receiving this mail because:

  *   You reported the bug.

Created attachment 88950 [details]
attachment-29197-1.dat

Created attachment 88951 [details]
Konqueror-dmesg.txt

OK, well it seems like eventually you start getting allocation errors. This could mean that we're somehow leaking video memory, or just not using it wisely, or that there's just not enough to go 'round. Modern applications will want a lot more 3d than this card can provide anyways... unless you plan on using it for 3d activities, you could set LIBGL_ALWAYS_SOFTWARE=1 somewhere in your environment which will cause mesa to fall back on llvmpipe or swrast.

Another possibility is that you're being bitten by the libdrm bug that was exposed by gcc-4.8 (which I see your kernel is compiled with). Try a libdrm with commit http://cgit.freedesktop.org/mesa/drm/commit/?id=482abbfafb56cbceaf5355c026434e638cddd0f1 applied and see if it helps.

-- 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/71.