Bug 15639 - [GM965 WINE] assert in dri_bufmgr_fake.c:983: dri_fake_emit_reloc() triggered
Summary: [GM965 WINE] assert in dri_bufmgr_fake.c:983: dri_fake_emit_reloc() triggered
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Eric Anholt
QA Contact:
URL:
Whiteboard:
Keywords:
: 16190 (view as bug list)
Depends on:
Blocks:
 
Reported: 2008-04-21 17:12 UTC by David L.
Modified: 2009-08-24 12:30 UTC (History)
5 users (show)

See Also:
i915 platform:
i915 features:


Attachments
log from when trying to start civilization 4 beyond the sword (16.02 KB, text/plain)
2008-07-04 08:31 UTC, Krister Bäckman
Details

Description David L. 2008-04-21 17:12:12 UTC
EVE Online on wine triggers an assert in the fake buffer manager:

$ INTEL_NO_TTM=1 wine eve.exe
(... snip various wine cruft ...)
osplash: ../common/dri_bufmgr_fake.c:983: dri_fake_emit_reloc: Assertion `target_fake->is_static || target_fake->size_accounted' failed.
wine: Assertion failed at address 0xffffe430 (thread 001b), starting debugger...
storage.c:86: pool_alloc: Assertion `sizeof(struct pool_arena) + len <= pool->arena_size && len' failed.
wine: Assertion failed at address 0xffffe430 (thread 002d), starting debugger...

Might be related to textures failing to load, got some wine warnings about GL_INVALID_ENUM earlier - they're gone now though. (seems to be http://www.winehq.org/pipermail/wine-bugs/2008-January/084809.html, no idea if this is related)

System:
Linux arkology 2.6.25-rc5-wl-02174-g68961eb-dirty #6 SMP Mon Mar 31 11:13:52
CEST 2008 x86_64 Intel(R) Core(TM)2 Duo CPU T7300 @ 2.00GHz GenuineIntel
GNU/Linux

drm f0e38f521790becbf9ca13ef5c579d12c6985d52
mesa 711222bb7b4fd7ea8e7f801ac26532360bc5440f
(mesa & drm versions are the same for 64- and 32bit. drm.ko and i915.ko from
drm git used.)

xserver (64-bit) c14f5dc237a31b13d98ae2d0d6143bd91083cf13
xf86-video-i810-2.2.99.903 (Gentoo; had various other problems with git)

DRI2 disabled per Option "DRI2" "false" in xorg.conf, AIGLX enabled.

Hardware:
00:02.0 0300: 8086:2a02 VGA compatible controller: Intel Corporation Mobile
GM965/GL960 Integrated Graphics Controller (rev 03)
00:02.1 0380: 8086:2a03 Display controller: Intel Corporation Mobile
GM965/GL960 Integrated Graphics Controller (rev 03)

(See #15638 about INTEL_NO_TTM. Also, this is 32bit compat too.)
Comment 1 David L. 2008-04-21 17:40:54 UTC
Removing the assert makes the game run for another 60 seconds, then the screen turns into very funny pixel gibberish, the X server locks up and i need to reboot the box with Ctrl+Alt+Del (which astonishingly works still.)

The pixel gibberish/lockup might be unrelated though, i kind of already had some experiences with it ;)
Comment 2 WuNian 2008-07-02 20:37:47 UTC
Was the wine warning of GL_INVALID_ENUM for DXT1 texture format? Mesa does not support DXTI because of patent issue. You can use "driconf" to open the support.

For the X server lockup, please you debug it and give the backtrace info.
Comment 3 Krister Bäckman 2008-07-04 08:31:16 UTC
Created attachment 17532 [details]
log from when trying to start civilization 4 beyond the sword

I'm having the same problem with Civilization 4 in wine.

I'm using gentoo x86 with only xorg-server, mesa and their deps from the x11 overlay which pulls things from git.

The version of the xf86-video-i810 didn't seem to matter. I tried with 2.2.0.90, 2.2.1, 2.3.2 and the git one. They all gave the same error.
Comment 4 Michael Fu 2008-08-05 21:06:47 UTC
*** Bug 16190 has been marked as a duplicate of this bug. ***
Comment 5 Hedayat Vatankhah 2008-08-12 06:13:01 UTC
I have the same problem (Fedora 9, latest mesa dri drivers and xorg-x11-drv-i810) running applications like: bzflag, rcssserver3d.
(I've event tried intel i810 2.4 with no success)
Comment 6 Kenneth Graunke 2008-09-24 00:35:32 UTC
I get essentially the same assert a few seconds after moving about the world in the 3D MMORPG PlaneShift:
(psclient.bin: ../common/dri_bufmgr_fake.c:982: dri_fake_emit_reloc: Assertion `target_fake->is_static || target_fake->size_accounted' failed.)

I have a G965/GMA X3000, and am running Arch Linux, Kernel 2.6.26.5, xorg-server 1.5.0, libdrm 2.3.1, mesa/libgl/intel-dri 7.2 (upgraded myself), and xf86-video-intel 2.4.2.  Doesn't matter if I use EXA or XAA.  I have _not_ set LIBGL_ALWAYS_INDIRECT.

By the way, I highly recommend testing with PlaneShift - being free and Linux/X native, it's readily available, and triggers an awful lot of 3D driver bugs.
Comment 7 Eric Anholt 2008-10-09 13:12:19 UTC
Please try with current mesa.  The interface that that assertion was about has been changed significantly, and I don't think you'll see this bug any more.

(also, to the person who suggested PlaneShift, it failed to start at all on my system)
Comment 8 Hedayat Vatankhah 2008-10-11 00:21:32 UTC
(In reply to comment #7)
> Please try with current mesa.  The interface that that assertion was about has
> been changed significantly, and I don't think you'll see this bug any more.
> 
> (also, to the person who suggested PlaneShift, it failed to start at all on my
> system)
> 

I've tried Fedora 10 beta which uses mesa-dri-drivers 7.2 and rcssserver3d works in it :)
Comment 9 Krister Bäckman 2008-10-12 07:04:42 UTC
(In reply to comment #8)
> (In reply to comment #7)
> > Please try with current mesa.  The interface that that assertion was about has
> > been changed significantly, and I don't think you'll see this bug any more.
> > 
> > (also, to the person who suggested PlaneShift, it failed to start at all on my
> > system)
> > 
> 
> I've tried Fedora 10 beta which uses mesa-dri-drivers 7.2 and rcssserver3d
> works in it :)
> 

Works on Gentoo with mesa-7.2 too!
Comment 10 Kenneth Graunke 2008-10-30 03:44:15 UTC
I still had this issue with Mesa 7.2 on Arch Linux.  However, I just tried Mesa master + linux drm-intel-next + xf86-video-intel 2.5.0 + xorg-server 1.5.2, and I haven't had any crashes or asserts.  Seems fixed to me!
Comment 11 Eric Anholt 2008-12-09 13:51:22 UTC
Reported fixed with current GEM code.
Comment 12 Adam Jackson 2009-08-24 12:30:03 UTC
Mass version move, cvs -> git


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.