Bug 21814

Created attachment 26000 [details]
full console output, including wine debug info and backtrace

I've just checked in a Mesa change to print more info when this failure occurs.  However, it would be great if you could also use gdb to print *tex in valid_texture_object() when the assertion fails.

(In reply to comment #2)
> I've just checked in a Mesa change to print more info when this failure occurs.

I get this now:

Mesa 7.6-devel implementation error: invalid texture object Target 0xcc8acc8b, Id = 3431648395
Please report at bugzilla.freedesktop.org
main/texobj.c:335: _mesa_reference_texobj: Asserzione 'valid_texture_object(tex)' fallita.

>  However, it would be great if you could also use gdb to print *tex in
> valid_texture_object() when the assertion fails.

I tried it under gdb but the game doesn't even load (I killed it after 5 minutes, it loads instantly without gdb).

Another way to use gdb would be to start the game normally, then use 'ps -elf' to find the PID for the game, then run gdb and use "attach <pid>" to attach gdb to the running program.

Otherwise, this looks like a probable memory corruption bug.  Could you try running with valgrind?

(In reply to comment #3)
> (In reply to comment #2)
> > I've just checked in a Mesa change to print more info when this failure occurs.
> 
> I get this now:
> 
> Mesa 7.6-devel implementation error: invalid texture object Target 0xcc8acc8b,
> Id = 3431648395
> Please report at bugzilla.freedesktop.org
> main/texobj.c:335: _mesa_reference_texobj: Asserzione
> 'valid_texture_object(tex)' fallita.


Mmh, the printed code changed on second run...

Mesa 7.6-devel implementation error: invalid texture object Target 0xdd0cdd0c, Id = 3708607756
Please report at bugzilla.freedesktop.org
main/texobj.c:312: _mesa_reference_texobj: Asserzione 'valid_texture_object(oldTex)' fallita.

Created attachment 26004 [details]
first run under valgrind - crash but don't print the mesa warning

Created attachment 26006 [details]
second run under valgrind - crash with the mesa warning

Can you provide a direct link to the game so I can try it?  I tried several of the so-called "download" sites but I just get redirected to advertisements and surveys and junk like that.

(In reply to comment #8)
> Can you provide a direct link to the game so I can try it?  I tried several of
> the so-called "download" sites but I just get redirected to advertisements and
> surveys and junk like that.
> 

I tried installing the 2 available demos, but I can't get working under wine ( one of the two can be downloaded from http://files.filefront.com/Codename+Panzers+Phase+Two+Demo/;3932312;/fileinfo.html ). I used the full game.

Note that the demos are also buggy on windows (they go too fast) on some dual core CPUs (as well as the original game before installing a patch available only for the full game version).

Also note that the game runs fine (at least up to the menu screen, including background with 3D action) with software rendering (altough slowly and too dark).

Created attachment 26710 [details]
full console output after radeon-rewrite merge

I am getting a different crash with the recent merged radeon-rewrite. Full log attached.

Backtrace:
=>1 0xb803f430 (0x0032dd34)
  2 0xb7d84098 abort+0x188() in libc.so.6 (0x0032de5c)
  3 0xb7d7b5ce __assert_fail+0xee() in libc.so.6 (0x0032dea0)
  4 0x7d64db34 in r300_dri.so (+0x50b34) (0x0032ded0)
  5 0x7d650a8b radeonRefillCurrentDmaRegion+0x10b(rmesa=<register ESI not in topmost frame>, size=0x1200) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/drivers/dri/r300/radeon_dma.c:203] in r300_dri.so (0x0032df10)
  6 0x7d650cca radeonAllocDmaRegion+0xba(rmesa=<register ESI not in topmost frame>, pbo=0x7d1eb610, poffset=0x7d1eb614, bytes=<register EDI not in topmost frame>, alignment=<is not available>) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/drivers/dri/r300/radeon_dma.c:234] in r300_dri.so (0x0032df40)
  7 0x7d650e8f rcommon_emit_vector+0x12f(ctx=0x7cfea620, aos=0x7d1eb610, data=<register ESI not in topmost frame>, size=0x3, stride=0x34, count=0x180) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/drivers/dri/r300/radeon_dma.c:146] in r300_dri.so (0x0032df80)
  8 0x7d62744d r300DrawPrims+0x77d(ctx=0x7cfea620, arrays=0x7d006ecc, prim=0x7d0058c0, nr_prims=0x1, ib=0x0, min_index=0x0, max_index=0x17f) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/drivers/dri/r300/r300_draw.c:359] in r300_dri.so (0x0032e120)
  9 0x7d6f0a7c vbo_exec_vtx_flush+0xfc(exec=0x7d0057a0, unmap=0x1) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/vbo/vbo_exec_draw.c:375] in r300_dri.so (0x0032e1a0)
  10 0x7d6ed480 vbo_exec_FlushVertices_internal+0x40(ctx=0x7cfea620, unmap=<is not available>) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/vbo/vbo_exec_api.c:781] in r300_dri.so (0x0032e1c0)
  11 0x7d6ed530 vbo_exec_FlushVertices+0x50(ctx=<register EDI not in topmost frame>, flags=0x1) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/vbo/vbo_exec_api.c:804] in r300_dri.so (0x0032e1e0)
  12 0x7d6d9e51 _mesa_ActiveTextureARB+0x101(texture=0x84c0) [/home/fabio/sorgenti/ati-drivers/mesa-compiled/src/mesa/main/texstate.c:299] in r300_dri.so (0x0032e200)
  13 0x7e8ec9b9 in wined3d (+0x7c9b9) (0x0032e250)
  14 0x7e8ece95 in wined3d (+0x7ce95) (0x0032e2b0)
  15 0x7e88c6b8 ActivateContext+0x548() in wined3d (0x0032e570)
  16 0x7e8c7152 drawPrimitive+0x1b2() in wined3d (0x0032e8f0)
  17 0x7e89b04b in wined3d (+0x2b04b) (0x0032e960)
  18 0x7e9849de in d3d9 (+0x49de) (0x0032e990)
  19 0x0053f80f in panzers_phase_2 (+0x13f80f) (0x02d12660)
  20 0x0013a080 (0x007eee68)
  21 0x005cce90 in panzers_phase_2 (+0x1cce90) (0x00670cb0)

Your latest backtrace looks like bug 21582. Check also the gdm log for errors.

(In reply to comment #11)
> Your latest backtrace looks like bug 21582.

I noticed that as well, but it's better lo leave this bug opened, maybe the original bug is not yet fixed. I will retest when bug 21582 is fixed.

> Check also the gdm log for errors.

I don't use compiz, so I don't get the errors here.

> I will retest when bug 21582 is fixed.

Bug 21582 is marked as fixed but I am still having the same assertion of https://bugs.freedesktop.org/attachment.cgi?id=26710.

See also bug 22438.

Created attachment 27117 [details]
wine output with mesa master 2009-06-25 (commit cdbcb051)

(In reply to comment #13)
> > I will retest when bug 21582 is fixed.
> 
> Bug 21582 is marked as fixed but I am still having the same assertion of
> https://bugs.freedesktop.org/attachment.cgi?id=26710.
> 
> See also bug 22438.

OK, bug 22438 is fixed but now the game crashes with a different backtrace (attached).

Closing since the original problem is no longer reproducible.

Opening last issue as bug 22528.

Mass version move, cvs -> git