Bug 28443 - [r300g] HoN sometimes crashes when loading
Summary: [r300g] HoN sometimes crashes when loading
Status: RESOLVED NOTOURBUG
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/r300 (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact:
URL: http://www.savage2.com/en/download.php
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2010-06-08 05:57 UTC by Pavel Ondračka
Modified: 2010-06-15 16:17 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
dmesg (50.79 KB, text/plain)
2010-06-08 05:57 UTC, Pavel Ondračka
Details
dmesg with Savage 2 (67.08 KB, text/plain)
2010-06-13 02:42 UTC, Pavel Ondračka
Details

Description Pavel Ondračka 2010-06-08 05:57:12 UTC
Created attachment 36140 [details]
dmesg

I've been experiencing this random crash for some time, but this was the first time I was able to get a backtrace. Sometimes when HoN menu or new map is loading it crash like this:

Failed to allocate :
   size      : 8388608 bytes
   alignment : 2048 bytes
   domains   : 2

Program received signal SIGSEGV, Segmentation fault.
get_drm_buffer (_buf=0x0) at radeon_drm_buffer.c:255
255	    if (_buf->vtbl == &radeon_drm_buffer_vtbl) {
Missing separate debuginfos, use: debuginfo-install glibc-2.12-2.i686
(gdb) bt
#0  get_drm_buffer (_buf=0x0) at radeon_drm_buffer.c:255
#1  0x02bd7267 in radeon_drm_bufmgr_set_tiling (_buf=0x0, microtiled=R300_BUFFER_LINEAR, macrotiled=R300_BUFFER_LINEAR, pitch=8192)
    at radeon_drm_buffer.c:319
#2  0x02bd6392 in radeon_r300_winsys_buffer_set_tiling (rws=0x887e258, buf=0x0, pitch=8192, microtiled=R300_BUFFER_LINEAR, macrotiled=R300_BUFFER_LINEAR)
    at radeon_r300.c:81
#3  0x02be72fc in r300_texture_create (screen=0x887fa20, base=0xbfffdd18) at r300_texture.c:982
#4  0x02bf6eeb in r300_resource_create (screen=0x887fa20, templ=0xbfffdd18) at r300_resource.c:37
#5  0x02be7a4f in r300_texture_get_transfer (ctx=0x8932c58, texture=0xa353778, sr=..., usage=2, box=0xbfffddb8) at r300_transfer.c:157
#6  0x02db71ee in u_get_transfer_vtbl (context=0x8932c58, resource=0xa353778, sr=..., usage=PIPE_TRANSFER_WRITE, box=0xbfffddb8) at util/u_resource.c:42
#7  0x02d03586 in pipe_get_transfer (st=0x89b7340, stImage=0xa193ba8, zoffset=0, usage=PIPE_TRANSFER_WRITE, x=0, y=0, w=2048, h=1024)
    at ../../src/gallium/auxiliary/util/u_inlines.h:342
#8  st_texture_image_map (st=0x89b7340, stImage=0xa193ba8, zoffset=0, usage=PIPE_TRANSFER_WRITE, x=0, y=0, w=2048, h=1024) at state_tracker/st_texture.c:147
#9  0x02d4226e in st_TexImage (ctx=0x898be10, dims=2, target=3553, level=0, internalFormat=6408, width=2048, height=1024, depth=1, 
    border=<value optimized out>, format=6408, type=5121, pixels=0x6548008, unpack=0x899abe0, texObj=0xa353558, texImage=0xa193ba8, imageSize=0, 
    compressed_src=0 '\000') at state_tracker/st_cb_texture.c:681
#10 0x02d42c7e in st_TexImage2D (ctx=0x898be10, target=3553, level=0, internalFormat=6408, width=2048, height=1024, border=0, format=6408, type=5121, 
    pixels=0x6548008, unpack=0x899abe0, texObj=0xa353558, texImage=0xa193ba8) at state_tracker/st_cb_texture.c:798
#11 0x02c99b0f in _mesa_TexImage2D (target=3553, level=0, internalFormat=6408, width=2048, height=1024, border=0, format=6408, type=5121, pixels=0x6548008)
    at main/teximage.c:2251
#12 0x07c10307 in glTexImage2D (target=3553, level=0, internalformat=6408, width=2048, height=1024, border=0, format=6408, type=5121, pixels=0x6548008)
    at ../../src/mesa/glapi/glapitemp.h:1002
#13 0x01f37461 in ?? () from /home/Paulie/HoN/vid_gl2-x86.so
Backtrace stopped: previous frame inner to this frame (corrupt stack?)

mesa: a2817f6ae566b672f195cff22e14e2058d3617ea
kernel: 2.6.33.5-112.fc13
GPU: ATI Technologies Inc M56P [Radeon Mobility X1600] (RV530)
Comment 1 Pavel Ondračka 2010-06-13 02:42:50 UTC
Created attachment 36241 [details]
dmesg with Savage 2

Another game affected by this bug is Savage 2. However this one doesn't crash only prints info into the terminal (probably thanks to commit 	45fb47d50c08bc4c11e4454883641501713e5710):

Failed to allocate :
   size      : 16777216 bytes
   alignment : 2048 bytes
   domains   : 2

this is printed several times and after that 

r300: Failed to create a transfer object, praise.

Adding download URL since this game is free.
Comment 2 Marek Olšák 2010-06-13 08:28:03 UTC
It's the same error. The real problem is this (from your dmesg):

hon-x86: page allocation failure. order:0, mode:0x2

I was told that if a kernel fails to allocate a page, there is nothing we can do. This really shouldn't happen at all. I am not best positioned to give you any advice on this, it's sort of outside of my scope. Besides trying out the newest kernel, I think it would be better to file a bug in https://bugzilla.kernel.org/
Comment 3 Pavel Ondračka 2010-06-13 08:44:24 UTC
OK, I'll try to reproduce this with some never kernel and if I succeed I will open a bug at kernel Bugzilla. This may take some time because this issue isn't 100% reproducible.
Comment 4 Marek Olšák 2010-06-15 16:17:56 UTC
There is an ongoing discussion about page allocation failures in the kernel bugzilla. This looks like your problem:

https://bugzilla.kernel.org/show_bug.cgi?id=16148#c8

I am closing the bug here.


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.