Created attachment 20847 [details]
Very resent X server with dri2 support, unable to get commit # as I did a pull a few minutes ago and the current tree doesn't seem to build
GIT anholt kernel tree
0xb7858a67 in _tnl_draw_prims () from /usr/lib/dri/i965_dri.so
#0 0xb7858a67 in _tnl_draw_prims () from /usr/lib/dri/i965_dri.so
#1 0xb779a560 in brw_draw_prims () from /usr/lib/dri/i965_dri.so
#2 0xb77a03d4 in meta_draw_quad () from /usr/lib/dri/i965_dri.so
#3 0xb776fd6e in intelClear () from /usr/lib/dri/i965_dri.so
#4 0xb77c9324 in _mesa_Clear () from /usr/lib/dri/i965_dri.so
#5 0xb7ec72ba in glClear () from //usr//lib/opengl/xorg-x11/lib/libGL.so.1
#6 0x084311e8 in drawview3dspace ()
#7 0x082c6d05 in scrarea_do_windraw ()
#8 0x081fcee4 in force_draw ()
#9 0x082aecd2 in ?? ()
#10 0x00000000 in ?? ()
d13f00l-desktop blender-2.48a-linux-glibc236-py25-i386 # ldd blender
linux-gate.so.1 => (0xffffe000)
libpython2.5.so.1.0 => /usr/lib/libpython2.5.so.1.0 (0xb7ea1000)
libSDL-1.2.so.0 => /usr/lib/libSDL-1.2.so.0 (0xb7e39000)
libGL.so.1 => //usr//lib/opengl/xorg-x11/lib/libGL.so.1 (0xb7dc3000)
libGLU.so.1 => /usr/lib/libGLU.so.1 (0xb7d53000)
libX11.so.6 => /usr/lib/libX11.so.6 (0xb7c2b000)
libXi.so.6 => /usr/lib/libXi.so.6 (0xb7c20000)
libutil.so.1 => /lib/libutil.so.1 (0xb7c1b000)
libc.so.6 => /lib/libc.so.6 (0xb7af4000)
libm.so.6 => /lib/libm.so.6 (0xb7acc000)
libdl.so.2 => /lib/libdl.so.2 (0xb7ac8000)
libpthread.so.0 => /lib/libpthread.so.0 (0xb7ab1000)
libXext.so.6 => /usr/lib/libXext.so.6 (0xb7aa1000)
libXxf86vm.so.1 => /usr/lib/libXxf86vm.so.1 (0xb7a9a000)
libXdamage.so.1 => /usr/lib/libXdamage.so.1 (0xb7a96000)
libXfixes.so.3 => /usr/lib/libXfixes.so.3 (0xb7a90000)
libdrm.so.2 => /usr/lib/libdrm.so.2 (0xb7a87000)
libstdc++.so.6 => /usr/lib/gcc/i486-pc-linux-gnu/4.3.1/libstdc++.so.6 (0
libgcc_s.so.1 => /usr/lib/gcc/i486-pc-linux-gnu/4.3.1/libgcc_s.so.1 (0xb
libXau.so.6 => /usr/lib/libXau.so.6 (0xb7985000)
libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 (0xb797f000)
Running in UXA mode, happens also in EXA.
EXT3-fs: recovery complete.
EXT3-fs: mounted filesystem with ordered data mode.
Adding 1004052k swap on /dev/sda2. Priority:-1 extents:1 across:1004052k
r8169: eth2: link up
r8169: eth2: link up
eth2: no IPv6 routers present
libdrm.so.2: segfault at 1 ip 00000001 sp bfc538d4 error 14 in libdrm.so.2.3.0[b8050000+7000]
libdrm.so.2: segfault at 2 ip 00000002 sp bff363b4 error 14 in libdrm.so.2.3.0[b7f32000+7000]
libdrm.so.2: segfault at 2 ip 00000002 sp bfb627d4 error 14 in libdrm.so.2.3.0[b805d000+7000]
[drm] Initialized drm 1.1.0 20060810
pci 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
pci 0000:00:02.0: setting latency timer to 64
pci 0000:00:02.0: irq 1275 for MSI/MSI-X
[drm] Initialized i915 1.6.0 20080730 on minor 0
Steps to reproduce:
select an empty area, left click like you were going to select something and drag the mouse around. Will crash in < 2 seconds, 100% of the time. Will lock X. Will have to reboot, card gets wedged.
Xorg file attached.
This was a segfault, not a lockup, right?
On my system it behaves a little bit different. I get a segfault from Blender, all the time, but it will only hang X sometimes.
Jiewen, can you reproduce this on our G45?
(In reply to comment #3)
> Jiewen, can you reproduce this on our G45?
Our G45 get a segfault from Blender
all the time, after that, if you run other 3D apps X will be hung.
For me, it's hung X every time, tried maybe 8 times.
meta_draw_quad does nasty things for the case that brw_draw_prims decides to do a sw fallback. I've got code to fix this in the meta-clear branch of my repo, but it's not correct yet.
failure present as far back as 154a9e5317f890618932cea0129ef887e16baf84
Author: Eric Anholt <email@example.com>
Date: Mon Dec 15 13:25:20 2008 -0800
i965: Update state before checking for fallbacks in brw_try_draw_prims.
This got flipped around in 7855b2aef6bd9e9c2d73260b5cd166159b2525c6.
Bug #18907. Thanks to idr for pointing me at a nicer testcase than blender.