Bug 14873

Summary: [i915 i965] glean (around pbo case) crash X
Product: Mesa Reporter: Shuang He <shuang.he>
Component: Drivers/DRI/i915Assignee: Eric Anholt <eric>
Status: VERIFIED FIXED QA Contact:
Severity: critical    
Priority: high CC: dri-devel, haihao.xiang
Version: unspecified   
Hardware: Other   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments: xorg conf
xorg log
dmesg

Description Shuang He 2008-03-06 23:05:51 UTC
System Environment:
--------------------------

--Platform: i965

--Architecture(32-bit,64-bit,compatiblity): 32-bit

--2D driver:
commit 502e41f6dff91393cfeaf08c9ba312781d1b7267

--3D driver:
commit d76545de5730f52328a81b8371d3b08cd7511e05

--Xserver:
commit 34b69e3bc0e6462eb60029fbcb4f5479494007a2

--Drm
commit 92a30dd608c2838dea97efc04e1447056f37d0b5

--Kernel:
2.6.23


Bug detailed description:
-------------------------
glean crash X around pbo case.
but This issue doesn't happen when pbo case is run alone.
we need to run glean from beginning to reproduce this issue.



Reproduce steps:
----------------
1. start X
2. ./glean -r result


Current result:
----------------
GL_ARB_pixel_buffer_object is supported
FAILURE: glBitmap failed (at tpbo.cpp:507)
  got (0, 0) = [0.000000, 0.000000, 0.000000], should be [1.000000, 1.000000, 1.000000]
FAILURE: glGetTexImage failed (at tpbo.cpp:694)
  got (0) = [0.000000, 0.000000, 0.000000], should be intel_bufmgr_ttm.c:644: Error mapping buffer front: Cannot allocate memory .
intel_bufmgr_ttm.c:644: Error mapping buffer depth: Cannot allocate memory .
glean: ../../../../../src/mesa/drivers/dri/common/spantmp2.h:373: intelReadRGBASpan_ARGB8888: Assertion `irb->pfMap' failed.
[1.000000, 1.000000, 0.000000]
pbo:  FAIL rgba8, db, z24, s8, win+pmap, id 33
        8 tests passed, 2 tests failed.
GL_ARB_pixel_buffer_object is supported
run.Glean.sh: line 18:  5381 Killed                  Xorg



Expected result:
----------------
X should not be crashed by glean
Comment 1 Shuang He 2008-03-06 23:08:35 UTC
Created attachment 14902 [details]
xorg conf
Comment 2 Shuang He 2008-03-06 23:09:52 UTC
Created attachment 14903 [details]
xorg log
Comment 3 Shuang He 2008-03-06 23:15:09 UTC
Created attachment 14904 [details]
dmesg
Comment 4 Shuang He 2008-03-19 22:10:48 UTC
seems this issue has gone
Comment 5 Shuang He 2008-03-19 22:11:13 UTC
verified
Comment 6 Eric Anholt 2008-03-20 16:11:39 UTC
It's an out-of-memory condition, so I would expect it to vary depending on your environment at the time the test is run.  I was reproducing it yesterday, but didn't get far in tracking it down (need to try another run with blendfunc disabled, as the console spam from that one was killing test performance).
Comment 7 Gordon Jin 2008-03-24 01:14:04 UTC
So it should be reopened.
Comment 8 Brian Paul 2008-03-24 11:06:22 UTC
I fixed two bugs in the glean pbo test last week.  Retest?
Comment 9 Shuang He 2008-04-08 20:13:24 UTC
This issue seems gone.
Comment 10 Shuang He 2008-04-08 20:13:54 UTC
verified
Comment 11 Shuang He 2008-09-05 00:04:16 UTC
We still get a lot of failures on G33, pass/fail=15/21. 
Comment 12 Shuang He 2008-09-05 00:06:25 UTC
reopen the wrong bug
Sorry for the noise
Comment 13 Shuang He 2008-09-05 00:06:48 UTC
verified

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.