Bug 32624 - [SNB GT2] Glean/'ARL test' fail for the first run
[SNB GT2] Glean/'ARL test' fail for the first run
Status: VERIFIED WORKSFORME
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965
git
Other Linux (All)
: medium major
Assigned To: haihao
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2010-12-23 23:10 UTC by Yi Sun
Modified: 2012-09-25 02:20 UTC (History)
5 users (show)

See Also:
i915 platform:
i915 features:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Yi Sun 2010-12-23 23:10:03 UTC
System Environment:
--------------------------
Platform:    Hunronriver
Kernel_version:  (drm-intel-fixes)448f53a1ede54eb854d036abf54573281412d650
Libdrm:         (master)2.4.23-4-gbad5242a59aa8e31cf10749e2ac69b3c66ef7da0
Mesa:           (7.10)b7c187df9ac9bb778cf0f2abb3ad86775e42b07b
Xserver:       
(server-1.9-branch)xorg-server-1.9.3-9-gd819ff1cb521af77764af75e04645af4d4d06e3d
Xf86_video_intel: (master)2.13.902-9-g58d9912e447cf6e29b20747b5ed6ee595687ce8d

Bug detailed description:
-------------------------
The output is as following:

glean: brw_vs_emit.c:1113: get_reladdr_constant: Assertion `c->func.brw->intel.gen < 6' failed.

Reproduce steps:
-------------------------
PIGLIT_TEST='ARL test' bin/glean -t vertProg1 -o -r t
Comment 1 Wang Zhenyu 2010-12-27 21:58:35 UTC
Fix already on 7.10. Close.
Comment 2 Yi Sun 2010-12-29 17:17:12 UTC
On HunronRiver, when i run the case for the first time, the output is as following(one test fail). But if i run it again or more times, the case is passed. Until reboot the system, the failure appear again. 
The issue doesn't appear on SurgarBay.


vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 33
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 34
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 105
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, z24, s8, win+pmap, id 157
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 158
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 159
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, z24, s8, win+pmap, id 163
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 164
        45 tests passed, 0 tests failed.
FAILURE:
  Program: ARL test
  Expected color: 0.11, 0.22, 0.33, 0.44
  Observed color: 0.0509804, 0.105882, 0.156863, 1
vertProg1:  FAIL rgb8, db, z24, s8, accrgb16, win+pmap, slow, id 165
        44 tests passed, 1 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 166
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, accrgba16, win+pmap, slow, id 167
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, z24, s8, win+pmap, id 171
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 172
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 173
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, z24, s8, win+pmap, id 177
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 178
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, win+pmap, id 179
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgb8, db, z24, s8, accrgb16, win+pmap, slow, id 180
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, win+pmap, id 181
        45 tests passed, 0 tests failed.
vertProg1:  PASS rgba8, db, z24, s8, accrgba16, win+pmap, slow, id 182
        45 tests passed, 0 tests failed.
Comment 3 fangxun 2011-03-07 05:33:16 UTC
Tested with latest code, this issue as sunyi described is still happens on our huronriver(rev08 and rev09) and SurgarBay(rev09 GT2). It passes on SurgarBay(rev09 GT1).
Comment 4 Gordon Jin 2011-03-07 17:04:20 UTC
So it happens on the 3 GT2 machines, but not the 1 GT1 machine.
Can developers confirm?
Comment 5 Eric Anholt 2011-04-29 16:17:16 UTC
Works fine for me on Mesa master (plus the glsl-vs-max-varyings fix), 8086:0126.
Comment 6 Eric Anholt 2011-06-28 12:13:34 UTC
Haven't seen it on either of the 2 GT2 machines I've used since my last comment.  Are you still seeing it?
Comment 7 Eric Anholt 2011-07-20 13:13:02 UTC
Can you test on your hardware with Mesa master?
Comment 8 Kenneth Graunke 2011-11-08 21:52:08 UTC
No response since March, and development hasn't seen this behavior.  Closing.  (Also, our primary development machines are all GT2s these days...)