Bug 85270

Summary: [BYT/BDW]igt/gem_bad_reloc/negative-reloc is sporadical skip
Product: DRI Reporter: lu hua <huax.lu>
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: intel-gfx-bugs, jinxianx.guo
Version: unspecified   
Hardware: All   
OS: Linux (All)   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg none

Description lu hua 2014-10-21 05:30:53 UTC
Created attachment 108145 [details]
dmesg

==System Environment==
--------------------------
Regression:  not sure
Non-working platforms: BDW

==kernel==
--------------------------
drm-intel-nightly/f2bcbc348f1b20b47f98976aa812d4540c3d9951

==Bug detailed description==
-----------------------------
Run 5 cycles on BDW with -nightly kernel.
It's skip on the first cycle.

output:
root@x-bdw05:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-gbba1cd0 (x86_64) (Linux: 3.18.0-rc1_drm-intel-nightly_f2bcbc_20141021+ x86_64)
Found offset 8912896 for 4k batch
Test requirement not met in function negative_reloc, file gem_bad_reloc.c:90:
Test requirement: gem_exec[0].offset < BIAS
Subtest negative-reloc: SKIP (0.000s)
root@x-bdw05:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-gbba1cd0 (x86_64) (Linux: 3.18.0-rc1_drm-intel-nightly_f2bcbc_20141021+ x86_64)
Found offset 73728 for 4k batch
Batch is now at offset 266240
Subtest negative-reloc: SUCCESS (0.000s)
root@x-bdw05:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-gbba1cd0 (x86_64) (Linux: 3.18.0-rc1_drm-intel-nightly_f2bcbc_20141021+ x86_64)
Found offset 73728 for 4k batch
Batch is now at offset 262144
Subtest negative-reloc: SUCCESS (0.001s)
root@x-bdw05:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-gbba1cd0 (x86_64) (Linux: 3.18.0-rc1_drm-intel-nightly_f2bcbc_20141021+ x86_64)
Found offset 73728 for 4k batch
Batch is now at offset 262144
Subtest negative-reloc: SUCCESS (0.000s)
root@x-bdw05:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-gbba1cd0 (x86_64) (Linux: 3.18.0-rc1_drm-intel-nightly_f2bcbc_20141021+ x86_64)
Found offset 73728 for 4k batch
Batch is now at offset 266240
Subtest negative-reloc: SUCCESS (0.000s)

==Reproduce steps==
---------------------------- 
1. ./gem_bad_reloc --run-subtest negative-reloc
Comment 1 Rodrigo Vivi 2014-10-21 22:55:37 UTC
Just means that in that space of memory the object got allocated doesn't need to be tested. So skipping. not a failure.
Comment 2 lu hua 2014-10-23 02:49:48 UTC
(In reply to Rodrigo Vivi from comment #1)
> Just means that in that space of memory the object got allocated doesn't
> need to be tested. So skipping. not a failure.

It has different result when run multiple rounds on one machine.
It's difficult to track the result.
Comment 3 Guo Jinxian 2014-12-03 02:52:17 UTC
The failure is able to reproduce on BYT latest -nightly(691817ae2a8f37cb1df33f21a6dd1c15a8832693)

root@x-byt06:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# dmesg -r|egrep "<[1-4]>"|grep drm
root@x-byt06:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc
IGT-Version: 1.8-g4e5c16c (x86_64) (Linux: 3.18.0-rc7_drm-intel-nightly_691817_20141202+ x86_64)
Found offset 17035264 for 4k batch
Test requirement not met in function negative_reloc, file gem_bad_reloc.c:90:
Test requirement: gem_exec[0].offset < BIAS
Subtest negative-reloc: SKIP (0.000s)
root@x-byt06:/GFX/Test/Intel_gpu_tools/intel-gpu-tools/tests# ./gem_bad_reloc --run-subtest negative-reloc-lut
IGT-Version: 1.8-g4e5c16c (x86_64) (Linux: 3.18.0-rc7_drm-intel-nightly_691817_20141202+ x86_64)
Found offset 17035264 for 4k batch
Test requirement not met in function negative_reloc, file gem_bad_reloc.c:90:
Test requirement: gem_exec[0].offset < BIAS
Subtest negative-reloc-lut: SKIP (0.000s)
Comment 4 Damien Lespiau 2014-12-04 16:50:14 UTC
Pushed a work-around in i-g-t:

commit 88ff1cec3a830fc467b91efb7e1dcdb4d82a0e17
Author: Damien Lespiau <damien.lespiau@intel.com>
Date:   Thu Dec 4 13:42:12 2014 +0000

    gem_bad_reloc: Don't flip-flop between SKIP and PASS
    
    Here is a cheap way for this test to give consistent results. This
    doesn't change the usefulness of this test, hopefully.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=85270
    Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
    Signed-off-by: Damien Lespiau <damien.lespiau@intel.com>
Comment 5 lu hua 2014-12-08 01:47:51 UTC
Verified.Fixed.
Comment 6 Elizabeth 2017-10-06 14:34:36 UTC
Closing old 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.