Bug 78645

Summary: [HSW Regression]igt/gem_exec_bad_domains/conflicting-write-domain fails
Product: DRI Reporter: Guo Jinxian <jinxianx.guo>
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: highest CC: intel-gfx-bugs, yi.sun
Version: unspecified   
Hardware: Other   
OS: All   
Whiteboard:
i915 platform: i915 features:
Attachments:
Description Flags
dmesg none

Description Guo Jinxian 2014-05-13 03:46:04 UTC 
Created attachment 98955 [details]
dmesg

==System Environment==
--------------------------
Regression: Yes. 
But I didn't find out good commit on -next-queued branch

Non-working platforms: HSW

==kernel==
--------------------------
-nightly: ed60c2747e0be176f28dc3dadba564949d90d821 (fails)
-queued: 1cf0ba14740d96fbf6f58a201f000a34b74f4725 (fails)
    Author: Chris Wilson <chris@chris-wilson.co.uk>
    Date:   Mon May 5 09:07:33 2014 +0100

    drm/i915: Flush request queue when waiting for ring space

    During the review of

    commit 1f70999f9052f5a1b0ce1a55aff3808f2ec9fe42
    Author: Chris Wilson <chris@chris-wilson.co.uk>
    Date:   Mon Jan 27 22:43:07 2014 +0000

        drm/i915: Prevent recursion by retiring requests when the ring is full

    Ville raised the point that our interaction with request->tail was
    likely to foul up other uses elsewhere (such as hang check comparing
    ACTHD against requests).

    However, we also need to restore the implicit retire requests that certain
    test cases depend upon (e.g. igt/gem_exec_lut_handle), this raises the
    spectre that the ppgtt will randomly call i915_gpu_idle() and recurse
    back into intel_ring_begin().

    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=78023
    Reviewed-by: Brad Volkin <bradley.d.volkin@intel.com>
    [danvet: Remove now unused 'tail' variable as spotted by Brad.]
    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
-fixes: 05adaf1f101f25f40f12c29403e6488f0e45f6b6 (works)
    Author: Jani Nikula <jani.nikula@intel.com>
    Date:   Fri May 9 14:52:34 2014 +0300

    drm/i915/vlv: reset VLV media force wake request register

    Media force wake get hangs the machine when the system is booted without
    displays attached. The assumption is that (at least some versions of)
    the firmware has skipped some initialization in that case.

    Empirical evidence suggests we need to reset the media force wake
    request register in addition to the render one to avoid hangs.

    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=75895
    Reported-by: Imre Deak <imre.deak@intel.com>
    Reported-by: Darren Hart <dvhart@linux.intel.com>
    Tested-by: Darren Hart <dvhart@linux.intel.com>
    Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
    Cc: stable@vger.kernel.org
    Signed-off-by: Jani Nikula <jani.nikula@intel.com>

==Bug detailed description==
-----------------------------
igt/gem_exec_bad_domains/conflicting-write-domain fails

Output:
./gem_exec_bad_domains --run-subtest conflicting-write-domain
IGT-Version: 1.6-gd848a36 (x86_64) (Linux: 3.15.0-rc3_drm-intel-next-queued_1cf0ba_20140512+ x86_64)
Test assertion failure function __real_main153, file gem_exec_bad_domains.c:204:
Last errno: 22, Invalid argument
Failed assertion: run_batch() == 0
Subtest conflicting-write-domain: FAIL


==Reproduce steps==
---------------------------- 
1. ./gem_exec_bad_domains --run-subtest conflicting-write-domain
Comment 1 Daniel Vetter 2014-05-15 15:28:18 UTC 
Bisect result please. If it works on -fixes or some older kernel, then git bisect will find the regression.

Please also double-check that it isn't a regression in igt.
Comment 2 Guo Jinxian 2014-05-23 02:06:26 UTC 
(In reply to comment #1)
> Bisect result please. If it works on -fixes or some older kernel, then git
> bisect will find the regression.
> 
> Please also double-check that it isn't a regression in igt.

Sorry, I can not find good commit on either -enxt-queud or igt. but the result on -fixes always passed.
Comment 3 Daniel Vetter 2014-06-18 10:06:04 UTC 
On Fri, May 23, 2014 at 4:06 AM,  <bugzilla-daemon@freedesktop.org> wrote:
> --- Comment #2 from Guo Jinxian <jinxianx.guo@intel.com> ---
> (In reply to comment #1)
>> Bisect result please. If it works on -fixes or some older kernel, then git
>> bisect will find the regression.
>>
>> Please also double-check that it isn't a regression in igt.
>
> Sorry, I can not find good commit on either -enxt-queud or igt. but the result
> on -fixes always passed.

Then please bisect between -fixes and -nightly. Or does -nightly work?
-Daniel
Comment 4 Jani Nikula 2014-09-05 13:35:55 UTC 
NEEDINFO
Comment 5 Guo Jinxian 2014-09-17 05:19:35 UTC 
Test passes on latest -nightly(ab7f1ab9c4d1c8af46517d60e54eeda9d8673f6a)

[root@x-hsw24 tests]# ./gem_exec_bad_domains --run-subtest conflicting-write-domain
IGT-Version: 1.8-g137877f (x86_64) (Linux: 3.17.0-rc5_drm-intel-nightly_ab7f1a_20140917+ x86_64)
Subtest conflicting-write-domain: SUCCESS (0.000s)
[root@x-hsw24 tests]# echo $?
0
Comment 6 Jari Tahvanainen 2017-02-10 08:42:02 UTC 
Closing (>2 years) old Verified+fixed.

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.