Bug 104594

Summary: [CI] igt@drm_mm@* - dmesg-warn - BUG: using smp_processor_id() in preemptible [00000000] code: drm_mm
Product: DRI Reporter: Marta Löfstedt <marta.lofstedt>
Component: DRM/IntelAssignee: Marta Löfstedt <marta.lofstedt>
Status: CLOSED DUPLICATE QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: intel-gfx-bugs
Version: DRI git   
Hardware: Other   
OS: All   
Whiteboard: ReadyForDev
i915 platform: KBL i915 features: GEM/Other

Description Marta Löfstedt 2018-01-12 08:16:27 UTC
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3621/shard-kbl2/igt@drm_mm@align.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3621/shard-kbl2/igt@drm_mm@align64.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3621/shard-kbl2/igt@drm_mm@insert.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3621/shard-kbl2/igt@drm_mm@insert_range.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3621/shard-kbl2/igt@drm_mm@reserve.html

<7>[  190.908905] [IGT] drm_mm: executing
<7>[  190.911182] [IGT] drm_mm: starting subtest reserve
<3>[  190.912001] BUG: using smp_processor_id() in preemptible [00000000] code: drm_mm/6141
<4>[  190.912011] caller is vprintk_func+0x47/0x80
<4>[  190.912059] CPU: 3 PID: 6141 Comm: drm_mm Tainted: G     UD W        4.15.0-rc7-CI-CI_DRM_3621+ #1
<4>[  190.912060] Hardware name:                  /NUC7i5BNB, BIOS BNKBL357.86A.0054.2017.1025.1822 10/25/2017
<4>[  190.912060] Call Trace:
<4>[  190.912060]  dump_stack+0x5f/0x86
<4>[  190.912061]  check_preemption_disabled+0xd9/0xe0
<4>[  190.912061]  vprintk_func+0x47/0x80
<4>[  190.912062]  ? 0xffffffffa0041000
<4>[  190.912062]  printk+0x3e/0x46
<4>[  190.912062]  test_drm_mm_init+0x33/0x1000 [test_drm_mm]
<4>[  190.912063]  ? 0xffffffffa0041000
<4>[  190.912063]  do_one_initcall+0x39/0x150
<4>[  190.912064]  ? kmem_cache_alloc_trace+0x1e7/0x2b0
<4>[  190.912064]  do_init_module+0x56/0x1ef
<4>[  190.912064]  load_module+0x231c/0x2d70
<4>[  190.912065]  ? show_coresize+0x20/0x20
<4>[  190.912065]  ? vfs_read+0x126/0x150
<4>[  190.912065]  ? SyS_finit_module+0xa5/0xe0
<4>[  190.912066]  SyS_finit_module+0xa5/0xe0
<4>[  190.912066]  entry_SYSCALL_64_fastpath+0x1c/0x89
<4>[  190.912066] RIP: 0033:0x7f75d8be49f9
<4>[  190.912067] RSP: 002b:00007fff49fc4928 EFLAGS: 00000206
<6>[  190.912168] drm_mm: Testing DRM range manger (struct drm_mm), with random_seed=0xa0e0b9aa max_iterations=8192 max_prime=128
Comment 1 Chris Wilson 2018-01-12 10:47:17 UTC
commit 2bbba4e94eb987701abcacae241929fd13a196b2
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Jan 11 14:04:02 2018 +0000

    drm/i915/pmu: Initialise our dynamic sysfs attributes for use with lockdep
    
    As we kmalloc our dynamic sysfs attributes, we have to give them an
    external static lock_class_key for them to use with lockdep.
    
    Fixes: 109ec558370f ("drm/i915/pmu: Only enumerate available counters in sysfs")
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
    Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
    Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
    Link: https://patchwork.freedesktop.org/patch/msgid/20180111140402.3984-1-chris@chris-wilson.co.uk
Comment 2 Marta Löfstedt 2018-01-12 13:31:49 UTC
Fix integrated into CI_DRM_3622, the issue has not been reproduced, thanz Chris.
Comment 3 Marta Löfstedt 2018-01-24 06:57:45 UTC
Started again on CI_DRM_3673 KBL-shards:

https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@align.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@align64.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@color_evict.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@color_evict_range.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@evict.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@insert.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3673/shard-kbl1/igt@drm_mm@reserve.html

It appear to be a bit flip/floppy so eventhough CI_DRM_3675 I can say that it is fixed. 

[  232.521092] BUG: using smp_processor_id() in preemptible [00000000] code: drm_mm/6173
[  232.521102] caller is vprintk_func+0x47/0x80
[  232.521151] CPU: 3 PID: 6173 Comm: drm_mm Tainted: G     UD W        4.15.0-rc9-CI-CI_DRM_3673+ #1
[  232.521151] Hardware name:                  /NUC7i5BNB, BIOS BNKBL357.86A.0054.2017.1025.1822 10/25/2017
[  232.521152] Call Trace:
[  232.521152]  dump_stack+0x5f/0x86
[  232.521152]  check_preemption_disabled+0xd9/0xe0
[  232.521153]  vprintk_func+0x47/0x80
[  232.521153]  ? 0xffffffffa0041000
[  232.521153]  printk+0x3e/0x46
[  232.521154]  test_drm_mm_init+0x33/0x1000 [test_drm_mm]
[  232.521154]  ? 0xffffffffa0041000
[  232.521155]  do_one_initcall+0x39/0x150
[  232.521155]  ? kmem_cache_alloc_trace+0x1e7/0x2b0
[  232.521155]  do_init_module+0x56/0x1ef
[  232.521156]  load_module+0x231c/0x2d70
[  232.521156]  ? show_coresize+0x20/0x20
[  232.521156]  ? vfs_read+0x126/0x150
[  232.521156]  ? SyS_finit_module+0xa5/0xe0
[  232.521157]  SyS_finit_module+0xa5/0xe0
[  232.521157]  entry_SYSCALL_64_fastpath+0x22/0x8f
[  232.521157] RIP: 0033:0x7ff4181b89f9
[  232.521158] RSP: 002b:00007ffc69ffc8b8 EFLAGS: 00000206
Comment 5 Chris Wilson 2018-01-24 09:25:48 UTC
(In reply to Marta Löfstedt from comment #4)
> Nope, this was not fixed new test on CI_DRM_3677:
> 
> https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3677/shard-kbl5/
> igt@drm_mm@align32.html
> https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3677/shard-kbl5/
> igt@drm_mm@color.html

No, this is a cascade failure from live_hangcheck on kbl. See the stack page overflow double fault.
Comment 6 Marta Löfstedt 2018-01-24 09:34:19 UTC
(In reply to Chris Wilson from comment #5)
> (In reply to Marta Löfstedt from comment #4)
> > Nope, this was not fixed new test on CI_DRM_3677:
> > 
> > https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3677/shard-kbl5/
> > igt@drm_mm@align32.html
> > https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3677/shard-kbl5/
> > igt@drm_mm@color.html
> 
> No, this is a cascade failure from live_hangcheck on kbl. See the stack page
> overflow double fault.

Agreed, keep forgetting that selftest shard don't have randomized test ordering. igt@drv_selftest@live_hangcheck is covered by softdog bug 104262. But it has been flip/flopping pass/fail/incomplete for quite some time on KBL-shards. All fails seem to correspond to issues on drm_mm tests.
Comment 7 Marta Löfstedt 2018-01-24 10:50:31 UTC

*** This bug has been marked as a duplicate of bug 104262 ***
Comment 8 Chris Wilson 2018-02-01 15:54:54 UTC
Updating dup to link to kbl-only stack page overflow bug.

*** This bug has been marked as a duplicate of bug 104786 ***

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.