[Intel-gfx] ✗ Fi.CI.BAT: failure for drm: Pass along the hotplug connector to the uevent

[Chris Wilson]

On Fri, Oct 21, 2016 at 09:01:59AM -0000, Patchwork wrote:
> == Series Details ==
> 
> Series: drm: Pass along the hotplug connector to the uevent
> URL   : https://patchwork.freedesktop.org/series/14152/
> State : failure
> 
> == Summary ==
> 
> Series 14152v1 drm: Pass along the hotplug connector to the uevent
> https://patchwork.freedesktop.org/api/1.0/series/14152/revisions/1/mbox/
> 
> Test kms_flip:
>         Subgroup basic-flip-vs-dpms:
>                 pass       -> INCOMPLETE (fi-snb-2600)
>                 pass       -> INCOMPLETE (fi-ivb-3770)
>         Subgroup basic-flip-vs-modeset:
>                 pass       -> INCOMPLETE (fi-ivb-3520m)

mutex deadlock at

[  259.382365] CPU: 0 PID: 71 Comm: kworker/0:1 Tainted: G     U          4.9.0-rc1-CI-Patchwork_2781+ #1
[  259.382369] Hardware name: Dell Inc. XPS 8300  /0Y2MRG, BIOS A06 10/17/2011
[  259.382409] Workqueue: events i915_hotplug_work_func [i915]
[  259.382414]  ffffc9000027bb40 ffffffff8142dce5 ffffffff827de530 ffffffff827de530
[  259.382423]  ffffc9000027bbf8 ffffffff810d8624 ffff88012a37af68 000000000027bb90
[  259.382432]  ffffffff810d6c3f ffff880100000003 ffffffff00012000 b57c70f93abc779a
[  259.382440] Call Trace:
[  259.382448]  [<ffffffff8142dce5>] dump_stack+0x67/0x92 
[  259.382455]  [<ffffffff810d8624>] __lock_acquire+0x1544/0x1930
[  259.382461]  [<ffffffff810d6c3f>] ? mark_held_locks+0x6f/0xa0
[  259.382468]  [<ffffffff810d6d92>] ? trace_hardirqs_on_caller+0x122/0x1b0
[  259.382474]  [<ffffffff810d6e2d>] ? trace_hardirqs_on+0xd/0x10
[  259.382480]  [<ffffffff810d8e32>] lock_acquire+0xb2/0x200
[  259.382486]  [<ffffffff8154ce5e>] ? drm_fb_helper_hotplug_event+0x2e/0x170
[  259.382493]  [<ffffffff8154ce5e>] ? drm_fb_helper_hotplug_event+0x2e/0x170
[  259.382500]  [<ffffffff81810862>] mutex_lock_nested+0x62/0x400
[  259.382506]  [<ffffffff8154ce5e>] ? drm_fb_helper_hotplug_event+0x2e/0x170
[  259.382512]  [<ffffffff81558938>] ? drm_sysfs_hotplug_event+0x88/0xa0
[  259.382518]  [<ffffffff8154ce5e>] drm_fb_helper_hotplug_event+0x2e/0x170
[  259.382558]  [<ffffffffa00dffcf>] intel_fbdev_output_poll_changed+0x1f/0x30 [i915]
[  259.382566]  [<ffffffff8153e2c2>] drm_kms_helper_hotplug_event+0x22/0x30
[  259.382604]  [<ffffffffa00d6750>] i915_hotplug_work_func+0x230/0x280 [i915]
[  259.382611]  [<ffffffff8109c22c>] process_one_work+0x1ec/0x6b0
[  259.382616]  [<ffffffff8109c1a6>] ? process_one_work+0x166/0x6b0 
[  259.382623]  [<ffffffff8109c739>] worker_thread+0x49/0x490
[  259.382628]  [<ffffffff8109c6f0>] ? process_one_work+0x6b0/0x6b0 
[  259.382634]  [<ffffffff8109c6f0>] ? process_one_work+0x6b0/0x6b0 
[  259.382639]  [<ffffffff810a2a5b>] kthread+0xeb/0x110
[  259.382645]  [<ffffffff810a2970>] ? kthread_park+0x60/0x60 
[  259.382651]  [<ffffffff818160a7>] ret_from_fork+0x27/0x40

due to fbdev callback from the helper. So that has to remain outside of the lock.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre