Bug 103856 - [CI] igt@perf_pmu - dmesg-warn - WARNING: possible circular locking dependency detected
Summary: [CI] igt@perf_pmu - dmesg-warn - WARNING: possible circular locking dependenc...
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: DRI git
Hardware: Other All
: medium normal
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard: ReadyForDev
Keywords:
Depends on:
Blocks:
 
Reported: 2017-11-23 07:17 UTC by Marta Löfstedt
Modified: 2017-11-23 07:53 UTC (History)
1 user (show)

See Also:
i915 platform: BXT, GLK, HSW, KBL, SNB
i915 features:

Attachments

Description Marta Löfstedt 2017-11-23 07:17:56 UTC 
https://intel-gfx-ci.01.org/tree/drm-tip/IGT_4004/shard-snb3/igt@perf_pmu@other-read-0.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3372/shard-apl7/igt@perf_pmu@other-read-0.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3372/shard-glkb6/igt@perf_pmu@other-read-0.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3372/shard-hsw4/igt@perf_pmu@other-read-0.html
https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_3372/shard-kbl3/igt@perf_pmu@other-read-0.html

[  376.994271] ======================================================
[  376.994277] WARNING: possible circular locking dependency detected
[  376.994285] 4.14.0-CI-CI_DRM_3372+ #1 Tainted: G     U         
[  376.994290] ------------------------------------------------------
[  376.994296] perf_pmu/1704 is trying to acquire lock:
[  376.994301]  (&mm->mmap_sem){++++}, at: [<ffffffff811bfe1e>] __might_fault+0x3e/0x90
[  376.994318] 
               but task is already holding lock:
[  376.994323]  (&cpuctx_mutex){+.+.}, at: [<ffffffff8116fe8c>] perf_event_ctx_lock_nested+0xbc/0x1d0
[  376.994337] 
               which lock already depends on the new lock.

[  376.994344] 
               the existing dependency chain (in reverse order) is:
[  376.994351] 
               -> #4 (&cpuctx_mutex){+.+.}:
[  376.994363]        __mutex_lock+0x86/0x9b0
[  376.994369]        perf_event_init_cpu+0x5a/0x90
[  376.994376]        perf_event_init+0x178/0x1a4
[  376.994384]        start_kernel+0x27f/0x3f1
[  376.994391]        verify_cpu+0x0/0xfb
[  376.994395] 
               -> #3 (pmus_lock){+.+.}:
[  376.994405]        __mutex_lock+0x86/0x9b0
[  376.994412]        perf_event_init_cpu+0x21/0x90
[  376.994418]        cpuhp_invoke_callback+0xca/0xc00
[  376.994424]        _cpu_up+0xa7/0x170
[  376.994429]        do_cpu_up+0x57/0x70
[  376.994435]        smp_init+0x62/0xa6
[  376.994441]        kernel_init_freeable+0x97/0x193
[  376.994448]        kernel_init+0xa/0x100
[  376.994454]        ret_from_fork+0x27/0x40
[  376.994458] 
               -> #2 (cpu_hotplug_lock.rw_sem){++++}:
[  376.994469]        cpus_read_lock+0x39/0xa0
[  376.994476]        apply_workqueue_attrs+0x12/0x50
[  376.994483]        __alloc_workqueue_key+0x1d8/0x4d8
[  376.994548]        i915_gem_init_userptr+0x5f/0x80 [i915]
[  376.994600]        i915_gem_init+0x7c/0x390 [i915]
[  376.994645]        i915_driver_load+0x99e/0x15c0 [i915]
[  376.994691]        i915_pci_probe+0x33/0x90 [i915]
[  376.994700]        pci_device_probe+0xa1/0x130
[  376.994707]        driver_probe_device+0x293/0x440
[  376.994713]        __driver_attach+0xde/0xe0
[  376.994719]        bus_for_each_dev+0x5c/0x90
[  376.994726]        bus_add_driver+0x16d/0x260
[  376.994732]        driver_register+0x57/0xc0
[  376.994737]        do_one_initcall+0x3e/0x160
[  376.994744]        do_init_module+0x5b/0x1fa
[  376.994750]        load_module+0x2374/0x2dc0
[  376.994756]        SyS_finit_module+0xaa/0xe0
[  376.994762]        do_syscall_64+0x5e/0x170
[  376.994768]        return_from_SYSCALL_64+0x0/0x7a
[  376.994772] 
               -> #1 (&dev->struct_mutex){+.+.}:
[  376.994784]        __mutex_lock+0x86/0x9b0
[  376.994833]        i915_mutex_lock_interruptible+0x4c/0x130 [i915]
[  376.994908]        i915_gem_fault+0x206/0x760 [i915]
[  376.994920]        __do_fault+0x1a/0x70
[  376.994927]        __handle_mm_fault+0x9b0/0xdb0
[  376.994935]        handle_mm_fault+0x154/0x300
[  376.994943]        __do_page_fault+0x2d6/0x570
[  376.994951]        page_fault+0x22/0x30
[  376.994957] 
               -> #0 (&mm->mmap_sem){++++}:
[  376.994973]        lock_acquire+0xaf/0x200
[  376.994981]        __might_fault+0x68/0x90
[  376.994989]        _copy_to_user+0x1e/0x70
[  376.994997]        perf_read+0x1aa/0x290
[  376.995005]        __vfs_read+0x23/0x120
[  376.995012]        vfs_read+0xa3/0x150
[  376.995019]        SyS_read+0x45/0xb0
[  376.995027]        entry_SYSCALL_64_fastpath+0x1c/0xb1
[  376.995033] 
               other info that might help us debug this:

[  376.995045] Chain exists of:
                 &mm->mmap_sem --> pmus_lock --> &cpuctx_mutex

[  376.995064]  Possible unsafe locking scenario:

[  376.995072]        CPU0                    CPU1
[  376.995078]        ----                    ----
[  376.995084]   lock(&cpuctx_mutex);
[  376.995092]                                lock(pmus_lock);
[  376.995104]                                lock(&cpuctx_mutex);
[  376.995113]   lock(&mm->mmap_sem);
[  376.995121] 
                *** DEADLOCK ***

[  376.995131] 1 lock held by perf_pmu/1704:
[  376.995137]  #0:  (&cpuctx_mutex){+.+.}, at: [<ffffffff8116fe8c>] perf_event_ctx_lock_nested+0xbc/0x1d0
[  376.995155] 
               stack backtrace:
[  376.995165] CPU: 1 PID: 1704 Comm: perf_pmu Tainted: G     U          4.14.0-CI-CI_DRM_3372+ #1
[  376.995176] Hardware name:                  /NUC6CAYB, BIOS AYAPLCEL.86A.0040.2017.0619.1722 06/19/2017
[  376.995187] Call Trace:
[  376.995197]  dump_stack+0x5f/0x86
[  376.995207]  print_circular_bug.isra.18+0x1d0/0x2c0
[  376.995217]  __lock_acquire+0x19c3/0x1b60
[  376.995227]  ? generic_exec_single+0x77/0xe0
[  376.995237]  ? lock_acquire+0xaf/0x200
[  376.995245]  lock_acquire+0xaf/0x200
[  376.995254]  ? __might_fault+0x3e/0x90
[  376.995263]  __might_fault+0x68/0x90
[  376.995271]  ? __might_fault+0x3e/0x90
[  376.995280]  _copy_to_user+0x1e/0x70
[  376.995288]  perf_read+0x1aa/0x290
[  376.995298]  __vfs_read+0x23/0x120
[  376.995307]  ? entry_SYSCALL_64_fastpath+0x5/0xb1
[  376.995316]  vfs_read+0xa3/0x150
[  376.995324]  SyS_read+0x45/0xb0
[  376.995332]  entry_SYSCALL_64_fastpath+0x1c/0xb1
[  376.995341] RIP: 0033:0x7f4bf6e9a6d0
[  376.995347] RSP: 002b:00007fffbff1f938 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
[  376.995359] RAX: ffffffffffffffda RBX: ffffc900006bfff0 RCX: 00007f4bf6e9a6d0
[  376.995368] RDX: 0000000000000010 RSI: 00007fffbff1f940 RDI: 0000000000000005
[  376.995377] RBP: 0000000000000005 R08: 0000565219788560 R09: 0000000000000000
[  376.995386] R10: 0000000000000073 R11: 0000000000000246 R12: 0000000000000046
[  376.995395] R13: 00007fffbff21100 R14: 0000000000000000 R15: 0000000000000000
Comment 44 Chris Wilson 2017-11-23 07:51:32 UTC 
commit ee48700dd57d9ce783ec40f035b324d0b75632e4
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Wed Nov 22 17:26:21 2017 +0000

    drm/i915: Call i915_gem_init_userptr() before taking struct_mutex
    
    We don't need struct_mutex to initialise userptr (it just allocates a
    workqueue for itself etc), but we do need struct_mutex later on in
    i915_gem_init() in order to feed requests onto the HW.
Comment 45 Marta Löfstedt 2017-11-23 07:53:24 UTC 
The fix was integrated to CI_DRM_3373, the issue is not reproduced I will close and archive.

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.