Re: drm_vm.c:drm_mmap: possible circular locking dependency detected (was: Re: Linux 2.6.33-rc2 - Merry Christmas ...)

From: KOSAKI Motohiro
Date: Sun Dec 27 2009 - 19:40:41 EST


Hi

Quick analysis is here.

> Hi,
>
> this jumped into dmesg upon resume (.config and dmesg are attached in
> the previous "EHCI resume sysfs duplicates..." message in this thread):
>
> =======================================================
> [ INFO: possible circular locking dependency detected ]
> 2.6.33-rc2-00001-g6d7daec #1
> -------------------------------------------------------
> Xorg/3076 is trying to acquire lock:
> (&dev->struct_mutex){+.+.+.}, at: [<ffffffff81223fd4>] drm_mmap+0x38/0x5c
>
> but task is already holding lock:
> (&mm->mmap_sem){++++++}, at: [<ffffffff810b7509>] sys_mmap_pgoff+0xd6/0x1b4
>
> which lock already depends on the new lock.
>
>
> the existing dependency chain (in reverse order) is:
>
> -> #3 (&mm->mmap_sem){++++++}:
> [<ffffffff810694c0>] __lock_acquire+0x1373/0x16fd
> [<ffffffff8106993c>] lock_acquire+0xf2/0x116
> [<ffffffff810bb2b5>] might_fault+0x95/0xb8 <- mmap_sem
> [<ffffffff810e87d6>] filldir+0x75/0xd0 <- sysfs_mutex
> [<ffffffff8112be2a>] sysfs_readdir+0x10f/0x149
> [<ffffffff810e895b>] vfs_readdir+0x6b/0xa8
> [<ffffffff810e8ae1>] sys_getdents+0x81/0xd1
> [<ffffffff810022f2>] system_call_fastpath+0x16/0x1b
>
> -> #2 (sysfs_mutex){+.+.+.}:
> [<ffffffff810694c0>] __lock_acquire+0x1373/0x16fd
> [<ffffffff8106993c>] lock_acquire+0xf2/0x116
> [<ffffffff813e9e5c>] mutex_lock_nested+0x63/0x354
> [<ffffffff8112c488>] sysfs_addrm_start+0x26/0x28 <- sysfs_mutex
> [<ffffffff8112c940>] sysfs_remove_dir+0x52/0x8d
> [<ffffffff8118b6f9>] kobject_del+0x16/0x37
> [<ffffffff8118b758>] kobject_release+0x3e/0x66
> [<ffffffff8118c5b5>] kref_put+0x43/0x4d
> [<ffffffff8118b674>] kobject_put+0x47/0x4b
> [<ffffffff813e11c1>] cacheinfo_cpu_callback+0xa2/0xdb
> [<ffffffff8105c317>] notifier_call_chain+0x37/0x63
> [<ffffffff8105c3c7>] raw_notifier_call_chain+0x14/0x16
> [<ffffffff813d58ec>] _cpu_down+0x1a5/0x29a <- cpu_hotplug.lock
> [<ffffffff8103c851>] disable_nonboot_cpus+0x74/0x10d
> [<ffffffff8107793e>] hibernation_snapshot+0x99/0x1d3
> [<ffffffff81077b46>] hibernate+0xce/0x172
> [<ffffffff810768d4>] state_store+0x5c/0xd3
> [<ffffffff8118b48b>] kobj_attr_store+0x17/0x19
> [<ffffffff8112b4bd>] sysfs_write_file+0x108/0x144
> [<ffffffff810daf53>] vfs_write+0xb2/0x153
> [<ffffffff810db0b7>] sys_write+0x4a/0x71
> [<ffffffff810022f2>] system_call_fastpath+0x16/0x1b
>
> -> #1 (cpu_hotplug.lock){+.+.+.}:
> [<ffffffff810694c0>] __lock_acquire+0x1373/0x16fd
> [<ffffffff8106993c>] lock_acquire+0xf2/0x116
> [<ffffffff813e9e5c>] mutex_lock_nested+0x63/0x354
> [<ffffffff8103c980>] get_online_cpus+0x3c/0x50 <- cpu_hotplug.lock
> [<ffffffff81014c1a>] mtrr_del_page+0x3e/0x13c
> [<ffffffff81014d5f>] mtrr_del+0x47/0x4f
> [<ffffffff8121c23b>] drm_rmmap_locked+0xdc/0x1a2
> [<ffffffff812226e3>] drm_master_destroy+0x86/0x11f
> [<ffffffff8118c5b5>] kref_put+0x43/0x4d
> [<ffffffff812225c4>] drm_master_put+0x20/0x2b
> [<ffffffff8121ea71>] drm_release+0x54b/0x688 <- dev->struct_mutex
> [<ffffffff810dbb24>] __fput+0x125/0x1e7
> [<ffffffff810dbc00>] fput+0x1a/0x1c
> [<ffffffff810d8d02>] filp_close+0x5d/0x67
> [<ffffffff810d8db9>] sys_close+0xad/0xe7
> [<ffffffff810022f2>] system_call_fastpath+0x16/0x1b
>
> -> #0 (&dev->struct_mutex){+.+.+.}:
> [<ffffffff81069170>] __lock_acquire+0x1023/0x16fd
> [<ffffffff8106993c>] lock_acquire+0xf2/0x116
> [<ffffffff813e9e5c>] mutex_lock_nested+0x63/0x354
> [<ffffffff81223fd4>] drm_mmap+0x38/0x5c <- dev->struct_mutex
> [<ffffffff810c34f5>] mmap_region+0x2e0/0x4ff
> [<ffffffff810c39a4>] do_mmap_pgoff+0x290/0x2f3
> [<ffffffff810b7529>] sys_mmap_pgoff+0xf6/0x1b4
> [<ffffffff8100719b>] sys_mmap+0x22/0x27 <- mmap_sem
> [<ffffffff810022f2>] system_call_fastpath+0x16/0x1b

This output seems to suggest need to fix drm.



--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/