Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Fri, Oct 19, 2012 at 02:49:32PM +0200, Peter Zijlstra wrote: > Of course, if you do run out of lock classes, the next thing to do is > to find the offending lock classes. First, the following command gives > you the number of lock classes currently in use along with the maximum: > > grep "lock-classes" /proc/lockdep_stats > > This command produces the following output on a modest system: > > lock-classes: 748 [max: 8191] After the BUG gets hit.. lock-classes: 1726 [max: 8191] > If the number allocated (748 above) increases continually over time, > then there is likely a leak. The following command can be used to > identify the leaking lock classes: > > grep "BD" /proc/lockdep > > Run the command and save the output, then compare against the output from > a later run of this command to identify the leakers. This same output > can also help you find situations where runtime lock initialization has > been omitted. I've not had chance to do this, because after the BUG, lockdep turns itself off, and I've not rebooted. I'm probably not going to get to this until after the weekend. There's just a *lot* of dependancies. Here's the full output http://codemonkey.org.uk/junk/lockdep the top few backwards deps.. 81c8f218 FD:1 BD: 1201 -.-.-.: pool_lock 82ae1210 FD:2 BD: 1200 -.-.-.: _hash[i].lock 820677c1 FD:1 BD: 1131 -.-.-.: _rq->rt_runtime_lock 82066949 FD:3 BD: 1131 -.-.-.: _base->lock 820221c0 FD:1 BD: 1130 -.-.-.: >cputimer.lock 820677b8 FD:5 BD: 1129 -.-.-.: _b->rt_runtime_lock 82067675 FD:3 BD: 1129 ..-.-.: >lock/1 82067674 FD:8 BD: 1128 -.-.-.: >lock 8298bbd0 FD:1 BD: 1006 -.-.-.: &(>list_lock)->rlock Dave -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Fri, 2012-10-19 at 01:21 -0400, Dave Jones wrote: > > Not sure why you are CC'ing a call site, rather than the maintainers of > > the code. Just looks like lockdep is using too small a static value. > > Though it is pretty darn large... > > You're right, it's a huge chunk of memory. > It looks like I can trigger this from multiple callsites.. > Another different trace below. > > Not sure why this suddenly got a lot worse in 3.7 Did we add a static array of structures with locks in somewhere? Doing that is a great way of blowing up the number of lock classes and the resulting amount of lock dependency chains. >From Documentation/lockdep-design.txt; it talks about overflowing MAX_LOCKDEP_KEYS, but I suppose its a good starts for overflowing the dependency entries too, more classes means more dependencies after all. --- Of course, if you do run out of lock classes, the next thing to do is to find the offending lock classes. First, the following command gives you the number of lock classes currently in use along with the maximum: grep "lock-classes" /proc/lockdep_stats This command produces the following output on a modest system: lock-classes: 748 [max: 8191] If the number allocated (748 above) increases continually over time, then there is likely a leak. The following command can be used to identify the leaking lock classes: grep "BD" /proc/lockdep Run the command and save the output, then compare against the output from a later run of this command to identify the leakers. This same output can also help you find situations where runtime lock initialization has been omitted. -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Fri, 2012-10-19 at 01:21 -0400, Dave Jones wrote: Not sure why you are CC'ing a call site, rather than the maintainers of the code. Just looks like lockdep is using too small a static value. Though it is pretty darn large... You're right, it's a huge chunk of memory. It looks like I can trigger this from multiple callsites.. Another different trace below. Not sure why this suddenly got a lot worse in 3.7 Did we add a static array of structures with locks in somewhere? Doing that is a great way of blowing up the number of lock classes and the resulting amount of lock dependency chains. From Documentation/lockdep-design.txt; it talks about overflowing MAX_LOCKDEP_KEYS, but I suppose its a good starts for overflowing the dependency entries too, more classes means more dependencies after all. --- Of course, if you do run out of lock classes, the next thing to do is to find the offending lock classes. First, the following command gives you the number of lock classes currently in use along with the maximum: grep lock-classes /proc/lockdep_stats This command produces the following output on a modest system: lock-classes: 748 [max: 8191] If the number allocated (748 above) increases continually over time, then there is likely a leak. The following command can be used to identify the leaking lock classes: grep BD /proc/lockdep Run the command and save the output, then compare against the output from a later run of this command to identify the leakers. This same output can also help you find situations where runtime lock initialization has been omitted. -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Fri, Oct 19, 2012 at 02:49:32PM +0200, Peter Zijlstra wrote: Of course, if you do run out of lock classes, the next thing to do is to find the offending lock classes. First, the following command gives you the number of lock classes currently in use along with the maximum: grep lock-classes /proc/lockdep_stats This command produces the following output on a modest system: lock-classes: 748 [max: 8191] After the BUG gets hit.. lock-classes: 1726 [max: 8191] If the number allocated (748 above) increases continually over time, then there is likely a leak. The following command can be used to identify the leaking lock classes: grep BD /proc/lockdep Run the command and save the output, then compare against the output from a later run of this command to identify the leakers. This same output can also help you find situations where runtime lock initialization has been omitted. I've not had chance to do this, because after the BUG, lockdep turns itself off, and I've not rebooted. I'm probably not going to get to this until after the weekend. There's just a *lot* of dependancies. Here's the full output http://codemonkey.org.uk/junk/lockdep the top few backwards deps.. 81c8f218 FD:1 BD: 1201 -.-.-.: pool_lock 82ae1210 FD:2 BD: 1200 -.-.-.: obj_hash[i].lock 820677c1 FD:1 BD: 1131 -.-.-.: rt_rq-rt_runtime_lock 82066949 FD:3 BD: 1131 -.-.-.: cpu_base-lock 820221c0 FD:1 BD: 1130 -.-.-.: sig-cputimer.lock 820677b8 FD:5 BD: 1129 -.-.-.: rt_b-rt_runtime_lock 82067675 FD:3 BD: 1129 ..-.-.: rq-lock/1 82067674 FD:8 BD: 1128 -.-.-.: rq-lock 8298bbd0 FD:1 BD: 1006 -.-.-.: (n-list_lock)-rlock Dave -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Thu, Oct 18, 2012 at 07:53:08AM +0200, Jens Axboe wrote: > On 2012-10-18 03:53, Dave Jones wrote: > > Triggered while fuzz testing.. > > > > > > BUG: MAX_LOCKDEP_ENTRIES too low! > > turning off the locking correctness validator. > > Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 > > Call Trace: > > [] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 > > [] __lock_acquire+0x1121/0x1ba0 > > [] lock_acquire+0xa2/0x220 > > [] ? free_one_page+0x32/0x450 > > [] ? sub_preempt_count+0x79/0xd0 > > [] _raw_spin_lock+0x40/0x80 > > [] ? free_one_page+0x32/0x450 > > [] free_one_page+0x32/0x450 > > [] ? __free_pages_ok.part.58+0x51/0x110 > > [] __free_pages_ok.part.58+0xac/0x110 > > [] __free_pages+0x73/0x90 > > [] __free_slab+0xd3/0x1b0 > > [] discard_slab+0x39/0x50 > > [] __slab_free+0x378/0x3a3 > > [] ? ioc_release_fn+0x99/0xe0 > > [] ? ioc_release_fn+0x99/0xe0 > > [] kmem_cache_free+0x2f2/0x320 > > [] ? sub_preempt_count+0x79/0xd0 > > [] ioc_release_fn+0x99/0xe0 > > [] process_one_work+0x207/0x780 > > [] ? process_one_work+0x197/0x780 > > [] ? get_io_context+0x20/0x20 > > [] worker_thread+0x15e/0x440 > > [] ? rescuer_thread+0x240/0x240 > > [] kthread+0xed/0x100 > > [] ? put_lock_stats.isra.25+0xe/0x40 > > [] ? kthread_create_on_node+0x160/0x160 > > [] ret_from_fork+0x7c/0xb0 > > [] ? kthread_create_on_node+0x160/0x160 > > Not sure why you are CC'ing a call site, rather than the maintainers of > the code. Just looks like lockdep is using too small a static value. > Though it is pretty darn large... You're right, it's a huge chunk of memory. It looks like I can trigger this from multiple callsites.. Another different trace below. Not sure why this suddenly got a lot worse in 3.7 Peter, Ingo ? Dave BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22350, comm: trinity-child0 Not tainted 3.7.0-rc1+ #36 Call Trace: [] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [] __lock_acquire+0x1121/0x1ba0 [] ? local_clock+0x89/0xa0 [] ? __swap_duplicate+0xb5/0x190 [] ? trace_hardirqs_off_caller+0x28/0xd0 [] lock_acquire+0xa2/0x220 [] ? __add_to_swap_cache+0x6d/0x180 [] ? _raw_spin_lock_irq+0x29/0x90 [] _raw_spin_lock_irq+0x56/0x90 [] ? __add_to_swap_cache+0x6d/0x180 [] __add_to_swap_cache+0x6d/0x180 [] read_swap_cache_async+0xb5/0x220 [] swapin_readahead+0x9e/0xf0 [] handle_pte_fault+0x6d6/0xae0 [] ? sub_preempt_count+0x79/0xd0 [] ? delay_tsc+0xae/0x120 [] ? __const_udelay+0x28/0x30 [] handle_mm_fault+0x289/0x350 [] __do_page_fault+0x18e/0x530 [] ? local_clock+0x89/0xa0 [] ? __slab_free+0x32e/0x3a3 [] ? rcu_user_exit+0xc9/0xf0 [] ? 0xa01f [] ? 0xa01f [] do_page_fault+0x2b/0x50 [] page_fault+0x28/0x30 [] ? 0xa01f [] ? strncpy_from_user+0x6c/0x120 [] ? 0xa01f [] setxattr+0x6f/0x1d0 [] ? sub_preempt_count+0x79/0xd0 [] ? __percpu_counter_add+0x75/0xc0 [] ? __sb_start_write+0x101/0x1d0 [] ? mnt_want_write+0x24/0x50 [] ? mnt_want_write+0x24/0x50 [] ? get_parent_ip+0x11/0x50 [] ? sub_preempt_count+0x79/0xd0 [] ? __mnt_want_write+0x60/0xa0 [] ? 0xa01f [] sys_setxattr+0x95/0xb0 [] tracesys+0xe1/0xe6 [] ? 0xa01f -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On Thu, Oct 18, 2012 at 07:53:08AM +0200, Jens Axboe wrote: On 2012-10-18 03:53, Dave Jones wrote: Triggered while fuzz testing.. BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 Call Trace: [810decdd] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [810e2871] __lock_acquire+0x1121/0x1ba0 [810e3a12] lock_acquire+0xa2/0x220 [8117bad2] ? free_one_page+0x32/0x450 [816c5a59] ? sub_preempt_count+0x79/0xd0 [816c0800] _raw_spin_lock+0x40/0x80 [8117bad2] ? free_one_page+0x32/0x450 [8117bad2] free_one_page+0x32/0x450 [8117bf41] ? __free_pages_ok.part.58+0x51/0x110 [8117bf9c] __free_pages_ok.part.58+0xac/0x110 [8117cd73] __free_pages+0x73/0x90 [811cb4f3] __free_slab+0xd3/0x1b0 [811cb609] discard_slab+0x39/0x50 [816b77db] __slab_free+0x378/0x3a3 [81341289] ? ioc_release_fn+0x99/0xe0 [81341289] ? ioc_release_fn+0x99/0xe0 [811cd4e2] kmem_cache_free+0x2f2/0x320 [816c5a59] ? sub_preempt_count+0x79/0xd0 [81341289] ioc_release_fn+0x99/0xe0 [81095a37] process_one_work+0x207/0x780 [810959c7] ? process_one_work+0x197/0x780 [813411f0] ? get_io_context+0x20/0x20 [8109638e] worker_thread+0x15e/0x440 [81096230] ? rescuer_thread+0x240/0x240 [8109d0cd] kthread+0xed/0x100 [810de02e] ? put_lock_stats.isra.25+0xe/0x40 [8109cfe0] ? kthread_create_on_node+0x160/0x160 [816c9dac] ret_from_fork+0x7c/0xb0 [8109cfe0] ? kthread_create_on_node+0x160/0x160 Not sure why you are CC'ing a call site, rather than the maintainers of the code. Just looks like lockdep is using too small a static value. Though it is pretty darn large... You're right, it's a huge chunk of memory. It looks like I can trigger this from multiple callsites.. Another different trace below. Not sure why this suddenly got a lot worse in 3.7 Peter, Ingo ? Dave BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22350, comm: trinity-child0 Not tainted 3.7.0-rc1+ #36 Call Trace: [810decdd] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [810e2871] __lock_acquire+0x1121/0x1ba0 [810b6f89] ? local_clock+0x89/0xa0 [811b84a5] ? __swap_duplicate+0xb5/0x190 [810dea58] ? trace_hardirqs_off_caller+0x28/0xd0 [810e3a12] lock_acquire+0xa2/0x220 [811b663d] ? __add_to_swap_cache+0x6d/0x180 [816c0f79] ? _raw_spin_lock_irq+0x29/0x90 [816c0fa6] _raw_spin_lock_irq+0x56/0x90 [811b663d] ? __add_to_swap_cache+0x6d/0x180 [811b663d] __add_to_swap_cache+0x6d/0x180 [811b6c45] read_swap_cache_async+0xb5/0x220 [811b6e4e] swapin_readahead+0x9e/0xf0 [811a2c86] handle_pte_fault+0x6d6/0xae0 [816c5ad9] ? sub_preempt_count+0x79/0xd0 [8136d37e] ? delay_tsc+0xae/0x120 [8136d268] ? __const_udelay+0x28/0x30 [811a4919] handle_mm_fault+0x289/0x350 [816c538e] __do_page_fault+0x18e/0x530 [810b6f89] ? local_clock+0x89/0xa0 [816b77e1] ? __slab_free+0x32e/0x3a3 [8112d2e9] ? rcu_user_exit+0xc9/0xf0 [a020] ? 0xa01f [a020] ? 0xa01f [816c575b] do_page_fault+0x2b/0x50 [816c1e38] page_fault+0x28/0x30 [a020] ? 0xa01f [813885ac] ? strncpy_from_user+0x6c/0x120 [a020] ? 0xa01f [8120e94f] setxattr+0x6f/0x1d0 [816c5ad9] ? sub_preempt_count+0x79/0xd0 [8137e8b5] ? __percpu_counter_add+0x75/0xc0 [811e92d1] ? __sb_start_write+0x101/0x1d0 [81209d94] ? mnt_want_write+0x24/0x50 [81209d94] ? mnt_want_write+0x24/0x50 [810b0de1] ? get_parent_ip+0x11/0x50 [816c5ad9] ? sub_preempt_count+0x79/0xd0 [81209d30] ? __mnt_want_write+0x60/0xa0 [a020] ? 0xa01f [8120ecd5] sys_setxattr+0x95/0xb0 [816ca108] tracesys+0xe1/0xe6 [a020] ? 0xa01f -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On 2012-10-18 03:53, Dave Jones wrote: > Triggered while fuzz testing.. > > > BUG: MAX_LOCKDEP_ENTRIES too low! > turning off the locking correctness validator. > Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 > Call Trace: > [] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 > [] __lock_acquire+0x1121/0x1ba0 > [] lock_acquire+0xa2/0x220 > [] ? free_one_page+0x32/0x450 > [] ? sub_preempt_count+0x79/0xd0 > [] _raw_spin_lock+0x40/0x80 > [] ? free_one_page+0x32/0x450 > [] free_one_page+0x32/0x450 > [] ? __free_pages_ok.part.58+0x51/0x110 > [] __free_pages_ok.part.58+0xac/0x110 > [] __free_pages+0x73/0x90 > [] __free_slab+0xd3/0x1b0 > [] discard_slab+0x39/0x50 > [] __slab_free+0x378/0x3a3 > [] ? ioc_release_fn+0x99/0xe0 > [] ? ioc_release_fn+0x99/0xe0 > [] kmem_cache_free+0x2f2/0x320 > [] ? sub_preempt_count+0x79/0xd0 > [] ioc_release_fn+0x99/0xe0 > [] process_one_work+0x207/0x780 > [] ? process_one_work+0x197/0x780 > [] ? get_io_context+0x20/0x20 > [] worker_thread+0x15e/0x440 > [] ? rescuer_thread+0x240/0x240 > [] kthread+0xed/0x100 > [] ? put_lock_stats.isra.25+0xe/0x40 > [] ? kthread_create_on_node+0x160/0x160 > [] ret_from_fork+0x7c/0xb0 > [] ? kthread_create_on_node+0x160/0x160 Not sure why you are CC'ing a call site, rather than the maintainers of the code. Just looks like lockdep is using too small a static value. Though it is pretty darn large... -- Jens Axboe -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
Triggered while fuzz testing.. BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 Call Trace: [] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [] __lock_acquire+0x1121/0x1ba0 [] lock_acquire+0xa2/0x220 [] ? free_one_page+0x32/0x450 [] ? sub_preempt_count+0x79/0xd0 [] _raw_spin_lock+0x40/0x80 [] ? free_one_page+0x32/0x450 [] free_one_page+0x32/0x450 [] ? __free_pages_ok.part.58+0x51/0x110 [] __free_pages_ok.part.58+0xac/0x110 [] __free_pages+0x73/0x90 [] __free_slab+0xd3/0x1b0 [] discard_slab+0x39/0x50 [] __slab_free+0x378/0x3a3 [] ? ioc_release_fn+0x99/0xe0 [] ? ioc_release_fn+0x99/0xe0 [] kmem_cache_free+0x2f2/0x320 [] ? sub_preempt_count+0x79/0xd0 [] ioc_release_fn+0x99/0xe0 [] process_one_work+0x207/0x780 [] ? process_one_work+0x197/0x780 [] ? get_io_context+0x20/0x20 [] worker_thread+0x15e/0x440 [] ? rescuer_thread+0x240/0x240 [] kthread+0xed/0x100 [] ? put_lock_stats.isra.25+0xe/0x40 [] ? kthread_create_on_node+0x160/0x160 [] ret_from_fork+0x7c/0xb0 [] ? kthread_create_on_node+0x160/0x160 -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
Triggered while fuzz testing.. BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 Call Trace: [810decdd] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [810e2871] __lock_acquire+0x1121/0x1ba0 [810e3a12] lock_acquire+0xa2/0x220 [8117bad2] ? free_one_page+0x32/0x450 [816c5a59] ? sub_preempt_count+0x79/0xd0 [816c0800] _raw_spin_lock+0x40/0x80 [8117bad2] ? free_one_page+0x32/0x450 [8117bad2] free_one_page+0x32/0x450 [8117bf41] ? __free_pages_ok.part.58+0x51/0x110 [8117bf9c] __free_pages_ok.part.58+0xac/0x110 [8117cd73] __free_pages+0x73/0x90 [811cb4f3] __free_slab+0xd3/0x1b0 [811cb609] discard_slab+0x39/0x50 [816b77db] __slab_free+0x378/0x3a3 [81341289] ? ioc_release_fn+0x99/0xe0 [81341289] ? ioc_release_fn+0x99/0xe0 [811cd4e2] kmem_cache_free+0x2f2/0x320 [816c5a59] ? sub_preempt_count+0x79/0xd0 [81341289] ioc_release_fn+0x99/0xe0 [81095a37] process_one_work+0x207/0x780 [810959c7] ? process_one_work+0x197/0x780 [813411f0] ? get_io_context+0x20/0x20 [8109638e] worker_thread+0x15e/0x440 [81096230] ? rescuer_thread+0x240/0x240 [8109d0cd] kthread+0xed/0x100 [810de02e] ? put_lock_stats.isra.25+0xe/0x40 [8109cfe0] ? kthread_create_on_node+0x160/0x160 [816c9dac] ret_from_fork+0x7c/0xb0 [8109cfe0] ? kthread_create_on_node+0x160/0x160 -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
Re: MAX_LOCKDEP_ENTRIES too low (called from ioc_release_fn)
On 2012-10-18 03:53, Dave Jones wrote: Triggered while fuzz testing.. BUG: MAX_LOCKDEP_ENTRIES too low! turning off the locking correctness validator. Pid: 22788, comm: kworker/2:1 Not tainted 3.7.0-rc1+ #34 Call Trace: [810decdd] add_lock_to_list.isra.29.constprop.45+0xdd/0xf0 [810e2871] __lock_acquire+0x1121/0x1ba0 [810e3a12] lock_acquire+0xa2/0x220 [8117bad2] ? free_one_page+0x32/0x450 [816c5a59] ? sub_preempt_count+0x79/0xd0 [816c0800] _raw_spin_lock+0x40/0x80 [8117bad2] ? free_one_page+0x32/0x450 [8117bad2] free_one_page+0x32/0x450 [8117bf41] ? __free_pages_ok.part.58+0x51/0x110 [8117bf9c] __free_pages_ok.part.58+0xac/0x110 [8117cd73] __free_pages+0x73/0x90 [811cb4f3] __free_slab+0xd3/0x1b0 [811cb609] discard_slab+0x39/0x50 [816b77db] __slab_free+0x378/0x3a3 [81341289] ? ioc_release_fn+0x99/0xe0 [81341289] ? ioc_release_fn+0x99/0xe0 [811cd4e2] kmem_cache_free+0x2f2/0x320 [816c5a59] ? sub_preempt_count+0x79/0xd0 [81341289] ioc_release_fn+0x99/0xe0 [81095a37] process_one_work+0x207/0x780 [810959c7] ? process_one_work+0x197/0x780 [813411f0] ? get_io_context+0x20/0x20 [8109638e] worker_thread+0x15e/0x440 [81096230] ? rescuer_thread+0x240/0x240 [8109d0cd] kthread+0xed/0x100 [810de02e] ? put_lock_stats.isra.25+0xe/0x40 [8109cfe0] ? kthread_create_on_node+0x160/0x160 [816c9dac] ret_from_fork+0x7c/0xb0 [8109cfe0] ? kthread_create_on_node+0x160/0x160 Not sure why you are CC'ing a call site, rather than the maintainers of the code. Just looks like lockdep is using too small a static value. Though it is pretty darn large... -- Jens Axboe -- To unsubscribe from this list: send the line unsubscribe linux-kernel in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/