| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <4F9C6653.9030407@gmail.com>
Date: Sat, 28 Apr 2012 14:51:15 -0700
From: "Justin P. Mattock" <justinmattock@...il.com>
To: linux-kernel@...r.kernel.org
Subject: NEXT: INFO: possible irq lock inversion dependency detected
This is showing up with the current linux-next over here:
[ 6.633932]
[ 6.634261] =========================================================
[ 6.634263] [ INFO: possible irq lock inversion dependency detected ]
[ 6.634265] 3.4.0-rc4-next-20120427 #1 Not tainted
[ 6.634266] ---------------------------------------------------------
[ 6.634268] swapper/1/0 just changed the state of lock:
[ 6.634270] (tg_stats_alloc_lock){+.-...}, at: [<c1286da3>]
throtl_pd_exit+0x23/0x60
[ 6.634278] but this lock was taken by another, HARDIRQ-safe lock in
the past:
[ 6.634279] (&(&q->__queue_lock)->rlock){-.-...}
[ 6.634281]
[ 6.634281] and interrupts could create inverse lock ordering between
them.
[ 6.634282]
[ 6.634283]
[ 6.634283] other info that might help us debug this:
[ 6.634284] Chain exists of:
[ 6.634285] &(&q->__queue_lock)->rlock --> &(&blkcg->lock)->rlock
--> tg_stats_alloc_lock
[ 6.634287]
[ 6.634288] Possible interrupt unsafe locking scenario:
[ 6.634288]
[ 6.634289] CPU0 CPU1
[ 6.634289] ---- ----
[ 6.634290] lock(tg_stats_alloc_lock);
[ 6.634291] local_irq_disable();
[ 6.634292]
lock(&(&q->__queue_lock)->rlock);
[ 6.634294] lock(&(&blkcg->lock)->rlock);
[ 6.634295] <Interrupt>
[ 6.634296] lock(&(&q->__queue_lock)->rlock);
[ 6.634297]
[ 6.634297] *** DEADLOCK ***
[ 6.634297]
[ 6.634298] no locks held by swapper/1/0.
[ 6.634299]
[ 6.634300] the shortest dependencies between 2nd lock and 1st lock:
[ 6.634306] -> (&(&q->__queue_lock)->rlock){-.-...} ops: 7758 {
[ 6.634308] IN-HARDIRQ-W at:
[ 6.634310] [<c1098b6f>]
__lock_acquire+0x90f/0x1660
[ 6.634315] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634318] [<c15180a7>]
_raw_spin_lock_irqsave+0x57/0x70
[ 6.634325] [<c13a8b28>]
ata_qc_schedule_eh+0x48/0x80
[ 6.634332] [<c139e9b0>]
ata_qc_complete+0x180/0x1f0
[ 6.634335] [<c13a8293>]
ata_do_link_abort+0x83/0xc0
[ 6.634338] [<c13a8b9f>]
ata_port_abort+0xf/0x20
[ 6.634341] [<c13b70e5>]
ahci_interrupt+0x425/0x640
[ 6.634346] [<c10be6cc>]
handle_irq_event_percpu+0x4c/0x240
[ 6.634351] [<c10be8fa>]
handle_irq_event+0x3a/0x60
[ 6.634353] [<c10c12e6>]
handle_edge_irq+0x66/0xf0
[ 6.634356] IN-SOFTIRQ-W at:
[ 6.634357] [<c109875d>]
__lock_acquire+0x4fd/0x1660
[ 6.634359] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634361] [<c15176d3>]
_raw_spin_lock+0x43/0x50
[ 6.634364] [<c1379bb9>]
scsi_device_unbusy+0x89/0xb0
[ 6.634370] [<c1370f5a>]
scsi_finish_command+0x2a/0x100
[ 6.634372] [<c1379e28>]
scsi_softirq_done+0x118/0x140
[ 6.634374] [<c1277d2b>]
blk_done_softirq+0x6b/0x80
[ 6.634378] [<c1047599>]
__do_softirq+0x99/0x1f0
[ 6.634383] [<c10477e7>]
run_ksoftirqd+0xf7/0x1a0
[ 6.634386] [<c1060ccd>]
kthread+0x7d/0x90
[ 6.634390] [<c151f7c2>]
kernel_thread_helper+0x6/0x10
[ 6.634394] INITIAL USE at:
[ 6.634395] [<c10984c5>]
__lock_acquire+0x265/0x1660
[ 6.634398] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634401] [<c1517eee>]
_raw_spin_lock_irq+0x4e/0x60
[ 6.634404] [<c1272807>]
blk_queue_bypass_start+0x17/0x80
[ 6.634410] [<c12864a0>]
blkcg_activate_policy+0x40/0x260
[ 6.634416] [<c1288c27>]
blk_throtl_init+0xb7/0xe0
[ 6.634419] [<c1286878>]
blkcg_init_queue+0x28/0x30
[ 6.634422] [<c126e438>]
blk_alloc_queue_node+0x1f8/0x240
[ 6.634425] [<c126e492>]
blk_alloc_queue+0x12/0x20
[ 6.634428] [<c1368909>]
brd_alloc+0x69/0x180
[ 6.634434] [<c17c59c2>]
brd_init+0xb0/0x185
[ 6.634440] [<c1003034>]
do_one_initcall+0x34/0x170
[ 6.634443] [<c1796b37>]
kernel_init+0x111/0x1a6
[ 6.634447] [<c151f7c2>]
kernel_thread_helper+0x6/0x10
[ 6.634450] }
[ 6.634451] ... key at: [<c1d943ec>] __key.34385+0x0/0x8
[ 6.634457] ... acquired at:
[ 6.634458] [<c1099de0>] lock_acquire+0x80/0x120
[ 6.634460] [<c15180a7>] _raw_spin_lock_irqsave+0x57/0x70
[ 6.634463] [<c12912ab>] get_from_free_list+0x1b/0x50
[ 6.634468] [<c12916ad>] idr_get_empty_slot+0x26d/0x290
[ 6.634471] [<c1291739>] ida_get_new_above+0x69/0x230
[ 6.634473] [<c1292048>] ida_simple_get+0x58/0xe0
[ 6.634476] [<c126e27c>] blk_alloc_queue_node+0x3c/0x240
[ 6.634478] [<c126e492>] blk_alloc_queue+0x12/0x20
[ 6.634481] [<c1368909>] brd_alloc+0x69/0x180
[ 6.634484] [<c17c59c2>] brd_init+0xb0/0x185
[ 6.634486] [<c1003034>] do_one_initcall+0x34/0x170
[ 6.634489] [<c1796b37>] kernel_init+0x111/0x1a6
[ 6.634491] [<c151f7c2>] kernel_thread_helper+0x6/0x10
[ 6.634494]
[ 6.634495] -> (&(&blkcg->lock)->rlock){......} ops: 89 {
[ 6.634498] INITIAL USE at:
[ 6.634499] [<c10984c5>]
__lock_acquire+0x265/0x1660
[ 6.634503] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634505] [<c15176d3>]
_raw_spin_lock+0x43/0x50
[ 6.634508] [<c128618c>]
__blkg_lookup_create+0x17c/0x2a0
[ 6.634512] [<c12864f3>]
blkcg_activate_policy+0x93/0x260
[ 6.634515] [<c1288c27>]
blk_throtl_init+0xb7/0xe0
[ 6.634517] [<c1286878>]
blkcg_init_queue+0x28/0x30
[ 6.634520] [<c126e438>]
blk_alloc_queue_node+0x1f8/0x240
[ 6.634523] [<c126e492>]
blk_alloc_queue+0x12/0x20
[ 6.634526] [<c1368909>]
brd_alloc+0x69/0x180
[ 6.634529] [<c17c59c2>]
brd_init+0xb0/0x185
[ 6.634532] [<c1003034>]
do_one_initcall+0x34/0x170
[ 6.634535] [<c1796b37>]
kernel_init+0x111/0x1a6
[ 6.634538] [<c151f7c2>]
kernel_thread_helper+0x6/0x10
[ 6.634540] }
[ 6.634541] ... key at: [<c1d949e8>] __key.29231+0x0/0x8
[ 6.634543] ... acquired at:
[ 6.634544] [<c1099de0>] lock_acquire+0x80/0x120
[ 6.634546] [<c15176d3>] _raw_spin_lock+0x43/0x50
[ 6.634548] [<c1286e5e>] throtl_pd_init+0x7e/0xc0
[ 6.634550] [<c128666b>] blkcg_activate_policy+0x20b/0x260
[ 6.634552] [<c1288c27>] blk_throtl_init+0xb7/0xe0
[ 6.634553] [<c1286878>] blkcg_init_queue+0x28/0x30
[ 6.634555] [<c126e438>] blk_alloc_queue_node+0x1f8/0x240
[ 6.634557] [<c126e492>] blk_alloc_queue+0x12/0x20
[ 6.634559] [<c1368909>] brd_alloc+0x69/0x180
[ 6.634561] [<c17c59c2>] brd_init+0xb0/0x185
[ 6.634562] [<c1003034>] do_one_initcall+0x34/0x170
[ 6.634564] [<c1796b37>] kernel_init+0x111/0x1a6
[ 6.634565] [<c151f7c2>] kernel_thread_helper+0x6/0x10
[ 6.634567]
[ 6.634568] -> (tg_stats_alloc_lock){+.-...} ops: 71 {
[ 6.634570] HARDIRQ-ON-W at:
[ 6.634571] [<c1098781>]
__lock_acquire+0x521/0x1660
[ 6.634573] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634575] [<c15176d3>]
_raw_spin_lock+0x43/0x50
[ 6.634577] [<c1286da3>]
throtl_pd_exit+0x23/0x60
[ 6.634579] [<c128568b>]
blkg_free+0x2b/0x60
[ 6.634581] [<c12856d0>]
blkg_rcu_free+0x10/0x20
[ 6.634583] [<c10c4996>]
__rcu_process_callbacks+0x1c6/0x430
[ 6.634586] [<c10c4c26>]
rcu_process_callbacks+0x26/0xe0
[ 6.634588] [<c1047599>]
__do_softirq+0x99/0x1f0
[ 6.634591] IN-SOFTIRQ-W at:
[ 6.634592] [<c109875d>]
__lock_acquire+0x4fd/0x1660
[ 6.634594] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634596] [<c15176d3>]
_raw_spin_lock+0x43/0x50
[ 6.634598] [<c1286da3>]
throtl_pd_exit+0x23/0x60
[ 6.634600] [<c128568b>]
blkg_free+0x2b/0x60
[ 6.634602] [<c12856d0>]
blkg_rcu_free+0x10/0x20
[ 6.634604] [<c10c4996>]
__rcu_process_callbacks+0x1c6/0x430
[ 6.634606] [<c10c4c26>]
rcu_process_callbacks+0x26/0xe0
[ 6.634608] [<c1047599>]
__do_softirq+0x99/0x1f0
[ 6.634611] INITIAL USE at:
[ 6.634612] [<c10984c5>]
__lock_acquire+0x265/0x1660
[ 6.634614] [<c1099de0>]
lock_acquire+0x80/0x120
[ 6.634616] [<c15176d3>]
_raw_spin_lock+0x43/0x50
[ 6.634618] [<c1286e5e>]
throtl_pd_init+0x7e/0xc0
[ 6.634619] [<c128666b>]
blkcg_activate_policy+0x20b/0x260
[ 6.634622] [<c1288c27>]
blk_throtl_init+0xb7/0xe0
[ 6.634624] [<c1286878>]
blkcg_init_queue+0x28/0x30
[ 6.634626] [<c126e438>]
blk_alloc_queue_node+0x1f8/0x240
[ 6.634628] [<c126e492>]
blk_alloc_queue+0x12/0x20
[ 6.634630] [<c1368909>]
brd_alloc+0x69/0x180
[ 6.634632] [<c17c59c2>]
brd_init+0xb0/0x185
[ 6.634634] [<c1003034>]
do_one_initcall+0x34/0x170
[ 6.634636] [<c1796b37>]
kernel_init+0x111/0x1a6
[ 6.634637] [<c151f7c2>]
kernel_thread_helper+0x6/0x10
[ 6.634639] }
[ 6.634640] ... key at: [<c1762010>] tg_stats_alloc_lock+0x10/0x20
[ 6.634644] ... acquired at:
[ 6.634645] [<c10979b0>] check_usage_backwards+0xf0/0x130
[ 6.634647] [<c1098110>] mark_lock+0x130/0x280
[ 6.634649] [<c1098781>] __lock_acquire+0x521/0x1660
[ 6.634651] [<c1099de0>] lock_acquire+0x80/0x120
[ 6.634653] [<c15176d3>] _raw_spin_lock+0x43/0x50
[ 6.634655] [<c1286da3>] throtl_pd_exit+0x23/0x60
[ 6.634657] [<c128568b>] blkg_free+0x2b/0x60
[ 6.634659] [<c12856d0>] blkg_rcu_free+0x10/0x20
[ 6.634661] [<c10c4996>] __rcu_process_callbacks+0x1c6/0x430
[ 6.634663] [<c10c4c26>] rcu_process_callbacks+0x26/0xe0
[ 6.634665] [<c1047599>] __do_softirq+0x99/0x1f0
[ 6.634666]
[ 6.634667]
[ 6.634667] stack backtrace:
[ 6.634669] Pid: 0, comm: swapper/1 Not tainted
3.4.0-rc4-next-20120427 #1
[ 6.634671] Call Trace:
[ 6.634674] [<c150e0b4>] ? printk+0x1d/0x1f
[ 6.634676] [<c1097784>] print_irq_inversion_bug.part.36+0x1a4/0x1b0
[ 6.634679] [<c10979b0>] check_usage_backwards+0xf0/0x130
[ 6.634681] [<c1098110>] mark_lock+0x130/0x280
[ 6.634682] [<c10978c0>] ? check_usage_forwards+0x130/0x130
[ 6.634685] [<c1098781>] __lock_acquire+0x521/0x1660
[ 6.634686] [<c10984c5>] ? __lock_acquire+0x265/0x1660
[ 6.634688] [<c10984c5>] ? __lock_acquire+0x265/0x1660
[ 6.634691] [<c1099de0>] lock_acquire+0x80/0x120
[ 6.634692] [<c1286da3>] ? throtl_pd_exit+0x23/0x60
[ 6.634695] [<c15176d3>] _raw_spin_lock+0x43/0x50
[ 6.634696] [<c1286da3>] ? throtl_pd_exit+0x23/0x60
[ 6.634698] [<c1286da3>] throtl_pd_exit+0x23/0x60
[ 6.634700] [<c128568b>] blkg_free+0x2b/0x60
[ 6.634702] [<c12856d0>] blkg_rcu_free+0x10/0x20
[ 6.634704] [<c10c4996>] __rcu_process_callbacks+0x1c6/0x430
[ 6.634706] [<c1047500>] ? local_bh_enable_ip+0xd0/0xd0
[ 6.634708] [<c10c4c26>] rcu_process_callbacks+0x26/0xe0
[ 6.634710] [<c1047500>] ? local_bh_enable_ip+0xd0/0xd0
[ 6.634712] [<c1047599>] __do_softirq+0x99/0x1f0
[ 6.634714] [<c1047500>] ? local_bh_enable_ip+0xd0/0xd0
[ 6.634715] <IRQ> [<c1047996>] ? irq_exit+0x86/0xb0
[ 6.634718] [<c151fa39>] ? smp_apic_timer_interrupt+0x59/0x88
[ 6.634722] [<c129c808>] ? trace_hardirqs_off_thunk+0xc/0x14
[ 6.634724] [<c15186b2>] ? apic_timer_interrupt+0x36/0x3c
[ 6.634728] [<c109007b>] ? timer_list_show+0x99b/0xec0
[ 6.634733] [<c12e3097>] ? intel_idle+0xc7/0x120
[ 6.634736] [<c1429545>] ? cpuidle_enter+0x15/0x20
[ 6.634738] [<c1429ada>] ? cpuidle_idle_call+0x8a/0x1e0
[ 6.634741] [<c10199bf>] ? cpu_idle+0xaf/0x100
[ 6.634743] [<c15073c0>] ? start_secondary+0x1f4/0x1fa
[ 13.057885] EXT4-fs (sda6): mounted filesystem with ordered data
mode. Opts: (null)
full dmesg here:
http://fpaste.org/IDFW/
Justin P. Mattock
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists