* [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
@ 2014-03-26 18:11 Jeff Mahoney
2014-03-26 18:19 ` David Sterba
2014-04-09 16:05 ` Filipe David Manana
0 siblings, 2 replies; 6+ messages in thread
From: Jeff Mahoney @ 2014-03-26 18:11 UTC (permalink / raw)
To: linux-btrfs
When encountering memory pressure, testers have run into the following
lockdep warning. It was caused by __link_block_group calling kobject_add
with the groups_sem held. kobject_add calls kvasprintf with GFP_KERNEL,
which gets us into reclaim context. The kobject doesn't actually need
to be added under the lock -- it just needs to ensure that it's only
added for the first block group to be linked.
=========================================================
[ INFO: possible irq lock inversion dependency detected ]
3.14.0-rc8-default #1 Not tainted
---------------------------------------------------------
kswapd0/169 just changed the state of lock:
(&delayed_node->mutex){+.+.-.}, at: [<ffffffffa018baea>] __btrfs_release_delayed_node+0x3a/0x200 [btrfs]
but this lock took another, RECLAIM_FS-unsafe lock in the past:
(&found->groups_sem){+++++.}
and interrupts could create inverse lock ordering between them.
other info that might help us debug this:
Possible interrupt unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&found->groups_sem);
local_irq_disable();
lock(&delayed_node->mutex);
lock(&found->groups_sem);
<Interrupt>
lock(&delayed_node->mutex);
*** DEADLOCK ***
2 locks held by kswapd0/169:
#0: (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>] shrink_slab+0x3a/0x160
#1: (&type->s_umount_key#27){++++..}, at: [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
---
fs/btrfs/extent-tree.c | 10 +++++++---
1 file changed, 7 insertions(+), 3 deletions(-)
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -8343,9 +8343,15 @@ static void __link_block_group(struct bt
struct btrfs_block_group_cache *cache)
{
int index = get_block_group_index(cache);
+ bool first = false;
down_write(&space_info->groups_sem);
- if (list_empty(&space_info->block_groups[index])) {
+ if (list_empty(&space_info->block_groups[index]))
+ first = true;
+ list_add_tail(&cache->list, &space_info->block_groups[index]);
+ up_write(&space_info->groups_sem);
+
+ if (first) {
struct kobject *kobj = &space_info->block_group_kobjs[index];
int ret;
@@ -8357,8 +8363,6 @@ static void __link_block_group(struct bt
kobject_put(&space_info->kobj);
}
}
- list_add_tail(&cache->list, &space_info->block_groups[index]);
- up_write(&space_info->groups_sem);
}
static struct btrfs_block_group_cache *
--
Jeff Mahoney
SUSE Labs
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
2014-03-26 18:11 [PATCH] btrfs: fix lockdep warning with reclaim lock inversion Jeff Mahoney
@ 2014-03-26 18:19 ` David Sterba
2014-04-09 16:05 ` Filipe David Manana
1 sibling, 0 replies; 6+ messages in thread
From: David Sterba @ 2014-03-26 18:19 UTC (permalink / raw)
To: Jeff Mahoney; +Cc: linux-btrfs
On Wed, Mar 26, 2014 at 02:11:26PM -0400, Jeff Mahoney wrote:
> When encountering memory pressure, testers have run into the following
> lockdep warning. It was caused by __link_block_group calling kobject_add
> with the groups_sem held. kobject_add calls kvasprintf with GFP_KERNEL,
> which gets us into reclaim context. The kobject doesn't actually need
> to be added under the lock -- it just needs to ensure that it's only
> added for the first block group to be linked.
>
> =========================================================
> [ INFO: possible irq lock inversion dependency detected ]
> 3.14.0-rc8-default #1 Not tainted
> ---------------------------------------------------------
> kswapd0/169 just changed the state of lock:
> (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa018baea>] __btrfs_release_delayed_node+0x3a/0x200 [btrfs]
> but this lock took another, RECLAIM_FS-unsafe lock in the past:
> (&found->groups_sem){+++++.}
>
> and interrupts could create inverse lock ordering between them.
>
> other info that might help us debug this:
> Possible interrupt unsafe locking scenario:
> CPU0 CPU1
> ---- ----
> lock(&found->groups_sem);
> local_irq_disable();
> lock(&delayed_node->mutex);
> lock(&found->groups_sem);
> <Interrupt>
> lock(&delayed_node->mutex);
>
> *** DEADLOCK ***
> 2 locks held by kswapd0/169:
> #0: (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>] shrink_slab+0x3a/0x160
> #1: (&type->s_umount_key#27){++++..}, at: [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
>
> Signed-off-by: Jeff Mahoney <jeffm@suse.com>
For the record, fixes https://bugzilla.kernel.org/show_bug.cgi?id=72501
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
2014-03-26 18:11 [PATCH] btrfs: fix lockdep warning with reclaim lock inversion Jeff Mahoney
2014-03-26 18:19 ` David Sterba
@ 2014-04-09 16:05 ` Filipe David Manana
2014-04-09 21:22 ` Jeff Mahoney
1 sibling, 1 reply; 6+ messages in thread
From: Filipe David Manana @ 2014-04-09 16:05 UTC (permalink / raw)
To: Jeff Mahoney; +Cc: linux-btrfs
On Wed, Mar 26, 2014 at 6:11 PM, Jeff Mahoney <jeffm@suse.com> wrote:
> When encountering memory pressure, testers have run into the following
> lockdep warning. It was caused by __link_block_group calling kobject_add
> with the groups_sem held. kobject_add calls kvasprintf with GFP_KERNEL,
> which gets us into reclaim context. The kobject doesn't actually need
> to be added under the lock -- it just needs to ensure that it's only
> added for the first block group to be linked.
>
> =========================================================
> [ INFO: possible irq lock inversion dependency detected ]
> 3.14.0-rc8-default #1 Not tainted
> ---------------------------------------------------------
> kswapd0/169 just changed the state of lock:
> (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa018baea>] __btrfs_release_delayed_node+0x3a/0x200 [btrfs]
> but this lock took another, RECLAIM_FS-unsafe lock in the past:
> (&found->groups_sem){+++++.}
>
> and interrupts could create inverse lock ordering between them.
>
> other info that might help us debug this:
> Possible interrupt unsafe locking scenario:
> CPU0 CPU1
> ---- ----
> lock(&found->groups_sem);
> local_irq_disable();
> lock(&delayed_node->mutex);
> lock(&found->groups_sem);
> <Interrupt>
> lock(&delayed_node->mutex);
>
> *** DEADLOCK ***
> 2 locks held by kswapd0/169:
> #0: (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>] shrink_slab+0x3a/0x160
> #1: (&type->s_umount_key#27){++++..}, at: [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
>
> Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Hi Jeff,
The same kind of problem happens when deleting the kobj. See the trace
below, thanks.
[23510.196996] BTRFS: device fsid 6b4518c2-60e8-4990-8c6e-fb15054b1e46
devid 1 transid 4 /dev/sdd
[23510.203212] BTRFS info (device sdd): disk space caching is enabled
[23510.203214] BTRFS: flagging fs with big metadata feature
[23510.227365] BTRFS: creating UUID tree
[23661.587069]
[23661.587073] =========================================================
[23661.587074] [ INFO: possible irq lock inversion dependency detected ]
[23661.587076] 3.13.0-fdm-btrfs-next-24+ #7 Not tainted
[23661.587077] ---------------------------------------------------------
[23661.587078] kswapd0/42 just changed the state of lock:
[23661.587079] (&delayed_node->mutex){+.+.-.}, at:
[<ffffffffa036d73c>] __btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587095] but this lock took another, RECLAIM_FS-unsafe lock in the past:
[23661.587096] (sysfs_mutex){+.+.+.}
[23661.587096]
[23661.587096] and interrupts could create inverse lock ordering between them.
[23661.587096]
[23661.587098]
[23661.587098] other info that might help us debug this:
[23661.587099] Chain exists of:
[23661.587099] &delayed_node->mutex --> &found->groups_sem --> sysfs_mutex
[23661.587099]
[23661.587102] Possible interrupt unsafe locking scenario:
[23661.587102]
[23661.587103] CPU0 CPU1
[23661.587103] ---- ----
[23661.587104] lock(sysfs_mutex);
[23661.587105] local_irq_disable();
[23661.587106] lock(&delayed_node->mutex);
[23661.587107] lock(&found->groups_sem);
[23661.587108] <Interrupt>
[23661.587109] lock(&delayed_node->mutex);
[23661.587110]
[23661.587110] *** DEADLOCK ***
[23661.587110]
[23661.587111] 2 locks held by kswapd0/42:
[23661.587112] #0: (shrinker_rwsem){++++..}, at:
[<ffffffff8114d0ec>] shrink_slab+0x3c/0x3a0
[23661.587117] #1: (&type->s_umount_key#31){+++++.}, at:
[<ffffffff811a43d4>] grab_super_passive+0x44/0x90
[23661.587122]
[23661.587122] the shortest dependencies between 2nd lock and 1st lock:
[23661.587126] -> (sysfs_mutex){+.+.+.} ops: 12918522 {
[23661.587129] HARDIRQ-ON-W at:
[23661.587130] [<ffffffff81099900>]
__lock_acquire+0x670/0x1e40
[23661.587133] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587134] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587137] [<ffffffff8121f1b6>]
sysfs_mount+0x166/0x210
[23661.587140] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587142] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587144] [<ffffffff811c2509>]
kern_mount_data+0x19/0x30
[23661.587145] [<ffffffff81f16082>] sysfs_init+0x55/0xb5
[23661.587148] [<ffffffff81f13435>] mnt_init+0xcd/0x1d8
[23661.587149] [<ffffffff81f130d7>]
vfs_caches_init+0x99/0x11c
[23661.587151] [<ffffffff81eebe1e>]
start_kernel+0x3ab/0x40c
[23661.587154] [<ffffffff81eeb579>]
x86_64_start_reservations+0x2a/0x2c
[23661.587155] [<ffffffff81eeb672>]
x86_64_start_kernel+0xf7/0xfb
[23661.587157] SOFTIRQ-ON-W at:
[23661.587158] [<ffffffff8109993c>]
__lock_acquire+0x6ac/0x1e40
[23661.587160] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587162] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587164] [<ffffffff8121f1b6>]
sysfs_mount+0x166/0x210
[23661.587166] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587167] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587169] [<ffffffff811c2509>]
kern_mount_data+0x19/0x30
[23661.587171] [<ffffffff81f16082>] sysfs_init+0x55/0xb5
[23661.587173] [<ffffffff81f13435>] mnt_init+0xcd/0x1d8
[23661.587174] [<ffffffff81f130d7>]
vfs_caches_init+0x99/0x11c
[23661.587175] [<ffffffff81eebe1e>]
start_kernel+0x3ab/0x40c
[23661.587177] [<ffffffff81eeb579>]
x86_64_start_reservations+0x2a/0x2c
[23661.587178] [<ffffffff81eeb672>]
x86_64_start_kernel+0xf7/0xfb
[23661.587180] RECLAIM_FS-ON-W at:
[23661.587181] [<ffffffff8109c039>]
mark_held_locks+0xb9/0x140
[23661.587183] [<ffffffff8109c7c5>]
lockdep_trace_alloc+0x65/0xd0
[23661.587185] [<ffffffff8118e7e3>]
kmem_cache_alloc+0x33/0x1e0
[23661.587187] [<ffffffff811bcc51>]
alloc_inode+0x61/0xa0
[23661.587190] [<ffffffff811be178>]
iget_locked+0xd8/0x1d0
[23661.587192] [<ffffffff8121b54a>]
sysfs_get_inode+0x1a/0x170
[23661.587193] [<ffffffff8121f1c5>]
sysfs_mount+0x175/0x210
[23661.587195] [<ffffffff811a4d53>]
mount_fs+0x43/0x1b0
[23661.587196] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587198] [<ffffffff811c4d49>]
do_mount+0x259/0xb70
[23661.587199] [<ffffffff811c594e>]
SyS_mount+0x8e/0xe0
[23661.587201] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587204] INITIAL USE at:
[23661.587204] [<ffffffff810995bd>]
__lock_acquire+0x32d/0x1e40
[23661.587206] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587207] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587209] [<ffffffff8121f1b6>]
sysfs_mount+0x166/0x210
[23661.587211] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587212] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587213] [<ffffffff811c2509>]
kern_mount_data+0x19/0x30
[23661.587215] [<ffffffff81f16082>] sysfs_init+0x55/0xb5
[23661.587216] [<ffffffff81f13435>] mnt_init+0xcd/0x1d8
[23661.587217] [<ffffffff81f130d7>]
vfs_caches_init+0x99/0x11c
[23661.587218] [<ffffffff81eebe1e>]
start_kernel+0x3ab/0x40c
[23661.587220] [<ffffffff81eeb579>]
x86_64_start_reservations+0x2a/0x2c
[23661.587221] [<ffffffff81eeb672>]
x86_64_start_kernel+0xf7/0xfb
[23661.587222] }
[23661.587223] ... key at: [<ffffffff81c79918>] sysfs_mutex+0x78/0xc0
[23661.587226] ... acquired at:
[23661.587227] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587228] [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390
[23661.587230] [<ffffffff8121defe>] sysfs_addrm_start+0x1e/0x20
[23661.587232] [<ffffffff8121e59a>] sysfs_remove+0x1a/0x40
[23661.587233] [<ffffffff8121e6c3>] sysfs_remove_dir+0x53/0x80
[23661.587234] [<ffffffff81361d11>] kobject_del+0x21/0x50
[23661.587237] [<ffffffffa030b060>]
btrfs_remove_block_group+0x420/0x540 [btrfs]
[23661.587246] [<ffffffffa034087b>]
btrfs_relocate_chunk.isra.61+0x60b/0x760 [btrfs]
[23661.587254] [<ffffffffa034497f>] btrfs_balance+0x86f/0xfd0 [btrfs]
[23661.587262] [<ffffffffa034cb45>] btrfs_ioctl_balance+0x1c5/0x530 [btrfs]
[23661.587269] [<ffffffffa034fcfc>] btrfs_ioctl+0x68c/0x28a0 [btrfs]
[23661.587276] [<ffffffff811b4938>] do_vfs_ioctl+0x88/0x570
[23661.587278] [<ffffffff811b4eb1>] SyS_ioctl+0x91/0xb0
[23661.587279] [<ffffffff81703212>] system_call_fastpath+0x16/0x1b
[23661.587281]
[23661.587282] -> (&found->groups_sem){++++..} ops: 897759 {
[23661.587284] HARDIRQ-ON-W at:
[23661.587285] [<ffffffff81099900>]
__lock_acquire+0x670/0x1e40
[23661.587287] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587288] [<ffffffff816f8a81>] down_write+0x61/0xc0
[23661.587290] [<ffffffffa0306004>]
__link_block_group+0x44/0x100 [btrfs]
[23661.587295] [<ffffffffa0309bf6>]
btrfs_read_block_groups+0x306/0x650 [btrfs]
[23661.587301] [<ffffffffa0318376>]
open_ctree+0x1676/0x2070 [btrfs]
[23661.587308] [<ffffffffa02ede54>]
btrfs_mount+0x6b4/0x910 [btrfs]
[23661.587312] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587313] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587315] [<ffffffff811c4d49>] do_mount+0x259/0xb70
[23661.587316] [<ffffffff811c594e>] SyS_mount+0x8e/0xe0
[23661.587318] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587320] HARDIRQ-ON-R at:
[23661.587320] [<ffffffff810997bd>]
__lock_acquire+0x52d/0x1e40
[23661.587322] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587323] [<ffffffff816f8b31>] down_read+0x51/0xa0
[23661.587325] [<ffffffffa0314098>]
btrfs_calc_num_tolerated_disk_barrier_failures+0x118/0x1e0 [btrfs]
[23661.587331] [<ffffffffa0318386>]
open_ctree+0x1686/0x2070 [btrfs]
[23661.587337] [<ffffffffa02ede54>]
btrfs_mount+0x6b4/0x910 [btrfs]
[23661.587341] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587342] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587344] [<ffffffff811c4d49>] do_mount+0x259/0xb70
[23661.587345] [<ffffffff811c594e>] SyS_mount+0x8e/0xe0
[23661.587346] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587348] SOFTIRQ-ON-W at:
[23661.587349] [<ffffffff8109993c>]
__lock_acquire+0x6ac/0x1e40
[23661.587350] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587352] [<ffffffff816f8a81>] down_write+0x61/0xc0
[23661.587353] [<ffffffffa0306004>]
__link_block_group+0x44/0x100 [btrfs]
[23661.587359] [<ffffffffa0309bf6>]
btrfs_read_block_groups+0x306/0x650 [btrfs]
[23661.587364] [<ffffffffa0318376>]
open_ctree+0x1676/0x2070 [btrfs]
[23661.587370] [<ffffffffa02ede54>]
btrfs_mount+0x6b4/0x910 [btrfs]
[23661.587374] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587375] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587376] [<ffffffff811c4d49>] do_mount+0x259/0xb70
[23661.587378] [<ffffffff811c594e>] SyS_mount+0x8e/0xe0
[23661.587379] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587381] SOFTIRQ-ON-R at:
[23661.587382] [<ffffffff8109993c>]
__lock_acquire+0x6ac/0x1e40
[23661.587383] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587384] [<ffffffff816f8b31>] down_read+0x51/0xa0
[23661.587386] [<ffffffffa0314098>]
btrfs_calc_num_tolerated_disk_barrier_failures+0x118/0x1e0 [btrfs]
[23661.587392] [<ffffffffa0318386>]
open_ctree+0x1686/0x2070 [btrfs]
[23661.587397] [<ffffffffa02ede54>]
btrfs_mount+0x6b4/0x910 [btrfs]
[23661.587401] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587402] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587404] [<ffffffff811c4d49>] do_mount+0x259/0xb70
[23661.587405] [<ffffffff811c594e>] SyS_mount+0x8e/0xe0
[23661.587406] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587408] INITIAL USE at:
[23661.587409] [<ffffffff810995bd>]
__lock_acquire+0x32d/0x1e40
[23661.587410] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587412] [<ffffffff816f8a81>] down_write+0x61/0xc0
[23661.587413] [<ffffffffa0306004>]
__link_block_group+0x44/0x100 [btrfs]
[23661.587418] [<ffffffffa0309bf6>]
btrfs_read_block_groups+0x306/0x650 [btrfs]
[23661.587424] [<ffffffffa0318376>]
open_ctree+0x1676/0x2070 [btrfs]
[23661.587429] [<ffffffffa02ede54>]
btrfs_mount+0x6b4/0x910 [btrfs]
[23661.587433] [<ffffffff811a4d53>] mount_fs+0x43/0x1b0
[23661.587435] [<ffffffff811c2403>]
vfs_kern_mount+0x73/0x160
[23661.587436] [<ffffffff811c4d49>] do_mount+0x259/0xb70
[23661.587437] [<ffffffff811c594e>] SyS_mount+0x8e/0xe0
[23661.587439] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587441] }
[23661.587441] ... key at: [<ffffffffa03b5ab8>]
__key.40358+0x0/0xfffffffffffe2548 [btrfs]
[23661.587447] ... acquired at:
[23661.587448] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587449] [<ffffffff816f8b31>] down_read+0x51/0xa0
[23661.587451] [<ffffffffa0306891>] find_free_extent+0x7d1/0xc30 [btrfs]
[23661.587456] [<ffffffffa0306d98>] btrfs_reserve_extent+0xa8/0x140 [btrfs]
[23661.587461] [<ffffffffa0308236>] btrfs_alloc_free_block+0xe6/0x460 [btrfs]
[23661.587467] [<ffffffffa02f6d30>] split_leaf+0x1b0/0x770 [btrfs]
[23661.587471] [<ffffffffa02f7f23>] btrfs_search_slot+0x963/0x9e0 [btrfs]
[23661.587476] [<ffffffffa02f9658>]
btrfs_insert_empty_items+0x78/0xd0 [btrfs]
[23661.587481] [<ffffffffa036db37>]
btrfs_insert_delayed_items+0x87/0x420 [btrfs]
[23661.587489] [<ffffffffa036e8c2>]
btrfs_async_run_delayed_root+0xf2/0x200 [btrfs]
[23661.587496] [<ffffffffa0349fc7>] normal_work_helper+0x137/0x2c0 [btrfs]
[23661.587503] [<ffffffff810681f5>] process_one_work+0x1f5/0x530
[23661.587505] [<ffffffff810689dc>] worker_thread+0x11c/0x370
[23661.587506] [<ffffffff810703c6>] kthread+0xf6/0x110
[23661.587508] [<ffffffff8170316c>] ret_from_fork+0x7c/0xb0
[23661.587510]
[23661.587510] -> (&delayed_node->mutex){+.+.-.} ops: 5556073 {
[23661.587513] HARDIRQ-ON-W at:
[23661.587514] [<ffffffff81099900>]
__lock_acquire+0x670/0x1e40
[23661.587515] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587516] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587518] [<ffffffffa036f7bf>]
btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
[23661.587525] [<ffffffffa0321107>]
btrfs_update_inode+0x67/0xf0 [btrfs]
[23661.587532] [<ffffffffa03293a2>]
btrfs_mkdir+0x182/0x1e0 [btrfs]
[23661.587539] [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170
[23661.587541] [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0
[23661.587542] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587544] SOFTIRQ-ON-W at:
[23661.587544] [<ffffffff8109993c>]
__lock_acquire+0x6ac/0x1e40
[23661.587546] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587547] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587549] [<ffffffffa036f7bf>]
btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
[23661.587556] [<ffffffffa0321107>]
btrfs_update_inode+0x67/0xf0 [btrfs]
[23661.587562] [<ffffffffa03293a2>]
btrfs_mkdir+0x182/0x1e0 [btrfs]
[23661.587568] [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170
[23661.587570] [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0
[23661.587571] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587573] IN-RECLAIM_FS-W at:
[23661.587574] [<ffffffff81099838>]
__lock_acquire+0x5a8/0x1e40
[23661.587575] [<ffffffff8109b6e5>]
lock_acquire+0x85/0x110
[23661.587577] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587578] [<ffffffffa036d73c>]
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587585] [<ffffffffa036ed04>]
btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[23661.587592] [<ffffffffa032650d>]
btrfs_evict_inode+0x1cd/0x520 [btrfs]
[23661.587599] [<ffffffff811bdd17>] evict+0xa7/0x1a0
[23661.587601] [<ffffffff811bde4e>]
dispose_list+0x3e/0x60
[23661.587602] [<ffffffff811bed3c>]
prune_icache_sb+0x4c/0x60
[23661.587604] [<ffffffff811a4641>]
super_cache_scan+0x111/0x170
[23661.587605] [<ffffffff8114d28a>]
shrink_slab+0x1da/0x3a0
[23661.587608] [<ffffffff81150905>]
balance_pgdat+0x3d5/0x5e0
[23661.587609] [<ffffffff81150cba>] kswapd+0x1aa/0x4a0
[23661.587610] [<ffffffff810703c6>] kthread+0xf6/0x110
[23661.587611] [<ffffffff8170316c>]
ret_from_fork+0x7c/0xb0
[23661.587613] INITIAL USE at:
[23661.587614] [<ffffffff810995bd>]
__lock_acquire+0x32d/0x1e40
[23661.587615] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587617] [<ffffffff816f7f36>]
mutex_lock_nested+0x76/0x390
[23661.587618] [<ffffffffa036f7bf>]
btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
[23661.587625] [<ffffffffa0321107>]
btrfs_update_inode+0x67/0xf0 [btrfs]
[23661.587631] [<ffffffffa03293a2>]
btrfs_mkdir+0x182/0x1e0 [btrfs]
[23661.587638] [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170
[23661.587640] [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0
[23661.587641] [<ffffffff81703212>]
system_call_fastpath+0x16/0x1b
[23661.587643] }
[23661.587643] ... key at: [<ffffffffa03b9a88>]
__key.35650+0x0/0xfffffffffffde578 [btrfs]
[23661.587649] ... acquired at:
[23661.587649] [<ffffffff810988f6>] check_usage_forwards+0x146/0x150
[23661.587651] [<ffffffff81099156>] mark_lock+0x186/0x2c0
[23661.587652] [<ffffffff81099838>] __lock_acquire+0x5a8/0x1e40
[23661.587653] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587655] [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390
[23661.587656] [<ffffffffa036d73c>]
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587663] [<ffffffffa036ed04>]
btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[23661.587670] [<ffffffffa032650d>] btrfs_evict_inode+0x1cd/0x520 [btrfs]
[23661.587676] [<ffffffff811bdd17>] evict+0xa7/0x1a0
[23661.587678] [<ffffffff811bde4e>] dispose_list+0x3e/0x60
[23661.587679] [<ffffffff811bed3c>] prune_icache_sb+0x4c/0x60
[23661.587681] [<ffffffff811a4641>] super_cache_scan+0x111/0x170
[23661.587682] [<ffffffff8114d28a>] shrink_slab+0x1da/0x3a0
[23661.587684] [<ffffffff81150905>] balance_pgdat+0x3d5/0x5e0
[23661.587685] [<ffffffff81150cba>] kswapd+0x1aa/0x4a0
[23661.587686] [<ffffffff810703c6>] kthread+0xf6/0x110
[23661.587687] [<ffffffff8170316c>] ret_from_fork+0x7c/0xb0
[23661.587689]
[23661.587690]
[23661.587690] stack backtrace:
[23661.587692] CPU: 1 PID: 42 Comm: kswapd0 Not tainted
3.13.0-fdm-btrfs-next-24+ #7
[23661.587693] Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011
[23661.587694] ffffffff825db980 ffff8801f5e176a8 ffffffff816f1da6
0000000000000007
[23661.587696] ffffffff825db980 ffff8801f5e176f8 ffffffff816ed6b8
ffffffff825db980
[23661.587698] ffff880213f2ac40 ffff8801f5e1770c 0000000000000000
ffff880213f2ac40
[23661.587700] Call Trace:
[23661.587702] [<ffffffff816f1da6>] dump_stack+0x4e/0x68
[23661.587705] [<ffffffff816ed6b8>] print_irq_inversion_bug.part.38+0x1b0/0x1bf
[23661.587706] [<ffffffff810988f6>] check_usage_forwards+0x146/0x150
[23661.587708] [<ffffffff810987b0>] ? check_usage_backwards+0x150/0x150
[23661.587710] [<ffffffff81099156>] mark_lock+0x186/0x2c0
[23661.587711] [<ffffffff81099838>] __lock_acquire+0x5a8/0x1e40
[23661.587713] [<ffffffff81095b87>] ? __bfs+0x157/0x250
[23661.587714] [<ffffffff8109631b>] ?
add_lock_to_list.isra.26.constprop.42+0x7b/0xc0
[23661.587716] [<ffffffff8109a5ed>] ? __lock_acquire+0x135d/0x1e40
[23661.587718] [<ffffffff816fa94f>] ? _raw_spin_unlock_irqrestore+0x3f/0x70
[23661.587719] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
[23661.587726] [<ffffffffa036d73c>] ?
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587729] [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390
[23661.587735] [<ffffffffa036d73c>] ?
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587737] [<ffffffff81095d08>] ? trace_hardirqs_off_caller+0x28/0xd0
[23661.587744] [<ffffffffa036d73c>] ?
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587747] [<ffffffff81082dff>] ? local_clock+0x4f/0x60
[23661.587754] [<ffffffffa036d73c>]
__btrfs_release_delayed_node+0x4c/0x1f0 [btrfs]
[23661.587761] [<ffffffffa036ed04>] btrfs_remove_delayed_node+0x24/0x30 [btrfs]
[23661.587767] [<ffffffffa032650d>] btrfs_evict_inode+0x1cd/0x520 [btrfs]
[23661.587769] [<ffffffff816fa9ab>] ? _raw_spin_unlock+0x2b/0x40
[23661.587771] [<ffffffff811bdd17>] evict+0xa7/0x1a0
[23661.587773] [<ffffffff811bde4e>] dispose_list+0x3e/0x60
[23661.587775] [<ffffffff811bed3c>] prune_icache_sb+0x4c/0x60
[23661.587776] [<ffffffff811a4641>] super_cache_scan+0x111/0x170
[23661.587778] [<ffffffff8114d28a>] shrink_slab+0x1da/0x3a0
[23661.587780] [<ffffffff81150905>] balance_pgdat+0x3d5/0x5e0
[23661.587782] [<ffffffff81150cba>] kswapd+0x1aa/0x4a0
[23661.587784] [<ffffffff8108e8e0>] ? bit_waitqueue+0xc0/0xc0
[23661.587785] [<ffffffff81150b10>] ? balance_pgdat+0x5e0/0x5e0
[23661.587787] [<ffffffff810703c6>] kthread+0xf6/0x110
[23661.587788] [<ffffffff810702d0>] ? kthread_create_on_node+0x250/0x250
[23661.587790] [<ffffffff8170316c>] ret_from_fork+0x7c/0xb0
[23661.587792] [<ffffffff810702d0>] ? kthread_create_on_node+0x250/0x250
> ---
> fs/btrfs/extent-tree.c | 10 +++++++---
> 1 file changed, 7 insertions(+), 3 deletions(-)
>
> --- a/fs/btrfs/extent-tree.c
> +++ b/fs/btrfs/extent-tree.c
> @@ -8343,9 +8343,15 @@ static void __link_block_group(struct bt
> struct btrfs_block_group_cache *cache)
> {
> int index = get_block_group_index(cache);
> + bool first = false;
>
> down_write(&space_info->groups_sem);
> - if (list_empty(&space_info->block_groups[index])) {
> + if (list_empty(&space_info->block_groups[index]))
> + first = true;
> + list_add_tail(&cache->list, &space_info->block_groups[index]);
> + up_write(&space_info->groups_sem);
> +
> + if (first) {
> struct kobject *kobj = &space_info->block_group_kobjs[index];
> int ret;
>
> @@ -8357,8 +8363,6 @@ static void __link_block_group(struct bt
> kobject_put(&space_info->kobj);
> }
> }
> - list_add_tail(&cache->list, &space_info->block_groups[index]);
> - up_write(&space_info->groups_sem);
> }
>
> static struct btrfs_block_group_cache *
>
>
> --
> Jeff Mahoney
> SUSE Labs
> --
> To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at http://vger.kernel.org/majordomo-info.html
--
Filipe David Manana,
"Reasonable men adapt themselves to the world.
Unreasonable men adapt the world to themselves.
That's why all progress depends on unreasonable men."
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
2014-04-09 16:05 ` Filipe David Manana
@ 2014-04-09 21:22 ` Jeff Mahoney
0 siblings, 0 replies; 6+ messages in thread
From: Jeff Mahoney @ 2014-04-09 21:22 UTC (permalink / raw)
To: fdmanana; +Cc: linux-btrfs
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
On 4/9/14, 12:05 PM, Filipe David Manana wrote:
> On Wed, Mar 26, 2014 at 6:11 PM, Jeff Mahoney <jeffm@suse.com>
> wrote:
>> When encountering memory pressure, testers have run into the
>> following lockdep warning. It was caused by __link_block_group
>> calling kobject_add with the groups_sem held. kobject_add calls
>> kvasprintf with GFP_KERNEL, which gets us into reclaim context.
>> The kobject doesn't actually need to be added under the lock --
>> it just needs to ensure that it's only added for the first block
>> group to be linked.
>>
>> ========================================================= [ INFO:
>> possible irq lock inversion dependency detected ]
>> 3.14.0-rc8-default #1 Not tainted
>> ---------------------------------------------------------
>> kswapd0/169 just changed the state of lock:
>> (&delayed_node->mutex){+.+.-.}, at: [<ffffffffa018baea>]
>> __btrfs_release_delayed_node+0x3a/0x200 [btrfs] but this lock
>> took another, RECLAIM_FS-unsafe lock in the past:
>> (&found->groups_sem){+++++.}
>>
>> and interrupts could create inverse lock ordering between them.
>>
>> other info that might help us debug this: Possible interrupt
>> unsafe locking scenario: CPU0 CPU1 ----
>> ---- lock(&found->groups_sem); local_irq_disable();
>> lock(&delayed_node->mutex); lock(&found->groups_sem);
>> <Interrupt> lock(&delayed_node->mutex);
>>
>> *** DEADLOCK *** 2 locks held by kswapd0/169: #0:
>> (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>]
>> shrink_slab+0x3a/0x160 #1: (&type->s_umount_key#27){++++..}, at:
>> [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
>>
>> Signed-off-by: Jeff Mahoney <jeffm@suse.com>
>
> Hi Jeff,
>
> The same kind of problem happens when deleting the kobj. See the
> trace below, thanks.
Thanks, Filipe. I have an updated patch I'll post in a minute.
- -Jeff
> [23510.196996] BTRFS: device fsid
> 6b4518c2-60e8-4990-8c6e-fb15054b1e46 devid 1 transid 4 /dev/sdd
> [23510.203212] BTRFS info (device sdd): disk space caching is
> enabled [23510.203214] BTRFS: flagging fs with big metadata
> feature [23510.227365] BTRFS: creating UUID tree [23661.587069]
> [23661.587073]
> =========================================================
> [23661.587074] [ INFO: possible irq lock inversion dependency
> detected ] [23661.587076] 3.13.0-fdm-btrfs-next-24+ #7 Not tainted
> [23661.587077]
> ---------------------------------------------------------
> [23661.587078] kswapd0/42 just changed the state of lock:
> [23661.587079] (&delayed_node->mutex){+.+.-.}, at:
> [<ffffffffa036d73c>] __btrfs_release_delayed_node+0x4c/0x1f0
> [btrfs] [23661.587095] but this lock took another,
> RECLAIM_FS-unsafe lock in the past: [23661.587096]
> (sysfs_mutex){+.+.+.} [23661.587096] [23661.587096] and interrupts
> could create inverse lock ordering between them. [23661.587096]
> [23661.587098] [23661.587098] other info that might help us debug
> this: [23661.587099] Chain exists of: [23661.587099]
> &delayed_node->mutex --> &found->groups_sem --> sysfs_mutex
> [23661.587099] [23661.587102] Possible interrupt unsafe locking
> scenario: [23661.587102] [23661.587103] CPU0
> CPU1 [23661.587103] ---- ----
> [23661.587104] lock(sysfs_mutex); [23661.587105]
> local_irq_disable(); [23661.587106]
> lock(&delayed_node->mutex); [23661.587107]
> lock(&found->groups_sem); [23661.587108] <Interrupt>
> [23661.587109] lock(&delayed_node->mutex); [23661.587110]
> [23661.587110] *** DEADLOCK *** [23661.587110] [23661.587111] 2
> locks held by kswapd0/42: [23661.587112] #0:
> (shrinker_rwsem){++++..}, at: [<ffffffff8114d0ec>]
> shrink_slab+0x3c/0x3a0 [23661.587117] #1:
> (&type->s_umount_key#31){+++++.}, at: [<ffffffff811a43d4>]
> grab_super_passive+0x44/0x90 [23661.587122] [23661.587122] the
> shortest dependencies between 2nd lock and 1st lock: [23661.587126]
> -> (sysfs_mutex){+.+.+.} ops: 12918522 { [23661.587129]
> HARDIRQ-ON-W at: [23661.587130]
> [<ffffffff81099900>] __lock_acquire+0x670/0x1e40 [23661.587133]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587134]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587137]
> [<ffffffff8121f1b6>] sysfs_mount+0x166/0x210 [23661.587140]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587142]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587144]
> [<ffffffff811c2509>] kern_mount_data+0x19/0x30 [23661.587145]
> [<ffffffff81f16082>] sysfs_init+0x55/0xb5 [23661.587148]
> [<ffffffff81f13435>] mnt_init+0xcd/0x1d8 [23661.587149]
> [<ffffffff81f130d7>] vfs_caches_init+0x99/0x11c [23661.587151]
> [<ffffffff81eebe1e>] start_kernel+0x3ab/0x40c [23661.587154]
> [<ffffffff81eeb579>] x86_64_start_reservations+0x2a/0x2c
> [23661.587155] [<ffffffff81eeb672>]
> x86_64_start_kernel+0xf7/0xfb [23661.587157] SOFTIRQ-ON-W at:
> [23661.587158] [<ffffffff8109993c>]
> __lock_acquire+0x6ac/0x1e40 [23661.587160]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587162]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587164]
> [<ffffffff8121f1b6>] sysfs_mount+0x166/0x210 [23661.587166]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587167]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587169]
> [<ffffffff811c2509>] kern_mount_data+0x19/0x30 [23661.587171]
> [<ffffffff81f16082>] sysfs_init+0x55/0xb5 [23661.587173]
> [<ffffffff81f13435>] mnt_init+0xcd/0x1d8 [23661.587174]
> [<ffffffff81f130d7>] vfs_caches_init+0x99/0x11c [23661.587175]
> [<ffffffff81eebe1e>] start_kernel+0x3ab/0x40c [23661.587177]
> [<ffffffff81eeb579>] x86_64_start_reservations+0x2a/0x2c
> [23661.587178] [<ffffffff81eeb672>]
> x86_64_start_kernel+0xf7/0xfb [23661.587180] RECLAIM_FS-ON-W
> at: [23661.587181] [<ffffffff8109c039>]
> mark_held_locks+0xb9/0x140 [23661.587183]
> [<ffffffff8109c7c5>] lockdep_trace_alloc+0x65/0xd0 [23661.587185]
> [<ffffffff8118e7e3>] kmem_cache_alloc+0x33/0x1e0 [23661.587187]
> [<ffffffff811bcc51>] alloc_inode+0x61/0xa0 [23661.587190]
> [<ffffffff811be178>] iget_locked+0xd8/0x1d0 [23661.587192]
> [<ffffffff8121b54a>] sysfs_get_inode+0x1a/0x170 [23661.587193]
> [<ffffffff8121f1c5>] sysfs_mount+0x175/0x210 [23661.587195]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587196]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587198]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587199]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587201]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587204]
> INITIAL USE at: [23661.587204]
> [<ffffffff810995bd>] __lock_acquire+0x32d/0x1e40 [23661.587206]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587207]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587209]
> [<ffffffff8121f1b6>] sysfs_mount+0x166/0x210 [23661.587211]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587212]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587213]
> [<ffffffff811c2509>] kern_mount_data+0x19/0x30 [23661.587215]
> [<ffffffff81f16082>] sysfs_init+0x55/0xb5 [23661.587216]
> [<ffffffff81f13435>] mnt_init+0xcd/0x1d8 [23661.587217]
> [<ffffffff81f130d7>] vfs_caches_init+0x99/0x11c [23661.587218]
> [<ffffffff81eebe1e>] start_kernel+0x3ab/0x40c [23661.587220]
> [<ffffffff81eeb579>] x86_64_start_reservations+0x2a/0x2c
> [23661.587221] [<ffffffff81eeb672>]
> x86_64_start_kernel+0xf7/0xfb [23661.587222] } [23661.587223]
> ... key at: [<ffffffff81c79918>] sysfs_mutex+0x78/0xc0
> [23661.587226] ... acquired at: [23661.587227]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587228]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587230]
> [<ffffffff8121defe>] sysfs_addrm_start+0x1e/0x20 [23661.587232]
> [<ffffffff8121e59a>] sysfs_remove+0x1a/0x40 [23661.587233]
> [<ffffffff8121e6c3>] sysfs_remove_dir+0x53/0x80 [23661.587234]
> [<ffffffff81361d11>] kobject_del+0x21/0x50 [23661.587237]
> [<ffffffffa030b060>] btrfs_remove_block_group+0x420/0x540 [btrfs]
> [23661.587246] [<ffffffffa034087b>]
> btrfs_relocate_chunk.isra.61+0x60b/0x760 [btrfs] [23661.587254]
> [<ffffffffa034497f>] btrfs_balance+0x86f/0xfd0 [btrfs]
> [23661.587262] [<ffffffffa034cb45>]
> btrfs_ioctl_balance+0x1c5/0x530 [btrfs] [23661.587269]
> [<ffffffffa034fcfc>] btrfs_ioctl+0x68c/0x28a0 [btrfs]
> [23661.587276] [<ffffffff811b4938>] do_vfs_ioctl+0x88/0x570
> [23661.587278] [<ffffffff811b4eb1>] SyS_ioctl+0x91/0xb0
> [23661.587279] [<ffffffff81703212>]
> system_call_fastpath+0x16/0x1b [23661.587281] [23661.587282] ->
> (&found->groups_sem){++++..} ops: 897759 { [23661.587284]
> HARDIRQ-ON-W at: [23661.587285]
> [<ffffffff81099900>] __lock_acquire+0x670/0x1e40 [23661.587287]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587288]
> [<ffffffff816f8a81>] down_write+0x61/0xc0 [23661.587290]
> [<ffffffffa0306004>] __link_block_group+0x44/0x100 [btrfs]
> [23661.587295] [<ffffffffa0309bf6>]
> btrfs_read_block_groups+0x306/0x650 [btrfs] [23661.587301]
> [<ffffffffa0318376>] open_ctree+0x1676/0x2070 [btrfs]
> [23661.587308] [<ffffffffa02ede54>]
> btrfs_mount+0x6b4/0x910 [btrfs] [23661.587312]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587313]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587315]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587316]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587318]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587320]
> HARDIRQ-ON-R at: [23661.587320]
> [<ffffffff810997bd>] __lock_acquire+0x52d/0x1e40 [23661.587322]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587323]
> [<ffffffff816f8b31>] down_read+0x51/0xa0 [23661.587325]
> [<ffffffffa0314098>]
> btrfs_calc_num_tolerated_disk_barrier_failures+0x118/0x1e0 [btrfs]
> [23661.587331] [<ffffffffa0318386>]
> open_ctree+0x1686/0x2070 [btrfs] [23661.587337]
> [<ffffffffa02ede54>] btrfs_mount+0x6b4/0x910 [btrfs] [23661.587341]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587342]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587344]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587345]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587346]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587348]
> SOFTIRQ-ON-W at: [23661.587349]
> [<ffffffff8109993c>] __lock_acquire+0x6ac/0x1e40 [23661.587350]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587352]
> [<ffffffff816f8a81>] down_write+0x61/0xc0 [23661.587353]
> [<ffffffffa0306004>] __link_block_group+0x44/0x100 [btrfs]
> [23661.587359] [<ffffffffa0309bf6>]
> btrfs_read_block_groups+0x306/0x650 [btrfs] [23661.587364]
> [<ffffffffa0318376>] open_ctree+0x1676/0x2070 [btrfs]
> [23661.587370] [<ffffffffa02ede54>]
> btrfs_mount+0x6b4/0x910 [btrfs] [23661.587374]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587375]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587376]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587378]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587379]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587381]
> SOFTIRQ-ON-R at: [23661.587382]
> [<ffffffff8109993c>] __lock_acquire+0x6ac/0x1e40 [23661.587383]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587384]
> [<ffffffff816f8b31>] down_read+0x51/0xa0 [23661.587386]
> [<ffffffffa0314098>]
> btrfs_calc_num_tolerated_disk_barrier_failures+0x118/0x1e0 [btrfs]
> [23661.587392] [<ffffffffa0318386>]
> open_ctree+0x1686/0x2070 [btrfs] [23661.587397]
> [<ffffffffa02ede54>] btrfs_mount+0x6b4/0x910 [btrfs] [23661.587401]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587402]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587404]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587405]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587406]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587408]
> INITIAL USE at: [23661.587409]
> [<ffffffff810995bd>] __lock_acquire+0x32d/0x1e40 [23661.587410]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587412]
> [<ffffffff816f8a81>] down_write+0x61/0xc0 [23661.587413]
> [<ffffffffa0306004>] __link_block_group+0x44/0x100 [btrfs]
> [23661.587418] [<ffffffffa0309bf6>]
> btrfs_read_block_groups+0x306/0x650 [btrfs] [23661.587424]
> [<ffffffffa0318376>] open_ctree+0x1676/0x2070 [btrfs]
> [23661.587429] [<ffffffffa02ede54>]
> btrfs_mount+0x6b4/0x910 [btrfs] [23661.587433]
> [<ffffffff811a4d53>] mount_fs+0x43/0x1b0 [23661.587435]
> [<ffffffff811c2403>] vfs_kern_mount+0x73/0x160 [23661.587436]
> [<ffffffff811c4d49>] do_mount+0x259/0xb70 [23661.587437]
> [<ffffffff811c594e>] SyS_mount+0x8e/0xe0 [23661.587439]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587441]
> } [23661.587441] ... key at: [<ffffffffa03b5ab8>]
> __key.40358+0x0/0xfffffffffffe2548 [btrfs] [23661.587447] ...
> acquired at: [23661.587448] [<ffffffff8109b6e5>]
> lock_acquire+0x85/0x110 [23661.587449] [<ffffffff816f8b31>]
> down_read+0x51/0xa0 [23661.587451] [<ffffffffa0306891>]
> find_free_extent+0x7d1/0xc30 [btrfs] [23661.587456]
> [<ffffffffa0306d98>] btrfs_reserve_extent+0xa8/0x140 [btrfs]
> [23661.587461] [<ffffffffa0308236>]
> btrfs_alloc_free_block+0xe6/0x460 [btrfs] [23661.587467]
> [<ffffffffa02f6d30>] split_leaf+0x1b0/0x770 [btrfs] [23661.587471]
> [<ffffffffa02f7f23>] btrfs_search_slot+0x963/0x9e0 [btrfs]
> [23661.587476] [<ffffffffa02f9658>]
> btrfs_insert_empty_items+0x78/0xd0 [btrfs] [23661.587481]
> [<ffffffffa036db37>] btrfs_insert_delayed_items+0x87/0x420 [btrfs]
> [23661.587489] [<ffffffffa036e8c2>]
> btrfs_async_run_delayed_root+0xf2/0x200 [btrfs] [23661.587496]
> [<ffffffffa0349fc7>] normal_work_helper+0x137/0x2c0 [btrfs]
> [23661.587503] [<ffffffff810681f5>]
> process_one_work+0x1f5/0x530 [23661.587505] [<ffffffff810689dc>]
> worker_thread+0x11c/0x370 [23661.587506] [<ffffffff810703c6>]
> kthread+0xf6/0x110 [23661.587508] [<ffffffff8170316c>]
> ret_from_fork+0x7c/0xb0 [23661.587510] [23661.587510] ->
> (&delayed_node->mutex){+.+.-.} ops: 5556073 { [23661.587513]
> HARDIRQ-ON-W at: [23661.587514]
> [<ffffffff81099900>] __lock_acquire+0x670/0x1e40 [23661.587515]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587516]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587518]
> [<ffffffffa036f7bf>] btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
> [23661.587525] [<ffffffffa0321107>]
> btrfs_update_inode+0x67/0xf0 [btrfs] [23661.587532]
> [<ffffffffa03293a2>] btrfs_mkdir+0x182/0x1e0 [btrfs] [23661.587539]
> [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170 [23661.587541]
> [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0 [23661.587542]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587544]
> SOFTIRQ-ON-W at: [23661.587544]
> [<ffffffff8109993c>] __lock_acquire+0x6ac/0x1e40 [23661.587546]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587547]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587549]
> [<ffffffffa036f7bf>] btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
> [23661.587556] [<ffffffffa0321107>]
> btrfs_update_inode+0x67/0xf0 [btrfs] [23661.587562]
> [<ffffffffa03293a2>] btrfs_mkdir+0x182/0x1e0 [btrfs] [23661.587568]
> [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170 [23661.587570]
> [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0 [23661.587571]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587573]
> IN-RECLAIM_FS-W at: [23661.587574]
> [<ffffffff81099838>] __lock_acquire+0x5a8/0x1e40 [23661.587575]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587577]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587578]
> [<ffffffffa036d73c>] __btrfs_release_delayed_node+0x4c/0x1f0
> [btrfs] [23661.587585] [<ffffffffa036ed04>]
> btrfs_remove_delayed_node+0x24/0x30 [btrfs] [23661.587592]
> [<ffffffffa032650d>] btrfs_evict_inode+0x1cd/0x520 [btrfs]
> [23661.587599] [<ffffffff811bdd17>]
> evict+0xa7/0x1a0 [23661.587601]
> [<ffffffff811bde4e>] dispose_list+0x3e/0x60 [23661.587602]
> [<ffffffff811bed3c>] prune_icache_sb+0x4c/0x60 [23661.587604]
> [<ffffffff811a4641>] super_cache_scan+0x111/0x170 [23661.587605]
> [<ffffffff8114d28a>] shrink_slab+0x1da/0x3a0 [23661.587608]
> [<ffffffff81150905>] balance_pgdat+0x3d5/0x5e0 [23661.587609]
> [<ffffffff81150cba>] kswapd+0x1aa/0x4a0 [23661.587610]
> [<ffffffff810703c6>] kthread+0xf6/0x110 [23661.587611]
> [<ffffffff8170316c>] ret_from_fork+0x7c/0xb0 [23661.587613]
> INITIAL USE at: [23661.587614]
> [<ffffffff810995bd>] __lock_acquire+0x32d/0x1e40 [23661.587615]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587617]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587618]
> [<ffffffffa036f7bf>] btrfs_delayed_update_inode+0x4f/0x500 [btrfs]
> [23661.587625] [<ffffffffa0321107>]
> btrfs_update_inode+0x67/0xf0 [btrfs] [23661.587631]
> [<ffffffffa03293a2>] btrfs_mkdir+0x182/0x1e0 [btrfs] [23661.587638]
> [<ffffffff811b03ca>] vfs_mkdir+0xca/0x170 [23661.587640]
> [<ffffffff811b2a8b>] SyS_mkdir+0xab/0xf0 [23661.587641]
> [<ffffffff81703212>] system_call_fastpath+0x16/0x1b [23661.587643]
> } [23661.587643] ... key at: [<ffffffffa03b9a88>]
> __key.35650+0x0/0xfffffffffffde578 [btrfs] [23661.587649] ...
> acquired at: [23661.587649] [<ffffffff810988f6>]
> check_usage_forwards+0x146/0x150 [23661.587651]
> [<ffffffff81099156>] mark_lock+0x186/0x2c0 [23661.587652]
> [<ffffffff81099838>] __lock_acquire+0x5a8/0x1e40 [23661.587653]
> [<ffffffff8109b6e5>] lock_acquire+0x85/0x110 [23661.587655]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587656]
> [<ffffffffa036d73c>] __btrfs_release_delayed_node+0x4c/0x1f0
> [btrfs] [23661.587663] [<ffffffffa036ed04>]
> btrfs_remove_delayed_node+0x24/0x30 [btrfs] [23661.587670]
> [<ffffffffa032650d>] btrfs_evict_inode+0x1cd/0x520 [btrfs]
> [23661.587676] [<ffffffff811bdd17>] evict+0xa7/0x1a0
> [23661.587678] [<ffffffff811bde4e>] dispose_list+0x3e/0x60
> [23661.587679] [<ffffffff811bed3c>] prune_icache_sb+0x4c/0x60
> [23661.587681] [<ffffffff811a4641>]
> super_cache_scan+0x111/0x170 [23661.587682] [<ffffffff8114d28a>]
> shrink_slab+0x1da/0x3a0 [23661.587684] [<ffffffff81150905>]
> balance_pgdat+0x3d5/0x5e0 [23661.587685] [<ffffffff81150cba>]
> kswapd+0x1aa/0x4a0 [23661.587686] [<ffffffff810703c6>]
> kthread+0xf6/0x110 [23661.587687] [<ffffffff8170316c>]
> ret_from_fork+0x7c/0xb0 [23661.587689] [23661.587690]
> [23661.587690] stack backtrace: [23661.587692] CPU: 1 PID: 42 Comm:
> kswapd0 Not tainted 3.13.0-fdm-btrfs-next-24+ #7 [23661.587693]
> Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 [23661.587694]
> ffffffff825db980 ffff8801f5e176a8 ffffffff816f1da6
> 0000000000000007 [23661.587696] ffffffff825db980 ffff8801f5e176f8
> ffffffff816ed6b8 ffffffff825db980 [23661.587698] ffff880213f2ac40
> ffff8801f5e1770c 0000000000000000 ffff880213f2ac40 [23661.587700]
> Call Trace: [23661.587702] [<ffffffff816f1da6>]
> dump_stack+0x4e/0x68 [23661.587705] [<ffffffff816ed6b8>]
> print_irq_inversion_bug.part.38+0x1b0/0x1bf [23661.587706]
> [<ffffffff810988f6>] check_usage_forwards+0x146/0x150
> [23661.587708] [<ffffffff810987b0>] ?
> check_usage_backwards+0x150/0x150 [23661.587710]
> [<ffffffff81099156>] mark_lock+0x186/0x2c0 [23661.587711]
> [<ffffffff81099838>] __lock_acquire+0x5a8/0x1e40 [23661.587713]
> [<ffffffff81095b87>] ? __bfs+0x157/0x250 [23661.587714]
> [<ffffffff8109631b>] ?
> add_lock_to_list.isra.26.constprop.42+0x7b/0xc0 [23661.587716]
> [<ffffffff8109a5ed>] ? __lock_acquire+0x135d/0x1e40 [23661.587718]
> [<ffffffff816fa94f>] ? _raw_spin_unlock_irqrestore+0x3f/0x70
> [23661.587719] [<ffffffff8109b6e5>] lock_acquire+0x85/0x110
> [23661.587726] [<ffffffffa036d73c>] ?
> __btrfs_release_delayed_node+0x4c/0x1f0 [btrfs] [23661.587729]
> [<ffffffff816f7f36>] mutex_lock_nested+0x76/0x390 [23661.587735]
> [<ffffffffa036d73c>] ? __btrfs_release_delayed_node+0x4c/0x1f0
> [btrfs] [23661.587737] [<ffffffff81095d08>] ?
> trace_hardirqs_off_caller+0x28/0xd0 [23661.587744]
> [<ffffffffa036d73c>] ? __btrfs_release_delayed_node+0x4c/0x1f0
> [btrfs] [23661.587747] [<ffffffff81082dff>] ?
> local_clock+0x4f/0x60 [23661.587754] [<ffffffffa036d73c>]
> __btrfs_release_delayed_node+0x4c/0x1f0 [btrfs] [23661.587761]
> [<ffffffffa036ed04>] btrfs_remove_delayed_node+0x24/0x30 [btrfs]
> [23661.587767] [<ffffffffa032650d>] btrfs_evict_inode+0x1cd/0x520
> [btrfs] [23661.587769] [<ffffffff816fa9ab>] ?
> _raw_spin_unlock+0x2b/0x40 [23661.587771] [<ffffffff811bdd17>]
> evict+0xa7/0x1a0 [23661.587773] [<ffffffff811bde4e>]
> dispose_list+0x3e/0x60 [23661.587775] [<ffffffff811bed3c>]
> prune_icache_sb+0x4c/0x60 [23661.587776] [<ffffffff811a4641>]
> super_cache_scan+0x111/0x170 [23661.587778] [<ffffffff8114d28a>]
> shrink_slab+0x1da/0x3a0 [23661.587780] [<ffffffff81150905>]
> balance_pgdat+0x3d5/0x5e0 [23661.587782] [<ffffffff81150cba>]
> kswapd+0x1aa/0x4a0 [23661.587784] [<ffffffff8108e8e0>] ?
> bit_waitqueue+0xc0/0xc0 [23661.587785] [<ffffffff81150b10>] ?
> balance_pgdat+0x5e0/0x5e0 [23661.587787] [<ffffffff810703c6>]
> kthread+0xf6/0x110 [23661.587788] [<ffffffff810702d0>] ?
> kthread_create_on_node+0x250/0x250 [23661.587790]
> [<ffffffff8170316c>] ret_from_fork+0x7c/0xb0 [23661.587792]
> [<ffffffff810702d0>] ? kthread_create_on_node+0x250/0x250
>
>
>
>> --- fs/btrfs/extent-tree.c | 10 +++++++--- 1 file changed, 7
>> insertions(+), 3 deletions(-)
>>
>> --- a/fs/btrfs/extent-tree.c +++ b/fs/btrfs/extent-tree.c @@
>> -8343,9 +8343,15 @@ static void __link_block_group(struct bt
>> struct btrfs_block_group_cache *cache) { int index =
>> get_block_group_index(cache); + bool first = false;
>>
>> down_write(&space_info->groups_sem); - if
>> (list_empty(&space_info->block_groups[index])) { + if
>> (list_empty(&space_info->block_groups[index])) +
>> first = true; + list_add_tail(&cache->list,
>> &space_info->block_groups[index]); +
>> up_write(&space_info->groups_sem); + + if (first) { struct
>> kobject *kobj = &space_info->block_group_kobjs[index]; int ret;
>>
>> @@ -8357,8 +8363,6 @@ static void __link_block_group(struct bt
>> kobject_put(&space_info->kobj); } } -
>> list_add_tail(&cache->list, &space_info->block_groups[index]); -
>> up_write(&space_info->groups_sem); }
>>
>> static struct btrfs_block_group_cache *
>>
>>
>> -- Jeff Mahoney SUSE Labs -- To unsubscribe from this list: send
>> the line "unsubscribe linux-btrfs" in the body of a message to
>> majordomo@vger.kernel.org More majordomo info at
>> http://vger.kernel.org/majordomo-info.html
>
>
>
- --
Jeff Mahoney
SUSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG/MacGPG2 v2.0.22 (Darwin)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=34Is
-----END PGP SIGNATURE-----
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
2014-04-09 21:23 Jeff Mahoney
@ 2014-04-14 16:55 ` David Sterba
0 siblings, 0 replies; 6+ messages in thread
From: David Sterba @ 2014-04-14 16:55 UTC (permalink / raw)
To: Jeff Mahoney; +Cc: linux-btrfs
On Wed, Apr 09, 2014 at 05:23:23PM -0400, Jeff Mahoney wrote:
> 2 locks held by kswapd0/169:
> #0: (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>] shrink_slab+0x3a/0x160
> #1: (&type->s_umount_key#27){++++..}, at: [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
>
> Signed-off-by: Jeff Mahoney <jeffm@suse.com>
The first version of patch has been merged meanwhile, can you please
send an incremental? (the 3rd hunk)
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH] btrfs: fix lockdep warning with reclaim lock inversion
@ 2014-04-09 21:23 Jeff Mahoney
2014-04-14 16:55 ` David Sterba
0 siblings, 1 reply; 6+ messages in thread
From: Jeff Mahoney @ 2014-04-09 21:23 UTC (permalink / raw)
To: linux-btrfs
When encountering memory pressure, testers have run into the following
lockdep warning. It was caused by __link_block_group calling kobject_add
with the groups_sem held. kobject_add calls kvasprintf with GFP_KERNEL,
which gets us into reclaim context. The kobject doesn't actually need
to be added under the lock -- it just needs to ensure that it's only
added for the first block group to be linked. We also need to release
the lock before removing the kobjects.
=========================================================
[ INFO: possible irq lock inversion dependency detected ]
3.14.0-rc8-default #1 Not tainted
---------------------------------------------------------
kswapd0/169 just changed the state of lock:
(&delayed_node->mutex){+.+.-.}, at: [<ffffffffa018baea>] __btrfs_release_delayed_node+0x3a/0x200 [btrfs]
but this lock took another, RECLAIM_FS-unsafe lock in the past:
(&found->groups_sem){+++++.}
and interrupts could create inverse lock ordering between them.
other info that might help us debug this:
Possible interrupt unsafe locking scenario:
CPU0 CPU1
---- ----
lock(&found->groups_sem);
local_irq_disable();
lock(&delayed_node->mutex);
lock(&found->groups_sem);
<Interrupt>
lock(&delayed_node->mutex);
*** DEADLOCK ***
2 locks held by kswapd0/169:
#0: (shrinker_rwsem){++++..}, at: [<ffffffff81159e8a>] shrink_slab+0x3a/0x160
#1: (&type->s_umount_key#27){++++..}, at: [<ffffffff811bac6f>] grab_super_passive+0x3f/0x90
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
---
fs/btrfs/extent-tree.c | 20 ++++++++++++++------
1 file changed, 14 insertions(+), 6 deletions(-)
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -8343,9 +8343,15 @@ static void __link_block_group(struct bt
struct btrfs_block_group_cache *cache)
{
int index = get_block_group_index(cache);
+ bool first = false;
down_write(&space_info->groups_sem);
- if (list_empty(&space_info->block_groups[index])) {
+ if (list_empty(&space_info->block_groups[index]))
+ first = true;
+ list_add_tail(&cache->list, &space_info->block_groups[index]);
+ up_write(&space_info->groups_sem);
+
+ if (first) {
struct kobject *kobj = &space_info->block_group_kobjs[index];
int ret;
@@ -8357,8 +8363,6 @@ static void __link_block_group(struct bt
kobject_put(&space_info->kobj);
}
}
- list_add_tail(&cache->list, &space_info->block_groups[index]);
- up_write(&space_info->groups_sem);
}
static struct btrfs_block_group_cache *
@@ -8693,6 +8697,7 @@ int btrfs_remove_block_group(struct btrf
struct btrfs_root *tree_root = root->fs_info->tree_root;
struct btrfs_key key;
struct inode *inode;
+ bool cleanup_needed = false;
int ret;
int index;
int factor;
@@ -8791,12 +8796,15 @@ int btrfs_remove_block_group(struct btrf
* are still on the list after taking the semaphore
*/
list_del_init(&block_group->list);
- if (list_empty(&block_group->space_info->block_groups[index])) {
+ if (list_empty(&block_group->space_info->block_groups[index]))
+ cleanup_needed = true;
+ up_write(&block_group->space_info->groups_sem);
+
+ if (cleanup_needed) {
+ clear_avail_alloc_bits(root->fs_info, block_group->flags);
kobject_del(&block_group->space_info->block_group_kobjs[index]);
kobject_put(&block_group->space_info->block_group_kobjs[index]);
- clear_avail_alloc_bits(root->fs_info, block_group->flags);
}
- up_write(&block_group->space_info->groups_sem);
if (block_group->cached == BTRFS_CACHE_STARTED)
wait_block_group_cache_done(block_group);
--
Jeff Mahoney
SUSE Labs
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2014-04-14 16:55 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2014-03-26 18:11 [PATCH] btrfs: fix lockdep warning with reclaim lock inversion Jeff Mahoney
2014-03-26 18:19 ` David Sterba
2014-04-09 16:05 ` Filipe David Manana
2014-04-09 21:22 ` Jeff Mahoney
2014-04-09 21:23 Jeff Mahoney
2014-04-14 16:55 ` David Sterba
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.