blkcg: tg_stats_alloc_lock is an irq lock
commitff26eaadf4d914e397872b99885d45756104e9ae
authorTejun Heo <tj@kernel.org>
Wed, 23 May 2012 10:16:21 +0000 (23 12:16 +0200)
committerJens Axboe <axboe@kernel.dk>
Wed, 23 May 2012 10:16:21 +0000 (23 12:16 +0200)
tree5020f3e4a14ab68f6d027366e719b7fb5193123c
parent0b7877d4eea3f93e3dd941999522bbd8c538cb53
blkcg: tg_stats_alloc_lock is an irq lock

tg_stats_alloc_lock nests inside queue lock and should always be held
with irq disabled.  throtl_pd_{init|exit}() were using non-irqsafe
spinlock ops which triggered inverse lock ordering via irq warning via
RCU freeing of blkg invoking throtl_pd_exit() w/o disabling IRQ.

Update both functions to use irq safe operations.

Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Sasha Levin <sasha.levin@oracle.com>
LKML-Reference: <1335339396.16988.80.camel@lappy>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
block/blk-throttle.c