]> git.karo-electronics.de Git - karo-tx-linux.git/commit
include/linux/fs.h: disable preempt when acquire i_size_seqcount write lock
authorFan Du <fan.du@windriver.com>
Wed, 20 Mar 2013 04:07:32 +0000 (15:07 +1100)
committerStephen Rothwell <sfr@canb.auug.org.au>
Fri, 22 Mar 2013 04:33:16 +0000 (15:33 +1100)
commit313e0f2f327ab7f66a2e6647439e0a20994f5427
tree7ef34a979367e5db09ff3437297855c5a10d8a5d
parent5a85e3d01003fe2a6a31bb0ca7b92ae159c78858
include/linux/fs.h: disable preempt when acquire i_size_seqcount write lock

Two rt tasks bind to one CPU core.

The higher priority rt task A preempts a lower priority rt task B which
has already taken the write seq lock, and then the higher priority rt task
A try to acquire read seq lock, it's doomed to lockup.

rt task A with lower priority: call write
i_size_write                                        rt task B with higher priority: call sync, and preempt task A
  write_seqcount_begin(&inode->i_size_seqcount);    i_size_read
  inode->i_size = i_size;                             read_seqcount_begin <-- lockup here...

So disable preempt when acquiring every i_size_seqcount *write* lock will
cure the problem.

Signed-off-by: Fan Du <fan.du@windriver.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
include/linux/fs.h