nfs: don't call __mark_inode_dirty while holding i_lock
authorDave Chinner <dchinner@redhat.com>
Tue, 12 Apr 2011 09:18:08 +0000 (19:18 +1000)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 12 Apr 2011 21:17:24 +0000 (14:17 -0700)
commit0d88f6e804c824454b5ed0d3034ed3dcf7467a87
tree3323e2860f7bf892c889270cf936a3c4a45f3b80
parent95042f9eb78a8d9a17455e2ef263f2f310ecef15
nfs: don't call __mark_inode_dirty while holding i_lock

nfs_scan_commit() is called with the inode->i_lock held, but it then
calls __mark_inode_dirty() while still holding the lock. This causes
a deadlock.

Push the inode->i_lock into nfs_scan_commit() so it can protect only
the parts of the code it needs to and can be dropped before the call
to __mark_inode_dirty() to avoid the deadlock.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Tested-by: Will Simoneau <simoneau@ele.uri.edu>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/nfs/write.c