diff options
| author | Kirill A. Shutemov <kirill.shutemov@linux.intel.com> | 2012-09-26 11:33:07 +1000 |
|---|---|---|
| committer | Al Viro <viro@zeniv.linux.org.uk> | 2012-10-02 21:35:55 -0400 |
| commit | 8c0a85377048b64c880e76ec7368904fe46d0b94 (patch) | |
| tree | 12e075f4159e88b09a89e35866017f82f80bcc7d /fs/nilfs2/super.c | |
| parent | 99621b44aa194eab594e1f17217231c02b519211 (diff) | |
| download | lwn-8c0a85377048b64c880e76ec7368904fe46d0b94.tar.gz lwn-8c0a85377048b64c880e76ec7368904fe46d0b94.zip | |
fs: push rcu_barrier() from deactivate_locked_super() to filesystems
There's no reason to call rcu_barrier() on every
deactivate_locked_super(). We only need to make sure that all delayed rcu
free inodes are flushed before we destroy related cache.
Removing rcu_barrier() from deactivate_locked_super() affects some fast
paths. E.g. on my machine exit_group() of a last process in IPC
namespace takes 0.07538s. rcu_barrier() takes 0.05188s of that time.
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Diffstat (limited to 'fs/nilfs2/super.c')
| -rw-r--r-- | fs/nilfs2/super.c | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/fs/nilfs2/super.c b/fs/nilfs2/super.c index 6a10812711c1..3c991dc84f2f 100644 --- a/fs/nilfs2/super.c +++ b/fs/nilfs2/super.c @@ -1382,6 +1382,12 @@ static void nilfs_segbuf_init_once(void *obj) static void nilfs_destroy_cachep(void) { + /* + * Make sure all delayed rcu free inodes are flushed before we + * destroy cache. + */ + rcu_barrier(); + if (nilfs_inode_cachep) kmem_cache_destroy(nilfs_inode_cachep); if (nilfs_transaction_cachep) |