sys_sync(): fix 16% performance regression in ffsb create_4k test

I run many ffsb test cases on JBODs (typically 13/12 disks).  Comparing
with kernel 2.6.30, 2.6.31-rc1 has about 16% regression with
ffsb_create_4k.  The sub test case creates files continuously for 10
minitues and every file is 1MB.

Bisect located below patch.

5cee581 is first bad commit
commit 5cee581
Author: Jan Kara <jack@suse.cz>
Date:   Mon Apr 27 16:43:51 2009 +0200

    vfs: Make sys_sync() use fsync_super() (version 4)

    It is unnecessarily fragile to have two places (fsync_super() and do_sync())
    doing data integrity sync of the filesystem. Alter __fsync_super() to
    accommodate needs of both callers and use it. So after this patch
    __fsync_super() is the only place where we gather all the calls needed to
    properly send all data on a filesystem to disk.

As a matter of fact, ffsb calls sys_sync in the end to make sure all data
is flushed to disks and the flushing is counted into the result.  vmstat
shows ffsb is blocked when syncing for a long time.  With 2.6.30, ffsb is
blocked for a short time.

I checked the patch and did experiments to recover the original methods.
Eventually, the root cause is the patch deletes the calling to
wakeup_pdflush when syncing, so only ffsb is blocked on disk I/O.
wakeup_pdflush could ask pdflush to write back pages with ffsb at the
same time.

[akpm@linux-foundation.org: restore comment too]
Signed-off-by: Zhang Yanmin <yanmin_zhang@linux.intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: Jens Axboe <jens.axboe@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

fs/sync.c

@@ -112,8 +112,13 @@ static void sync_filesystems(int wait)
 	mutex_unlock(&mutex);
 }
 
+/*
+ * sync everything.  Start out by waking pdflush, because that writes back
+ * all queues in parallel.
+ */
 SYSCALL_DEFINE0(sync)
 {
+	wakeup_pdflush(0);
 	sync_filesystems(0);
 	sync_filesystems(1);
 	if (unlikely(laptop_mode))