From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cuda.sgi.com (cuda3.sgi.com [192.48.176.15]) by oss.sgi.com (8.14.3/8.14.3/SuSE Linux 0.8) with ESMTP id p044lbpl176146 for ; Mon, 3 Jan 2011 22:47:38 -0600 Received: from mail.internode.on.net (localhost [127.0.0.1]) by cuda.sgi.com (Spam Firewall) with ESMTP id E0BF61D08392 for ; Mon, 3 Jan 2011 20:49:34 -0800 (PST) Received: from mail.internode.on.net (bld-mail15.adl6.internode.on.net [150.101.137.100]) by cuda.sgi.com with ESMTP id 5HFGpiplVOOwXdNj for ; Mon, 03 Jan 2011 20:49:34 -0800 (PST) Received: from dastard (unverified [121.44.135.206]) by mail.internode.on.net (SurgeMail 3.8f2) with ESMTP id 40314813-1927428 for ; Tue, 04 Jan 2011 15:19:33 +1030 (CDT) Received: from chute ([192.168.1.1] helo=disappointment) by dastard with esmtp (Exim 4.72) (envelope-from ) id 1PZypv-0000TW-Nk for xfs@oss.sgi.com; Tue, 04 Jan 2011 15:49:31 +1100 Received: from dave by disappointment with local (Exim 4.72) (envelope-from ) id 1PZypt-0003vI-Vs for xfs@oss.sgi.com; Tue, 04 Jan 2011 15:49:29 +1100 From: Dave Chinner Subject: [PATCH] xfs: ensure log covering transactions are synchronous Date: Tue, 4 Jan 2011 15:49:29 +1100 Message-Id: <1294116569-15052-1-git-send-email-david@fromorbit.com> List-Id: XFS Filesystem from SGI List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Sender: xfs-bounces@oss.sgi.com Errors-To: xfs-bounces@oss.sgi.com To: xfs@oss.sgi.com From: Dave Chinner To ensure the log is covered and the filesystem idles correctly, we need to ensure that dummy transactions hit the disk and do not stay pinned in memory. If the superblock is pinned in memory, it can't be flushed so the log covering cannot make progress. The result is dependent on timing - more oftent han not we continue to issues a log covering transaction every 36s rather than idling after ~90s. Fix this by making the log covering transaction synchronous. To avoid additional log force from xfssyncd, make the log covering transaction take the place of the existing log force in the xfssyncd background sync process. Signed-off-by: Dave Chinner --- fs/xfs/linux-2.6/xfs_super.c | 2 +- fs/xfs/linux-2.6/xfs_sync.c | 11 ++++++----- fs/xfs/xfs_fsops.c | 10 +++++----- fs/xfs/xfs_fsops.h | 2 +- 4 files changed, 13 insertions(+), 12 deletions(-) diff --git a/fs/xfs/linux-2.6/xfs_super.c b/fs/xfs/linux-2.6/xfs_super.c index c51faaa..af32f37 100644 --- a/fs/xfs/linux-2.6/xfs_super.c +++ b/fs/xfs/linux-2.6/xfs_super.c @@ -1413,7 +1413,7 @@ xfs_fs_freeze( xfs_save_resvblks(mp); xfs_quiesce_attr(mp); - return -xfs_fs_log_dummy(mp, SYNC_WAIT); + return -xfs_fs_log_dummy(mp); } STATIC int diff --git a/fs/xfs/linux-2.6/xfs_sync.c b/fs/xfs/linux-2.6/xfs_sync.c index a02480d..e22f005 100644 --- a/fs/xfs/linux-2.6/xfs_sync.c +++ b/fs/xfs/linux-2.6/xfs_sync.c @@ -362,7 +362,7 @@ xfs_quiesce_data( /* mark the log as covered if needed */ if (xfs_log_need_covered(mp)) - error2 = xfs_fs_log_dummy(mp, SYNC_WAIT); + error2 = xfs_fs_log_dummy(mp); /* flush data-only devices */ if (mp->m_rtdev_targp) @@ -503,13 +503,14 @@ xfs_sync_worker( int error; if (!(mp->m_flags & XFS_MOUNT_RDONLY)) { - xfs_log_force(mp, 0); - xfs_reclaim_inodes(mp, 0); /* dgc: errors ignored here */ - error = xfs_qm_sync(mp, SYNC_TRYLOCK); if (mp->m_super->s_frozen == SB_UNFROZEN && xfs_log_need_covered(mp)) - error = xfs_fs_log_dummy(mp, 0); + error = xfs_fs_log_dummy(mp); + else + xfs_log_force(mp, 0); + xfs_reclaim_inodes(mp, 0); + error = xfs_qm_sync(mp, SYNC_TRYLOCK); } mp->m_sync_seq++; wake_up(&mp->m_wait_single_sync_task); diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c index f56d30e..cec89dd 100644 --- a/fs/xfs/xfs_fsops.c +++ b/fs/xfs/xfs_fsops.c @@ -612,12 +612,13 @@ out: * * We cannot use an inode here for this - that will push dirty state back up * into the VFS and then periodic inode flushing will prevent log covering from - * making progress. Hence we log a field in the superblock instead. + * making progress. Hence we log a field in the superblock instead and use a + * synchronous transaction to ensure the superblock is immediately unpinned + * and can be written back. */ int xfs_fs_log_dummy( - xfs_mount_t *mp, - int flags) + xfs_mount_t *mp) { xfs_trans_t *tp; int error; @@ -632,8 +633,7 @@ xfs_fs_log_dummy( /* log the UUID because it is an unchanging field */ xfs_mod_sb(tp, XFS_SB_UUID); - if (flags & SYNC_WAIT) - xfs_trans_set_sync(tp); + xfs_trans_set_sync(tp); return xfs_trans_commit(tp, 0); } diff --git a/fs/xfs/xfs_fsops.h b/fs/xfs/xfs_fsops.h index a786c52..1b6a98b 100644 --- a/fs/xfs/xfs_fsops.h +++ b/fs/xfs/xfs_fsops.h @@ -25,6 +25,6 @@ extern int xfs_fs_counts(xfs_mount_t *mp, xfs_fsop_counts_t *cnt); extern int xfs_reserve_blocks(xfs_mount_t *mp, __uint64_t *inval, xfs_fsop_resblks_t *outval); extern int xfs_fs_goingdown(xfs_mount_t *mp, __uint32_t inflags); -extern int xfs_fs_log_dummy(xfs_mount_t *mp, int flags); +extern int xfs_fs_log_dummy(struct xfs_mount *mp); #endif /* __XFS_FSOPS_H__ */ -- 1.7.2.3 _______________________________________________ xfs mailing list xfs@oss.sgi.com http://oss.sgi.com/mailman/listinfo/xfs