From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wu Fengguang Subject: Re: [PATCH 02/17] writeback: update dirtied_when for synced inode to prevent livelock Date: Fri, 13 May 2011 11:08:10 +0800 Message-ID: <20110513030810.GB8016@localhost> References: <20110512135706.937596128@intel.com> <20110512140030.883635923@intel.com> <20110512224211.GI19446@dastard> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: Andrew Morton , Jan Kara , Christoph Hellwig , "linux-fsdevel@vger.kernel.org" , LKML To: Dave Chinner Return-path: Content-Disposition: inline In-Reply-To: <20110512224211.GI19446@dastard> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On Fri, May 13, 2011 at 06:42:11AM +0800, Dave Chinner wrote: > On Thu, May 12, 2011 at 09:57:08PM +0800, Wu Fengguang wrote: > > Explicitly update .dirtied_when on synced inodes, so that they are no > > longer considered for writeback in the next round. > > > > We'll do more aggressive "keep writeback as long as we wrote something" > > logic in wb_writeback(). The "use LONG_MAX .nr_to_write" trick in commit > > b9543dac5bbc ("writeback: avoid livelocking WB_SYNC_ALL writeback") will > > no longer be enough to stop sync livelock. > > > > Reviewed-by: Jan Kara > > Signed-off-by: Wu Fengguang > > --- > > fs/fs-writeback.c | 9 +++++++++ > > 1 file changed, 9 insertions(+) > > > > ext3/ext4 are working fine now, however tests show that XFS may still > > livelock inside the XFS routines: > > > > [ 3581.181253] sync D ffff8800b6ca15d8 4560 4403 4392 0x00000000 > > [ 3581.181734] ffff88006f775bc8 0000000000000046 ffff8800b6ca12b8 00000001b6ca1938 > > [ 3581.182411] ffff88006f774000 00000000001d2e40 00000000001d2e40 ffff8800b6ca1280 > > [ 3581.183088] 00000000001d2e40 ffff88006f775fd8 00000340af111ef2 00000000001d2e40 > > [ 3581.183765] Call Trace: > > [ 3581.184008] [] ? lock_release_holdtime+0xa3/0xab > > [ 3581.184392] [] ? prepare_to_wait+0x6c/0x79 > > [ 3581.184756] [] ? prepare_to_wait+0x6c/0x79 > > [ 3581.185120] [] xfs_ioend_wait+0x87/0x9f > > [ 3581.185474] [] ? wake_up_bit+0x2a/0x2a > > [ 3581.185827] [] xfs_sync_inode_data+0x92/0x9d > > [ 3581.186198] [] xfs_inode_ag_walk+0x1a5/0x287 > > [ 3581.186569] [] ? xfs_inode_ag_walk+0x25e/0x287 > > [ 3581.186946] [] ? xfs_sync_worker+0x69/0x69 > > [ 3581.187311] [] ? xfs_perag_get+0x68/0xd0 > > [ 3581.187669] [] ? local_clock+0x41/0x5a > > [ 3581.188020] [] ? lock_release_holdtime+0xa3/0xab > > [ 3581.188403] [] ? xfs_check_sizes+0x160/0x160 > > [ 3581.188773] [] ? xfs_perag_get+0x68/0xd0 > > [ 3581.189130] [] ? xfs_perag_get+0x80/0xd0 > > [ 3581.189488] [] ? xfs_check_sizes+0x160/0x160 > > [ 3581.189858] [] ? xfs_inode_ag_iterator+0x6d/0x8f > > [ 3581.190241] [] ? xfs_sync_worker+0x69/0x69 > > [ 3581.190606] [] xfs_inode_ag_iterator+0x47/0x8f > > [ 3581.190982] [] ? __sync_filesystem+0x7a/0x7a > > [ 3581.191352] [] xfs_sync_data+0x24/0x43 > > [ 3581.191703] [] xfs_quiesce_data+0x2c/0x88 > > [ 3581.192065] [] xfs_fs_sync_fs+0x21/0x48 > > [ 3581.192419] [] __sync_filesystem+0x66/0x7a > > [ 3581.192783] [] sync_one_sb+0x16/0x18 > > [ 3581.193128] [] iterate_supers+0x72/0xce > > [ 3581.193482] [] sync_filesystems+0x20/0x22 > > [ 3581.193842] [] sys_sync+0x21/0x33 > > [ 3581.194177] [] system_call_fastpath+0x16/0x1b > > > > --- linux-next.orig/fs/fs-writeback.c 2011-05-10 09:50:07.000000000 +0800 > > +++ linux-next/fs/fs-writeback.c 2011-05-10 10:03:00.000000000 +0800 > > @@ -419,6 +419,15 @@ writeback_single_inode(struct inode *ino > > spin_lock(&inode->i_lock); > > inode->i_state &= ~I_SYNC; > > if (!(inode->i_state & I_FREEING)) { > > + /* > > + * Sync livelock prevention. Each inode is tagged and synced in > > + * one shot, so we can unconditionally update its dirty time to > > + * prevent syncing it again. Note that time ordering of b_dirty > > + * list will be kept because the following code either removes > > + * the inode from b_dirty or calls redirty_tail(). > > + */ > > + if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_sync) > > + inode->dirtied_when = jiffies; > > Shouldn't this update only ocur if the inode is still dirty? Yeah, that would be better even though the current form won't lead to errors. Let's add one more test (inode->i_state & I_DIRTY)? (I was actually aware of the trade offs and didn't bother to add it..) Thanks, Fengguang