From: Ken Chen Date: Wed, 17 Oct 2007 06:30:38 +0000 (-0700) Subject: writeback: fix periodic superblock dirty inode flushing X-Git-Tag: v2.6.24-rc1~569 X-Git-Url: https://err.no/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=0e0f4fc22ece8e593167eccbb1a4154565c11faa;p=linux-2.6 writeback: fix periodic superblock dirty inode flushing Current -mm tree has bucketful of bug fixes in periodic writeback path. However, we still hit a glitch where dirty pages on a given inode aren't completely flushed to the disk, and system will accumulate large amount of dirty pages beyond what dirty_expire_interval is designed for. The problem is __sync_single_inode() will move an inode to sb->s_dirty list even when there are more pending dirty pages on that inode. If there is another inode with a small number of dirty pages, we hit a case where the loop iteration in wb_kupdate() terminates prematurely because wbc.nr_to_write > 0. Thus leaving the inode that has large amount of dirty pages behind and it has to wait for another dirty_writeback_interval before we flush it again. We effectively only write out MAX_WRITEBACK_PAGES every dirty_writeback_interval. If the rate of dirtying is sufficiently high, the system will start accumulate a large number of dirty pages. So fix it by having another sb->s_more_io list on which to park the inode while we iterate through sb->s_io and to allow each dirty inode which resides on that sb to have an equal chance of flushing some amount of dirty pages. Signed-off-by: Ken Chen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 39fadfad86..c9d105ff79 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -165,25 +165,11 @@ static void redirty_tail(struct inode *inode) } /* - * Redirty an inode, but mark it as the very next-to-be-written inode on its - * superblock's dirty-inode list. - * We need to preserve s_dirty's reverse-time-orderedness, so we cheat by - * setting this inode's dirtied_when to the same value as that of the inode - * which is presently head-of-list, if present head-of-list is newer than this - * inode. (head-of-list is the least-recently-dirtied inode: the oldest one). + * requeue inode for re-scanning after sb->s_io list is exhausted. */ -static void redirty_head(struct inode *inode) +static void requeue_io(struct inode *inode) { - struct super_block *sb = inode->i_sb; - - if (!list_empty(&sb->s_dirty)) { - struct inode *head_inode; - - head_inode = list_entry(sb->s_dirty.prev, struct inode, i_list); - if (time_after(inode->dirtied_when, head_inode->dirtied_when)) - inode->dirtied_when = head_inode->dirtied_when; - } - list_move_tail(&inode->i_list, &sb->s_dirty); + list_move(&inode->i_list, &inode->i_sb->s_more_io); } /* @@ -255,7 +241,7 @@ __sync_single_inode(struct inode *inode, struct writeback_control *wbc) * uncongested. */ inode->i_state |= I_DIRTY_PAGES; - redirty_head(inode); + requeue_io(inode); } else { /* * Otherwise fully redirty the inode so that @@ -315,7 +301,7 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc) * on s_io. We'll have another go at writing back this inode * when the s_dirty iodes get moved back onto s_io. */ - redirty_head(inode); + requeue_io(inode); /* * Even if we don't actually write the inode itself here, @@ -410,14 +396,14 @@ sync_sb_inodes(struct super_block *sb, struct writeback_control *wbc) wbc->encountered_congestion = 1; if (!sb_is_blkdev_sb(sb)) break; /* Skip a congested fs */ - redirty_head(inode); + requeue_io(inode); continue; /* Skip a congested blockdev */ } if (wbc->bdi && bdi != wbc->bdi) { if (!sb_is_blkdev_sb(sb)) break; /* fs has the wrong queue */ - redirty_head(inode); + requeue_io(inode); continue; /* blockdev has wrong queue */ } @@ -427,8 +413,10 @@ sync_sb_inodes(struct super_block *sb, struct writeback_control *wbc) /* Was this inode dirtied too recently? */ if (wbc->older_than_this && time_after(inode->dirtied_when, - *wbc->older_than_this)) + *wbc->older_than_this)) { + list_splice_init(&sb->s_io, sb->s_dirty.prev); break; + } /* Is another pdflush already flushing this queue? */ if (current_is_pdflush() && !writeback_acquire(bdi)) @@ -458,6 +446,10 @@ sync_sb_inodes(struct super_block *sb, struct writeback_control *wbc) if (wbc->nr_to_write <= 0) break; } + + if (list_empty(&sb->s_io)) + list_splice_init(&sb->s_more_io, &sb->s_io); + return; /* Leave any unwritten inodes on s_io */ } diff --git a/fs/super.c b/fs/super.c index fc8ebedc6b..1bfcca2104 100644 --- a/fs/super.c +++ b/fs/super.c @@ -67,6 +67,7 @@ static struct super_block *alloc_super(struct file_system_type *type) } INIT_LIST_HEAD(&s->s_dirty); INIT_LIST_HEAD(&s->s_io); + INIT_LIST_HEAD(&s->s_more_io); INIT_LIST_HEAD(&s->s_files); INIT_LIST_HEAD(&s->s_instances); INIT_HLIST_HEAD(&s->s_anon); diff --git a/include/linux/fs.h b/include/linux/fs.h index 30aca33994..0b38a897c1 100644 --- a/include/linux/fs.h +++ b/include/linux/fs.h @@ -1002,6 +1002,7 @@ struct super_block { struct list_head s_inodes; /* all inodes */ struct list_head s_dirty; /* dirty inodes */ struct list_head s_io; /* parked for writeback */ + struct list_head s_more_io; /* parked for more writeback */ struct hlist_head s_anon; /* anonymous dentries for (nfs) exporting */ struct list_head s_files;