Skip to contentSkip to conversation

Sync and VFS scalability improvements V2

PATCH In review Josef Bacik 2 participants 14 messages Mar 20, 2015 – Apr 7, 2015

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

cover Sync and VFS scalability improvements V2

Here is the update to the patches based on comments from Al and Jan. The changes since V1 are

- Dropped the IOP_NOTHASHED patch and replaced it wit the hlist_fake patch as per Al's suggestion.
- Dropped "inode: convert per-sb inode list to a list_lru" since Jan has found some issues with it and it doesn't seem as awesome as we want.
- Fixed "inode: don't softlockup when evicting inodes" since it was based on the list_lru patch.
- Added my SOB to all the patches as per Jan's suggestion and added his relevant Reviewed-by's.

I'm getting new performance numbers for these patches now, hopefully that stuff will finish soon and I can get them posted. I wanted to get the updated patchset up as quickly as possible since we're getting close to the merge window and I'm getting ready to disappear into vacation for a little bit. Thanks,

Josef

#1 lore raw

PATCH 1/8 writeback: plug writeback at a high level

From: Dave Chinner <[email protected]>

Doing writeback on lots of little files causes terrible IOPS storms because of the per-mapping writeback plugging we do. This essentially causes imeediate dispatch of IO for each mapping, regardless of the context in which writeback is occurring.

IOWs, running a concurrent write-lots-of-small 4k files using fsmark on XFS results in a huge number of IOPS being issued for data writes. Metadata writes are sorted and plugged at a high level by XFS, so aggregate nicely into large IOs. However, data writeback IOs are dispatched in individual 4k IOs, even when the blocks of two consecutively written files are adjacent.

Test VM: 8p, 8GB RAM, 4xSSD in RAID0, 100TB sparse XFS filesystem, metadata CRCs enabled.

Kernel: 3.10-rc5 + xfsdev + my 3.11 xfs queue (~70 patches)

Test:

$ ./fs_mark  -D  10000  -S0  -n  10000  -s  4096  -L  120  -d
/mnt/scratch/0  -d  /mnt/scratch/1  -d  /mnt/scratch/2  -d
/mnt/scratch/3  -d  /mnt/scratch/4  -d  /mnt/scratch/5  -d
/mnt/scratch/6  -d  /mnt/scratch/7

Result:

wall sys create rate Physical write IO
time CPU (avg files/s) IOPS Bandwidth
----- ----- ------------ ------ ---------
unpatched 6m56s 15m47s 24,000+/-500 26,000 130MB/s
patched 5m06s 13m28s 32,800+/-600 1,500 180MB/s
improvement -26.44% -14.68% +36.67% -94.23% +38.46%

If I use zero length files, this workload at about 500 IOPS, so plugging drops the data IOs from roughly 25,500/s to 1000/s. 3 lines of code, 35% better throughput for 15% less CPU.

The benefits of plugging at this layer are likely to be higher for spinning media as the IO patterns for this workload are going make a much bigger difference on high IO latency devices.....

---
 fs/fs-writeback.c | 3 +++
 1 file changed, 3 insertions(+)
fs/fs-writeback.c +3 -0
index e907052..a9ff2b7 100644
@@ -659,7 +659,9 @@ static long writeback_sb_inodes(struct super_block *sb,
 	unsigned long start_time = jiffies;
 	long write_chunk;
 	long wrote = 0;  /* count both pages and inodes */
+	struct blk_plug plug;
 
+	blk_start_plug(&plug);
 	while (!list_empty(&wb->b_io)) {
 		struct inode *inode = wb_inode(wb->b_io.prev);
 
@@ -756,6 +758,7 @@ static long writeback_sb_inodes(struct super_block *sb,
 				break;
 		}
 	}
+	blk_finish_plug(&plug);
 	return wrote;
 }
 
Signature · 1 line
1.8.3.1
#2 lore raw

PATCH 2/8 inode: add hlist_fake to avoid the inode hash lock in evict

In reply to Josef Bacik

Some filesystems don't use the VFS inode hash and fake the fact they are hashed so that all the writeback code works correctly. However, this means the evict() path still tries to remove the inode from the hash, meaning that the inode_hash_lock() needs to be taken unnecessarily. Hence under certain workloads the inode_hash_lock can be contended even if the inode is never actually hashed.

To avoid this add hlist_fake to test if the inode isn't actually hashed to avoid taking the hash lock on inodes that have never been hashed. Based on Dave Chinner's

inode: add IOP_NOTHASHED to avoid inode hash lock in evict

basd on Al's suggestions. Thanks,

---
 include/linux/fs.h   | 2 +-
 include/linux/list.h | 5 +++++
 2 files changed, 6 insertions(+), 1 deletion(-)
include/linux/fs.h +1 -1
index b4d71b5..31da757 100644
@@ -2528,7 +2528,7 @@ static inline void insert_inode_hash(struct inode *inode)
 extern void __remove_inode_hash(struct inode *);
 static inline void remove_inode_hash(struct inode *inode)
 {
-	if (!inode_unhashed(inode))
+	if (!inode_unhashed(inode) && !hlist_fake(&inode->i_hash))
 		__remove_inode_hash(inode);
 }
 
include/linux/list.h +5 -0
index feb773c..3e3e64a 100644
@@ -672,6 +672,11 @@ static inline void hlist_add_fake(struct hlist_node *n)
 	n->pprev = &n->next;
 }
 
+static inline bool hlist_fake(struct hlist_node *h)
+{
+	return h->pprev == &h->next;
+}
+
 /*
  * Move a list from one list head to another. Fixup the pprev
  * reference of the first entry if it exists.
Signature · 1 line
1.8.3.1
#3 lore raw

PATCH 3/8 inode: convert inode_sb_list_lock to per-sb

In reply to Josef Bacik

From: Dave Chinner <[email protected]>

The process of reducing contention on per-superblock inode lists starts with moving the locking to match the per-superblock inode list. This takes the global lock out of the picture and reduces the contention problems to within a single filesystem. This doesn't get rid of contention as the locks still have global CPU scope, but it does isolate operations on different superblocks form each other.

---
 fs/block_dev.c                   | 12 ++++++------
 fs/drop_caches.c                 | 10 ++++++----
 fs/fs-writeback.c                | 12 ++++++------
 fs/inode.c                       | 28 +++++++++++++---------------
 fs/internal.h                    |  1 -
 fs/notify/inode_mark.c           | 20 ++++++++++----------
 fs/quota/dquot.c                 | 16 ++++++++--------
 fs/super.c                       |  3 ++-
 include/linux/fs.h               |  5 ++++-
 include/linux/fsnotify_backend.h |  4 ++--
 10 files changed, 57 insertions(+), 54 deletions(-)
fs/block_dev.c +6 -6
index 975266b..2eb2436 100644
@@ -1753,7 +1753,7 @@ void iterate_bdevs(void (*func)(struct block_device *, void *), void *arg)
 {
 	struct inode *inode, *old_inode = NULL;
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&blockdev_superblock->s_inode_list_lock);
 	list_for_each_entry(inode, &blockdev_superblock->s_inodes, i_sb_list) {
 		struct address_space *mapping = inode->i_mapping;
 
@@ -1765,13 +1765,13 @@ void iterate_bdevs(void (*func)(struct block_device *, void *), void *arg)
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&blockdev_superblock->s_inode_list_lock);
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
 		 * removed from s_inodes list while we dropped the
-		 * inode_sb_list_lock.  We cannot iput the inode now as we can
+		 * s_inode_list_lock  We cannot iput the inode now as we can
 		 * be holding the last reference and we cannot iput it under
-		 * inode_sb_list_lock. So we keep the reference and iput it
+		 * s_inode_list_lock. So we keep the reference and iput it
 		 * later.
 		 */
 		iput(old_inode);
@@ -1779,8 +1779,8 @@ void iterate_bdevs(void (*func)(struct block_device *, void *), void *arg)
 
 		func(I_BDEV(inode), arg);
 
-		spin_lock(&inode_sb_list_lock);
+		spin_lock(&blockdev_superblock->s_inode_list_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&blockdev_superblock->s_inode_list_lock);
 	iput(old_inode);
 }
fs/drop_caches.c +6 -4
index 5718cb9..d72d52b 100644
@@ -17,7 +17,7 @@ static void drop_pagecache_sb(struct super_block *sb, void *unused)
 {
 	struct inode *inode, *toput_inode = NULL;
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		spin_lock(&inode->i_lock);
 		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
@@ -27,13 +27,15 @@ static void drop_pagecache_sb(struct super_block *sb, void *unused)
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&sb->s_inode_list_lock);
+
 		invalidate_mapping_pages(inode->i_mapping, 0, -1);
 		iput(toput_inode);
 		toput_inode = inode;
-		spin_lock(&inode_sb_list_lock);
+
+		spin_lock(&sb->s_inode_list_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 	iput(toput_inode);
 }
 
fs/fs-writeback.c +6 -6
index a9ff2b7..9780f6c 100644
@@ -1298,7 +1298,7 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 
 	/*
 	 * Data integrity sync. Must wait for all pages under writeback,
@@ -1318,14 +1318,14 @@ static void wait_sb_inodes(struct super_block *sb)
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&sb->s_inode_list_lock);
 
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
 		 * removed from s_inodes list while we dropped the
-		 * inode_sb_list_lock.  We cannot iput the inode now as we can
+		 * s_inode_list_lock.  We cannot iput the inode now as we can
 		 * be holding the last reference and we cannot iput it under
-		 * inode_sb_list_lock. So we keep the reference and iput it
+		 * s_inode_list_lock. So we keep the reference and iput it
 		 * later.
 		 */
 		iput(old_inode);
@@ -1335,9 +1335,9 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&inode_sb_list_lock);
+		spin_lock(&sb->s_inode_list_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 	iput(old_inode);
 }
 
fs/inode.c +13 -15
index f00b16f..bd61843 100644
@@ -28,8 +28,8 @@
  *   inode->i_state, inode->i_hash, __iget()
  * Inode LRU list locks protect:
  *   inode->i_sb->s_inode_lru, inode->i_lru
- * inode_sb_list_lock protects:
- *   sb->s_inodes, inode->i_sb_list
+ * inode->i_sb->s_inode_list_lock protects:
+ *   inode->i_sb->s_inodes, inode->i_sb_list
  * bdi->wb.list_lock protects:
  *   bdi->wb.b_{dirty,io,more_io,dirty_time}, inode->i_wb_list
  * inode_hash_lock protects:
@@ -37,7 +37,7 @@
  *
  * Lock ordering:
  *
- * inode_sb_list_lock
+ * inode->i_sb->s_inode_list_lock
  *   inode->i_lock
  *     Inode LRU list locks
  *
@@ -45,7 +45,7 @@
  *   inode->i_lock
  *
  * inode_hash_lock
- *   inode_sb_list_lock
+ *   inode->i_sb->s_inode_list_lock
  *   inode->i_lock
  *
  * iunique_lock
@@ -57,8 +57,6 @@ static unsigned int i_hash_shift __read_mostly;
 static struct hlist_head *inode_hashtable __read_mostly;
 static __cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_hash_lock);
 
-__cacheline_aligned_in_smp DEFINE_SPINLOCK(inode_sb_list_lock);
-
 /*
  * Empty aops. Can be used for the cases where the user does not
  * define any of the address_space operations.
@@ -424,18 +422,18 @@ static void inode_lru_list_del(struct inode *inode)
  */
 void inode_sb_list_add(struct inode *inode)
 {
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&inode->i_sb->s_inode_list_lock);
 	list_add(&inode->i_sb_list, &inode->i_sb->s_inodes);
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&inode->i_sb->s_inode_list_lock);
 }
 EXPORT_SYMBOL_GPL(inode_sb_list_add);
 
 static inline void inode_sb_list_del(struct inode *inode)
 {
 	if (!list_empty(&inode->i_sb_list)) {
-		spin_lock(&inode_sb_list_lock);
+		spin_lock(&inode->i_sb->s_inode_list_lock);
 		list_del_init(&inode->i_sb_list);
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&inode->i_sb->s_inode_list_lock);
 	}
 }
 
@@ -592,7 +590,7 @@ void evict_inodes(struct super_block *sb)
 	struct inode *inode, *next;
 	LIST_HEAD(dispose);
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
 		if (atomic_read(&inode->i_count))
 			continue;
@@ -608,7 +606,7 @@ void evict_inodes(struct super_block *sb)
 		spin_unlock(&inode->i_lock);
 		list_add(&inode->i_lru, &dispose);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 
 	dispose_list(&dispose);
 }
@@ -629,7 +627,7 @@ int invalidate_inodes(struct super_block *sb, bool kill_dirty)
 	struct inode *inode, *next;
 	LIST_HEAD(dispose);
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
 		spin_lock(&inode->i_lock);
 		if (inode->i_state & (I_NEW | I_FREEING | I_WILL_FREE)) {
@@ -652,7 +650,7 @@ int invalidate_inodes(struct super_block *sb, bool kill_dirty)
 		spin_unlock(&inode->i_lock);
 		list_add(&inode->i_lru, &dispose);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 
 	dispose_list(&dispose);
 
@@ -884,7 +882,7 @@ struct inode *new_inode(struct super_block *sb)
 {
 	struct inode *inode;
 
-	spin_lock_prefetch(&inode_sb_list_lock);
+	spin_lock_prefetch(&sb->s_inode_list_lock);
 
 	inode = new_inode_pseudo(sb);
 	if (inode)
fs/internal.h +0 -1
index 01dce1d..b6c350d 100644
@@ -111,7 +111,6 @@ extern int open_check_o_direct(struct file *f);
 /*
  * inode.c
  */
-extern spinlock_t inode_sb_list_lock;
 extern long prune_icache_sb(struct super_block *sb, struct shrink_control *sc);
 extern void inode_add_lru(struct inode *inode);
 
fs/notify/inode_mark.c +10 -10
index 3daf513..a4e1a8f 100644
@@ -163,17 +163,17 @@ int fsnotify_add_inode_mark(struct fsnotify_mark *mark,
 
 /**
  * fsnotify_unmount_inodes - an sb is unmounting.  handle any watched inodes.
- * @list: list of inodes being unmounted (sb->s_inodes)
+ * @sb: superblock being unmounted.
  *
  * Called during unmount with no locks held, so needs to be safe against
- * concurrent modifiers. We temporarily drop inode_sb_list_lock and CAN block.
+ * concurrent modifiers. We temporarily drop sb->s_inode_list_lock and CAN block.
  */
-void fsnotify_unmount_inodes(struct list_head *list)
+void fsnotify_unmount_inodes(struct super_block *sb)
 {
 	struct inode *inode, *next_i, *need_iput = NULL;
 
-	spin_lock(&inode_sb_list_lock);
-	list_for_each_entry_safe(inode, next_i, list, i_sb_list) {
+	spin_lock(&sb->s_inode_list_lock);
+	list_for_each_entry_safe(inode, next_i, &sb->s_inodes, i_sb_list) {
 		struct inode *need_iput_tmp;
 
 		/*
@@ -209,7 +209,7 @@ void fsnotify_unmount_inodes(struct list_head *list)
 		spin_unlock(&inode->i_lock);
 
 		/* In case the dropping of a reference would nuke next_i. */
-		while (&next_i->i_sb_list != list) {
+		while (&next_i->i_sb_list != &sb->s_inodes) {
 			spin_lock(&next_i->i_lock);
 			if (!(next_i->i_state & (I_FREEING | I_WILL_FREE)) &&
 						atomic_read(&next_i->i_count)) {
@@ -224,12 +224,12 @@ void fsnotify_unmount_inodes(struct list_head *list)
 		}
 
 		/*
-		 * We can safely drop inode_sb_list_lock here because either
+		 * We can safely drop s_inode_list_lock here because either
 		 * we actually hold references on both inode and next_i or
 		 * end of list.  Also no new inodes will be added since the
 		 * umount has begun.
 		 */
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&sb->s_inode_list_lock);
 
 		if (need_iput_tmp)
 			iput(need_iput_tmp);
@@ -241,7 +241,7 @@ void fsnotify_unmount_inodes(struct list_head *list)
 
 		iput(inode);
 
-		spin_lock(&inode_sb_list_lock);
+		spin_lock(&sb->s_inode_list_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 }
fs/quota/dquot.c +8 -8
index 0ccd4ba..68c7ae3 100644
@@ -920,7 +920,7 @@ static void add_dquot_ref(struct super_block *sb, int type)
 	int reserved = 0;
 #endif
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		spin_lock(&inode->i_lock);
 		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
@@ -931,7 +931,7 @@ static void add_dquot_ref(struct super_block *sb, int type)
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&inode_sb_list_lock);
+		spin_unlock(&sb->s_inode_list_lock);
 
 #ifdef CONFIG_QUOTA_DEBUG
 		if (unlikely(inode_get_rsv_space(inode) > 0))
@@ -943,15 +943,15 @@ static void add_dquot_ref(struct super_block *sb, int type)
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
 		 * removed from s_inodes list while we dropped the
-		 * inode_sb_list_lock We cannot iput the inode now as we can be
+		 * s_inode_list_lock. We cannot iput the inode now as we can be
 		 * holding the last reference and we cannot iput it under
-		 * inode_sb_list_lock. So we keep the reference and iput it
+		 * s_inode_list_lock. So we keep the reference and iput it
 		 * later.
 		 */
 		old_inode = inode;
-		spin_lock(&inode_sb_list_lock);
+		spin_lock(&sb->s_inode_list_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 	iput(old_inode);
 
 #ifdef CONFIG_QUOTA_DEBUG
@@ -1019,7 +1019,7 @@ static void remove_dquot_ref(struct super_block *sb, int type,
 	struct inode *inode;
 	int reserved = 0;
 
-	spin_lock(&inode_sb_list_lock);
+	spin_lock(&sb->s_inode_list_lock);
 	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
 		/*
 		 *  We have to scan also I_NEW inodes because they can already
@@ -1035,7 +1035,7 @@ static void remove_dquot_ref(struct super_block *sb, int type,
 		}
 		spin_unlock(&dq_data_lock);
 	}
-	spin_unlock(&inode_sb_list_lock);
+	spin_unlock(&sb->s_inode_list_lock);
 #ifdef CONFIG_QUOTA_DEBUG
 	if (reserved) {
 		printk(KERN_WARNING "VFS (%s): Writes happened after quota"
fs/super.c +2 -1
index 2b7dc90..85d6a62 100644
@@ -191,6 +191,7 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags)
 	INIT_HLIST_NODE(&s->s_instances);
 	INIT_HLIST_BL_HEAD(&s->s_anon);
 	INIT_LIST_HEAD(&s->s_inodes);
+	spin_lock_init(&s->s_inode_list_lock);
 
 	if (list_lru_init_memcg(&s->s_dentry_lru))
 		goto fail;
@@ -399,7 +400,7 @@ void generic_shutdown_super(struct super_block *sb)
 		sync_filesystem(sb);
 		sb->s_flags &= ~MS_ACTIVE;
 
-		fsnotify_unmount_inodes(&sb->s_inodes);
+		fsnotify_unmount_inodes(sb);
 
 		evict_inodes(sb);
 
include/linux/fs.h +4 -1
index 31da757..35953b5 100644
@@ -1244,7 +1244,6 @@ struct super_block {
 #endif
 	const struct xattr_handler **s_xattr;
 
-	struct list_head	s_inodes;	/* all inodes */
 	struct hlist_bl_head	s_anon;		/* anonymous dentries for (nfs) exporting */
 	struct list_head	s_mounts;	/* list of mounts; _not_ for fs use */
 	struct block_device	*s_bdev;
@@ -1315,6 +1314,10 @@ struct super_block {
 	 * Indicates how deep in a filesystem stack this SB is
 	 */
 	int s_stack_depth;
+
+	/* s_inode_list_lock protects s_inodes */
+	spinlock_t		s_inode_list_lock ____cacheline_aligned_in_smp;
+	struct list_head	s_inodes;	/* all inodes */
 };
 
 extern struct timespec current_fs_time(struct super_block *sb);
include/linux/fsnotify_backend.h +2 -2
index 0f313f9..236cbc4 100644
@@ -359,7 +359,7 @@ extern void fsnotify_clear_marks_by_group_flags(struct fsnotify_group *group, un
 extern void fsnotify_clear_marks_by_group(struct fsnotify_group *group);
 extern void fsnotify_get_mark(struct fsnotify_mark *mark);
 extern void fsnotify_put_mark(struct fsnotify_mark *mark);
-extern void fsnotify_unmount_inodes(struct list_head *list);
+extern void fsnotify_unmount_inodes(struct super_block *sb);
 
 /* put here because inotify does some weird stuff when destroying watches */
 extern void fsnotify_init_event(struct fsnotify_event *event,
@@ -395,7 +395,7 @@ static inline u32 fsnotify_get_cookie(void)
 	return 0;
 }
 
-static inline void fsnotify_unmount_inodes(struct list_head *list)
+static inline void fsnotify_unmount_inodes(struct super_block *sb)
 {}
 
 #endif	/* CONFIG_FSNOTIFY */
Signature · 1 line
1.8.3.1
#4 lore raw

PATCH 4/8 sync: serialise per-superblock sync operations

In reply to Josef Bacik

From: Dave Chinner <[email protected]>

When competing sync(2) calls walk the same filesystem, they need to walk the list of inodes on the superblock to find all the inodes that we need to wait for IO completion on. However, when multiple wait_sb_inodes() calls do this at the same time, they contend on the the inode_sb_list_lock and the contention causes system wide slowdowns. In effect, concurrent sync(2) calls can take longer and burn more CPU than if they were serialised.

Stop the worst of the contention by adding a per-sb mutex to wrap around wait_sb_inodes() so that we only execute one sync(2) IO completion walk per superblock superblock at a time and hence avoid contention being triggered by concurrent sync(2) calls.

---
 fs/fs-writeback.c  | 11 +++++++++++
 fs/super.c         |  1 +
 include/linux/fs.h |  2 ++
 3 files changed, 14 insertions(+)
fs/fs-writeback.c +11 -0
index 9780f6c..dcfe21c 100644
@@ -1288,6 +1288,15 @@ out_unlock_inode:
 }
 EXPORT_SYMBOL(__mark_inode_dirty);
 
+/*
+ * The @s_sync_lock is used to serialise concurrent sync operations
+ * to avoid lock contention problems with concurrent wait_sb_inodes() calls.
+ * Concurrent callers will block on the s_sync_lock rather than doing contending
+ * walks. The queueing maintains sync(2) required behaviour as all the IO that
+ * has been issued up to the time this function is enter is guaranteed to be
+ * completed by the time we have gained the lock and waited for all IO that is
+ * in progress regardless of the order callers are granted the lock.
+ */
 static void wait_sb_inodes(struct super_block *sb)
 {
 	struct inode *inode, *old_inode = NULL;
@@ -1298,6 +1307,7 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
+	mutex_lock(&sb->s_sync_lock);
 	spin_lock(&sb->s_inode_list_lock);
 
 	/*
@@ -1339,6 +1349,7 @@ static void wait_sb_inodes(struct super_block *sb)
 	}
 	spin_unlock(&sb->s_inode_list_lock);
 	iput(old_inode);
+	mutex_unlock(&sb->s_sync_lock);
 }
 
 /**
fs/super.c +1 -0
index 85d6a62..6a05d94 100644
@@ -190,6 +190,7 @@ static struct super_block *alloc_super(struct file_system_type *type, int flags)
 	s->s_flags = flags;
 	INIT_HLIST_NODE(&s->s_instances);
 	INIT_HLIST_BL_HEAD(&s->s_anon);
+	mutex_init(&s->s_sync_lock);
 	INIT_LIST_HEAD(&s->s_inodes);
 	spin_lock_init(&s->s_inode_list_lock);
 
include/linux/fs.h +2 -0
index 35953b5..10cdf24 100644
@@ -1310,6 +1310,8 @@ struct super_block {
 	struct list_lru		s_inode_lru ____cacheline_aligned_in_smp;
 	struct rcu_head		rcu;
 
+	struct mutex		s_sync_lock;	/* sync serialisation lock */
+
 	/*
 	 * Indicates how deep in a filesystem stack this SB is
 	 */
Signature · 1 line
1.8.3.1
#5 lore raw

PATCH 5/8 inode: rename i_wb_list to i_io_list

In reply to Josef Bacik

From: Dave Chinner <[email protected]>

There's a small consistency problem between the inode and writeback naming. Writeback calls the "for IO" inode queues b_io and b_more_io, but the inode calls these the "writeback list" or i_wb_list. This makes it hard to an new "under writeback" list to the inode, or call it an "under IO" list on the bdi because either way we'll have writeback on IO and IO on writeback and it'll just be confusing. I'm getting confused just writing this!

So, rename the inode "for IO" list variable to i_io_list so we can add a new "writeback list" in a subsequent patch.

---
 fs/fs-writeback.c  | 20 ++++++++++----------
 fs/inode.c         |  6 +++---
 include/linux/fs.h |  2 +-
 mm/backing-dev.c   |  8 ++++----
 4 files changed, 18 insertions(+), 18 deletions(-)
fs/fs-writeback.c +10 -10
index dcfe21c..cabebde 100644
@@ -84,7 +84,7 @@ EXPORT_SYMBOL_GPL(inode_to_bdi);
 
 static inline struct inode *wb_inode(struct list_head *head)
 {
-	return list_entry(head, struct inode, i_wb_list);
+	return list_entry(head, struct inode, i_io_list);
 }
 
 /*
@@ -193,7 +193,7 @@ void inode_wb_list_del(struct inode *inode)
 	struct backing_dev_info *bdi = inode_to_bdi(inode);
 
 	spin_lock(&bdi->wb.list_lock);
-	list_del_init(&inode->i_wb_list);
+	list_del_init(&inode->i_io_list);
 	spin_unlock(&bdi->wb.list_lock);
 }
 
@@ -216,7 +216,7 @@ static void redirty_tail(struct inode *inode, struct bdi_writeback *wb)
 		if (time_before(inode->dirtied_when, tail->dirtied_when))
 			inode->dirtied_when = jiffies;
 	}
-	list_move(&inode->i_wb_list, &wb->b_dirty);
+	list_move(&inode->i_io_list, &wb->b_dirty);
 }
 
 /*
@@ -225,7 +225,7 @@ static void redirty_tail(struct inode *inode, struct bdi_writeback *wb)
 static void requeue_io(struct inode *inode, struct bdi_writeback *wb)
 {
 	assert_spin_locked(&wb->list_lock);
-	list_move(&inode->i_wb_list, &wb->b_more_io);
+	list_move(&inode->i_io_list, &wb->b_more_io);
 }
 
 static void inode_sync_complete(struct inode *inode)
@@ -284,7 +284,7 @@ static int move_expired_inodes(struct list_head *delaying_queue,
 		if (older_than_this &&
 		    inode_dirtied_after(inode, *older_than_this))
 			break;
-		list_move(&inode->i_wb_list, &tmp);
+		list_move(&inode->i_io_list, &tmp);
 		moved++;
 		if (flags & EXPIRE_DIRTY_ATIME)
 			set_bit(__I_DIRTY_TIME_EXPIRED, &inode->i_state);
@@ -307,7 +307,7 @@ static int move_expired_inodes(struct list_head *delaying_queue,
 		list_for_each_prev_safe(pos, node, &tmp) {
 			inode = wb_inode(pos);
 			if (inode->i_sb == sb)
-				list_move(&inode->i_wb_list, dispatch_queue);
+				list_move(&inode->i_io_list, dispatch_queue);
 		}
 	}
 out:
@@ -458,10 +458,10 @@ static void requeue_inode(struct inode *inode, struct bdi_writeback *wb,
 		 */
 		redirty_tail(inode, wb);
 	} else if (inode->i_state & I_DIRTY_TIME) {
-		list_move(&inode->i_wb_list, &wb->b_dirty_time);
+		list_move(&inode->i_io_list, &wb->b_dirty_time);
 	} else {
 		/* The inode is clean. Remove from writeback lists. */
-		list_del_init(&inode->i_wb_list);
+		list_del_init(&inode->i_io_list);
 	}
 }
 
@@ -598,7 +598,7 @@ writeback_single_inode(struct inode *inode, struct bdi_writeback *wb,
 	 * touch it. See comment above for explanation.
 	 */
 	if (!(inode->i_state & I_DIRTY_ALL))
-		list_del_init(&inode->i_wb_list);
+		list_del_init(&inode->i_io_list);
 	spin_unlock(&wb->list_lock);
 	inode_sync_complete(inode);
 out:
@@ -1272,7 +1272,7 @@ void __mark_inode_dirty(struct inode *inode, int flags)
 			}
 
 			inode->dirtied_when = jiffies;
-			list_move(&inode->i_wb_list, dirtytime ?
+			list_move(&inode->i_io_list, dirtytime ?
 				  &bdi->wb.b_dirty_time : &bdi->wb.b_dirty);
 			spin_unlock(&bdi->wb.list_lock);
 			trace_writeback_dirty_inode_enqueue(inode);
fs/inode.c +3 -3
index bd61843..e0f2a60 100644
@@ -31,7 +31,7 @@
  * inode->i_sb->s_inode_list_lock protects:
  *   inode->i_sb->s_inodes, inode->i_sb_list
  * bdi->wb.list_lock protects:
- *   bdi->wb.b_{dirty,io,more_io,dirty_time}, inode->i_wb_list
+ *   bdi->wb.b_{dirty,io,more_io,dirty_time}, inode->i_io_list
  * inode_hash_lock protects:
  *   inode_hashtable, inode->i_hash
  *
@@ -355,7 +355,7 @@ void inode_init_once(struct inode *inode)
 	memset(inode, 0, sizeof(*inode));
 	INIT_HLIST_NODE(&inode->i_hash);
 	INIT_LIST_HEAD(&inode->i_devices);
-	INIT_LIST_HEAD(&inode->i_wb_list);
+	INIT_LIST_HEAD(&inode->i_io_list);
 	INIT_LIST_HEAD(&inode->i_lru);
 	address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
@@ -523,7 +523,7 @@ static void evict(struct inode *inode)
 	BUG_ON(!(inode->i_state & I_FREEING));
 	BUG_ON(!list_empty(&inode->i_lru));
 
-	if (!list_empty(&inode->i_wb_list))
+	if (!list_empty(&inode->i_io_list))
 		inode_wb_list_del(inode);
 
 	inode_sb_list_del(inode);
include/linux/fs.h +1 -1
index 10cdf24..2a28bbd 100644
@@ -606,7 +606,7 @@ struct inode {
 	unsigned long		dirtied_when;	/* jiffies of first dirtying */
 
 	struct hlist_node	i_hash;
-	struct list_head	i_wb_list;	/* backing dev IO list */
+	struct list_head	i_io_list;	/* backing dev IO list */
 	struct list_head	i_lru;		/* inode LRU list */
 	struct list_head	i_sb_list;
 	union {
mm/backing-dev.c +4 -4
index 6dc4580..6e7a644 100644
@@ -54,13 +54,13 @@ static int bdi_debug_stats_show(struct seq_file *m, void *v)
 
 	nr_dirty = nr_io = nr_more_io = nr_dirty_time = 0;
 	spin_lock(&wb->list_lock);
-	list_for_each_entry(inode, &wb->b_dirty, i_wb_list)
+	list_for_each_entry(inode, &wb->b_dirty, i_io_list)
 		nr_dirty++;
-	list_for_each_entry(inode, &wb->b_io, i_wb_list)
+	list_for_each_entry(inode, &wb->b_io, i_io_list)
 		nr_io++;
-	list_for_each_entry(inode, &wb->b_more_io, i_wb_list)
+	list_for_each_entry(inode, &wb->b_more_io, i_io_list)
 		nr_more_io++;
-	list_for_each_entry(inode, &wb->b_dirty_time, i_wb_list)
+	list_for_each_entry(inode, &wb->b_dirty_time, i_io_list)
 		if (inode->i_state & I_DIRTY_TIME)
 			nr_dirty_time++;
 	spin_unlock(&wb->list_lock);
Signature · 1 line
1.8.3.1
#6 lore raw

PATCH 6/8 bdi: add a new writeback list for sync

In reply to Josef Bacik

From: Dave Chinner <[email protected]>

wait_sb_inodes() current does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on.

To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the bdi when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for.

To avoid needing all the realted locking to be safe against interrupts, Jan Kara suggested that we be lazy about removal from the writeback list. That is, we don't remove inodes from the writeback list on IO completion, but do it directly during a wait_sb_inodes() walk.

This means that the a rare sync(2) call will have some work to do skipping clean inodes However, in the current problem case of concurrent sync workloads, concurrent wait_sb_inodes() calls only walk the very recently dispatched inodes and hence should have very little work to do.

This also means that we have to remove the inodes from the writeback list during eviction. Do this in inode_wait_for_writeback() once all writeback on the inode is complete.

---
 fs/fs-writeback.c           | 115 ++++++++++++++++++++++++++++++++++++++------
 fs/inode.c                  |   1 +
 include/linux/backing-dev.h |   3 ++
 include/linux/fs.h          |   1 +
 mm/backing-dev.c            |   1 +
 mm/page-writeback.c         |  14 ++++++
 6 files changed, 119 insertions(+), 16 deletions(-)
fs/fs-writeback.c +99 -16
index cabebde..82b0f43 100644
@@ -198,6 +198,34 @@ void inode_wb_list_del(struct inode *inode)
 }
 
 /*
+ * mark an inode as under writeback on the given bdi
+ */
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi, struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		if (list_empty(&inode->i_wb_list))
+			list_add_tail(&inode->i_wb_list, &bdi->wb.b_writeback);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
+ * clear an inode as under writeback on the given bdi
+ */
+static void bdi_clear_inode_writeback(struct backing_dev_info *bdi,
+				      struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (!list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		list_del_init(&inode->i_wb_list);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
  * Redirty an inode: set its when-it-was dirtied timestamp and move it to the
  * furthest end of its superblock's dirty-inode list.
  *
@@ -371,13 +399,23 @@ static void __inode_wait_for_writeback(struct inode *inode)
 }
 
 /*
- * Wait for writeback on an inode to complete. Caller must have inode pinned.
+ * Wait for writeback on an inode to complete during eviction. Caller must have
+ * inode pinned.
  */
 void inode_wait_for_writeback(struct inode *inode)
 {
+	BUG_ON(!(inode->i_state & I_FREEING));
+
 	spin_lock(&inode->i_lock);
 	__inode_wait_for_writeback(inode);
 	spin_unlock(&inode->i_lock);
+
+	/*
+	 * bd_inode's will have already had the bdev free'd so don't bother
+	 * doing the bdi_clear_inode_writeback.
+	 */
+	if (!sb_is_blkdev_sb(inode->i_sb))
+		bdi_clear_inode_writeback(inode_to_bdi(inode), inode);
 }
 
 /*
@@ -1299,7 +1337,9 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  */
 static void wait_sb_inodes(struct super_block *sb)
 {
-	struct inode *inode, *old_inode = NULL;
+	struct backing_dev_info *bdi = sb->s_bdi;
+	struct inode *old_inode = NULL;
+	LIST_HEAD(sync_list);
 
 	/*
 	 * We need to be protected against the filesystem going from
@@ -1307,28 +1347,59 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
-	mutex_lock(&sb->s_sync_lock);
-	spin_lock(&sb->s_inode_list_lock);
-
 	/*
-	 * Data integrity sync. Must wait for all pages under writeback,
-	 * because there may have been pages dirtied before our sync
-	 * call, but which had writeout started before we write it out.
-	 * In which case, the inode may not be on the dirty list, but
-	 * we still have to wait for that writeout.
+	 * Data integrity sync. Must wait for all pages under writeback, because
+	 * there may have been pages dirtied before our sync call, but which had
+	 * writeout started before we write it out.  In which case, the inode
+	 * may not be on the dirty list, but we still have to wait for that
+	 * writeout.
+	 *
+	 * To avoid syncing inodes put under IO after we have started here,
+	 * splice the io list to a temporary list head and walk that. Newly
+	 * dirtied inodes will go onto the primary list so we won't wait for
+	 * them. This is safe to do as we are serialised by the s_sync_lock,
+	 * so we'll complete processing the complete list before the next
+	 * sync operation repeats the splice-and-walk process.
+	 *
+	 * Inodes that have pages under writeback after we've finished the wait
+	 * may or may not be on the primary list. They had pages under put IO
+	 * after we started our wait, so we need to make sure the next sync or
+	 * IO completion treats them correctly, Move them back to the primary
+	 * list and restart the walk.
+	 *
+	 * Inodes that are clean after we have waited for them don't belong on
+	 * any list, so simply remove them from the sync list and move onwards.
 	 */
-	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+	mutex_lock(&sb->s_sync_lock);
+	spin_lock(&bdi->wb.list_lock);
+	list_splice_init(&bdi->wb.b_writeback, &sync_list);
+
+	while (!list_empty(&sync_list)) {
+		struct inode *inode = list_first_entry(&sync_list, struct inode,
+						       i_wb_list);
 		struct address_space *mapping = inode->i_mapping;
 
+		/*
+		 * We are lazy on IO completion and don't remove inodes from the
+		 * list when they are clean. Detect that immediately and skip
+		 * inodes we don't ahve to wait on.
+		 */
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			list_del_init(&inode->i_wb_list);
+			cond_resched_lock(&bdi->wb.list_lock);
+			continue;
+		}
+
 		spin_lock(&inode->i_lock);
-		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
-		    (mapping->nrpages == 0)) {
+		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) {
+                        list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
 			spin_unlock(&inode->i_lock);
+			cond_resched_lock(&bdi->wb.list_lock);
 			continue;
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&sb->s_inode_list_lock);
+		spin_unlock(&bdi->wb.list_lock);
 
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
@@ -1345,9 +1416,21 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&sb->s_inode_list_lock);
+		/*
+		 * the inode has been written back now, so check whether we
+		 * still have pages under IO and move it back to the primary
+		 * list if necessary
+		 */
+		spin_lock_irq(&mapping->tree_lock);
+		spin_lock(&bdi->wb.list_lock);
+		if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			WARN_ON(list_empty(&inode->i_wb_list));
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
+                } else
+			list_del_init(&inode->i_wb_list);
+		spin_unlock_irq(&mapping->tree_lock);
 	}
-	spin_unlock(&sb->s_inode_list_lock);
+	spin_unlock(&bdi->wb.list_lock);
 	iput(old_inode);
 	mutex_unlock(&sb->s_sync_lock);
 }
fs/inode.c +1 -0
index e0f2a60..b961e5a 100644
@@ -356,6 +356,7 @@ void inode_init_once(struct inode *inode)
 	INIT_HLIST_NODE(&inode->i_hash);
 	INIT_LIST_HEAD(&inode->i_devices);
 	INIT_LIST_HEAD(&inode->i_io_list);
+	INIT_LIST_HEAD(&inode->i_wb_list);
 	INIT_LIST_HEAD(&inode->i_lru);
 	address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
include/linux/backing-dev.h +3 -0
index aff923a..12d8224 100644
@@ -56,6 +56,7 @@ struct bdi_writeback {
 	struct list_head b_io;		/* parked for writeback */
 	struct list_head b_more_io;	/* parked for more writeback */
 	struct list_head b_dirty_time;	/* time stamps are dirty */
+	struct list_head b_writeback;	/* inodes under writeback */
 	spinlock_t list_lock;		/* protects the b_* lists */
 };
 
@@ -124,6 +125,8 @@ void bdi_start_background_writeback(struct backing_dev_info *bdi);
 void bdi_writeback_workfn(struct work_struct *work);
 int bdi_has_dirty_io(struct backing_dev_info *bdi);
 void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi);
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi,
+			      struct inode *inode);
 
 extern spinlock_t bdi_lock;
 extern struct list_head bdi_list;
include/linux/fs.h +1 -0
index 2a28bbd..bdf0735 100644
@@ -609,6 +609,7 @@ struct inode {
 	struct list_head	i_io_list;	/* backing dev IO list */
 	struct list_head	i_lru;		/* inode LRU list */
 	struct list_head	i_sb_list;
+	struct list_head	i_wb_list;	/* backing dev writeback list */
 	union {
 		struct hlist_head	i_dentry;
 		struct rcu_head		i_rcu;
mm/backing-dev.c +1 -0
index 6e7a644..df31958 100644
@@ -386,6 +386,7 @@ static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
 	INIT_LIST_HEAD(&wb->b_io);
 	INIT_LIST_HEAD(&wb->b_more_io);
 	INIT_LIST_HEAD(&wb->b_dirty_time);
+	INIT_LIST_HEAD(&wb->b_writeback);
 	spin_lock_init(&wb->list_lock);
 	INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
 }
mm/page-writeback.c +14 -0
index 45e187b..4f21498 100644
@@ -2359,11 +2359,25 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 		spin_lock_irqsave(&mapping->tree_lock, flags);
 		ret = TestSetPageWriteback(page);
 		if (!ret) {
+			bool on_wblist;
+
+			on_wblist = mapping_tagged(mapping,
+						   PAGECACHE_TAG_WRITEBACK);
+
 			radix_tree_tag_set(&mapping->page_tree,
 						page_index(page),
 						PAGECACHE_TAG_WRITEBACK);
 			if (bdi_cap_account_writeback(bdi))
 				__inc_bdi_stat(bdi, BDI_WRITEBACK);
+
+			/*
+			 * we can come through here when swapping anonymous
+			 * pages, so we don't necessarily have an inode to
+			 * track for sync. Inodes are remove lazily, so there is
+			 * no equivalent in test_clear_page_writeback().
+			 */
+			if (!on_wblist && mapping->host)
+				bdi_mark_inode_writeback(bdi, mapping->host);
 		}
 		if (!PageDirty(page))
 			radix_tree_tag_clear(&mapping->page_tree,
Signature · 1 line
1.8.3.1
#7 lore raw

PATCH 7/8 writeback: periodically trim the writeback list

In reply to Josef Bacik

From: Dave Chinner <[email protected]>

Inodes are removed lazily from the bdi writeback list, so in the absence of sync(2) work inodes will build up on the bdi writback list even though they are no longer under IO. Use the periodic kupdate work check to remove inodes no longer under IO from the writeback list.

---
 fs/fs-writeback.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
fs/fs-writeback.c +19 -0
index 82b0f43..aa0de0f 100644
@@ -1037,6 +1037,23 @@ static long wb_check_background_flush(struct bdi_writeback *wb)
 	return 0;
 }
 
+/*
+ * clean out writeback list for all inodes that don't have IO in progress
+ */
+static void wb_trim_writeback_list(struct bdi_writeback *wb)
+{
+	struct inode *inode;
+	struct inode *tmp;
+
+	spin_lock(&wb->list_lock);
+	list_for_each_entry_safe(inode, tmp, &wb->b_writeback, i_wb_list) {
+		if (!mapping_tagged(inode->i_mapping, PAGECACHE_TAG_WRITEBACK))
+			list_del_init(&inode->i_wb_list);
+	}
+	spin_unlock(&wb->list_lock);
+
+}
+
 static long wb_check_old_data_flush(struct bdi_writeback *wb)
 {
 	unsigned long expired;
@@ -1053,6 +1070,8 @@ static long wb_check_old_data_flush(struct bdi_writeback *wb)
 	if (time_before(jiffies, expired))
 		return 0;
 
+	wb_trim_writeback_list(wb);
+
 	wb->last_old_flush = jiffies;
 	nr_pages = get_nr_dirty_pages();
 
Signature · 1 line
1.8.3.1
#8 lore raw

PATCH 7/8 bdi: add a new writeback list for sync V3

From: Dave Chinner <[email protected]>

wait_sb_inodes() current does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on.

To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the bdi when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for.

To avoid needing all the realted locking to be safe against interrupts, Jan Kara suggested that we be lazy about removal from the writeback list. That is, we don't remove inodes from the writeback list on IO completion, but do it directly during a wait_sb_inodes() walk.

This means that the a rare sync(2) call will have some work to do skipping clean inodes However, in the current problem case of concurrent sync workloads, concurrent wait_sb_inodes() calls only walk the very recently dispatched inodes and hence should have very little work to do.

This also means that we have to remove the inodes from the writeback list during eviction. Do this in inode_wait_for_writeback() once all writeback on the inode is complete.

---
V2->V3:
-noticed a lockdep warning because of mapping->tree_lock-->wb.list_lock
dependancy and the need for tree_lock to be under IRQ. I re-arranged it so the dependancy is now wb.list_lock-->mapping->tree_lock and it seems to be ok, Jan could you look hard at this?

 fs/fs-writeback.c           | 115 ++++++++++++++++++++++++++++++++++++++------
 fs/inode.c                  |   1 +
 include/linux/backing-dev.h |   3 ++
 include/linux/fs.h          |   1 +
 mm/backing-dev.c            |   1 +
 mm/page-writeback.c         |  19 ++++++++
 6 files changed, 124 insertions(+), 16 deletions(-)
fs/fs-writeback.c +99 -16
index cabebde..e70d2ea 100644
@@ -198,6 +198,34 @@ void inode_wb_list_del(struct inode *inode)
 }
 
 /*
+ * mark an inode as under writeback on the given bdi
+ */
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi, struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		if (list_empty(&inode->i_wb_list))
+			list_add_tail(&inode->i_wb_list, &bdi->wb.b_writeback);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
+ * clear an inode as under writeback on the given bdi
+ */
+static void bdi_clear_inode_writeback(struct backing_dev_info *bdi,
+				      struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (!list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		list_del_init(&inode->i_wb_list);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
  * Redirty an inode: set its when-it-was dirtied timestamp and move it to the
  * furthest end of its superblock's dirty-inode list.
  *
@@ -371,13 +399,23 @@ static void __inode_wait_for_writeback(struct inode *inode)
 }
 
 /*
- * Wait for writeback on an inode to complete. Caller must have inode pinned.
+ * Wait for writeback on an inode to complete during eviction. Caller must have
+ * inode pinned.
  */
 void inode_wait_for_writeback(struct inode *inode)
 {
+	BUG_ON(!(inode->i_state & I_FREEING));
+
 	spin_lock(&inode->i_lock);
 	__inode_wait_for_writeback(inode);
 	spin_unlock(&inode->i_lock);
+
+	/*
+	 * bd_inode's will have already had the bdev free'd so don't bother
+	 * doing the bdi_clear_inode_writeback.
+	 */
+	if (!sb_is_blkdev_sb(inode->i_sb))
+		bdi_clear_inode_writeback(inode_to_bdi(inode), inode);
 }
 
 /*
@@ -1299,7 +1337,9 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  */
 static void wait_sb_inodes(struct super_block *sb)
 {
-	struct inode *inode, *old_inode = NULL;
+	struct backing_dev_info *bdi = sb->s_bdi;
+	struct inode *old_inode = NULL;
+	LIST_HEAD(sync_list);
 
 	/*
 	 * We need to be protected against the filesystem going from
@@ -1307,28 +1347,59 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
-	mutex_lock(&sb->s_sync_lock);
-	spin_lock(&sb->s_inode_list_lock);
-
 	/*
-	 * Data integrity sync. Must wait for all pages under writeback,
-	 * because there may have been pages dirtied before our sync
-	 * call, but which had writeout started before we write it out.
-	 * In which case, the inode may not be on the dirty list, but
-	 * we still have to wait for that writeout.
+	 * Data integrity sync. Must wait for all pages under writeback, because
+	 * there may have been pages dirtied before our sync call, but which had
+	 * writeout started before we write it out.  In which case, the inode
+	 * may not be on the dirty list, but we still have to wait for that
+	 * writeout.
+	 *
+	 * To avoid syncing inodes put under IO after we have started here,
+	 * splice the io list to a temporary list head and walk that. Newly
+	 * dirtied inodes will go onto the primary list so we won't wait for
+	 * them. This is safe to do as we are serialised by the s_sync_lock,
+	 * so we'll complete processing the complete list before the next
+	 * sync operation repeats the splice-and-walk process.
+	 *
+	 * Inodes that have pages under writeback after we've finished the wait
+	 * may or may not be on the primary list. They had pages under put IO
+	 * after we started our wait, so we need to make sure the next sync or
+	 * IO completion treats them correctly, Move them back to the primary
+	 * list and restart the walk.
+	 *
+	 * Inodes that are clean after we have waited for them don't belong on
+	 * any list, so simply remove them from the sync list and move onwards.
 	 */
-	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+	mutex_lock(&sb->s_sync_lock);
+	spin_lock(&bdi->wb.list_lock);
+	list_splice_init(&bdi->wb.b_writeback, &sync_list);
+
+	while (!list_empty(&sync_list)) {
+		struct inode *inode = list_first_entry(&sync_list, struct inode,
+						       i_wb_list);
 		struct address_space *mapping = inode->i_mapping;
 
+		/*
+		 * We are lazy on IO completion and don't remove inodes from the
+		 * list when they are clean. Detect that immediately and skip
+		 * inodes we don't ahve to wait on.
+		 */
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			list_del_init(&inode->i_wb_list);
+			cond_resched_lock(&bdi->wb.list_lock);
+			continue;
+		}
+
 		spin_lock(&inode->i_lock);
-		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
-		    (mapping->nrpages == 0)) {
+		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) {
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
 			spin_unlock(&inode->i_lock);
+			cond_resched_lock(&bdi->wb.list_lock);
 			continue;
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&sb->s_inode_list_lock);
+		spin_unlock(&bdi->wb.list_lock);
 
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
@@ -1345,9 +1416,21 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&sb->s_inode_list_lock);
+		/*
+		 * the inode has been written back now, so check whether we
+		 * still have pages under IO and move it back to the primary
+		 * list if necessary.
+		 */
+		spin_lock(&bdi->wb.list_lock);
+		spin_lock_irq(&mapping->tree_lock);
+		if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			WARN_ON(list_empty(&inode->i_wb_list));
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
+                } else
+			list_del_init(&inode->i_wb_list);
+		spin_unlock_irq(&mapping->tree_lock);
 	}
-	spin_unlock(&sb->s_inode_list_lock);
+	spin_unlock(&bdi->wb.list_lock);
 	iput(old_inode);
 	mutex_unlock(&sb->s_sync_lock);
 }
fs/inode.c +1 -0
index e0f2a60..b961e5a 100644
@@ -356,6 +356,7 @@ void inode_init_once(struct inode *inode)
 	INIT_HLIST_NODE(&inode->i_hash);
 	INIT_LIST_HEAD(&inode->i_devices);
 	INIT_LIST_HEAD(&inode->i_io_list);
+	INIT_LIST_HEAD(&inode->i_wb_list);
 	INIT_LIST_HEAD(&inode->i_lru);
 	address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
include/linux/backing-dev.h +3 -0
index aff923a..12d8224 100644
@@ -56,6 +56,7 @@ struct bdi_writeback {
 	struct list_head b_io;		/* parked for writeback */
 	struct list_head b_more_io;	/* parked for more writeback */
 	struct list_head b_dirty_time;	/* time stamps are dirty */
+	struct list_head b_writeback;	/* inodes under writeback */
 	spinlock_t list_lock;		/* protects the b_* lists */
 };
 
@@ -124,6 +125,8 @@ void bdi_start_background_writeback(struct backing_dev_info *bdi);
 void bdi_writeback_workfn(struct work_struct *work);
 int bdi_has_dirty_io(struct backing_dev_info *bdi);
 void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi);
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi,
+			      struct inode *inode);
 
 extern spinlock_t bdi_lock;
 extern struct list_head bdi_list;
include/linux/fs.h +1 -0
index 2a28bbd..bdf0735 100644
@@ -609,6 +609,7 @@ struct inode {
 	struct list_head	i_io_list;	/* backing dev IO list */
 	struct list_head	i_lru;		/* inode LRU list */
 	struct list_head	i_sb_list;
+	struct list_head	i_wb_list;	/* backing dev writeback list */
 	union {
 		struct hlist_head	i_dentry;
 		struct rcu_head		i_rcu;
mm/backing-dev.c +1 -0
index 6e7a644..df31958 100644
@@ -386,6 +386,7 @@ static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
 	INIT_LIST_HEAD(&wb->b_io);
 	INIT_LIST_HEAD(&wb->b_more_io);
 	INIT_LIST_HEAD(&wb->b_dirty_time);
+	INIT_LIST_HEAD(&wb->b_writeback);
 	spin_lock_init(&wb->list_lock);
 	INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
 }
mm/page-writeback.c +19 -0
index 45e187b..b1f1d48 100644
@@ -2355,10 +2355,14 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 	if (mapping) {
 		struct backing_dev_info *bdi = inode_to_bdi(mapping->host);
 		unsigned long flags;
+		bool on_wblist = true;
 
 		spin_lock_irqsave(&mapping->tree_lock, flags);
 		ret = TestSetPageWriteback(page);
 		if (!ret) {
+			on_wblist = mapping_tagged(mapping,
+						   PAGECACHE_TAG_WRITEBACK);
+
 			radix_tree_tag_set(&mapping->page_tree,
 						page_index(page),
 						PAGECACHE_TAG_WRITEBACK);
@@ -2374,6 +2378,21 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 						page_index(page),
 						PAGECACHE_TAG_TOWRITE);
 		spin_unlock_irqrestore(&mapping->tree_lock, flags);
+
+		/*
+		 * we can come through here when swapping anonymous pages, so
+		 * we don't necessarily have an inode to track for sync. Inodes
+		 * are remove lazily, so there is no equivalent in
+		 * test_clear_page_writeback().
+		 *
+		 * We do this outside the mapping->tree_lock because we don't
+		 * want to force the bdi list lock to have to be used with irq's
+		 * disabled.  Anybody who cares about synchronization here
+		 * should take the bdi's list lock and then take the inodes
+		 * mapping tree_lock.
+		 */
+		if (!on_wblist && mapping->host)
+			bdi_mark_inode_writeback(bdi, mapping->host);
 	} else {
 		ret = TestSetPageWriteback(page);
 	}
Signature · 1 line
1.8.3.1
maintainer
#9 lore raw

Re: [PATCH 7/8] bdi: add a new writeback list for sync V3

Hi Josef,

Quoted 39 lines · Josef Bacik

From: Dave Chinner <[email protected]>

wait_sb_inodes() current does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on.

To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the bdi when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for.

To avoid needing all the realted locking to be safe against interrupts, Jan Kara suggested that we be lazy about removal from the writeback list. That is, we don't remove inodes from the writeback list on IO completion, but do it directly during a wait_sb_inodes() walk.

This means that the a rare sync(2) call will have some work to do skipping clean inodes However, in the current problem case of concurrent sync workloads, concurrent wait_sb_inodes() calls only walk the very recently dispatched inodes and hence should have very little work to do.

This also means that we have to remove the inodes from the writeback list during eviction. Do this in inode_wait_for_writeback() once all writeback on the inode is complete.

---
V2->V3:
-noticed a lockdep warning because of mapping->tree_lock-->wb.list_lock
dependancy and the need for tree_lock to be under IRQ. I re-arranged it so the dependancy is now wb.list_lock-->mapping->tree_lock and it seems to be ok, Jan could you look hard at this?

...

Quoted 70 lines
fs/fs-writeback.c +42 -12
@@ -1299,7 +1337,9 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  */
 static void wait_sb_inodes(struct super_block *sb)
 {
-	struct inode *inode, *old_inode = NULL;
+	struct backing_dev_info *bdi = sb->s_bdi;
+	struct inode *old_inode = NULL;
+	LIST_HEAD(sync_list);
 
 	/*
 	 * We need to be protected against the filesystem going from
@@ -1307,28 +1347,59 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
-	mutex_lock(&sb->s_sync_lock);
-	spin_lock(&sb->s_inode_list_lock);
-
 	/*
-	 * Data integrity sync. Must wait for all pages under writeback,
-	 * because there may have been pages dirtied before our sync
-	 * call, but which had writeout started before we write it out.
-	 * In which case, the inode may not be on the dirty list, but
-	 * we still have to wait for that writeout.
+	 * Data integrity sync. Must wait for all pages under writeback, because
+	 * there may have been pages dirtied before our sync call, but which had
+	 * writeout started before we write it out.  In which case, the inode
+	 * may not be on the dirty list, but we still have to wait for that
+	 * writeout.
+	 *
+	 * To avoid syncing inodes put under IO after we have started here,
+	 * splice the io list to a temporary list head and walk that. Newly
+	 * dirtied inodes will go onto the primary list so we won't wait for
+	 * them. This is safe to do as we are serialised by the s_sync_lock,
+	 * so we'll complete processing the complete list before the next
+	 * sync operation repeats the splice-and-walk process.
+	 *
+	 * Inodes that have pages under writeback after we've finished the wait
+	 * may or may not be on the primary list. They had pages under put IO
+	 * after we started our wait, so we need to make sure the next sync or
+	 * IO completion treats them correctly, Move them back to the primary
+	 * list and restart the walk.
+	 *
+	 * Inodes that are clean after we have waited for them don't belong on
+	 * any list, so simply remove them from the sync list and move onwards.
 	 */
-	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+	mutex_lock(&sb->s_sync_lock);
+	spin_lock(&bdi->wb.list_lock);
+	list_splice_init(&bdi->wb.b_writeback, &sync_list);
+
+	while (!list_empty(&sync_list)) {
+		struct inode *inode = list_first_entry(&sync_list, struct inode,
+						       i_wb_list);
 		struct address_space *mapping = inode->i_mapping;
 
+		/*
+		 * We are lazy on IO completion and don't remove inodes from the
+		 * list when they are clean. Detect that immediately and skip
+		 * inodes we don't ahve to wait on.
+		 */
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			list_del_init(&inode->i_wb_list);
+			cond_resched_lock(&bdi->wb.list_lock);
+			continue;
+		}
+
 		spin_lock(&inode->i_lock);
-		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
-		    (mapping->nrpages == 0)) {

fs/fs-writeback.c static void wait_sb_inodes(struct super_block *sb) · line 1402

Quoted 10 lines
fs/fs-writeback.c +4 -1
+		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) {
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
 			spin_unlock(&inode->i_lock);
+			cond_resched_lock(&bdi->wb.list_lock);
 			continue;
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&sb->s_inode_list_lock);
+		spin_unlock(&bdi->wb.list_lock);

Hum, so you seem to keep that dance with inode references around (keeping
refs to current and prev inode). However that isn't necessary anymore once you always pull the first inode from the list. So just grab reference to the inode before dropping i_lock and put it when you are done with it...

fs/fs-writeback.c static void wait_sb_inodes(struct super_block *sb) · line 1425

Quoted 12 lines
fs/fs-writeback.c +7 -1
@@ -1345,9 +1416,21 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&sb->s_inode_list_lock);
+		/*
+		 * the inode has been written back now, so check whether we
+		 * still have pages under IO and move it back to the primary
+		 * list if necessary.
+		 */
+		spin_lock(&bdi->wb.list_lock);
+		spin_lock_irq(&mapping->tree_lock);

Please comment that tree_lock is really necessary here. We need it
because bdi_mark_inode_writeback() needn't have bothered with grabbing list_lock and may have exited without doing anything when it saw inode on our spliced list.

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
maintainer
#10 lore raw

Re: [PATCH 7/8] bdi: add a new writeback list for sync V3

In reply to Josef Bacik

BTW, this should have been patch 6/8, not 7/8...

Honza

Quoted 319 lines · Josef Bacik

From: Dave Chinner <[email protected]>

wait_sb_inodes() current does a walk of all inodes in the filesystem to find dirty one to wait on during sync. This is highly inefficient and wastes a lot of CPU when there are lots of clean cached inodes that we don't need to wait on.

To avoid this "all inode" walk, we need to track inodes that are currently under writeback that we need to wait for. We do this by adding inodes to a writeback list on the bdi when the mapping is first tagged as having pages under writeback. wait_sb_inodes() can then walk this list of "inodes under IO" and wait specifically just for the inodes that the current sync(2) needs to wait for.

To avoid needing all the realted locking to be safe against interrupts, Jan Kara suggested that we be lazy about removal from the writeback list. That is, we don't remove inodes from the writeback list on IO completion, but do it directly during a wait_sb_inodes() walk.

This means that the a rare sync(2) call will have some work to do skipping clean inodes However, in the current problem case of concurrent sync workloads, concurrent wait_sb_inodes() calls only walk the very recently dispatched inodes and hence should have very little work to do.

This also means that we have to remove the inodes from the writeback list during eviction. Do this in inode_wait_for_writeback() once all writeback on the inode is complete.

---
V2->V3:

(quoted diff) +0 -1
-noticed a lockdep warning because of mapping->tree_lock-->wb.list_lock
 dependancy and the need for tree_lock to be under IRQ.  I re-arranged it so the
 dependancy is now wb.list_lock-->mapping->tree_lock and it seems to be ok, Jan
 could you look hard at this?
 
 fs/fs-writeback.c           | 115 ++++++++++++++++++++++++++++++++++++++------
 fs/inode.c                  |   1 +
 include/linux/backing-dev.h |   3 ++
 include/linux/fs.h          |   1 +
 mm/backing-dev.c            |   1 +
 mm/page-writeback.c         |  19 ++++++++
 6 files changed, 124 insertions(+), 16 deletions(-)
fs/fs-writeback.c +99 -16
index cabebde..e70d2ea 100644
@@ -198,6 +198,34 @@ void inode_wb_list_del(struct inode *inode)
 }
 
 /*
+ * mark an inode as under writeback on the given bdi
+ */
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi, struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		if (list_empty(&inode->i_wb_list))
+			list_add_tail(&inode->i_wb_list, &bdi->wb.b_writeback);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
+ * clear an inode as under writeback on the given bdi
+ */
+static void bdi_clear_inode_writeback(struct backing_dev_info *bdi,
+				      struct inode *inode)
+{
+	WARN_ON_ONCE(bdi != inode_to_bdi(inode));
+	if (!list_empty(&inode->i_wb_list)) {
+		spin_lock(&bdi->wb.list_lock);
+		list_del_init(&inode->i_wb_list);
+		spin_unlock(&bdi->wb.list_lock);
+	}
+}
+
+/*
  * Redirty an inode: set its when-it-was dirtied timestamp and move it to the
  * furthest end of its superblock's dirty-inode list.
  *
@@ -371,13 +399,23 @@ static void __inode_wait_for_writeback(struct inode *inode)
 }
 
 /*
- * Wait for writeback on an inode to complete. Caller must have inode pinned.
+ * Wait for writeback on an inode to complete during eviction. Caller must have
+ * inode pinned.
  */
 void inode_wait_for_writeback(struct inode *inode)
 {
+	BUG_ON(!(inode->i_state & I_FREEING));
+
 	spin_lock(&inode->i_lock);
 	__inode_wait_for_writeback(inode);
 	spin_unlock(&inode->i_lock);
+
+	/*
+	 * bd_inode's will have already had the bdev free'd so don't bother
+	 * doing the bdi_clear_inode_writeback.
+	 */
+	if (!sb_is_blkdev_sb(inode->i_sb))
+		bdi_clear_inode_writeback(inode_to_bdi(inode), inode);
 }
 
 /*
@@ -1299,7 +1337,9 @@ EXPORT_SYMBOL(__mark_inode_dirty);
  */
 static void wait_sb_inodes(struct super_block *sb)
 {
-	struct inode *inode, *old_inode = NULL;
+	struct backing_dev_info *bdi = sb->s_bdi;
+	struct inode *old_inode = NULL;
+	LIST_HEAD(sync_list);
 
 	/*
 	 * We need to be protected against the filesystem going from
@@ -1307,28 +1347,59 @@ static void wait_sb_inodes(struct super_block *sb)
 	 */
 	WARN_ON(!rwsem_is_locked(&sb->s_umount));
 
-	mutex_lock(&sb->s_sync_lock);
-	spin_lock(&sb->s_inode_list_lock);
-
 	/*
-	 * Data integrity sync. Must wait for all pages under writeback,
-	 * because there may have been pages dirtied before our sync
-	 * call, but which had writeout started before we write it out.
-	 * In which case, the inode may not be on the dirty list, but
-	 * we still have to wait for that writeout.
+	 * Data integrity sync. Must wait for all pages under writeback, because
+	 * there may have been pages dirtied before our sync call, but which had
+	 * writeout started before we write it out.  In which case, the inode
+	 * may not be on the dirty list, but we still have to wait for that
+	 * writeout.
+	 *
+	 * To avoid syncing inodes put under IO after we have started here,
+	 * splice the io list to a temporary list head and walk that. Newly
+	 * dirtied inodes will go onto the primary list so we won't wait for
+	 * them. This is safe to do as we are serialised by the s_sync_lock,
+	 * so we'll complete processing the complete list before the next
+	 * sync operation repeats the splice-and-walk process.
+	 *
+	 * Inodes that have pages under writeback after we've finished the wait
+	 * may or may not be on the primary list. They had pages under put IO
+	 * after we started our wait, so we need to make sure the next sync or
+	 * IO completion treats them correctly, Move them back to the primary
+	 * list and restart the walk.
+	 *
+	 * Inodes that are clean after we have waited for them don't belong on
+	 * any list, so simply remove them from the sync list and move onwards.
 	 */
-	list_for_each_entry(inode, &sb->s_inodes, i_sb_list) {
+	mutex_lock(&sb->s_sync_lock);
+	spin_lock(&bdi->wb.list_lock);
+	list_splice_init(&bdi->wb.b_writeback, &sync_list);
+
+	while (!list_empty(&sync_list)) {
+		struct inode *inode = list_first_entry(&sync_list, struct inode,
+						       i_wb_list);
 		struct address_space *mapping = inode->i_mapping;
 
+		/*
+		 * We are lazy on IO completion and don't remove inodes from the
+		 * list when they are clean. Detect that immediately and skip
+		 * inodes we don't ahve to wait on.
+		 */
+		if (!mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			list_del_init(&inode->i_wb_list);
+			cond_resched_lock(&bdi->wb.list_lock);
+			continue;
+		}
+
 		spin_lock(&inode->i_lock);
-		if ((inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) ||
-		    (mapping->nrpages == 0)) {
+		if (inode->i_state & (I_FREEING|I_WILL_FREE|I_NEW)) {
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
 			spin_unlock(&inode->i_lock);
+			cond_resched_lock(&bdi->wb.list_lock);
 			continue;
 		}
 		__iget(inode);
 		spin_unlock(&inode->i_lock);
-		spin_unlock(&sb->s_inode_list_lock);
+		spin_unlock(&bdi->wb.list_lock);
 
 		/*
 		 * We hold a reference to 'inode' so it couldn't have been
@@ -1345,9 +1416,21 @@ static void wait_sb_inodes(struct super_block *sb)
 
 		cond_resched();
 
-		spin_lock(&sb->s_inode_list_lock);
+		/*
+		 * the inode has been written back now, so check whether we
+		 * still have pages under IO and move it back to the primary
+		 * list if necessary.
+		 */
+		spin_lock(&bdi->wb.list_lock);
+		spin_lock_irq(&mapping->tree_lock);
+		if (mapping_tagged(mapping, PAGECACHE_TAG_WRITEBACK)) {
+			WARN_ON(list_empty(&inode->i_wb_list));
+			list_move(&inode->i_wb_list, &bdi->wb.b_writeback);
+                } else
+			list_del_init(&inode->i_wb_list);
+		spin_unlock_irq(&mapping->tree_lock);
 	}
-	spin_unlock(&sb->s_inode_list_lock);
+	spin_unlock(&bdi->wb.list_lock);
 	iput(old_inode);
 	mutex_unlock(&sb->s_sync_lock);
 }
fs/inode.c +1 -0
index e0f2a60..b961e5a 100644
@@ -356,6 +356,7 @@ void inode_init_once(struct inode *inode)
 	INIT_HLIST_NODE(&inode->i_hash);
 	INIT_LIST_HEAD(&inode->i_devices);
 	INIT_LIST_HEAD(&inode->i_io_list);
+	INIT_LIST_HEAD(&inode->i_wb_list);
 	INIT_LIST_HEAD(&inode->i_lru);
 	address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
include/linux/backing-dev.h +3 -0
index aff923a..12d8224 100644
@@ -56,6 +56,7 @@ struct bdi_writeback {
 	struct list_head b_io;		/* parked for writeback */
 	struct list_head b_more_io;	/* parked for more writeback */
 	struct list_head b_dirty_time;	/* time stamps are dirty */
+	struct list_head b_writeback;	/* inodes under writeback */
 	spinlock_t list_lock;		/* protects the b_* lists */
 };
 
@@ -124,6 +125,8 @@ void bdi_start_background_writeback(struct backing_dev_info *bdi);
 void bdi_writeback_workfn(struct work_struct *work);
 int bdi_has_dirty_io(struct backing_dev_info *bdi);
 void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi);
+void bdi_mark_inode_writeback(struct backing_dev_info *bdi,
+			      struct inode *inode);
 
 extern spinlock_t bdi_lock;
 extern struct list_head bdi_list;
include/linux/fs.h +1 -0
index 2a28bbd..bdf0735 100644
@@ -609,6 +609,7 @@ struct inode {
 	struct list_head	i_io_list;	/* backing dev IO list */
 	struct list_head	i_lru;		/* inode LRU list */
 	struct list_head	i_sb_list;
+	struct list_head	i_wb_list;	/* backing dev writeback list */
 	union {
 		struct hlist_head	i_dentry;
 		struct rcu_head		i_rcu;
mm/backing-dev.c +1 -0
index 6e7a644..df31958 100644
@@ -386,6 +386,7 @@ static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
 	INIT_LIST_HEAD(&wb->b_io);
 	INIT_LIST_HEAD(&wb->b_more_io);
 	INIT_LIST_HEAD(&wb->b_dirty_time);
+	INIT_LIST_HEAD(&wb->b_writeback);
 	spin_lock_init(&wb->list_lock);
 	INIT_DELAYED_WORK(&wb->dwork, bdi_writeback_workfn);
 }
mm/page-writeback.c +19 -0
index 45e187b..b1f1d48 100644
@@ -2355,10 +2355,14 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 	if (mapping) {
 		struct backing_dev_info *bdi = inode_to_bdi(mapping->host);
 		unsigned long flags;
+		bool on_wblist = true;
 
 		spin_lock_irqsave(&mapping->tree_lock, flags);
 		ret = TestSetPageWriteback(page);
 		if (!ret) {
+			on_wblist = mapping_tagged(mapping,
+						   PAGECACHE_TAG_WRITEBACK);
+
 			radix_tree_tag_set(&mapping->page_tree,
 						page_index(page),
 						PAGECACHE_TAG_WRITEBACK);
@@ -2374,6 +2378,21 @@ int __test_set_page_writeback(struct page *page, bool keep_write)
 						page_index(page),
 						PAGECACHE_TAG_TOWRITE);
 		spin_unlock_irqrestore(&mapping->tree_lock, flags);
+
+		/*
+		 * we can come through here when swapping anonymous pages, so
+		 * we don't necessarily have an inode to track for sync. Inodes
+		 * are remove lazily, so there is no equivalent in
+		 * test_clear_page_writeback().
+		 *
+		 * We do this outside the mapping->tree_lock because we don't
+		 * want to force the bdi list lock to have to be used with irq's
+		 * disabled.  Anybody who cares about synchronization here
+		 * should take the bdi's list lock and then take the inodes
+		 * mapping tree_lock.
+		 */
+		if (!on_wblist && mapping->host)
+			bdi_mark_inode_writeback(bdi, mapping->host);
 	} else {
 		ret = TestSetPageWriteback(page);
 	}
Signature · 1 line
1.8.3.1
Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#11 lore raw

PATCH 8/8 inode: don't softlockup when evicting inodes

In reply to Josef Bacik

On a box with a lot of ram (148gb) I can make the box softlockup after running an fs_mark job that creates hundreds of millions of empty files. This is because we never generate enough memory pressure to keep the number of inodes on our unused list low, so when we go to unmount we have to evict ~100 million inodes. This makes one processor a very unhappy person, so add a cond_resched() in dispose_list() and cond_resched_lock() in the eviction isolation function to combat this. Thanks,

---
 fs/inode.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
fs/inode.c +10 -0
index b961e5a..c58dbd3 100644
@@ -574,6 +574,7 @@ static void dispose_list(struct list_head *head)
 		list_del_init(&inode->i_lru);
 
 		evict(inode);
+		cond_resched();
 	}
 }
 
@@ -592,6 +593,7 @@ void evict_inodes(struct super_block *sb)
 	LIST_HEAD(dispose);
 
 	spin_lock(&sb->s_inode_list_lock);
+again:
 	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
 		if (atomic_read(&inode->i_count))
 			continue;
@@ -606,6 +608,14 @@ void evict_inodes(struct super_block *sb)
 		inode_lru_list_del(inode);
 		spin_unlock(&inode->i_lock);
 		list_add(&inode->i_lru, &dispose);
+
+		/*
+		 * We can have a ton of inodes to evict at unmount time given
+		 * enough memory, check to see if we need to go to sleep for a
+		 * bit so we don't livelock.
+		 */
+		if (cond_resched_lock(&sb->s_inode_list_lock))
+			goto again;
 	}
 	spin_unlock(&sb->s_inode_list_lock);
 
Signature · 1 line
1.8.3.1
maintainer
#12 lore raw

Re: [PATCH 8/8] inode: don't softlockup when evicting inodes

Sorry for a late reply. I was ill last week...

Quoted 18 lines · Josef Bacik

On a box with a lot of ram (148gb) I can make the box softlockup after running an fs_mark job that creates hundreds of millions of empty files. This is because we never generate enough memory pressure to keep the number of inodes on our unused list low, so when we go to unmount we have to evict ~100 million inodes. This makes one processor a very unhappy person, so add a cond_resched() in dispose_list() and cond_resched_lock() in the eviction isolation function to combat this. Thanks,

---
 fs/inode.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
fs/inode.c +0 -0
index b961e5a..c58dbd3 100644

fs/inode.c static void dispose_list(struct list_head *head) · line 577

Quoted 4 lines
fs/inode.c +1 -0
@@ -574,6 +574,7 @@ static void dispose_list(struct list_head *head)
 		list_del_init(&inode->i_lru);
 
 		evict(inode);
+		cond_resched();

Fine.

Quoted 13 lines
fs/inode.c +1 -0
 	}
 }
 
@@ -592,6 +593,7 @@ void evict_inodes(struct super_block *sb)
 	LIST_HEAD(dispose);
 
 	spin_lock(&sb->s_inode_list_lock);
+again:
 	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
 		if (atomic_read(&inode->i_count))
 			continue;
@@ -606,6 +608,14 @@ void evict_inodes(struct super_block *sb)
 		inode_lru_list_del(inode);

fs/inode.c void evict_inodes(struct super_block *sb) · line 618

Quoted 10 lines
fs/inode.c +8 -0
 		spin_unlock(&inode->i_lock);
 		list_add(&inode->i_lru, &dispose);
+
+		/*
+		 * We can have a ton of inodes to evict at unmount time given
+		 * enough memory, check to see if we need to go to sleep for a
+		 * bit so we don't livelock.
+		 */
+		if (cond_resched_lock(&sb->s_inode_list_lock))
+			goto again;

Not so fine. How this is ever guaranteed to finish? We don't move inodes
from the i_sb_list in this loop so if we ever take 'goto again' we just start doing all the work from the beginning...

What needs to happen is that if we need to resched, we drop sb->s_inode_list_lock, call dispose_list(&dispose) and *then* restart from the beginning since we have freed all the inodes that we isolated...

Honza

Signature · 2 lines
Jan Kara <[email protected]>
SUSE Labs, CR
#13 lore raw

Re: [PATCH 8/8] inode: don't softlockup when evicting inodes

Quoted 2 lines · Jan Kara

Sorry for a late reply. I was ill last week...

That's ok, I was on vacation for the last two weeks ;).

Quoted 52 lines · Josef Bacik
Quoted 18 lines

On a box with a lot of ram (148gb) I can make the box softlockup after running an fs_mark job that creates hundreds of millions of empty files. This is because we never generate enough memory pressure to keep the number of inodes on our unused list low, so when we go to unmount we have to evict ~100 million inodes. This makes one processor a very unhappy person, so add a cond_resched() in dispose_list() and cond_resched_lock() in the eviction isolation function to combat this. Thanks,

---
  fs/inode.c | 10 ++++++++++
  1 file changed, 10 insertions(+)
fs/inode.c +0 -0
index b961e5a..c58dbd3 100644

fs/inode.c static void dispose_list(struct list_head *head) · line 577

Quoted 4 lines
fs/inode.c +1 -0
@@ -574,6 +574,7 @@ static void dispose_list(struct list_head *head)
  		list_del_init(&inode->i_lru);
 
  		evict(inode);
+		cond_resched();

Fine.

Quoted 13 lines
fs/inode.c +1 -0
  	}
  }
 
@@ -592,6 +593,7 @@ void evict_inodes(struct super_block *sb)
  	LIST_HEAD(dispose);
 
  	spin_lock(&sb->s_inode_list_lock);
+again:
  	list_for_each_entry_safe(inode, next, &sb->s_inodes, i_sb_list) {
  		if (atomic_read(&inode->i_count))
  			continue;
@@ -606,6 +608,14 @@ void evict_inodes(struct super_block *sb)
  		inode_lru_list_del(inode);

fs/inode.c void evict_inodes(struct super_block *sb) · line 618

Quoted 10 lines
fs/inode.c +8 -0
  		spin_unlock(&inode->i_lock);
  		list_add(&inode->i_lru, &dispose);
+
+		/*
+		 * We can have a ton of inodes to evict at unmount time given
+		 * enough memory, check to see if we need to go to sleep for a
+		 * bit so we don't livelock.
+		 */
+		if (cond_resched_lock(&sb->s_inode_list_lock))
+			goto again;

Not so fine. How this is ever guaranteed to finish? We don't move inodes
from the i_sb_list in this loop so if we ever take 'goto again' we just start doing all the work from the beginning...

Quoted 4 lines

What needs to happen is that if we need to resched, we drop sb->s_inode_list_lock, call dispose_list(&dispose) and *then* restart from the beginning since we have freed all the inodes that we isolated...

Ooops, good point. I'll get this fixed up, thanks,

Josef