From f9cae926f35e8230330f28c7b743ad088611a8de Mon Sep 17 00:00:00 2001
From: Jan Kara <jack@suse.cz>
Date: Fri, 29 May 2020 16:08:58 +0200
Subject: writeback: Fix sync livelock due to b_dirty_time processing

When we are processing writeback for sync(2), move_expired_inodes()
didn't set any inode expiry value (older_than_this). This can result in
writeback never completing if there's steady stream of inodes added to
b_dirty_time list as writeback rechecks dirty lists after each writeback
round whether there's more work to be done. Fix the problem by using
sync(2) start time is inode expiry value when processing b_dirty_time
list similarly as for ordinarily dirtied inodes. This requires some
refactoring of older_than_this handling which simplifies the code
noticeably as a bonus.

Fixes: 0ae45f63d4ef ("vfs: add support for a lazytime mount option")
CC: stable@vger.kernel.org
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 include/trace/events/writeback.h | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

(limited to 'include/trace')

diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index 10f5d1fa7347..7565dcd59697 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -498,8 +498,9 @@ DEFINE_WBC_EVENT(wbc_writepage);
 TRACE_EVENT(writeback_queue_io,
 	TP_PROTO(struct bdi_writeback *wb,
 		 struct wb_writeback_work *work,
+		 unsigned long dirtied_before,
 		 int moved),
-	TP_ARGS(wb, work, moved),
+	TP_ARGS(wb, work, dirtied_before, moved),
 	TP_STRUCT__entry(
 		__array(char,		name, 32)
 		__field(unsigned long,	older)
@@ -509,19 +510,17 @@ TRACE_EVENT(writeback_queue_io,
 		__field(ino_t,		cgroup_ino)
 	),
 	TP_fast_assign(
-		unsigned long *older_than_this = work->older_than_this;
 		strscpy_pad(__entry->name, bdi_dev_name(wb->bdi), 32);
-		__entry->older	= older_than_this ?  *older_than_this : 0;
-		__entry->age	= older_than_this ?
-				  (jiffies - *older_than_this) * 1000 / HZ : -1;
+		__entry->older	= dirtied_before;
+		__entry->age	= (jiffies - dirtied_before) * 1000 / HZ;
 		__entry->moved	= moved;
 		__entry->reason	= work->reason;
 		__entry->cgroup_ino	= __trace_wb_assign_cgroup(wb);
 	),
 	TP_printk("bdi %s: older=%lu age=%ld enqueue=%d reason=%s cgroup_ino=%lu",
 		__entry->name,
-		__entry->older,	/* older_than_this in jiffies */
-		__entry->age,	/* older_than_this in relative milliseconds */
+		__entry->older,	/* dirtied_before in jiffies */
+		__entry->age,	/* dirtied_before in relative milliseconds */
 		__entry->moved,
 		__print_symbolic(__entry->reason, WB_WORK_REASON),
 		(unsigned long)__entry->cgroup_ino
-- 
cgit v1.2.3


From 5fcd57505c002efc5823a7355e21f48dd02d5a51 Mon Sep 17 00:00:00 2001
From: Jan Kara <jack@suse.cz>
Date: Fri, 29 May 2020 16:24:43 +0200
Subject: writeback: Drop I_DIRTY_TIME_EXPIRE

The only use of I_DIRTY_TIME_EXPIRE is to detect in
__writeback_single_inode() that inode got there because flush worker
decided it's time to writeback the dirty inode time stamps (either
because we are syncing or because of age). However we can detect this
directly in __writeback_single_inode() and there's no need for the
strange propagation with I_DIRTY_TIME_EXPIRE flag.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
---
 fs/ext4/inode.c                  |  2 +-
 fs/fs-writeback.c                | 28 +++++++++++-----------------
 fs/xfs/libxfs/xfs_trans_inode.c  |  4 ++--
 include/linux/fs.h               |  1 -
 include/trace/events/writeback.h |  1 -
 5 files changed, 14 insertions(+), 22 deletions(-)

(limited to 'include/trace')

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 40ec5c7ef0d3..4db497f02ffb 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -4887,7 +4887,7 @@ static void __ext4_update_other_inode_time(struct super_block *sb,
 	    (inode->i_state & I_DIRTY_TIME)) {
 		struct ext4_inode_info	*ei = EXT4_I(inode);
 
-		inode->i_state &= ~(I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED);
+		inode->i_state &= ~I_DIRTY_TIME;
 		spin_unlock(&inode->i_lock);
 
 		spin_lock(&ei->i_raw_lock);
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index ae17d64a3e18..149227160ff0 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -1238,7 +1238,7 @@ static bool inode_dirtied_after(struct inode *inode, unsigned long t)
  */
 static int move_expired_inodes(struct list_head *delaying_queue,
 			       struct list_head *dispatch_queue,
-			       int flags, unsigned long dirtied_before)
+			       unsigned long dirtied_before)
 {
 	LIST_HEAD(tmp);
 	struct list_head *pos, *node;
@@ -1254,8 +1254,6 @@ static int move_expired_inodes(struct list_head *delaying_queue,
 		list_move(&inode->i_io_list, &tmp);
 		moved++;
 		spin_lock(&inode->i_lock);
-		if (flags & EXPIRE_DIRTY_ATIME)
-			inode->i_state |= I_DIRTY_TIME_EXPIRED;
 		inode->i_state |= I_SYNC_QUEUED;
 		spin_unlock(&inode->i_lock);
 		if (sb_is_blkdev_sb(inode->i_sb))
@@ -1303,11 +1301,11 @@ static void queue_io(struct bdi_writeback *wb, struct wb_writeback_work *work,
 
 	assert_spin_locked(&wb->list_lock);
 	list_splice_init(&wb->b_more_io, &wb->b_io);
-	moved = move_expired_inodes(&wb->b_dirty, &wb->b_io, 0, dirtied_before);
+	moved = move_expired_inodes(&wb->b_dirty, &wb->b_io, dirtied_before);
 	if (!work->for_sync)
 		time_expire_jif = jiffies - dirtytime_expire_interval * HZ;
 	moved += move_expired_inodes(&wb->b_dirty_time, &wb->b_io,
-				     EXPIRE_DIRTY_ATIME, time_expire_jif);
+				     time_expire_jif);
 	if (moved)
 		wb_io_lists_populated(wb);
 	trace_writeback_queue_io(wb, work, dirtied_before, moved);
@@ -1483,18 +1481,14 @@ __writeback_single_inode(struct inode *inode, struct writeback_control *wbc)
 	spin_lock(&inode->i_lock);
 
 	dirty = inode->i_state & I_DIRTY;
-	if (inode->i_state & I_DIRTY_TIME) {
-		if ((dirty & I_DIRTY_INODE) ||
-		    wbc->sync_mode == WB_SYNC_ALL ||
-		    unlikely(inode->i_state & I_DIRTY_TIME_EXPIRED) ||
-		    unlikely(time_after(jiffies,
-					(inode->dirtied_time_when +
-					 dirtytime_expire_interval * HZ)))) {
-			dirty |= I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED;
-			trace_writeback_lazytime(inode);
-		}
-	} else
-		inode->i_state &= ~I_DIRTY_TIME_EXPIRED;
+	if ((inode->i_state & I_DIRTY_TIME) &&
+	    ((dirty & I_DIRTY_INODE) ||
+	     wbc->sync_mode == WB_SYNC_ALL || wbc->for_sync ||
+	     time_after(jiffies, inode->dirtied_time_when +
+			dirtytime_expire_interval * HZ))) {
+		dirty |= I_DIRTY_TIME;
+		trace_writeback_lazytime(inode);
+	}
 	inode->i_state &= ~dirty;
 
 	/*
diff --git a/fs/xfs/libxfs/xfs_trans_inode.c b/fs/xfs/libxfs/xfs_trans_inode.c
index b5dfb6654842..1b4df6636944 100644
--- a/fs/xfs/libxfs/xfs_trans_inode.c
+++ b/fs/xfs/libxfs/xfs_trans_inode.c
@@ -96,9 +96,9 @@ xfs_trans_log_inode(
 	 * to log the timestamps, or will clear already cleared fields in the
 	 * worst case.
 	 */
-	if (inode->i_state & (I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED)) {
+	if (inode->i_state & I_DIRTY_TIME) {
 		spin_lock(&inode->i_lock);
-		inode->i_state &= ~(I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED);
+		inode->i_state &= ~I_DIRTY_TIME;
 		spin_unlock(&inode->i_lock);
 	}
 
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 48556efcdcf0..45eadf5bea5d 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2178,7 +2178,6 @@ static inline void kiocb_clone(struct kiocb *kiocb, struct kiocb *kiocb_src,
 #define I_DIO_WAKEUP		(1 << __I_DIO_WAKEUP)
 #define I_LINKABLE		(1 << 10)
 #define I_DIRTY_TIME		(1 << 11)
-#define I_DIRTY_TIME_EXPIRED	(1 << 12)
 #define I_WB_SWITCH		(1 << 13)
 #define I_OVL_INUSE		(1 << 14)
 #define I_CREATING		(1 << 15)
diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h
index 7565dcd59697..e7cbccc7c14c 100644
--- a/include/trace/events/writeback.h
+++ b/include/trace/events/writeback.h
@@ -20,7 +20,6 @@
 		{I_CLEAR,		"I_CLEAR"},		\
 		{I_SYNC,		"I_SYNC"},		\
 		{I_DIRTY_TIME,		"I_DIRTY_TIME"},	\
-		{I_DIRTY_TIME_EXPIRED,	"I_DIRTY_TIME_EXPIRED"}, \
 		{I_REFERENCED,		"I_REFERENCED"}		\
 	)
 
-- 
cgit v1.2.3


From ab74c7b23f3770935016e3eb3ecdf1e42b73efaa Mon Sep 17 00:00:00 2001
From: Theodore Ts'o <tytso@mit.edu>
Date: Wed, 15 Jul 2020 11:48:55 -0400
Subject: ext4: indicate via a block bitmap read is prefetched via a tracepoint

Modify the ext4_read_block_bitmap_load tracepoint so that it tells us
whether a block bitmap is being prefetched.

Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Artem Blagodarenko <artem.blagodarenko@gmail.com>
---
 fs/ext4/balloc.c            |  2 +-
 include/trace/events/ext4.h | 24 ++++++++++++++++++++----
 2 files changed, 21 insertions(+), 5 deletions(-)

(limited to 'include/trace')

diff --git a/fs/ext4/balloc.c b/fs/ext4/balloc.c
index 1e2b1b4093aa..48c3df47748d 100644
--- a/fs/ext4/balloc.c
+++ b/fs/ext4/balloc.c
@@ -494,7 +494,7 @@ ext4_read_block_bitmap_nowait(struct super_block *sb, ext4_group_t block_group,
 	 * submit the buffer_head for reading
 	 */
 	set_buffer_new(bh);
-	trace_ext4_read_block_bitmap_load(sb, block_group);
+	trace_ext4_read_block_bitmap_load(sb, block_group, ignore_locked);
 	bh->b_end_io = ext4_end_bitmap_read;
 	get_bh(bh);
 	submit_bh(REQ_OP_READ, REQ_META | REQ_PRIO |
diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h
index cc41d692ae8e..cbcd2e1a608d 100644
--- a/include/trace/events/ext4.h
+++ b/include/trace/events/ext4.h
@@ -1312,18 +1312,34 @@ DEFINE_EVENT(ext4__bitmap_load, ext4_mb_buddy_bitmap_load,
 	TP_ARGS(sb, group)
 );
 
-DEFINE_EVENT(ext4__bitmap_load, ext4_read_block_bitmap_load,
+DEFINE_EVENT(ext4__bitmap_load, ext4_load_inode_bitmap,
 
 	TP_PROTO(struct super_block *sb, unsigned long group),
 
 	TP_ARGS(sb, group)
 );
 
-DEFINE_EVENT(ext4__bitmap_load, ext4_load_inode_bitmap,
+TRACE_EVENT(ext4_read_block_bitmap_load,
+	TP_PROTO(struct super_block *sb, unsigned long group, bool prefetch),
 
-	TP_PROTO(struct super_block *sb, unsigned long group),
+	TP_ARGS(sb, group, prefetch),
 
-	TP_ARGS(sb, group)
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	__u32,	group			)
+		__field(	bool,	prefetch		)
+
+	),
+
+	TP_fast_assign(
+		__entry->dev	= sb->s_dev;
+		__entry->group	= group;
+		__entry->prefetch = prefetch;
+	),
+
+	TP_printk("dev %d,%d group %u prefetch %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->group, __entry->prefetch)
 );
 
 TRACE_EVENT(ext4_direct_IO_enter,
-- 
cgit v1.2.3


From 3d392b2676bf3199863a1e5efb2c087ad9d442a4 Mon Sep 17 00:00:00 2001
From: Theodore Ts'o <tytso@mit.edu>
Date: Fri, 17 Jul 2020 00:14:40 -0400
Subject: ext4: add prefetch_block_bitmaps mount option

For file systems where we can afford to keep the buddy bitmaps cached,
we can speed up initial writes to large file systems by starting to
load the block allocation bitmaps as soon as the file system is
mounted.  This won't work well for _super_ large file systems, or
memory constrained systems, so we only enable this when it is
requested via a mount option.

Addresses-Google-Bug: 159488342
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Andreas Dilger <adilger@dilger.ca>
---
 fs/ext4/ext4.h              | 15 +++++++++++-
 fs/ext4/mballoc.c           | 10 ++++----
 fs/ext4/super.c             | 57 ++++++++++++++++++++++++++++++++-------------
 include/trace/events/ext4.h | 44 ++++++++++++++++++++++++++++++++++
 4 files changed, 103 insertions(+), 23 deletions(-)

(limited to 'include/trace')

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 26ae31a994a2..e5f0c66a6156 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1188,6 +1188,7 @@ struct ext4_inode_info {
 #define EXT4_MOUNT_JOURNAL_CHECKSUM	0x800000 /* Journal checksums */
 #define EXT4_MOUNT_JOURNAL_ASYNC_COMMIT	0x1000000 /* Journal Async Commit */
 #define EXT4_MOUNT_WARN_ON_ERROR	0x2000000 /* Trigger WARN_ON on error */
+#define EXT4_MOUNT_PREFETCH_BLOCK_BITMAPS 0x4000000
 #define EXT4_MOUNT_DELALLOC		0x8000000 /* Delalloc support */
 #define EXT4_MOUNT_DATA_ERR_ABORT	0x10000000 /* Abort on file data write */
 #define EXT4_MOUNT_BLOCK_VALIDITY	0x20000000 /* Block validity checking */
@@ -2334,9 +2335,15 @@ struct ext4_lazy_init {
 	struct mutex		li_list_mtx;
 };
 
+enum ext4_li_mode {
+	EXT4_LI_MODE_PREFETCH_BBITMAP,
+	EXT4_LI_MODE_ITABLE,
+};
+
 struct ext4_li_request {
 	struct super_block	*lr_super;
-	struct ext4_sb_info	*lr_sbi;
+	enum ext4_li_mode	lr_mode;
+	ext4_group_t		lr_first_not_zeroed;
 	ext4_group_t		lr_next_group;
 	struct list_head	lr_request;
 	unsigned long		lr_next_sched;
@@ -2676,6 +2683,12 @@ extern int ext4_mb_reserve_blocks(struct super_block *, int);
 extern void ext4_discard_preallocations(struct inode *);
 extern int __init ext4_init_mballoc(void);
 extern void ext4_exit_mballoc(void);
+extern ext4_group_t ext4_mb_prefetch(struct super_block *sb,
+				     ext4_group_t group,
+				     unsigned int nr, int *cnt);
+extern void ext4_mb_prefetch_fini(struct super_block *sb, ext4_group_t group,
+				  unsigned int nr);
+
 extern void ext4_free_blocks(handle_t *handle, struct inode *inode,
 			     struct buffer_head *bh, ext4_fsblk_t block,
 			     unsigned long count, int flags);
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 9a07da53ab7b..8ecd49dd3906 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -2232,9 +2232,8 @@ out:
  * Start prefetching @nr block bitmaps starting at @group.
  * Return the next group which needs to be prefetched.
  */
-static ext4_group_t
-ext4_mb_prefetch(struct super_block *sb, ext4_group_t group,
-		 unsigned int nr, int *cnt)
+ext4_group_t ext4_mb_prefetch(struct super_block *sb, ext4_group_t group,
+			      unsigned int nr, int *cnt)
 {
 	ext4_group_t ngroups = ext4_get_groups_count(sb);
 	struct buffer_head *bh;
@@ -2284,9 +2283,8 @@ ext4_mb_prefetch(struct super_block *sb, ext4_group_t group,
  * waiting for the block allocation bitmap read to finish when
  * ext4_mb_prefetch_fini is called from ext4_mb_regular_allocator().
  */
-static void
-ext4_mb_prefetch_fini(struct super_block *sb, ext4_group_t group,
-		      unsigned int nr)
+void ext4_mb_prefetch_fini(struct super_block *sb, ext4_group_t group,
+			   unsigned int nr)
 {
 	while (nr-- > 0) {
 		struct ext4_group_desc *gdp = ext4_get_group_desc(sb, group,
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index 8c00f0a09f4d..1cbe3f248faf 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1521,6 +1521,7 @@ enum {
 	Opt_dioread_nolock, Opt_dioread_lock,
 	Opt_discard, Opt_nodiscard, Opt_init_itable, Opt_noinit_itable,
 	Opt_max_dir_size_kb, Opt_nojournal_checksum, Opt_nombcache,
+	Opt_prefetch_block_bitmaps,
 };
 
 static const match_table_t tokens = {
@@ -1612,6 +1613,7 @@ static const match_table_t tokens = {
 	{Opt_test_dummy_encryption, "test_dummy_encryption"},
 	{Opt_nombcache, "nombcache"},
 	{Opt_nombcache, "no_mbcache"},	/* for backward compatibility */
+	{Opt_prefetch_block_bitmaps, "prefetch_block_bitmaps"},
 	{Opt_removed, "check=none"},	/* mount option from ext2/3 */
 	{Opt_removed, "nocheck"},	/* mount option from ext2/3 */
 	{Opt_removed, "reservation"},	/* mount option from ext2/3 */
@@ -1829,6 +1831,8 @@ static const struct mount_opts {
 	{Opt_max_dir_size_kb, 0, MOPT_GTE0},
 	{Opt_test_dummy_encryption, 0, MOPT_STRING},
 	{Opt_nombcache, EXT4_MOUNT_NO_MBCACHE, MOPT_SET},
+	{Opt_prefetch_block_bitmaps, EXT4_MOUNT_PREFETCH_BLOCK_BITMAPS,
+	 MOPT_SET},
 	{Opt_err, 0, 0}
 };
 
@@ -3201,15 +3205,34 @@ static void print_daily_error_info(struct timer_list *t)
 static int ext4_run_li_request(struct ext4_li_request *elr)
 {
 	struct ext4_group_desc *gdp = NULL;
-	ext4_group_t group, ngroups;
-	struct super_block *sb;
+	struct super_block *sb = elr->lr_super;
+	ext4_group_t ngroups = EXT4_SB(sb)->s_groups_count;
+	ext4_group_t group = elr->lr_next_group;
 	unsigned long timeout = 0;
+	unsigned int prefetch_ios = 0;
 	int ret = 0;
 
-	sb = elr->lr_super;
-	ngroups = EXT4_SB(sb)->s_groups_count;
+	if (elr->lr_mode == EXT4_LI_MODE_PREFETCH_BBITMAP) {
+		elr->lr_next_group = ext4_mb_prefetch(sb, group,
+				EXT4_SB(sb)->s_mb_prefetch, &prefetch_ios);
+		if (prefetch_ios)
+			ext4_mb_prefetch_fini(sb, elr->lr_next_group,
+					      prefetch_ios);
+		trace_ext4_prefetch_bitmaps(sb, group, elr->lr_next_group,
+					    prefetch_ios);
+		if (group >= elr->lr_next_group) {
+			ret = 1;
+			if (elr->lr_first_not_zeroed != ngroups &&
+			    !sb_rdonly(sb) && test_opt(sb, INIT_INODE_TABLE)) {
+				elr->lr_next_group = elr->lr_first_not_zeroed;
+				elr->lr_mode = EXT4_LI_MODE_ITABLE;
+				ret = 0;
+			}
+		}
+		return ret;
+	}
 
-	for (group = elr->lr_next_group; group < ngroups; group++) {
+	for (; group < ngroups; group++) {
 		gdp = ext4_get_group_desc(sb, group, NULL);
 		if (!gdp) {
 			ret = 1;
@@ -3227,9 +3250,10 @@ static int ext4_run_li_request(struct ext4_li_request *elr)
 		timeout = jiffies;
 		ret = ext4_init_inode_table(sb, group,
 					    elr->lr_timeout ? 0 : 1);
+		trace_ext4_lazy_itable_init(sb, group);
 		if (elr->lr_timeout == 0) {
 			timeout = (jiffies - timeout) *
-				  elr->lr_sbi->s_li_wait_mult;
+				EXT4_SB(elr->lr_super)->s_li_wait_mult;
 			elr->lr_timeout = timeout;
 		}
 		elr->lr_next_sched = jiffies + elr->lr_timeout;
@@ -3244,15 +3268,11 @@ static int ext4_run_li_request(struct ext4_li_request *elr)
  */
 static void ext4_remove_li_request(struct ext4_li_request *elr)
 {
-	struct ext4_sb_info *sbi;
-
 	if (!elr)
 		return;
 
-	sbi = elr->lr_sbi;
-
 	list_del(&elr->lr_request);
-	sbi->s_li_request = NULL;
+	EXT4_SB(elr->lr_super)->s_li_request = NULL;
 	kfree(elr);
 }
 
@@ -3461,7 +3481,6 @@ static int ext4_li_info_new(void)
 static struct ext4_li_request *ext4_li_request_new(struct super_block *sb,
 					    ext4_group_t start)
 {
-	struct ext4_sb_info *sbi = EXT4_SB(sb);
 	struct ext4_li_request *elr;
 
 	elr = kzalloc(sizeof(*elr), GFP_KERNEL);
@@ -3469,8 +3488,13 @@ static struct ext4_li_request *ext4_li_request_new(struct super_block *sb,
 		return NULL;
 
 	elr->lr_super = sb;
-	elr->lr_sbi = sbi;
-	elr->lr_next_group = start;
+	elr->lr_first_not_zeroed = start;
+	if (test_opt(sb, PREFETCH_BLOCK_BITMAPS))
+		elr->lr_mode = EXT4_LI_MODE_PREFETCH_BBITMAP;
+	else {
+		elr->lr_mode = EXT4_LI_MODE_ITABLE;
+		elr->lr_next_group = start;
+	}
 
 	/*
 	 * Randomize first schedule time of the request to
@@ -3500,8 +3524,9 @@ int ext4_register_li_request(struct super_block *sb,
 		goto out;
 	}
 
-	if (first_not_zeroed == ngroups || sb_rdonly(sb) ||
-	    !test_opt(sb, INIT_INODE_TABLE))
+	if (!test_opt(sb, PREFETCH_BLOCK_BITMAPS) &&
+	    (first_not_zeroed == ngroups || sb_rdonly(sb) ||
+	     !test_opt(sb, INIT_INODE_TABLE)))
 		goto out;
 
 	elr = ext4_li_request_new(sb, first_not_zeroed);
diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h
index cbcd2e1a608d..8008d2e116b9 100644
--- a/include/trace/events/ext4.h
+++ b/include/trace/events/ext4.h
@@ -2742,6 +2742,50 @@ TRACE_EVENT(ext4_error,
 		  __entry->function, __entry->line)
 );
 
+TRACE_EVENT(ext4_prefetch_bitmaps,
+	    TP_PROTO(struct super_block *sb, ext4_group_t group,
+		     ext4_group_t next, unsigned int prefetch_ios),
+
+	TP_ARGS(sb, group, next, prefetch_ios),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	__u32,	group			)
+		__field(	__u32,	next			)
+		__field(	__u32,	ios			)
+	),
+
+	TP_fast_assign(
+		__entry->dev	= sb->s_dev;
+		__entry->group	= group;
+		__entry->next	= next;
+		__entry->ios	= prefetch_ios;
+	),
+
+	TP_printk("dev %d,%d group %u next %u ios %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->group, __entry->next, __entry->ios)
+);
+
+TRACE_EVENT(ext4_lazy_itable_init,
+	    TP_PROTO(struct super_block *sb, ext4_group_t group),
+
+	TP_ARGS(sb, group),
+
+	TP_STRUCT__entry(
+		__field(	dev_t,	dev			)
+		__field(	__u32,	group			)
+	),
+
+	TP_fast_assign(
+		__entry->dev	= sb->s_dev;
+		__entry->group	= group;
+	),
+
+	TP_printk("dev %d,%d group %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev), __entry->group)
+);
+
 #endif /* _TRACE_EXT4_H */
 
 /* This part must be outside protection */
-- 
cgit v1.2.3


From 27bc446e2def38db3244a6eb4bb1d6312936610a Mon Sep 17 00:00:00 2001
From: brookxu <brookxu.cn@gmail.com>
Date: Mon, 17 Aug 2020 15:36:15 +0800
Subject: ext4: limit the length of per-inode prealloc list
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

In the scenario of writing sparse files, the per-inode prealloc list may
be very long, resulting in high overhead for ext4_mb_use_preallocated().
To circumvent this problem, we limit the maximum length of per-inode
prealloc list to 512 and allow users to modify it.

After patching, we observed that the sys ratio of cpu has dropped, and
the system throughput has increased significantly. We created a process
to write the sparse file, and the running time of the process on the
fixed kernel was significantly reduced, as follows:

Running time on unfixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real    0m2.051s
user    0m0.008s
sys     0m2.026s

Running time on fixed kernel：
[root@TENCENT64 ~]# time taskset 0x01 ./sparse /data1/sparce.dat
real    0m0.471s
user    0m0.004s
sys     0m0.395s

Signed-off-by: Chunguang Xu <brookxu@tencent.com>
Link: https://lore.kernel.org/r/d7a98178-056b-6db5-6bce-4ead23f4a257@gmail.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
---
 Documentation/admin-guide/ext4.rst |  3 ++
 fs/ext4/ext4.h                     |  4 ++-
 fs/ext4/extents.c                  | 10 +++---
 fs/ext4/file.c                     |  2 +-
 fs/ext4/indirect.c                 |  2 +-
 fs/ext4/inode.c                    |  6 ++--
 fs/ext4/ioctl.c                    |  2 +-
 fs/ext4/mballoc.c                  | 74 +++++++++++++++++++++++++++++++++-----
 fs/ext4/mballoc.h                  |  4 +++
 fs/ext4/move_extent.c              |  4 +--
 fs/ext4/super.c                    |  3 +-
 fs/ext4/sysfs.c                    |  2 ++
 include/trace/events/ext4.h        | 17 +++++----
 13 files changed, 104 insertions(+), 29 deletions(-)

(limited to 'include/trace')

diff --git a/Documentation/admin-guide/ext4.rst b/Documentation/admin-guide/ext4.rst
index 7fc6a72920c9..59bcc4a92602 100644
--- a/Documentation/admin-guide/ext4.rst
+++ b/Documentation/admin-guide/ext4.rst
@@ -482,6 +482,9 @@ Files in /sys/fs/ext4/<devname>:
         multiple of this tuning parameter if the stripe size is not set in the
         ext4 superblock
 
+  mb_max_inode_prealloc
+        The maximum length of per-inode ext4_prealloc_space list.
+
   mb_max_to_scan
         The maximum number of extents the multiblock allocator will search to
         find the best extent.
diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index 71b4370a3f91..523e00d7b392 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1070,6 +1070,7 @@ struct ext4_inode_info {
 	struct timespec64 i_crtime;
 
 	/* mballoc */
+	atomic_t i_prealloc_active;
 	struct list_head i_prealloc_list;
 	spinlock_t i_prealloc_lock;
 
@@ -1518,6 +1519,7 @@ struct ext4_sb_info {
 	unsigned int s_mb_stats;
 	unsigned int s_mb_order2_reqs;
 	unsigned int s_mb_group_prealloc;
+	unsigned int s_mb_max_inode_prealloc;
 	unsigned int s_max_dir_size_kb;
 	/* where last allocation was done - for stream allocation */
 	unsigned long s_mb_last_group;
@@ -2682,7 +2684,7 @@ extern int ext4_mb_release(struct super_block *);
 extern ext4_fsblk_t ext4_mb_new_blocks(handle_t *,
 				struct ext4_allocation_request *, int *);
 extern int ext4_mb_reserve_blocks(struct super_block *, int);
-extern void ext4_discard_preallocations(struct inode *);
+extern void ext4_discard_preallocations(struct inode *, unsigned int);
 extern int __init ext4_init_mballoc(void);
 extern void ext4_exit_mballoc(void);
 extern ext4_group_t ext4_mb_prefetch(struct super_block *sb,
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 0eea09aa0f26..a0481582187a 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -100,7 +100,7 @@ static int ext4_ext_trunc_restart_fn(struct inode *inode, int *dropped)
 	 * i_mutex. So we can safely drop the i_data_sem here.
 	 */
 	BUG_ON(EXT4_JOURNAL(inode) == NULL);
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 	up_write(&EXT4_I(inode)->i_data_sem);
 	*dropped = 1;
 	return 0;
@@ -4266,7 +4266,7 @@ got_allocated_blocks:
 			 * not a good idea to call discard here directly,
 			 * but otherwise we'd need to call it every free().
 			 */
-			ext4_discard_preallocations(inode);
+			ext4_discard_preallocations(inode, 0);
 			if (flags & EXT4_GET_BLOCKS_DELALLOC_RESERVE)
 				fb_flags = EXT4_FREE_BLOCKS_NO_QUOT_UPDATE;
 			ext4_free_blocks(handle, inode, NULL, newblock,
@@ -5293,7 +5293,7 @@ static int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
 	}
 
 	down_write(&EXT4_I(inode)->i_data_sem);
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 
 	ret = ext4_es_remove_extent(inode, punch_start,
 				    EXT_MAX_BLOCKS - punch_start);
@@ -5307,7 +5307,7 @@ static int ext4_collapse_range(struct inode *inode, loff_t offset, loff_t len)
 		up_write(&EXT4_I(inode)->i_data_sem);
 		goto out_stop;
 	}
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 
 	ret = ext4_ext_shift_extents(inode, handle, punch_stop,
 				     punch_stop - punch_start, SHIFT_LEFT);
@@ -5439,7 +5439,7 @@ static int ext4_insert_range(struct inode *inode, loff_t offset, loff_t len)
 		goto out_stop;
 
 	down_write(&EXT4_I(inode)->i_data_sem);
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 
 	path = ext4_find_extent(inode, offset_lblk, NULL, 0);
 	if (IS_ERR(path)) {
diff --git a/fs/ext4/file.c b/fs/ext4/file.c
index 7a2720517bbb..e608ce3fb535 100644
--- a/fs/ext4/file.c
+++ b/fs/ext4/file.c
@@ -147,7 +147,7 @@ static int ext4_release_file(struct inode *inode, struct file *filp)
 			(atomic_read(&inode->i_writecount) == 1) &&
 			!EXT4_I(inode)->i_reserved_data_blocks) {
 		down_write(&EXT4_I(inode)->i_data_sem);
-		ext4_discard_preallocations(inode);
+		ext4_discard_preallocations(inode, 0);
 		up_write(&EXT4_I(inode)->i_data_sem);
 	}
 	if (is_dx(inode) && filp->private_data)
diff --git a/fs/ext4/indirect.c b/fs/ext4/indirect.c
index 433ca8415c5a..80c9f33800be 100644
--- a/fs/ext4/indirect.c
+++ b/fs/ext4/indirect.c
@@ -696,7 +696,7 @@ static int ext4_ind_trunc_restart_fn(handle_t *handle, struct inode *inode,
 	 * i_mutex. So we can safely drop the i_data_sem here.
 	 */
 	BUG_ON(EXT4_JOURNAL(inode) == NULL);
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 	up_write(&EXT4_I(inode)->i_data_sem);
 	*dropped = 1;
 	return 0;
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 0b07576af3bf..77543f988258 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -383,7 +383,7 @@ void ext4_da_update_reserve_space(struct inode *inode,
 	 */
 	if ((ei->i_reserved_data_blocks == 0) &&
 	    !inode_is_open_for_write(inode))
-		ext4_discard_preallocations(inode);
+		ext4_discard_preallocations(inode, 0);
 }
 
 static int __check_block_validity(struct inode *inode, const char *func,
@@ -4055,7 +4055,7 @@ int ext4_punch_hole(struct inode *inode, loff_t offset, loff_t length)
 	if (stop_block > first_block) {
 
 		down_write(&EXT4_I(inode)->i_data_sem);
-		ext4_discard_preallocations(inode);
+		ext4_discard_preallocations(inode, 0);
 
 		ret = ext4_es_remove_extent(inode, first_block,
 					    stop_block - first_block);
@@ -4210,7 +4210,7 @@ int ext4_truncate(struct inode *inode)
 
 	down_write(&EXT4_I(inode)->i_data_sem);
 
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 
 	if (ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS))
 		err = ext4_ext_truncate(handle, inode);
diff --git a/fs/ext4/ioctl.c b/fs/ext4/ioctl.c
index 6e70a63dcca7..36eca3bc036a 100644
--- a/fs/ext4/ioctl.c
+++ b/fs/ext4/ioctl.c
@@ -202,7 +202,7 @@ static long swap_inode_boot_loader(struct super_block *sb,
 	reset_inode_seed(inode);
 	reset_inode_seed(inode_bl);
 
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 
 	err = ext4_mark_inode_dirty(handle, inode);
 	if (err < 0) {
diff --git a/fs/ext4/mballoc.c b/fs/ext4/mballoc.c
index 45ac6088b4ac..132c118d12e1 100644
--- a/fs/ext4/mballoc.c
+++ b/fs/ext4/mballoc.c
@@ -2878,6 +2878,7 @@ int ext4_mb_init(struct super_block *sb)
 	sbi->s_mb_stats = MB_DEFAULT_STATS;
 	sbi->s_mb_stream_request = MB_DEFAULT_STREAM_THRESHOLD;
 	sbi->s_mb_order2_reqs = MB_DEFAULT_ORDER2_REQS;
+	sbi->s_mb_max_inode_prealloc = MB_DEFAULT_MAX_INODE_PREALLOC;
 	/*
 	 * The default group preallocation is 512, which for 4k block
 	 * sizes translates to 2 megabytes.  However for bigalloc file
@@ -3816,6 +3817,26 @@ void ext4_mb_generate_from_pa(struct super_block *sb, void *bitmap,
 	mb_debug(sb, "preallocated %d for group %u\n", preallocated, group);
 }
 
+static void ext4_mb_mark_pa_deleted(struct super_block *sb,
+				    struct ext4_prealloc_space *pa)
+{
+	struct ext4_inode_info *ei;
+
+	if (pa->pa_deleted) {
+		ext4_warning(sb, "deleted pa, type:%d, pblk:%llu, lblk:%u, len:%d\n",
+			     pa->pa_type, pa->pa_pstart, pa->pa_lstart,
+			     pa->pa_len);
+		return;
+	}
+
+	pa->pa_deleted = 1;
+
+	if (pa->pa_type == MB_INODE_PA) {
+		ei = EXT4_I(pa->pa_inode);
+		atomic_dec(&ei->i_prealloc_active);
+	}
+}
+
 static void ext4_mb_pa_callback(struct rcu_head *head)
 {
 	struct ext4_prealloc_space *pa;
@@ -3848,7 +3869,7 @@ static void ext4_mb_put_pa(struct ext4_allocation_context *ac,
 		return;
 	}
 
-	pa->pa_deleted = 1;
+	ext4_mb_mark_pa_deleted(sb, pa);
 	spin_unlock(&pa->pa_lock);
 
 	grp_blk = pa->pa_pstart;
@@ -3972,6 +3993,7 @@ ext4_mb_new_inode_pa(struct ext4_allocation_context *ac)
 	spin_lock(pa->pa_obj_lock);
 	list_add_rcu(&pa->pa_inode_list, &ei->i_prealloc_list);
 	spin_unlock(pa->pa_obj_lock);
+	atomic_inc(&ei->i_prealloc_active);
 }
 
 /*
@@ -4182,7 +4204,7 @@ repeat:
 		}
 
 		/* seems this one can be freed ... */
-		pa->pa_deleted = 1;
+		ext4_mb_mark_pa_deleted(sb, pa);
 
 		/* we can trust pa_free ... */
 		free += pa->pa_free;
@@ -4245,7 +4267,7 @@ out_dbg:
  *
  * FIXME!! Make sure it is valid at all the call sites
  */
-void ext4_discard_preallocations(struct inode *inode)
+void ext4_discard_preallocations(struct inode *inode, unsigned int needed)
 {
 	struct ext4_inode_info *ei = EXT4_I(inode);
 	struct super_block *sb = inode->i_sb;
@@ -4263,15 +4285,19 @@ void ext4_discard_preallocations(struct inode *inode)
 
 	mb_debug(sb, "discard preallocation for inode %lu\n",
 		 inode->i_ino);
-	trace_ext4_discard_preallocations(inode);
+	trace_ext4_discard_preallocations(inode,
+			atomic_read(&ei->i_prealloc_active), needed);
 
 	INIT_LIST_HEAD(&list);
 
+	if (needed == 0)
+		needed = UINT_MAX;
+
 repeat:
 	/* first, collect all pa's in the inode */
 	spin_lock(&ei->i_prealloc_lock);
-	while (!list_empty(&ei->i_prealloc_list)) {
-		pa = list_entry(ei->i_prealloc_list.next,
+	while (!list_empty(&ei->i_prealloc_list) && needed) {
+		pa = list_entry(ei->i_prealloc_list.prev,
 				struct ext4_prealloc_space, pa_inode_list);
 		BUG_ON(pa->pa_obj_lock != &ei->i_prealloc_lock);
 		spin_lock(&pa->pa_lock);
@@ -4288,10 +4314,11 @@ repeat:
 
 		}
 		if (pa->pa_deleted == 0) {
-			pa->pa_deleted = 1;
+			ext4_mb_mark_pa_deleted(sb, pa);
 			spin_unlock(&pa->pa_lock);
 			list_del_rcu(&pa->pa_inode_list);
 			list_add(&pa->u.pa_tmp_list, &list);
+			needed--;
 			continue;
 		}
 
@@ -4592,7 +4619,7 @@ ext4_mb_discard_lg_preallocations(struct super_block *sb,
 		BUG_ON(pa->pa_type != MB_GROUP_PA);
 
 		/* seems this one can be freed ... */
-		pa->pa_deleted = 1;
+		ext4_mb_mark_pa_deleted(sb, pa);
 		spin_unlock(&pa->pa_lock);
 
 		list_del_rcu(&pa->pa_inode_list);
@@ -4690,11 +4717,30 @@ static void ext4_mb_add_n_trim(struct ext4_allocation_context *ac)
 	return ;
 }
 
+/*
+ * if per-inode prealloc list is too long, trim some PA
+ */
+static void ext4_mb_trim_inode_pa(struct inode *inode)
+{
+	struct ext4_inode_info *ei = EXT4_I(inode);
+	struct ext4_sb_info *sbi = EXT4_SB(inode->i_sb);
+	int count, delta;
+
+	count = atomic_read(&ei->i_prealloc_active);
+	delta = (sbi->s_mb_max_inode_prealloc >> 2) + 1;
+	if (count > sbi->s_mb_max_inode_prealloc + delta) {
+		count -= sbi->s_mb_max_inode_prealloc;
+		ext4_discard_preallocations(inode, count);
+	}
+}
+
 /*
  * release all resource we used in allocation
  */
 static int ext4_mb_release_context(struct ext4_allocation_context *ac)
 {
+	struct inode *inode = ac->ac_inode;
+	struct ext4_inode_info *ei = EXT4_I(inode);
 	struct ext4_sb_info *sbi = EXT4_SB(ac->ac_sb);
 	struct ext4_prealloc_space *pa = ac->ac_pa;
 	if (pa) {
@@ -4720,6 +4766,17 @@ static int ext4_mb_release_context(struct ext4_allocation_context *ac)
 				ext4_mb_add_n_trim(ac);
 			}
 		}
+
+		if (pa->pa_type == MB_INODE_PA) {
+			/*
+			 * treat per-inode prealloc list as a lru list, then try
+			 * to trim the least recently used PA.
+			 */
+			spin_lock(pa->pa_obj_lock);
+			list_move(&pa->pa_inode_list, &ei->i_prealloc_list);
+			spin_unlock(pa->pa_obj_lock);
+		}
+
 		ext4_mb_put_pa(ac, ac->ac_sb, pa);
 	}
 	if (ac->ac_bitmap_page)
@@ -4729,6 +4786,7 @@ static int ext4_mb_release_context(struct ext4_allocation_context *ac)
 	if (ac->ac_flags & EXT4_MB_HINT_GROUP_ALLOC)
 		mutex_unlock(&ac->ac_lg->lg_mutex);
 	ext4_mb_collect_stats(ac);
+	ext4_mb_trim_inode_pa(inode);
 	return 0;
 }
 
diff --git a/fs/ext4/mballoc.h b/fs/ext4/mballoc.h
index 6b4d17c2935d..e75b4749aa1c 100644
--- a/fs/ext4/mballoc.h
+++ b/fs/ext4/mballoc.h
@@ -73,6 +73,10 @@
  */
 #define MB_DEFAULT_GROUP_PREALLOC	512
 
+/*
+ * maximum length of inode prealloc list
+ */
+#define MB_DEFAULT_MAX_INODE_PREALLOC	512
 
 struct ext4_free_data {
 	/* this links the free block information from sb_info */
diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
index 1ed86fb6c302..0d601b822875 100644
--- a/fs/ext4/move_extent.c
+++ b/fs/ext4/move_extent.c
@@ -686,8 +686,8 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp, __u64 orig_blk,
 
 out:
 	if (*moved_len) {
-		ext4_discard_preallocations(orig_inode);
-		ext4_discard_preallocations(donor_inode);
+		ext4_discard_preallocations(orig_inode, 0);
+		ext4_discard_preallocations(donor_inode, 0);
 	}
 
 	ext4_ext_drop_refs(path);
diff --git a/fs/ext4/super.c b/fs/ext4/super.c
index daa94c7f7271..13bdddc081e0 100644
--- a/fs/ext4/super.c
+++ b/fs/ext4/super.c
@@ -1127,6 +1127,7 @@ static struct inode *ext4_alloc_inode(struct super_block *sb)
 	inode_set_iversion(&ei->vfs_inode, 1);
 	spin_lock_init(&ei->i_raw_lock);
 	INIT_LIST_HEAD(&ei->i_prealloc_list);
+	atomic_set(&ei->i_prealloc_active, 0);
 	spin_lock_init(&ei->i_prealloc_lock);
 	ext4_es_init_tree(&ei->i_es_tree);
 	rwlock_init(&ei->i_es_lock);
@@ -1220,7 +1221,7 @@ void ext4_clear_inode(struct inode *inode)
 {
 	invalidate_inode_buffers(inode);
 	clear_inode(inode);
-	ext4_discard_preallocations(inode);
+	ext4_discard_preallocations(inode, 0);
 	ext4_es_remove_extent(inode, 0, EXT_MAX_BLOCKS);
 	dquot_drop(inode);
 	if (EXT4_I(inode)->jinode) {
diff --git a/fs/ext4/sysfs.c b/fs/ext4/sysfs.c
index 7fee11cc30e7..bfabb799fa45 100644
--- a/fs/ext4/sysfs.c
+++ b/fs/ext4/sysfs.c
@@ -218,6 +218,7 @@ EXT4_RW_ATTR_SBI_UI(mb_min_to_scan, s_mb_min_to_scan);
 EXT4_RW_ATTR_SBI_UI(mb_order2_req, s_mb_order2_reqs);
 EXT4_RW_ATTR_SBI_UI(mb_stream_req, s_mb_stream_request);
 EXT4_RW_ATTR_SBI_UI(mb_group_prealloc, s_mb_group_prealloc);
+EXT4_RW_ATTR_SBI_UI(mb_max_inode_prealloc, s_mb_max_inode_prealloc);
 EXT4_RW_ATTR_SBI_UI(extent_max_zeroout_kb, s_extent_max_zeroout_kb);
 EXT4_ATTR(trigger_fs_error, 0200, trigger_test_error);
 EXT4_RW_ATTR_SBI_UI(err_ratelimit_interval_ms, s_err_ratelimit_state.interval);
@@ -264,6 +265,7 @@ static struct attribute *ext4_attrs[] = {
 	ATTR_LIST(mb_order2_req),
 	ATTR_LIST(mb_stream_req),
 	ATTR_LIST(mb_group_prealloc),
+	ATTR_LIST(mb_max_inode_prealloc),
 	ATTR_LIST(max_writeback_mb_bump),
 	ATTR_LIST(extent_max_zeroout_kb),
 	ATTR_LIST(trigger_fs_error),
diff --git a/include/trace/events/ext4.h b/include/trace/events/ext4.h
index 8008d2e116b9..4c8b99ec8606 100644
--- a/include/trace/events/ext4.h
+++ b/include/trace/events/ext4.h
@@ -746,24 +746,29 @@ TRACE_EVENT(ext4_mb_release_group_pa,
 );
 
 TRACE_EVENT(ext4_discard_preallocations,
-	TP_PROTO(struct inode *inode),
+	TP_PROTO(struct inode *inode, unsigned int len, unsigned int needed),
 
-	TP_ARGS(inode),
+	TP_ARGS(inode, len, needed),
 
 	TP_STRUCT__entry(
-		__field(	dev_t,	dev			)
-		__field(	ino_t,	ino			)
+		__field(	dev_t,		dev		)
+		__field(	ino_t,		ino		)
+		__field(	unsigned int,	len		)
+		__field(	unsigned int,	needed		)
 
 	),
 
 	TP_fast_assign(
 		__entry->dev	= inode->i_sb->s_dev;
 		__entry->ino	= inode->i_ino;
+		__entry->len	= len;
+		__entry->needed	= needed;
 	),
 
-	TP_printk("dev %d,%d ino %lu",
+	TP_printk("dev %d,%d ino %lu len: %u needed %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  (unsigned long) __entry->ino)
+		  (unsigned long) __entry->ino, __entry->len,
+		  __entry->needed)
 );
 
 TRACE_EVENT(ext4_mb_discard_preallocations,
-- 
cgit v1.2.3


From 12564485ed8caac3c18572793ec01330792c7191 Mon Sep 17 00:00:00 2001
From: Shawn Anastasio <shawn@anastas.io>
Date: Fri, 21 Aug 2020 13:55:56 -0500
Subject: Revert "powerpc/64s: Remove PROT_SAO support"

This reverts commit 5c9fa16e8abd342ce04dc830c1ebb2a03abf6c05.

Since PROT_SAO can still be useful for certain classes of software,
reintroduce it. Concerns about guest migration for LPARs using SAO
will be addressed next.

Signed-off-by: Shawn Anastasio <shawn@anastas.io>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200821185558.35561-2-shawn@anastas.io
---
 arch/powerpc/include/asm/book3s/64/pgtable.h  |  8 ++---
 arch/powerpc/include/asm/cputable.h           | 10 +++----
 arch/powerpc/include/asm/mman.h               | 26 ++++++++++++++---
 arch/powerpc/include/asm/nohash/64/pgtable.h  |  2 ++
 arch/powerpc/include/uapi/asm/mman.h          |  2 +-
 arch/powerpc/kernel/dt_cpu_ftrs.c             |  2 +-
 arch/powerpc/mm/book3s64/hash_utils.c         |  2 ++
 include/linux/mm.h                            |  2 ++
 include/trace/events/mmflags.h                |  2 ++
 mm/ksm.c                                      |  4 +++
 tools/testing/selftests/powerpc/mm/.gitignore |  1 +
 tools/testing/selftests/powerpc/mm/Makefile   |  4 ++-
 tools/testing/selftests/powerpc/mm/prot_sao.c | 42 +++++++++++++++++++++++++++
 13 files changed, 90 insertions(+), 17 deletions(-)
 create mode 100644 tools/testing/selftests/powerpc/mm/prot_sao.c

(limited to 'include/trace')

diff --git a/arch/powerpc/include/asm/book3s/64/pgtable.h b/arch/powerpc/include/asm/book3s/64/pgtable.h
index 6de56c3b33c4..495fc0ccb453 100644
--- a/arch/powerpc/include/asm/book3s/64/pgtable.h
+++ b/arch/powerpc/include/asm/book3s/64/pgtable.h
@@ -20,13 +20,9 @@
 #define _PAGE_RW		(_PAGE_READ | _PAGE_WRITE)
 #define _PAGE_RWX		(_PAGE_READ | _PAGE_WRITE | _PAGE_EXEC)
 #define _PAGE_PRIVILEGED	0x00008 /* kernel access only */
-
-#define _PAGE_CACHE_CTL		0x00030 /* Bits for the folowing cache modes */
-			/*	No bits set is normal cacheable memory */
-			/*	0x00010 unused, is SAO bit on radix POWER9 */
+#define _PAGE_SAO		0x00010 /* Strong access order */
 #define _PAGE_NON_IDEMPOTENT	0x00020 /* non idempotent memory */
 #define _PAGE_TOLERANT		0x00030 /* tolerant memory, cache inhibited */
-
 #define _PAGE_DIRTY		0x00080 /* C: page changed */
 #define _PAGE_ACCESSED		0x00100 /* R: page referenced */
 /*
@@ -828,6 +824,8 @@ static inline void __set_pte_at(struct mm_struct *mm, unsigned long addr,
 	return hash__set_pte_at(mm, addr, ptep, pte, percpu);
 }
 
+#define _PAGE_CACHE_CTL	(_PAGE_SAO | _PAGE_NON_IDEMPOTENT | _PAGE_TOLERANT)
+
 #define pgprot_noncached pgprot_noncached
 static inline pgprot_t pgprot_noncached(pgprot_t prot)
 {
diff --git a/arch/powerpc/include/asm/cputable.h b/arch/powerpc/include/asm/cputable.h
index e005b4581023..32a15dc49e8c 100644
--- a/arch/powerpc/include/asm/cputable.h
+++ b/arch/powerpc/include/asm/cputable.h
@@ -196,7 +196,7 @@ static inline void cpu_feature_keys_init(void) { }
 #define CPU_FTR_SPURR			LONG_ASM_CONST(0x0000000001000000)
 #define CPU_FTR_DSCR			LONG_ASM_CONST(0x0000000002000000)
 #define CPU_FTR_VSX			LONG_ASM_CONST(0x0000000004000000)
-// Free					LONG_ASM_CONST(0x0000000008000000)
+#define CPU_FTR_SAO			LONG_ASM_CONST(0x0000000008000000)
 #define CPU_FTR_CP_USE_DCBTZ		LONG_ASM_CONST(0x0000000010000000)
 #define CPU_FTR_UNALIGNED_LD_STD	LONG_ASM_CONST(0x0000000020000000)
 #define CPU_FTR_ASYM_SMT		LONG_ASM_CONST(0x0000000040000000)
@@ -441,7 +441,7 @@ static inline void cpu_feature_keys_init(void) { }
 	    CPU_FTR_MMCRA | CPU_FTR_SMT | \
 	    CPU_FTR_COHERENT_ICACHE | \
 	    CPU_FTR_PURR | CPU_FTR_SPURR | CPU_FTR_REAL_LE | \
-	    CPU_FTR_DSCR | CPU_FTR_ASYM_SMT | \
+	    CPU_FTR_DSCR | CPU_FTR_SAO  | CPU_FTR_ASYM_SMT | \
 	    CPU_FTR_STCX_CHECKS_ADDRESS | CPU_FTR_POPCNTB | CPU_FTR_POPCNTD | \
 	    CPU_FTR_CFAR | CPU_FTR_HVMODE | \
 	    CPU_FTR_VMX_COPY | CPU_FTR_HAS_PPR | CPU_FTR_DABRX )
@@ -450,7 +450,7 @@ static inline void cpu_feature_keys_init(void) { }
 	    CPU_FTR_MMCRA | CPU_FTR_SMT | \
 	    CPU_FTR_COHERENT_ICACHE | \
 	    CPU_FTR_PURR | CPU_FTR_SPURR | CPU_FTR_REAL_LE | \
-	    CPU_FTR_DSCR | \
+	    CPU_FTR_DSCR | CPU_FTR_SAO  | \
 	    CPU_FTR_STCX_CHECKS_ADDRESS | CPU_FTR_POPCNTB | CPU_FTR_POPCNTD | \
 	    CPU_FTR_CFAR | CPU_FTR_HVMODE | CPU_FTR_VMX_COPY | \
 	    CPU_FTR_DBELL | CPU_FTR_HAS_PPR | CPU_FTR_DAWR | \
@@ -461,7 +461,7 @@ static inline void cpu_feature_keys_init(void) { }
 	    CPU_FTR_MMCRA | CPU_FTR_SMT | \
 	    CPU_FTR_COHERENT_ICACHE | \
 	    CPU_FTR_PURR | CPU_FTR_SPURR | CPU_FTR_REAL_LE | \
-	    CPU_FTR_DSCR | \
+	    CPU_FTR_DSCR | CPU_FTR_SAO  | \
 	    CPU_FTR_STCX_CHECKS_ADDRESS | CPU_FTR_POPCNTB | CPU_FTR_POPCNTD | \
 	    CPU_FTR_CFAR | CPU_FTR_HVMODE | CPU_FTR_VMX_COPY | \
 	    CPU_FTR_DBELL | CPU_FTR_HAS_PPR | CPU_FTR_ARCH_207S | \
@@ -479,7 +479,7 @@ static inline void cpu_feature_keys_init(void) { }
 	    CPU_FTR_MMCRA | CPU_FTR_SMT | \
 	    CPU_FTR_COHERENT_ICACHE | \
 	    CPU_FTR_PURR | CPU_FTR_SPURR | CPU_FTR_REAL_LE | \
-	    CPU_FTR_DSCR | \
+	    CPU_FTR_DSCR | CPU_FTR_SAO  | \
 	    CPU_FTR_STCX_CHECKS_ADDRESS | CPU_FTR_POPCNTB | CPU_FTR_POPCNTD | \
 	    CPU_FTR_CFAR | CPU_FTR_HVMODE | CPU_FTR_VMX_COPY | \
 	    CPU_FTR_DBELL | CPU_FTR_HAS_PPR | CPU_FTR_ARCH_207S | \
diff --git a/arch/powerpc/include/asm/mman.h b/arch/powerpc/include/asm/mman.h
index 7c07728af300..4ba303ea27f5 100644
--- a/arch/powerpc/include/asm/mman.h
+++ b/arch/powerpc/include/asm/mman.h
@@ -13,20 +13,38 @@
 #include <linux/pkeys.h>
 #include <asm/cpu_has_feature.h>
 
-#ifdef CONFIG_PPC_MEM_KEYS
 static inline unsigned long arch_calc_vm_prot_bits(unsigned long prot,
 		unsigned long pkey)
 {
-	return pkey_to_vmflag_bits(pkey);
+#ifdef CONFIG_PPC_MEM_KEYS
+	return (((prot & PROT_SAO) ? VM_SAO : 0) | pkey_to_vmflag_bits(pkey));
+#else
+	return ((prot & PROT_SAO) ? VM_SAO : 0);
+#endif
 }
 #define arch_calc_vm_prot_bits(prot, pkey) arch_calc_vm_prot_bits(prot, pkey)
 
 static inline pgprot_t arch_vm_get_page_prot(unsigned long vm_flags)
 {
-	return __pgprot(vmflag_to_pte_pkey_bits(vm_flags));
+#ifdef CONFIG_PPC_MEM_KEYS
+	return (vm_flags & VM_SAO) ?
+		__pgprot(_PAGE_SAO | vmflag_to_pte_pkey_bits(vm_flags)) :
+		__pgprot(0 | vmflag_to_pte_pkey_bits(vm_flags));
+#else
+	return (vm_flags & VM_SAO) ? __pgprot(_PAGE_SAO) : __pgprot(0);
+#endif
 }
 #define arch_vm_get_page_prot(vm_flags) arch_vm_get_page_prot(vm_flags)
-#endif
+
+static inline bool arch_validate_prot(unsigned long prot, unsigned long addr)
+{
+	if (prot & ~(PROT_READ | PROT_WRITE | PROT_EXEC | PROT_SEM | PROT_SAO))
+		return false;
+	if ((prot & PROT_SAO) && !cpu_has_feature(CPU_FTR_SAO))
+		return false;
+	return true;
+}
+#define arch_validate_prot arch_validate_prot
 
 #endif /* CONFIG_PPC64 */
 #endif	/* _ASM_POWERPC_MMAN_H */
diff --git a/arch/powerpc/include/asm/nohash/64/pgtable.h b/arch/powerpc/include/asm/nohash/64/pgtable.h
index 59ee9fa4ae09..6cb8aa357191 100644
--- a/arch/powerpc/include/asm/nohash/64/pgtable.h
+++ b/arch/powerpc/include/asm/nohash/64/pgtable.h
@@ -82,6 +82,8 @@
  */
 #include <asm/nohash/pte-book3e.h>
 
+#define _PAGE_SAO	0
+
 #define PTE_RPN_MASK	(~((1UL << PTE_RPN_SHIFT) - 1))
 
 /*
diff --git a/arch/powerpc/include/uapi/asm/mman.h b/arch/powerpc/include/uapi/asm/mman.h
index 3a700351feca..c0c737215b00 100644
--- a/arch/powerpc/include/uapi/asm/mman.h
+++ b/arch/powerpc/include/uapi/asm/mman.h
@@ -11,7 +11,7 @@
 #include <asm-generic/mman-common.h>
 
 
-#define PROT_SAO	0x10		/* Unsupported since v5.9 */
+#define PROT_SAO	0x10		/* Strong Access Ordering */
 
 #define MAP_RENAME      MAP_ANONYMOUS   /* In SunOS terminology */
 #define MAP_NORESERVE   0x40            /* don't reserve swap pages */
diff --git a/arch/powerpc/kernel/dt_cpu_ftrs.c b/arch/powerpc/kernel/dt_cpu_ftrs.c
index 8dc46f38680b..f204ad79b6b5 100644
--- a/arch/powerpc/kernel/dt_cpu_ftrs.c
+++ b/arch/powerpc/kernel/dt_cpu_ftrs.c
@@ -653,7 +653,7 @@ static struct dt_cpu_feature_match __initdata
 	{"processor-control-facility-v3", feat_enable_dbell, CPU_FTR_DBELL},
 	{"processor-utilization-of-resources-register", feat_enable_purr, 0},
 	{"no-execute", feat_enable, 0},
-	/* strong-access-ordering is unused */
+	{"strong-access-ordering", feat_enable, CPU_FTR_SAO},
 	{"cache-inhibited-large-page", feat_enable_large_ci, 0},
 	{"coprocessor-icswx", feat_enable, 0},
 	{"hypervisor-virtualization-interrupt", feat_enable_hvi, 0},
diff --git a/arch/powerpc/mm/book3s64/hash_utils.c b/arch/powerpc/mm/book3s64/hash_utils.c
index 890a71c5293e..c663e7ba801f 100644
--- a/arch/powerpc/mm/book3s64/hash_utils.c
+++ b/arch/powerpc/mm/book3s64/hash_utils.c
@@ -232,6 +232,8 @@ unsigned long htab_convert_pte_flags(unsigned long pteflags)
 		rflags |= HPTE_R_I;
 	else if ((pteflags & _PAGE_CACHE_CTL) == _PAGE_NON_IDEMPOTENT)
 		rflags |= (HPTE_R_I | HPTE_R_G);
+	else if ((pteflags & _PAGE_CACHE_CTL) == _PAGE_SAO)
+		rflags |= (HPTE_R_W | HPTE_R_I | HPTE_R_M);
 	else
 		/*
 		 * Add memory coherence if cache inhibited is not set
diff --git a/include/linux/mm.h b/include/linux/mm.h
index 1983e08f5906..5abe6df4247e 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -321,6 +321,8 @@ extern unsigned int kobjsize(const void *objp);
 
 #if defined(CONFIG_X86)
 # define VM_PAT		VM_ARCH_1	/* PAT reserves whole VMA at once (x86) */
+#elif defined(CONFIG_PPC)
+# define VM_SAO		VM_ARCH_1	/* Strong Access Ordering (powerpc) */
 #elif defined(CONFIG_PARISC)
 # define VM_GROWSUP	VM_ARCH_1
 #elif defined(CONFIG_IA64)
diff --git a/include/trace/events/mmflags.h b/include/trace/events/mmflags.h
index 939092dbcb8b..5fb752034386 100644
--- a/include/trace/events/mmflags.h
+++ b/include/trace/events/mmflags.h
@@ -114,6 +114,8 @@ IF_HAVE_PG_IDLE(PG_idle,		"idle"		)
 
 #if defined(CONFIG_X86)
 #define __VM_ARCH_SPECIFIC_1 {VM_PAT,     "pat"           }
+#elif defined(CONFIG_PPC)
+#define __VM_ARCH_SPECIFIC_1 {VM_SAO,     "sao"           }
 #elif defined(CONFIG_PARISC) || defined(CONFIG_IA64)
 #define __VM_ARCH_SPECIFIC_1 {VM_GROWSUP,	"growsup"	}
 #elif !defined(CONFIG_MMU)
diff --git a/mm/ksm.c b/mm/ksm.c
index 0aa2247bddd7..90a625b02a1d 100644
--- a/mm/ksm.c
+++ b/mm/ksm.c
@@ -2453,6 +2453,10 @@ int ksm_madvise(struct vm_area_struct *vma, unsigned long start,
 		if (vma_is_dax(vma))
 			return 0;
 
+#ifdef VM_SAO
+		if (*vm_flags & VM_SAO)
+			return 0;
+#endif
 #ifdef VM_SPARC_ADI
 		if (*vm_flags & VM_SPARC_ADI)
 			return 0;
diff --git a/tools/testing/selftests/powerpc/mm/.gitignore b/tools/testing/selftests/powerpc/mm/.gitignore
index 91c775c23c66..aac4a59f9e28 100644
--- a/tools/testing/selftests/powerpc/mm/.gitignore
+++ b/tools/testing/selftests/powerpc/mm/.gitignore
@@ -2,6 +2,7 @@
 hugetlb_vs_thp_test
 subpage_prot
 tempfile
+prot_sao
 segv_errors
 wild_bctr
 large_vm_fork_separation
diff --git a/tools/testing/selftests/powerpc/mm/Makefile b/tools/testing/selftests/powerpc/mm/Makefile
index 250ce172e0da..defe488d6bf1 100644
--- a/tools/testing/selftests/powerpc/mm/Makefile
+++ b/tools/testing/selftests/powerpc/mm/Makefile
@@ -2,7 +2,7 @@
 noarg:
 	$(MAKE) -C ../
 
-TEST_GEN_PROGS := hugetlb_vs_thp_test subpage_prot segv_errors wild_bctr \
+TEST_GEN_PROGS := hugetlb_vs_thp_test subpage_prot prot_sao segv_errors wild_bctr \
 		  large_vm_fork_separation bad_accesses pkey_exec_prot \
 		  pkey_siginfo stack_expansion_signal stack_expansion_ldst
 
@@ -14,6 +14,8 @@ include ../../lib.mk
 
 $(TEST_GEN_PROGS): ../harness.c ../utils.c
 
+$(OUTPUT)/prot_sao: ../utils.c
+
 $(OUTPUT)/wild_bctr: CFLAGS += -m64
 $(OUTPUT)/large_vm_fork_separation: CFLAGS += -m64
 $(OUTPUT)/bad_accesses: CFLAGS += -m64
diff --git a/tools/testing/selftests/powerpc/mm/prot_sao.c b/tools/testing/selftests/powerpc/mm/prot_sao.c
new file mode 100644
index 000000000000..e2eed65b7735
--- /dev/null
+++ b/tools/testing/selftests/powerpc/mm/prot_sao.c
@@ -0,0 +1,42 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2016, Michael Ellerman, IBM Corp.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/mman.h>
+
+#include <asm/cputable.h>
+
+#include "utils.h"
+
+#define SIZE (64 * 1024)
+
+int test_prot_sao(void)
+{
+	char *p;
+
+	/* 2.06 or later should support SAO */
+	SKIP_IF(!have_hwcap(PPC_FEATURE_ARCH_2_06));
+
+	/*
+	 * Ensure we can ask for PROT_SAO.
+	 * We can't really verify that it does the right thing, but at least we
+	 * confirm the kernel will accept it.
+	 */
+	p = mmap(NULL, SIZE, PROT_READ | PROT_WRITE | PROT_SAO,
+		 MAP_ANONYMOUS | MAP_PRIVATE, -1, 0);
+	FAIL_IF(p == MAP_FAILED);
+
+	/* Write to the mapping, to at least cause a fault */
+	memset(p, 0xaa, SIZE);
+
+	return 0;
+}
+
+int main(void)
+{
+	return test_harness(test_prot_sao, "prot-sao");
+}
-- 
cgit v1.2.3