[PATCH v2] ext4: place buffer head allocation before handle start

[Jinke Han]

From: Jinke Han <hanjinke.666@xxxxxxxxxxxxx>

In our product environment, we encounter some jbd hung waiting handles to
stop while several writters were doing memory reclaim for buffer head
allocation in delay alloc write path. Ext4 do buffer head allocation with
holding transaction handle which may be blocked too long if the reclaim
works not so smooth. According to our bcc trace, the reclaim time in
buffer head allocation can reach 258s and the jbd transaction commit also
take almost the same time meanwhile. Except for these extreme cases,
we often see several seconds delays for cgroup memory reclaim on our
servers. This is more likely to happen considering docker environment.

One thing to note, the allocation of buffer heand is as often as page
allocation or more often when blocksize less than page size. Just like
page cache allocation, we should also place the buffer head allocation
before startting the handle.

After commit:cc883236b792, no nore need to do for delay alloc path, just
do it for no delay alloc code.

Signed-off-by: Jinke Han <hanjinke.666@xxxxxxxxxxxxx>
---
 fs/ext4/inode.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 601214453c3a..0d6c4ec7c840 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -1188,6 +1188,13 @@ static int ext4_write_begin(struct file *file, struct address_space *mapping,
 	page = grab_cache_page_write_begin(mapping, index);
 	if (!page)
 		return -ENOMEM;
+	/*
+	 * The same as page allocation, we prealloc buffer heads before
+	 * starting the handle.
+	 */
+	if (!page_has_buffers(page))
+		create_empty_buffers(page, inode->i_sb->s_blocksize, 0);
+
 	unlock_page(page);
 
 retry_journal:
-- 
2.20.1