sync livelock fixes

Wu Fengguang linux-fsdevel lkml 1 version Apr 30, 2011 In review

Get series with b4

b4 am [email protected]

Versions newest first

#	Subject and files	Size	Replies
1/3	writeback: introduce wbc.for_sync to cover the two sync stages`fs/ext4/inode.c` `fs/fs-writeback.c` `include/linux/writeback.h` +2 more	+19 ~~-10~~	2 replies
2/3	writeback: update dirtied_when for synced inode to prevent livelock`fs/fs-writeback.c`	+8 -0	0 replies
3/3	writeback: avoid extra sync work at enqueue time`fs/fs-writeback.c` `include/linux/writeback.h`	+4 ~~-16~~	0 replies

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Cover letter [PATCH 0/3] sync livelock fixes

Andrew,

I wrote a simple script to test sync livelock and this patchset is working as expected:

sync time: 2
Dirty: 26492 kB
Writeback: 30260 kB
NFS_Unstable: 0 kB
WritebackTmp: 0 kB
sync NOT livelocked

In particular patch 2 fixes the sync livelock problem introduced by patch "writeback: try more writeback as long as something was written".

Thanks,
Fengguang
---

#!/bin/sh

umount /dev/sda7

# mkfs.xfs -f /dev/sda7
mkfs.ext4 /dev/sda7
mount /dev/sda7 /fs

echo $((50<<20)) > /proc/sys/vm/dirty_bytes

pid=
for i in seq 10
do

	dd if=/dev/zero of=/fs/zero-$i bs=1M count=1000 &
	pid="$pid $!"

done

sleep 1

tic=$(date +'%s')
sync
tac=$(date +'%s')

echo
echo sync time: $((tac-tic))
egrep '(Dirty|Writeback|NFS_Unstable)' /proc/meminfo

pidof dd > /dev/null && { kill -9 $pid; echo sync NOT livelocked; }