diff options
| author | Wu Fengguang <fengguang.wu@intel.com> | 2010-06-06 10:38:15 -0600 |
|---|---|---|
| committer | Greg Kroah-Hartman <gregkh@suse.de> | 2011-10-03 11:40:43 -0700 |
| commit | ac693061b11c33d5a5c5ec1925de7abd3fcb0971 (patch) | |
| tree | 968e4efb688a21caedd5a3e7baea3d6343abf4bd /fs | |
| parent | 0d39ea76fa3b9a97e9dc45282613bc239ffa8509 (diff) | |
writeback: introduce .tagged_writepages for the WB_SYNC_NONE sync stage
commit 6e6938b6d3130305a5960c86b1a9b21e58cf6144 upstream.
sync(2) is performed in two stages: the WB_SYNC_NONE sync and the
WB_SYNC_ALL sync. Identify the first stage with .tagged_writepages and
do livelock prevention for it, too.
Jan's commit f446daaea9 ("mm: implement writeback livelock avoidance
using page tagging") is a partial fix in that it only fixed the
WB_SYNC_ALL phase livelock.
Although ext4 is tested to no longer livelock with commit f446daaea9,
it may due to some "redirty_tail() after pages_skipped" effect which
is by no means a guarantee for _all_ the file systems.
Note that writeback_inodes_sb() is called by not only sync(), they are
treated the same because the other callers also need livelock prevention.
Impact: It changes the order in which pages/inodes are synced to disk.
Now in the WB_SYNC_NONE stage, it won't proceed to write the next inode
until finished with the current inode.
Acked-by: Jan Kara <jack@suse.cz>
CC: Dave Chinner <david@fromorbit.com>
Signed-off-by: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
Diffstat (limited to 'fs')
| -rw-r--r-- | fs/ext4/inode.c | 4 | ||||
| -rw-r--r-- | fs/fs-writeback.c | 17 |
2 files changed, 11 insertions, 10 deletions
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c index b864839e1d67..c94774c3276e 100644 --- a/fs/ext4/inode.c +++ b/fs/ext4/inode.c @@ -2756,7 +2756,7 @@ static int write_cache_pages_da(struct address_space *mapping, index = wbc->range_start >> PAGE_CACHE_SHIFT; end = wbc->range_end >> PAGE_CACHE_SHIFT; - if (wbc->sync_mode == WB_SYNC_ALL) + if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) tag = PAGECACHE_TAG_TOWRITE; else tag = PAGECACHE_TAG_DIRTY; @@ -2988,7 +2988,7 @@ static int ext4_da_writepages(struct address_space *mapping, } retry: - if (wbc->sync_mode == WB_SYNC_ALL) + if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) tag_pages_for_writeback(mapping, index, end); while (!ret && wbc->nr_to_write > 0) { diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 0f015a0468de..5ed2ce9a28d0 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -36,6 +36,7 @@ struct wb_writeback_work { long nr_pages; struct super_block *sb; enum writeback_sync_modes sync_mode; + unsigned int tagged_writepages:1; unsigned int for_kupdate:1; unsigned int range_cyclic:1; unsigned int for_background:1; @@ -650,6 +651,7 @@ static long wb_writeback(struct bdi_writeback *wb, { struct writeback_control wbc = { .sync_mode = work->sync_mode, + .tagged_writepages = work->tagged_writepages, .older_than_this = NULL, .for_kupdate = work->for_kupdate, .for_background = work->for_background, @@ -657,7 +659,7 @@ static long wb_writeback(struct bdi_writeback *wb, }; unsigned long oldest_jif; long wrote = 0; - long write_chunk; + long write_chunk = MAX_WRITEBACK_PAGES; struct inode *inode; if (wbc.for_kupdate) { @@ -683,9 +685,7 @@ static long wb_writeback(struct bdi_writeback *wb, * (quickly) tag currently dirty pages * (maybe slowly) sync all tagged pages */ - if (wbc.sync_mode == WB_SYNC_NONE) - write_chunk = MAX_WRITEBACK_PAGES; - else + if (wbc.sync_mode == WB_SYNC_ALL || wbc.tagged_writepages) write_chunk = LONG_MAX; wbc.wb_start = jiffies; /* livelock avoidance */ @@ -1188,10 +1188,11 @@ void writeback_inodes_sb_nr(struct super_block *sb, unsigned long nr) { DECLARE_COMPLETION_ONSTACK(done); struct wb_writeback_work work = { - .sb = sb, - .sync_mode = WB_SYNC_NONE, - .done = &done, - .nr_pages = nr, + .sb = sb, + .sync_mode = WB_SYNC_NONE, + .tagged_writepages = 1, + .done = &done, + .nr_pages = nr, }; WARN_ON(!rwsem_is_locked(&sb->s_umount)); |