From: Wu Fengguang <fengguang.wu@intel.com>
To: Andrew Morton <akpm@linux-foundation.org>
Cc: Jan Kara <jack@suse.cz>, Dave Chinner <david@fromorbit.com>,
Wu Fengguang <fengguang.wu@intel.com>
Cc: Christoph Hellwig <hch@infradead.org>
Cc: <linux-fsdevel@vger.kernel.org>
Cc: LKML <linux-kernel@vger.kernel.org>
Subject: [PATCH 01/15] writeback: introduce .tagged_writepages for the WB_SYNC_NONE sync stage
Date: Wed, 08 Jun 2011 05:32:37 +0800 [thread overview]
Message-ID: <20110607213853.513366237@intel.com> (raw)
In-Reply-To: 20110607213236.634026193@intel.com
[-- Attachment #1: writeback-for-sync.patch --]
[-- Type: text/plain, Size: 4593 bytes --]
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.
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>
---
fs/ext4/inode.c | 4 ++--
fs/fs-writeback.c | 17 +++++++++--------
include/linux/writeback.h | 1 +
mm/page-writeback.c | 4 ++--
4 files changed, 14 insertions(+), 12 deletions(-)
--- linux-next.orig/fs/fs-writeback.c 2011-06-08 05:20:05.000000000 +0800
+++ linux-next/fs/fs-writeback.c 2011-06-08 05:20:06.000000000 +0800
@@ -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_writ
{
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_writ
};
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_writ
* (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
{
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));
--- linux-next.orig/include/linux/writeback.h 2011-06-08 05:20:05.000000000 +0800
+++ linux-next/include/linux/writeback.h 2011-06-08 05:20:06.000000000 +0800
@@ -47,6 +47,7 @@ struct writeback_control {
unsigned encountered_congestion:1; /* An output: a queue is full */
unsigned for_kupdate:1; /* A kupdate writeback */
unsigned for_background:1; /* A background writeback */
+ unsigned tagged_writepages:1; /* tag-and-write to avoid livelock */
unsigned for_reclaim:1; /* Invoked from the page allocator */
unsigned range_cyclic:1; /* range_start is cyclic */
unsigned more_io:1; /* more io to be dispatched */
--- linux-next.orig/mm/page-writeback.c 2011-06-08 05:20:05.000000000 +0800
+++ linux-next/mm/page-writeback.c 2011-06-08 05:20:06.000000000 +0800
@@ -892,12 +892,12 @@ int write_cache_pages(struct address_spa
range_whole = 1;
cycled = 1; /* ignore range_cyclic tests */
}
- 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;
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);
done_index = index;
while (!done && (index <= end)) {
--- linux-next.orig/fs/ext4/inode.c 2011-06-08 05:20:05.000000000 +0800
+++ linux-next/fs/ext4/inode.c 2011-06-08 05:20:06.000000000 +0800
@@ -2741,7 +2741,7 @@ static int write_cache_pages_da(struct a
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;
@@ -2973,7 +2973,7 @@ static int ext4_da_writepages(struct add
}
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) {
next prev parent reply other threads:[~2011-06-07 21:32 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2011-06-07 21:32 [PATCH 00/15] writeback fixes and cleanups for 3.0 (v5) Wu Fengguang
2011-06-07 21:32 ` Wu Fengguang [this message]
2011-06-07 23:02 ` [PATCH 01/15] writeback: introduce .tagged_writepages for the WB_SYNC_NONE sync stage Andrew Morton
2011-06-07 23:24 ` Wu Fengguang
2011-06-07 21:32 ` [PATCH 02/15] writeback: update dirtied_when for synced inode to prevent livelock Wu Fengguang
2011-06-07 23:02 ` Andrew Morton
2011-06-07 23:51 ` Wu Fengguang
2011-06-07 21:32 ` [PATCH 03/15] writeback: introduce writeback_control.inodes_cleaned Wu Fengguang
2011-06-07 23:03 ` Andrew Morton
2011-06-08 0:10 ` Wu Fengguang
2011-06-07 21:32 ` [PATCH 04/15] writeback: try more writeback as long as something was written Wu Fengguang
2011-06-07 21:32 ` [PATCH 05/15] writeback: the kupdate expire timestamp should be a moving target Wu Fengguang
2011-06-07 21:32 ` [PATCH 06/15] writeback: refill b_io iff empty Wu Fengguang
2011-06-07 21:32 ` [PATCH 07/15] writeback: split inode_wb_list_lock into bdi_writeback.list_lock Wu Fengguang
2011-06-07 23:03 ` Andrew Morton
2011-06-08 0:20 ` Wu Fengguang
2011-06-08 0:35 ` Andrew Morton
2011-06-08 1:36 ` Wu Fengguang
2011-06-07 21:32 ` [PATCH 08/15] writeback: elevate queue_io() into wb_writeback() Wu Fengguang
2011-06-07 21:32 ` [PATCH 09/15] writeback: avoid extra sync work at enqueue time Wu Fengguang
2011-06-07 21:32 ` [PATCH 10/15] writeback: add bdi_dirty_limit() kernel-doc Wu Fengguang
2011-06-07 21:32 ` [PATCH 11/15] writeback: skip balance_dirty_pages() for in-memory fs Wu Fengguang
2011-06-11 13:07 ` Wu Fengguang
2011-06-13 13:42 ` Jan Kara
2011-06-07 21:32 ` [PATCH 12/15] writeback: remove writeback_control.more_io Wu Fengguang
2011-07-11 21:31 ` Hugh Dickins
2011-07-12 6:20 ` Wu Fengguang
2011-07-12 19:50 ` Hugh Dickins
2011-07-13 5:49 ` Hugh Dickins
2011-07-13 10:57 ` Hugh Dickins
2011-07-13 11:19 ` Jan Kara
2011-07-13 15:06 ` Hugh Dickins
2011-07-13 22:07 ` Wu Fengguang
2011-06-07 21:32 ` [PATCH 13/15] writeback: remove .nonblocking and .encountered_congestion Wu Fengguang
2011-06-07 21:32 ` [PATCH 14/15] writeback: trace event writeback_single_inode Wu Fengguang
2011-06-07 21:32 ` [PATCH 15/15] writeback: trace event writeback_queue_io Wu Fengguang
2011-06-07 23:04 ` [PATCH 00/15] writeback fixes and cleanups for 3.0 (v5) Andrew Morton
2011-06-08 2:01 ` Wu Fengguang
2011-06-08 6:21 ` Sedat Dilek
2011-06-08 13:45 ` Wu Fengguang
2011-06-09 1:16 ` Stephen Rothwell
2011-06-09 2:18 ` Wu Fengguang
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20110607213853.513366237@intel.com \
--to=fengguang.wu@intel.com \
--cc=akpm@linux-foundation.org \
--cc=david@fromorbit.com \
--cc=jack@suse.cz \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).