lk/linux-kernel-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Btrfs: use WRITE_SYNC for synchronous writes

Browse Source 
Part of reducing fsync/O_SYNC/O_DIRECT latencies is using WRITE_SYNC for
writes we plan on waiting on in the near future.  This patch
mirrors recent changes in other filesystems and the generic code to
use WRITE_SYNC when WB_SYNC_ALL is passed and to use WRITE_SYNC for
other latency critical writes.

Btrfs uses async worker threads for checksumming before the write is done,
and then again to actually submit the bios.  The bio submission code just
runs a per-device list of bios that need to be sent down the pipe.

This list is split into low priority and high priority lists so the
WRITE_SYNC IO happens first.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
This commit is contained in:
Chris Mason
2009-04-20 15:50:09 -04:00
parent 0882e8dd3a
commit ffbd517d5a
5 changed files with 142 additions and 47 deletions
Download Patch File Download Diff File Expand all files Collapse all files

46
fs/btrfs/extent_io.c
View File

@ -50,7 +50,10 @@ struct extent_page_data {
/* tells writepage not to lock the state bits for this range
* it still does the unlocking
*/
int extent_locked;
unsigned int extent_locked:1;
/* tells the submit_bio code to use a WRITE_SYNC */
unsigned int sync_io:1;
};
int __init extent_io_init(void)
@ -2136,8 +2139,14 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
u64 delalloc_end;
int page_started;
int compressed;
int write_flags;
unsigned long nr_written = 0;
if (wbc->sync_mode == WB_SYNC_ALL)
write_flags = WRITE_SYNC_PLUG;
else
write_flags = WRITE;
WARN_ON(!PageLocked(page));
pg_offset = i_size & (PAGE_CACHE_SIZE - 1);
if (page->index > end_index ||
@ -2314,9 +2323,9 @@ static int __extent_writepage(struct page *page, struct writeback_control *wbc,
(unsigned long long)end);
}
ret = submit_extent_page(WRITE, tree, page, sector,
iosize, pg_offset, bdev,
&epd->bio, max_nr,
ret = submit_extent_page(write_flags, tree, page,
sector, iosize, pg_offset,
bdev, &epd->bio, max_nr,
end_bio_extent_writepage,
0, 0, 0);
if (ret)
@ -2460,13 +2469,21 @@ retry:
return ret;
}
static void flush_epd_write_bio(struct extent_page_data *epd)
{
if (epd->bio) {
if (epd->sync_io)
submit_one_bio(WRITE_SYNC, epd->bio, 0, 0);
else
submit_one_bio(WRITE, epd->bio, 0, 0);
epd->bio = NULL;
}
}
static noinline void flush_write_bio(void *data)
{
struct extent_page_data *epd = data;
if (epd->bio) {
submit_one_bio(WRITE, epd->bio, 0, 0);
epd->bio = NULL;
}
flush_epd_write_bio(epd);
}
int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
@ -2480,6 +2497,7 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
.tree = tree,
.get_extent = get_extent,
.extent_locked = 0,
.sync_io = wbc->sync_mode == WB_SYNC_ALL,
};
struct writeback_control wbc_writepages = {
.bdi = wbc->bdi,
@ -2490,13 +2508,11 @@ int extent_write_full_page(struct extent_io_tree *tree, struct page *page,
.range_end = (loff_t)-1,
};
ret = __extent_writepage(page, wbc, &epd);
extent_write_cache_pages(tree, mapping, &wbc_writepages,
__extent_writepage, &epd, flush_write_bio);
if (epd.bio)
submit_one_bio(WRITE, epd.bio, 0, 0);
flush_epd_write_bio(&epd);
return ret;
}
@ -2515,6 +2531,7 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
.tree = tree,
.get_extent = get_extent,
.extent_locked = 1,
.sync_io = mode == WB_SYNC_ALL,
};
struct writeback_control wbc_writepages = {
.bdi = inode->i_mapping->backing_dev_info,
@ -2540,8 +2557,7 @@ int extent_write_locked_range(struct extent_io_tree *tree, struct inode *inode,
start += PAGE_CACHE_SIZE;
}
if (epd.bio)
submit_one_bio(WRITE, epd.bio, 0, 0);
flush_epd_write_bio(&epd);
return ret;
}
@ -2556,13 +2572,13 @@ int extent_writepages(struct extent_io_tree *tree,
.tree = tree,
.get_extent = get_extent,
.extent_locked = 0,
.sync_io = wbc->sync_mode == WB_SYNC_ALL,
};
ret = extent_write_cache_pages(tree, mapping, wbc,
__extent_writepage, &epd,
flush_write_bio);
if (epd.bio)
submit_one_bio(WRITE, epd.bio, 0, 0);
flush_epd_write_bio(&epd);
return ret;
}