When testing f2fs on an SSD, I found some 128 page IOs followed by 1 page IO
were issued by f2fs_write_node_pages.
This means that there were some mishandling flows which degrades performance.
Previous f2fs_write_node_pages determines the number of pages to be written,
nr_to_write, as follows.
1. The bio_get_nr_vecs returns 129 pages.
2. The bio_alloc makes a room for 128 pages.
3. The initial 128 pages go into one bio.
4. The existing bio is submitted, and a new bio is prepared for the last 1 page.
5. Finally, sync_node_pages submits the last 1 page bio.
The problem is from the use of bio_get_nr_vecs, so this patch replace it
with max_hw_blocks using queue_max_sectors.
Signed-off-by: Jaegeuk Kim <[email protected]>
---
fs/f2fs/node.c | 6 ++----
fs/f2fs/segment.c | 2 +-
fs/f2fs/segment.h | 11 +++++++++++
3 files changed, 14 insertions(+), 5 deletions(-)
diff --git a/fs/f2fs/node.c b/fs/f2fs/node.c
index f14eb7b..7209d63 100644
--- a/fs/f2fs/node.c
+++ b/fs/f2fs/node.c
@@ -1171,7 +1171,6 @@ static int f2fs_write_node_pages(struct address_space *mapping,
struct writeback_control *wbc)
{
struct f2fs_sb_info *sbi = F2FS_SB(mapping->host->i_sb);
- struct block_device *bdev = sbi->sb->s_bdev;
long nr_to_write = wbc->nr_to_write;
/* First check balancing cached NAT entries */
@@ -1185,10 +1184,9 @@ static int f2fs_write_node_pages(struct address_space *mapping,
return 0;
/* if mounting is failed, skip writing node pages */
- wbc->nr_to_write = bio_get_nr_vecs(bdev);
+ wbc->nr_to_write = max_hw_blocks(sbi);
sync_node_pages(sbi, 0, wbc);
- wbc->nr_to_write = nr_to_write -
- (bio_get_nr_vecs(bdev) - wbc->nr_to_write);
+ wbc->nr_to_write = nr_to_write - (max_hw_blocks(sbi) - wbc->nr_to_write);
return 0;
}
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 392ccb3..d8e84e4 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -734,7 +734,7 @@ static void submit_write_page(struct f2fs_sb_info *sbi, struct page *page,
do_submit_bio(sbi, type, false);
alloc_new:
if (sbi->bio[type] == NULL) {
- sbi->bio[type] = f2fs_bio_alloc(bdev, bio_get_nr_vecs(bdev));
+ sbi->bio[type] = f2fs_bio_alloc(bdev, max_hw_blocks(sbi));
sbi->bio[type]->bi_sector = SECTOR_FROM_BLOCK(sbi, blk_addr);
/*
* The end_io will be assigned at the sumbission phase.
diff --git a/fs/f2fs/segment.h b/fs/f2fs/segment.h
index 26fc054..062424a 100644
--- a/fs/f2fs/segment.h
+++ b/fs/f2fs/segment.h
@@ -8,6 +8,8 @@
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*/
+#include <linux/blkdev.h>
+
/* constant macro */
#define NULL_SEGNO ((unsigned int)(~0))
#define NULL_SECNO ((unsigned int)(~0))
@@ -86,6 +88,8 @@
#define SECTOR_FROM_BLOCK(sbi, blk_addr) \
(blk_addr << ((sbi)->log_blocksize - F2FS_LOG_SECTOR_SIZE))
+#define SECTOR_TO_BLOCK(sbi, sectors) \
+ (sectors >> ((sbi)->log_blocksize - F2FS_LOG_SECTOR_SIZE))
/* during checkpoint, bio_private is used to synchronize the last bio */
struct bio_private {
@@ -624,3 +628,10 @@ static inline bool sec_usage_check(struct f2fs_sb_info *sbi, unsigned int secno)
return true;
return false;
}
+
+static inline unsigned int max_hw_blocks(struct f2fs_sb_info *sbi)
+{
+ struct block_device *bdev = sbi->sb->s_bdev;
+ struct request_queue *q = bdev_get_queue(bdev);
+ return SECTOR_TO_BLOCK(sbi, queue_max_sectors(q));
+}
--
1.8.1.3.566.gaa39828