linux-fsdevel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Naohiro Aota <naohiro.aota@wdc.com>
To: linux-btrfs@vger.kernel.org, David Sterba <dsterba@suse.com>
Cc: Chris Mason <clm@fb.com>, Josef Bacik <josef@toxicpanda.com>,
	Nikolay Borisov <nborisov@suse.com>,
	Damien Le Moal <damien.lemoal@wdc.com>,
	Matias Bjorling <Matias.Bjorling@wdc.com>,
	Johannes Thumshirn <jthumshirn@suse.de>,
	Hannes Reinecke <hare@suse.com>,
	Anand Jain <anand.jain@oracle.com>,
	linux-fsdevel@vger.kernel.org,
	Naohiro Aota <naohiro.aota@wdc.com>
Subject: [PATCH v4 21/27] btrfs: avoid async checksum/submit on HMZONED mode
Date: Fri, 23 Aug 2019 19:10:30 +0900	[thread overview]
Message-ID: <20190823101036.796932-22-naohiro.aota@wdc.com> (raw)
In-Reply-To: <20190823101036.796932-1-naohiro.aota@wdc.com>

In HMZONED, btrfs use per-Block Group zone_io_lock to serialize the data
write IOs or use per-FS hmzoned_meta_io_lock to serialize the metadata
write IOs.

Even with these serialization, write bios sent from
{btree,btrfs}_write_cache_pages can be reordered by async checksum workers
as these workers are per CPU and not per zone.

To preserve write BIO ordering, we can disable async checksum on HMZONED.
This does not result in lower performance with HDDs as a single CPU core is
fast enough to do checksum for a single zone write stream with the maximum
possible bandwidth of the device. If multiple zones are being written
simultaneously, HDD seek overhead lowers the achievable maximum bandwidth,
resulting again in a per zone checksum serialization not affecting
performance.

Besides, this commit disable async_submit in
btrfs_submit_compressed_write() for the same reason. This part will be
unnecessary once btrfs get the "btrfs: fix cgroup writeback support"
series.

Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
---
 fs/btrfs/compression.c | 5 +++--
 fs/btrfs/disk-io.c     | 2 ++
 fs/btrfs/inode.c       | 9 ++++++---
 3 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/fs/btrfs/compression.c b/fs/btrfs/compression.c
index 60c47b417a4b..058dea5e432f 100644
--- a/fs/btrfs/compression.c
+++ b/fs/btrfs/compression.c
@@ -322,6 +322,7 @@ blk_status_t btrfs_submit_compressed_write(struct inode *inode, u64 start,
 	struct block_device *bdev;
 	blk_status_t ret;
 	int skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
+	int async_submit = !btrfs_fs_incompat(fs_info, HMZONED);
 
 	WARN_ON(!PAGE_ALIGNED(start));
 	cb = kmalloc(compressed_bio_size(fs_info, compressed_len), GFP_NOFS);
@@ -377,7 +378,7 @@ blk_status_t btrfs_submit_compressed_write(struct inode *inode, u64 start,
 				BUG_ON(ret); /* -ENOMEM */
 			}
 
-			ret = btrfs_map_bio(fs_info, bio, 0, 1);
+			ret = btrfs_map_bio(fs_info, bio, 0, async_submit);
 			if (ret) {
 				bio->bi_status = ret;
 				bio_endio(bio);
@@ -408,7 +409,7 @@ blk_status_t btrfs_submit_compressed_write(struct inode *inode, u64 start,
 		BUG_ON(ret); /* -ENOMEM */
 	}
 
-	ret = btrfs_map_bio(fs_info, bio, 0, 1);
+	ret = btrfs_map_bio(fs_info, bio, 0, async_submit);
 	if (ret) {
 		bio->bi_status = ret;
 		bio_endio(bio);
diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
index a9632e455eb5..9bae051f9f44 100644
--- a/fs/btrfs/disk-io.c
+++ b/fs/btrfs/disk-io.c
@@ -873,6 +873,8 @@ static blk_status_t btree_submit_bio_start(void *private_data, struct bio *bio,
 static int check_async_write(struct btrfs_fs_info *fs_info,
 			     struct btrfs_inode *bi)
 {
+	if (btrfs_fs_incompat(fs_info, HMZONED))
+		return 0;
 	if (atomic_read(&bi->sync_writers))
 		return 0;
 	if (test_bit(BTRFS_FS_CSUM_IMPL_FAST, &fs_info->flags))
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index 95f4ce8ac8d0..bb0ae3107e60 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -2075,7 +2075,8 @@ static blk_status_t btrfs_submit_bio_hook(struct inode *inode, struct bio *bio,
 	enum btrfs_wq_endio_type metadata = BTRFS_WQ_ENDIO_DATA;
 	blk_status_t ret = 0;
 	int skip_sum;
-	int async = !atomic_read(&BTRFS_I(inode)->sync_writers);
+	int async = !atomic_read(&BTRFS_I(inode)->sync_writers) &&
+		!btrfs_fs_incompat(fs_info, HMZONED);
 
 	skip_sum = BTRFS_I(inode)->flags & BTRFS_INODE_NODATASUM;
 
@@ -8383,7 +8384,8 @@ static inline blk_status_t btrfs_submit_dio_bio(struct bio *bio,
 
 	/* Check btrfs_submit_bio_hook() for rules about async submit. */
 	if (async_submit)
-		async_submit = !atomic_read(&BTRFS_I(inode)->sync_writers);
+		async_submit = !atomic_read(&BTRFS_I(inode)->sync_writers) &&
+			!btrfs_fs_incompat(fs_info, HMZONED);
 
 	if (!write) {
 		ret = btrfs_bio_wq_end_io(fs_info, bio, BTRFS_WQ_ENDIO_DATA);
@@ -8448,7 +8450,8 @@ static int btrfs_submit_direct_hook(struct btrfs_dio_private *dip)
 	}
 
 	/* async crcs make it difficult to collect full stripe writes. */
-	if (btrfs_data_alloc_profile(fs_info) & BTRFS_BLOCK_GROUP_RAID56_MASK)
+	if (btrfs_data_alloc_profile(fs_info) & BTRFS_BLOCK_GROUP_RAID56_MASK ||
+	    btrfs_fs_incompat(fs_info, HMZONED))
 		async_submit = 0;
 	else
 		async_submit = 1;
-- 
2.23.0


  parent reply	other threads:[~2019-08-23 10:11 UTC|newest]

Thread overview: 38+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-23 10:10 [PATCH v4 00/27] btrfs zoned block device support Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 01/27] btrfs: introduce HMZONED feature flag Naohiro Aota
2019-08-23 11:45   ` Johannes Thumshirn
2019-08-23 10:10 ` [PATCH v4 02/27] btrfs: Get zone information of zoned block devices Naohiro Aota
2019-08-23 11:57   ` Johannes Thumshirn
2019-08-26  6:29     ` Naohiro Aota
2019-08-24  9:22   ` kbuild test robot
2019-08-24 10:49   ` kbuild test robot
2019-08-23 10:10 ` [PATCH v4 03/27] btrfs: Check and enable HMZONED mode Naohiro Aota
2019-08-23 12:07   ` Johannes Thumshirn
2019-08-26  8:38     ` Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 04/27] btrfs: disallow RAID5/6 in " Naohiro Aota
2019-08-23 12:09   ` Johannes Thumshirn
2019-08-23 10:10 ` [PATCH v4 05/27] btrfs: disallow space_cache " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 06/27] btrfs: disallow NODATACOW " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 07/27] btrfs: disable tree-log " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 08/27] btrfs: disable fallocate " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 09/27] btrfs: align device extent allocation to zone boundary Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 10/27] btrfs: do sequential extent allocation in HMZONED mode Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 11/27] btrfs: make unmirroed BGs readonly only if we have at least one writable BG Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 12/27] btrfs: ensure metadata space available on/after degraded mount in HMZONED Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 13/27] btrfs: reset zones of unused block groups Naohiro Aota
2019-08-24 11:32   ` kbuild test robot
2019-08-25  4:56   ` kbuild test robot
2019-08-23 10:10 ` [PATCH v4 14/27] btrfs: limit super block locations in HMZONED mode Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 15/27] btrfs: redirty released extent buffers in sequential BGs Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 16/27] btrfs: serialize data allocation and submit IOs Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 17/27] btrfs: implement atomic compressed IO submission Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 18/27] btrfs: support direct write IO in HMZONED Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 19/27] btrfs: serialize meta IOs on HMZONED mode Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 20/27] btrfs: wait existing extents before truncating Naohiro Aota
2019-08-23 10:10 ` Naohiro Aota [this message]
2019-08-23 10:10 ` [PATCH v4 22/27] btrfs: disallow mixed-bg in HMZONED mode Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 23/27] btrfs: disallow inode_cache " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 24/27] btrfs: support dev-replace " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 25/27] btrfs: enable relocation " Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 26/27] btrfs: relocate block group to repair IO failure in HMZONED Naohiro Aota
2019-08-23 10:10 ` [PATCH v4 27/27] btrfs: enable to mount HMZONED incompat flag Naohiro Aota

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=20190823101036.796932-22-naohiro.aota@wdc.com \
    --to=naohiro.aota@wdc.com \
    --cc=Matias.Bjorling@wdc.com \
    --cc=anand.jain@oracle.com \
    --cc=clm@fb.com \
    --cc=damien.lemoal@wdc.com \
    --cc=dsterba@suse.com \
    --cc=hare@suse.com \
    --cc=josef@toxicpanda.com \
    --cc=jthumshirn@suse.de \
    --cc=linux-btrfs@vger.kernel.org \
    --cc=linux-fsdevel@vger.kernel.org \
    --cc=nborisov@suse.com \
    --subject='Re: [PATCH v4 21/27] btrfs: avoid async checksum/submit on HMZONED mode' \
    /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

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).