Linux-BTRFS Archive on lore.kernel.org
 help / color / 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>,
	linux-fsdevel@vger.kernel.org,
	Naohiro Aota <naohiro.aota@wdc.com>
Subject: [PATCH v3 16/27] btrfs: serialize data allocation and submit IOs
Date: Thu,  8 Aug 2019 18:30:27 +0900
Message-ID: <20190808093038.4163421-17-naohiro.aota@wdc.com> (raw)
In-Reply-To: <20190808093038.4163421-1-naohiro.aota@wdc.com>

To preserve sequential write pattern on the drives, we must serialize
allocation and submit_bio. This commit add per-block group mutex
"zone_io_lock" and find_free_extent_seq() hold the lock. The lock is kept
even after returning from find_free_extent(). It is released when submiting
IOs corresponding to the allocation is completed.

Implementing such behavior under __extent_writepage_io is almost impossible
because once pages are unlocked we are not sure when submiting IOs for an
allocated region is finished or not. Instead, this commit add
run_delalloc_hmzoned() to write out non-compressed data IOs at once using
extent_write_locked_rage(). After the write, we can call
btrfs_hmzoned_unlock_allocation() to unlock the block group for new
allocation.

Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
---
 fs/btrfs/ctree.h       |  1 +
 fs/btrfs/extent-tree.c |  5 +++++
 fs/btrfs/hmzoned.h     | 34 +++++++++++++++++++++++++++++++
 fs/btrfs/inode.c       | 45 ++++++++++++++++++++++++++++++++++++++++--
 4 files changed, 83 insertions(+), 2 deletions(-)

diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
index 3d31a1960c4d..1e924c0d1210 100644
--- a/fs/btrfs/ctree.h
+++ b/fs/btrfs/ctree.h
@@ -619,6 +619,7 @@ struct btrfs_block_group_cache {
 	 * zone.
 	 */
 	u64 alloc_offset;
+	struct mutex zone_io_lock;
 };
 
 /* delayed seq elem */
diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index bc95a73a762d..5b1a9e607555 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -5532,6 +5532,7 @@ static int find_free_extent_seq(struct btrfs_block_group_cache *cache,
 	if (cache->alloc_type != BTRFS_ALLOC_SEQ)
 		return 1;
 
+	btrfs_hmzoned_data_io_lock(cache);
 	spin_lock(&space_info->lock);
 	spin_lock(&cache->lock);
 
@@ -5563,6 +5564,9 @@ static int find_free_extent_seq(struct btrfs_block_group_cache *cache,
 out:
 	spin_unlock(&cache->lock);
 	spin_unlock(&space_info->lock);
+	/* if succeeds, unlock after submit_bio */
+	if (ret)
+		btrfs_hmzoned_data_io_unlock(cache);
 	return ret;
 }
 
@@ -8104,6 +8108,7 @@ btrfs_create_block_group_cache(struct btrfs_fs_info *fs_info,
 	btrfs_init_free_space_ctl(cache);
 	atomic_set(&cache->trimming, 0);
 	mutex_init(&cache->free_space_lock);
+	mutex_init(&cache->zone_io_lock);
 	btrfs_init_full_stripe_locks_tree(&cache->full_stripe_locks_root);
 	cache->alloc_type = BTRFS_ALLOC_FIT;
 
diff --git a/fs/btrfs/hmzoned.h b/fs/btrfs/hmzoned.h
index 3a73c3c5e1da..a8e7286708d4 100644
--- a/fs/btrfs/hmzoned.h
+++ b/fs/btrfs/hmzoned.h
@@ -39,6 +39,7 @@ int btrfs_hmzoned_check_metadata_space(struct btrfs_fs_info *fs_info);
 void btrfs_redirty_list_add(struct btrfs_transaction *trans,
 			    struct extent_buffer *eb);
 void btrfs_free_redirty_list(struct btrfs_transaction *trans);
+void btrfs_hmzoned_data_io_unlock_at(struct inode *inode, u64 start, u64 len);
 
 static inline bool btrfs_dev_is_sequential(struct btrfs_device *device, u64 pos)
 {
@@ -140,4 +141,37 @@ static inline bool btrfs_check_super_location(struct btrfs_device *device,
 		!btrfs_dev_is_sequential(device, pos);
 }
 
+
+static inline void btrfs_hmzoned_data_io_lock(
+	struct btrfs_block_group_cache *cache)
+{
+	/* No need to lock metadata BGs or non-sequential BGs */
+	if (!(cache->flags & BTRFS_BLOCK_GROUP_DATA) ||
+	    cache->alloc_type != BTRFS_ALLOC_SEQ)
+		return;
+	mutex_lock(&cache->zone_io_lock);
+}
+
+static inline void btrfs_hmzoned_data_io_unlock(
+	struct btrfs_block_group_cache *cache)
+{
+	if (!(cache->flags & BTRFS_BLOCK_GROUP_DATA) ||
+	    cache->alloc_type != BTRFS_ALLOC_SEQ)
+		return;
+	mutex_unlock(&cache->zone_io_lock);
+}
+
+static inline void btrfs_hmzoned_data_io_unlock_logical(
+	struct btrfs_fs_info *fs_info, u64 logical)
+{
+	struct btrfs_block_group_cache *cache;
+
+	if (!btrfs_fs_incompat(fs_info, HMZONED))
+		return;
+
+	cache = btrfs_lookup_block_group(fs_info, logical);
+	btrfs_hmzoned_data_io_unlock(cache);
+	btrfs_put_block_group(cache);
+}
+
 #endif
diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c
index ee582a36653d..d504200c9767 100644
--- a/fs/btrfs/inode.c
+++ b/fs/btrfs/inode.c
@@ -48,6 +48,7 @@
 #include "qgroup.h"
 #include "dedupe.h"
 #include "delalloc-space.h"
+#include "hmzoned.h"
 
 struct btrfs_iget_args {
 	struct btrfs_key *location;
@@ -1279,6 +1280,39 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page,
 	return 0;
 }
 
+static noinline int run_delalloc_hmzoned(struct inode *inode,
+					 struct page *locked_page, u64 start,
+					 u64 end, int *page_started,
+					 unsigned long *nr_written)
+{
+	struct extent_map *em;
+	u64 logical;
+	int ret;
+
+	ret = cow_file_range(inode, locked_page, start, end,
+			     end, page_started, nr_written, 0, NULL);
+	if (ret)
+		return ret;
+
+	if (*page_started)
+		return 0;
+
+	em = btrfs_get_extent(BTRFS_I(inode), NULL, 0, start, end - start + 1,
+			      0);
+	ASSERT(em != NULL && em->block_start < EXTENT_MAP_LAST_BYTE);
+	logical = em->block_start;
+	free_extent_map(em);
+
+	__set_page_dirty_nobuffers(locked_page);
+	account_page_redirty(locked_page);
+	extent_write_locked_range(inode, start, end, WB_SYNC_ALL);
+	*page_started = 1;
+
+	btrfs_hmzoned_data_io_unlock_logical(btrfs_sb(inode->i_sb), logical);
+
+	return 0;
+}
+
 static noinline int csum_exist_in_range(struct btrfs_fs_info *fs_info,
 					u64 bytenr, u64 num_bytes)
 {
@@ -1645,17 +1679,24 @@ int btrfs_run_delalloc_range(struct inode *inode, struct page *locked_page,
 	int ret;
 	int force_cow = need_force_cow(inode, start, end);
 	unsigned int write_flags = wbc_to_write_flags(wbc);
+	int do_compress = inode_can_compress(inode) &&
+		inode_need_compress(inode, start, end);
+	int hmzoned = btrfs_fs_incompat(btrfs_sb(inode->i_sb), HMZONED);
 
 	if (BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW && !force_cow) {
+		ASSERT(!hmzoned);
 		ret = run_delalloc_nocow(inode, locked_page, start, end,
 					 page_started, 1, nr_written);
 	} else if (BTRFS_I(inode)->flags & BTRFS_INODE_PREALLOC && !force_cow) {
+		ASSERT(!hmzoned);
 		ret = run_delalloc_nocow(inode, locked_page, start, end,
 					 page_started, 0, nr_written);
-	} else if (!inode_can_compress(inode) ||
-		   !inode_need_compress(inode, start, end)) {
+	} else if (!do_compress && !hmzoned) {
 		ret = cow_file_range(inode, locked_page, start, end, end,
 				      page_started, nr_written, 1, NULL);
+	} else if (!do_compress && hmzoned) {
+		ret = run_delalloc_hmzoned(inode, locked_page, start, end,
+					   page_started, nr_written);
 	} else {
 		set_bit(BTRFS_INODE_HAS_ASYNC_EXTENT,
 			&BTRFS_I(inode)->runtime_flags);
-- 
2.22.0


  parent reply index

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-08-08  9:30 [PATCH v3 00/27] btrfs zoned block device support Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 01/27] btrfs: introduce HMZONED feature flag Naohiro Aota
2019-08-16  4:49   ` Anand Jain
2019-08-08  9:30 ` [PATCH v3 02/27] btrfs: Get zone information of zoned block devices Naohiro Aota
2019-08-16  4:44   ` Anand Jain
2019-08-16 14:19     ` Damien Le Moal
2019-08-16 23:47       ` Anand Jain
2019-08-16 23:55         ` Damien Le Moal
2019-08-08  9:30 ` [PATCH v3 03/27] btrfs: Check and enable HMZONED mode Naohiro Aota
2019-08-16  5:46   ` Anand Jain
2019-08-16 14:23     ` Damien Le Moal
2019-08-16 23:56       ` Anand Jain
2019-08-17  0:05         ` Damien Le Moal
2019-08-20  5:07         ` Naohiro Aota
2019-08-20 13:05           ` David Sterba
2019-08-08  9:30 ` [PATCH v3 04/27] btrfs: disallow RAID5/6 in " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 05/27] btrfs: disallow space_cache " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 06/27] btrfs: disallow NODATACOW " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 07/27] btrfs: disable tree-log " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 08/27] btrfs: disable fallocate " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 09/27] btrfs: align device extent allocation to zone boundary Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 10/27] btrfs: do sequential extent allocation in HMZONED mode Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 11/27] btrfs: make unmirroed BGs readonly only if we have at least one writable BG Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 12/27] btrfs: ensure metadata space available on/after degraded mount in HMZONED Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 13/27] btrfs: reset zones of unused block groups Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 14/27] btrfs: limit super block locations in HMZONED mode Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 15/27] btrfs: redirty released extent buffers in sequential BGs Naohiro Aota
2019-08-08  9:30 ` Naohiro Aota [this message]
2019-08-08  9:30 ` [PATCH v3 17/27] btrfs: implement atomic compressed IO submission Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 18/27] btrfs: support direct write IO in HMZONED Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 19/27] btrfs: serialize meta IOs on HMZONED mode Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 20/27] btrfs: wait existing extents before truncating Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 21/27] btrfs: avoid async checksum/submit on HMZONED mode Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 22/27] btrfs: disallow mixed-bg in " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 23/27] btrfs: disallow inode_cache " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 24/27] btrfs: support dev-replace " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 25/27] btrfs: enable relocation " Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 26/27] btrfs: relocate block group to repair IO failure in HMZONED Naohiro Aota
2019-08-08  9:30 ` [PATCH v3 27/27] btrfs: enable to mount HMZONED incompat flag Naohiro Aota

Reply instructions:

You may reply publically 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=20190808093038.4163421-17-naohiro.aota@wdc.com \
    --to=naohiro.aota@wdc.com \
    --cc=Matias.Bjorling@wdc.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 \
    /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

Linux-BTRFS Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-btrfs/0 linux-btrfs/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-btrfs linux-btrfs/ https://lore.kernel.org/linux-btrfs \
		linux-btrfs@vger.kernel.org linux-btrfs@archiver.kernel.org
	public-inbox-index linux-btrfs


Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-btrfs


AGPL code for this site: git clone https://public-inbox.org/ public-inbox