From: Josef Bacik <josef@toxicpanda.com>
To: Naohiro Aota <naohiro.aota@wdc.com>,
linux-btrfs@vger.kernel.org, dsterba@suse.com
Cc: hare@suse.com, linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH v9 28/41] btrfs: serialize meta IOs on ZONED mode
Date: Tue, 3 Nov 2020 11:04:37 -0500 [thread overview]
Message-ID: <e3107657-aa88-7682-cece-83fac2b93d16@toxicpanda.com> (raw)
In-Reply-To: <61771fe28bda89abcdb55b2a00be05eb82d2216e.1604065695.git.naohiro.aota@wdc.com>
On 10/30/20 9:51 AM, Naohiro Aota wrote:
> We cannot use zone append for writing metadata, because the B-tree nodes
> have references to each other using the logical address. Without knowing
> the address in advance, we cannot construct the tree in the first place.
> So we need to serialize write IOs for metadata.
>
> We cannot add a mutex around allocation and submission because metadata
> blocks are allocated in an earlier stage to build up B-trees.
>
> Add a zoned_meta_io_lock and hold it during metadata IO submission in
> btree_write_cache_pages() to serialize IOs. Furthermore, this add a
> per-block group metadata IO submission pointer "meta_write_pointer" to
> ensure sequential writing, which can be caused when writing back blocks in
> an unfinished transaction.
>
> Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
> ---
> fs/btrfs/block-group.h | 1 +
> fs/btrfs/ctree.h | 1 +
> fs/btrfs/disk-io.c | 1 +
> fs/btrfs/extent_io.c | 27 ++++++++++++++++++++++-
> fs/btrfs/zoned.c | 50 ++++++++++++++++++++++++++++++++++++++++++
> fs/btrfs/zoned.h | 31 ++++++++++++++++++++++++++
> 6 files changed, 110 insertions(+), 1 deletion(-)
>
> diff --git a/fs/btrfs/block-group.h b/fs/btrfs/block-group.h
> index 401e9bcefaec..b2a8a3beceac 100644
> --- a/fs/btrfs/block-group.h
> +++ b/fs/btrfs/block-group.h
> @@ -190,6 +190,7 @@ struct btrfs_block_group {
> */
> u64 alloc_offset;
> u64 zone_unusable;
> + u64 meta_write_pointer;
> };
>
> static inline u64 btrfs_block_group_end(struct btrfs_block_group *block_group)
> diff --git a/fs/btrfs/ctree.h b/fs/btrfs/ctree.h
> index 383c83a1f5b5..736f679f1310 100644
> --- a/fs/btrfs/ctree.h
> +++ b/fs/btrfs/ctree.h
> @@ -955,6 +955,7 @@ struct btrfs_fs_info {
> };
> /* max size to emit ZONE_APPEND write command */
> u64 max_zone_append_size;
> + struct mutex zoned_meta_io_lock;
>
> #ifdef CONFIG_BTRFS_FS_REF_VERIFY
> spinlock_t ref_verify_lock;
> diff --git a/fs/btrfs/disk-io.c b/fs/btrfs/disk-io.c
> index 778716e223ff..f02b121d8213 100644
> --- a/fs/btrfs/disk-io.c
> +++ b/fs/btrfs/disk-io.c
> @@ -2652,6 +2652,7 @@ void btrfs_init_fs_info(struct btrfs_fs_info *fs_info)
> mutex_init(&fs_info->delete_unused_bgs_mutex);
> mutex_init(&fs_info->reloc_mutex);
> mutex_init(&fs_info->delalloc_root_mutex);
> + mutex_init(&fs_info->zoned_meta_io_lock);
> seqlock_init(&fs_info->profiles_lock);
>
> INIT_LIST_HEAD(&fs_info->dirty_cowonly_roots);
> diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
> index 3f49febafc69..3cce444d5dbb 100644
> --- a/fs/btrfs/extent_io.c
> +++ b/fs/btrfs/extent_io.c
> @@ -25,6 +25,7 @@
> #include "backref.h"
> #include "disk-io.h"
> #include "zoned.h"
> +#include "block-group.h"
>
> static struct kmem_cache *extent_state_cache;
> static struct kmem_cache *extent_buffer_cache;
> @@ -4001,6 +4002,7 @@ int btree_write_cache_pages(struct address_space *mapping,
> struct writeback_control *wbc)
> {
> struct extent_buffer *eb, *prev_eb = NULL;
> + struct btrfs_block_group *cache = NULL;
> struct extent_page_data epd = {
> .bio = NULL,
> .extent_locked = 0,
> @@ -4035,6 +4037,7 @@ int btree_write_cache_pages(struct address_space *mapping,
> tag = PAGECACHE_TAG_TOWRITE;
> else
> tag = PAGECACHE_TAG_DIRTY;
> + btrfs_zoned_meta_io_lock(fs_info);
> retry:
> if (wbc->sync_mode == WB_SYNC_ALL)
> tag_pages_for_writeback(mapping, index, end);
> @@ -4077,12 +4080,30 @@ int btree_write_cache_pages(struct address_space *mapping,
> if (!ret)
> continue;
>
> + if (!btrfs_check_meta_write_pointer(fs_info, eb,
> + &cache)) {
> + /*
> + * If for_sync, this hole will be filled with
> + * trasnsaction commit.
> + */
> + if (wbc->sync_mode == WB_SYNC_ALL &&
> + !wbc->for_sync)
> + ret = -EAGAIN;
> + else
> + ret = 0;
> + done = 1;
> + free_extent_buffer(eb);
> + break;
> + }
> +
> prev_eb = eb;
> ret = lock_extent_buffer_for_io(eb, &epd);
> if (!ret) {
> + btrfs_revert_meta_write_pointer(cache, eb);
> free_extent_buffer(eb);
> continue;
> } else if (ret < 0) {
> + btrfs_revert_meta_write_pointer(cache, eb);
> done = 1;
> free_extent_buffer(eb);
> break;
> @@ -4115,10 +4136,12 @@ int btree_write_cache_pages(struct address_space *mapping,
> index = 0;
> goto retry;
> }
> + if (cache)
> + btrfs_put_block_group(cache);
> ASSERT(ret <= 0);
> if (ret < 0) {
> end_write_bio(&epd, ret);
> - return ret;
> + goto out;
> }
> /*
> * If something went wrong, don't allow any metadata write bio to be
> @@ -4153,6 +4176,8 @@ int btree_write_cache_pages(struct address_space *mapping,
> ret = -EROFS;
> end_write_bio(&epd, ret);
> }
> +out:
> + btrfs_zoned_meta_io_unlock(fs_info);
> return ret;
> }
>
> diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
> index 50393d560c9a..15bc7d451348 100644
> --- a/fs/btrfs/zoned.c
> +++ b/fs/btrfs/zoned.c
> @@ -989,6 +989,9 @@ int btrfs_load_block_group_zone_info(struct btrfs_block_group *cache)
> ret = -EIO;
> }
>
> + if (!ret)
> + cache->meta_write_pointer = cache->alloc_offset + cache->start;
> +
> kfree(alloc_offsets);
> free_extent_map(em);
>
> @@ -1120,3 +1123,50 @@ void btrfs_rewrite_logical_zoned(struct btrfs_ordered_extent *ordered)
> kfree(logical);
> bdput(bdev);
> }
> +
> +bool btrfs_check_meta_write_pointer(struct btrfs_fs_info *fs_info,
> + struct extent_buffer *eb,
> + struct btrfs_block_group **cache_ret)
> +{
> + struct btrfs_block_group *cache;
> +
> + if (!btrfs_is_zoned(fs_info))
> + return true;
> +
> + cache = *cache_ret;
> +
> + if (cache && (eb->start < cache->start ||
> + cache->start + cache->length <= eb->start)) {
> + btrfs_put_block_group(cache);
> + cache = NULL;
> + *cache_ret = NULL;
> + }
> +
> + if (!cache)
> + cache = btrfs_lookup_block_group(fs_info, eb->start);
> +
> + if (cache) {
> + *cache_ret = cache;
Don't set this here, set it after the if statement.
> +
> + if (cache->meta_write_pointer != eb->start) {
> + btrfs_put_block_group(cache);
> + cache = NULL;
> + *cache_ret = NULL;
And delete these two lines. Then you can add
Reviewed-by: Josef Bacik <josef@toxicpanda.com>
Thanks,
Josef
next prev parent reply other threads:[~2020-11-03 16:04 UTC|newest]
Thread overview: 119+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-30 13:51 [PATCH v9 00/41] btrfs: zoned block device support Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 01/41] block: add bio_add_zone_append_page Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 02/41] iomap: support REQ_OP_ZONE_APPEND Naohiro Aota
2020-11-02 5:34 ` Naohiro Aota
2020-11-02 16:55 ` Darrick J. Wong
2020-11-02 17:39 ` Johannes Thumshirn
2020-10-30 13:51 ` [PATCH v9 03/41] btrfs: introduce ZONED feature flag Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 04/41] btrfs: Get zone information of zoned block devices Naohiro Aota
2020-11-02 16:53 ` Josef Bacik
2020-11-02 16:58 ` Johannes Thumshirn
2020-11-02 21:07 ` Naohiro Aota
2020-11-03 12:02 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 05/41] btrfs: Check and enable ZONED mode Naohiro Aota
2020-11-03 12:13 ` David Sterba
2020-11-06 9:36 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 06/41] btrfs: introduce max_zone_append_size Naohiro Aota
2020-11-02 16:57 ` Josef Bacik
2020-11-03 12:16 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 07/41] btrfs: disallow space_cache in ZONED mode Naohiro Aota
2020-11-02 17:02 ` Josef Bacik
2020-11-02 17:37 ` Johannes Thumshirn
2020-11-03 12:48 ` David Sterba
2020-11-10 10:14 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 08/41] btrfs: disallow NODATACOW " Naohiro Aota
2020-11-02 17:05 ` Josef Bacik
2020-11-03 12:57 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 09/41] btrfs: disable fallocate " Naohiro Aota
2020-11-03 13:00 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 10/41] btrfs: disallow mixed-bg " Naohiro Aota
2020-11-03 13:01 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 11/41] btrfs: implement log-structured superblock for " Naohiro Aota
2020-11-02 18:22 ` Josef Bacik
2020-11-02 18:53 ` Johannes Thumshirn
2020-11-02 19:01 ` Josef Bacik
2020-11-02 19:31 ` Johannes Thumshirn
2020-11-03 8:21 ` Naohiro Aota
2020-11-02 18:54 ` Josef Bacik
2020-11-03 3:31 ` Naohiro Aota
2020-11-03 13:15 ` David Sterba
2020-11-03 14:10 ` David Sterba
2020-11-06 10:37 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 12/41] btrfs: implement zoned chunk allocator Naohiro Aota
2020-11-02 20:09 ` Josef Bacik
2020-11-02 22:21 ` Naohiro Aota
2020-11-03 13:23 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 13/41] btrfs: verify device extent is aligned to zone Naohiro Aota
2020-11-02 20:14 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 14/41] btrfs: load zone's alloction offset Naohiro Aota
2020-11-02 20:25 ` Josef Bacik
2020-11-02 20:29 ` Josef Bacik
2020-11-02 22:43 ` Naohiro Aota
2020-11-02 22:40 ` Naohiro Aota
2020-11-03 13:28 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 15/41] btrfs: emulate write pointer for conventional zones Naohiro Aota
2020-11-02 20:37 ` Josef Bacik
2020-11-03 1:25 ` Naohiro Aota
2020-11-03 13:32 ` David Sterba
2020-10-30 13:51 ` [PATCH v9 16/41] btrfs: track unusable bytes for zones Naohiro Aota
2020-11-03 14:25 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 17/41] btrfs: do sequential extent allocation in ZONED mode Naohiro Aota
2020-11-03 14:28 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 18/41] btrfs: reset zones of unused block groups Naohiro Aota
2020-11-03 14:34 ` Josef Bacik
2020-11-10 10:40 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 19/41] btrfs: redirty released extent buffers in ZONED mode Naohiro Aota
2020-11-03 14:41 ` Josef Bacik
2020-11-06 9:11 ` Johannes Thumshirn
2020-11-06 15:01 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 20/41] btrfs: extract page adding function Naohiro Aota
2020-11-03 14:45 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 21/41] btrfs: use bio_add_zone_append_page for zoned btrfs Naohiro Aota
2020-11-03 14:55 ` Josef Bacik
2020-11-10 10:42 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 22/41] btrfs: handle REQ_OP_ZONE_APPEND as writing Naohiro Aota
2020-11-03 14:57 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 23/41] btrfs: split ordered extent when bio is sent Naohiro Aota
2020-11-03 15:29 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 24/41] btrfs: extend btrfs_rmap_block for specifying a device Naohiro Aota
2020-11-03 15:32 ` Josef Bacik
2020-11-06 10:52 ` Johannes Thumshirn
2020-10-30 13:51 ` [PATCH v9 25/41] btrfs: use ZONE_APPEND write for ZONED btrfs Naohiro Aota
2020-11-03 15:55 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 26/41] btrfs: enable zone append writing for direct IO Naohiro Aota
2020-11-03 15:56 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 27/41] btrfs: introduce dedicated data write path for ZONED mode Naohiro Aota
2020-11-03 15:57 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 28/41] btrfs: serialize meta IOs on " Naohiro Aota
2020-11-03 16:04 ` Josef Bacik [this message]
2020-10-30 13:51 ` [PATCH v9 29/41] btrfs: wait existing extents before truncating Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 30/41] btrfs: avoid async metadata checksum on ZONED mode Naohiro Aota
2020-11-03 16:05 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 31/41] btrfs: mark block groups to copy for device-replace Naohiro Aota
2020-11-03 17:09 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 32/41] btrfs: implement cloning for ZONED device-replace Naohiro Aota
2020-11-03 17:15 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 33/41] btrfs: implement copying " Naohiro Aota
2020-11-03 17:19 ` Josef Bacik
2020-11-10 11:09 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 34/41] btrfs: support dev-replace in ZONED mode Naohiro Aota
2020-11-03 20:34 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 35/41] btrfs: enable relocation " Naohiro Aota
2020-11-03 20:39 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 36/41] btrfs: relocate block group to repair IO failure in ZONED Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 37/41] btrfs: split alloc_log_tree() Naohiro Aota
2020-11-03 20:42 ` Josef Bacik
2020-11-03 22:10 ` Amy Parker
2020-11-10 11:12 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 38/41] btrfs: extend zoned allocator to use dedicated tree-log block group Naohiro Aota
2020-11-03 20:47 ` Josef Bacik
2020-11-10 6:37 ` Naohiro Aota
2020-10-30 13:51 ` [PATCH v9 39/41] btrfs: serialize log transaction on ZONED mode Naohiro Aota
2020-11-03 20:49 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 40/41] btrfs: reorder log node allocation Naohiro Aota
2020-11-03 20:49 ` Josef Bacik
2020-10-30 13:51 ` [PATCH v9 41/41] btrfs: enable to mount ZONED incompat flag Naohiro Aota
2020-10-31 3:40 ` [PATCH v9 01/41] block: add bio_add_zone_append_page Jens Axboe
2020-11-02 5:15 ` Naohiro Aota
2020-11-02 8:24 ` Johannes Thumshirn
2020-11-03 11:54 ` [PATCH v9 00/41] btrfs: zoned block device support David Sterba
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=e3107657-aa88-7682-cece-83fac2b93d16@toxicpanda.com \
--to=josef@toxicpanda.com \
--cc=dsterba@suse.com \
--cc=hare@suse.com \
--cc=linux-btrfs@vger.kernel.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=naohiro.aota@wdc.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
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).