From: Wang Yugui <wangyugui@e16-tech.com>
To: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: linux-fsdevel@vger.kernel.org, linux-xfs@vger.kernel.org,
Dave Chinner <david@fromorbit.com>,
Christoph Hellwig <hch@infradead.org>,
"Darrick J . Wong" <djwong@kernel.org>
Subject: Re: [PATCH v3 0/8] Create large folios in iomap buffered write path
Date: Wed, 21 Jun 2023 20:03:13 +0800 [thread overview]
Message-ID: <20230621200305.23CB.409509F4@e16-tech.com> (raw)
In-Reply-To: <20230612203910.724378-1-willy@infradead.org>
Hi,
> Commit ebb7fb1557b1 limited the length of ioend chains to 4096 entries
> to improve worst-case latency. Unfortunately, this had the effect of
> limiting the performance of:
>
> fio -name write-bandwidth -rw=write -bs=1024Ki -size=32Gi -runtime=30 \
> -iodepth 1 -ioengine sync -zero_buffers=1 -direct=0 -end_fsync=1 \
> -numjobs=4 -directory=/mnt/test
>
> The problem ends up being lock contention on the i_pages spinlock as we
> clear the writeback bit on each folio (and propagate that up through
> the tree). By using larger folios, we decrease the number of folios
> to be processed by a factor of 256 for this benchmark, eliminating the
> lock contention.
>
> It's also the right thing to do. This is a project that has been on
> the back burner for years, it just hasn't been important enough to do
> before now.
>
> I think it's probably best if this goes through the iomap tree since
> the changes outside iomap are either to the page cache or they're
> trivial.
>
> v3:
> - Fix the handling of compound highmem pages in copy_page_from_iter_atomic()
> - Rename fgp_t to fgf_t
> - Clarify some wording in the documentation
This v3 patches broken linux 6.4-rc7.
fstests(btrfs/007 and more) will fail with the v3 patches .
but it works well without the v3 patches.
Best Regards
Wang Yugui (wangyugui@e16-tech.com)
2023/06/21
> v2:
> - Fix misplaced semicolon
> - Rename fgp_order to fgp_set_order
> - Rename FGP_ORDER to FGP_GET_ORDER
> - Add fgp_t
> - Update the documentation for ->release_folio
> - Fix iomap_invalidate_folio()
> - Update iomap_release_folio()
>
> Matthew Wilcox (Oracle) (8):
> iov_iter: Handle compound highmem pages in
> copy_page_from_iter_atomic()
> iomap: Remove large folio handling in iomap_invalidate_folio()
> doc: Correct the description of ->release_folio
> iomap: Remove unnecessary test from iomap_release_folio()
> filemap: Add fgf_t typedef
> filemap: Allow __filemap_get_folio to allocate large folios
> iomap: Create large folios in the buffered write path
> iomap: Copy larger chunks from userspace
>
> Documentation/filesystems/locking.rst | 15 ++++--
> fs/btrfs/file.c | 6 +--
> fs/f2fs/compress.c | 2 +-
> fs/f2fs/f2fs.h | 2 +-
> fs/gfs2/bmap.c | 2 +-
> fs/iomap/buffered-io.c | 43 ++++++++--------
> include/linux/iomap.h | 2 +-
> include/linux/pagemap.h | 71 ++++++++++++++++++++++-----
> lib/iov_iter.c | 43 ++++++++++------
> mm/filemap.c | 61 ++++++++++++-----------
> mm/folio-compat.c | 2 +-
> mm/readahead.c | 13 -----
> 12 files changed, 159 insertions(+), 103 deletions(-)
>
> --
> 2.39.2
next prev parent reply other threads:[~2023-06-21 12:03 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-06-12 20:39 [PATCH v3 0/8] Create large folios in iomap buffered write path Matthew Wilcox (Oracle)
2023-06-12 20:39 ` [PATCH v3 1/8] iov_iter: Handle compound highmem pages in copy_page_from_iter_atomic() Matthew Wilcox (Oracle)
2023-06-13 4:52 ` Christoph Hellwig
2023-07-10 3:36 ` Matthew Wilcox
2023-06-12 20:39 ` [PATCH v3 2/8] iomap: Remove large folio handling in iomap_invalidate_folio() Matthew Wilcox (Oracle)
2023-06-12 20:39 ` [PATCH v3 3/8] doc: Correct the description of ->release_folio Matthew Wilcox (Oracle)
2023-06-13 4:53 ` Christoph Hellwig
2023-06-12 20:39 ` [PATCH v3 4/8] iomap: Remove unnecessary test from iomap_release_folio() Matthew Wilcox (Oracle)
2023-06-13 4:53 ` Christoph Hellwig
2023-06-13 16:19 ` Matthew Wilcox
2023-06-12 20:39 ` [PATCH v3 5/8] filemap: Add fgf_t typedef Matthew Wilcox (Oracle)
2023-06-13 4:53 ` Christoph Hellwig
2023-06-12 20:39 ` [PATCH v3 6/8] filemap: Allow __filemap_get_folio to allocate large folios Matthew Wilcox (Oracle)
2023-06-12 22:49 ` Dave Chinner
2023-06-13 0:42 ` Matthew Wilcox
2023-06-13 1:30 ` Dave Chinner
2023-06-13 2:00 ` Matthew Wilcox
2023-06-13 7:54 ` Dave Chinner
2023-06-13 13:34 ` Matthew Wilcox
2023-06-16 17:45 ` Matthew Wilcox
2023-06-16 22:40 ` Dave Chinner
2023-06-13 4:56 ` Christoph Hellwig
2023-06-12 20:39 ` [PATCH v3 7/8] iomap: Create large folios in the buffered write path Matthew Wilcox (Oracle)
2023-06-13 4:56 ` Christoph Hellwig
2023-06-12 20:39 ` [PATCH v3 8/8] iomap: Copy larger chunks from userspace Matthew Wilcox (Oracle)
2023-06-13 4:58 ` Christoph Hellwig
2023-06-13 19:43 ` Matthew Wilcox
2023-07-10 3:45 ` Matthew Wilcox
2023-06-17 7:13 ` Ritesh Harjani
2023-06-19 17:09 ` Matthew Wilcox
2023-07-10 3:57 ` Matthew Wilcox
2023-06-21 12:03 ` Wang Yugui [this message]
2023-07-10 3:55 ` [PATCH v3 0/8] Create large folios in iomap buffered write path Matthew Wilcox
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=20230621200305.23CB.409509F4@e16-tech.com \
--to=wangyugui@e16-tech.com \
--cc=david@fromorbit.com \
--cc=djwong@kernel.org \
--cc=hch@infradead.org \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-xfs@vger.kernel.org \
--cc=willy@infradead.org \
/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).