From: Ming Lei <ming.lei@redhat.com>
To: Jens Axboe <axboe@kernel.dk>, linux-block@vger.kernel.org
Cc: io-uring@vger.kernel.org,
Caleb Sander Mateos <csander@purestorage.com>,
Nitesh Shetty <nj.shetty@samsung.com>,
Ming Lei <ming.lei@redhat.com>
Subject: [PATCH V2 0/3] block: avoid to use bi_vcnt in bio_may_need_split()
Date: Wed, 31 Dec 2025 11:00:54 +0800 [thread overview]
Message-ID: <20251231030101.3093960-1-ming.lei@redhat.com> (raw)
This series cleans up bio handling to use bi_iter consistently for both
cloned and non-cloned bios, removing the reliance on bi_vcnt which is
only meaningful for non-cloned bios.
Currently, bio_may_need_split() uses bi_vcnt to check if a bio has a
single segment. While this works, it's inconsistent with how cloned bios
operate - they use bi_iter for iteration, not bi_vcnt. This inconsistency
led to io_uring needing to recalculate iov_iter.nr_segs to ensure bi_vcnt
gets a correct value when copied.
This series unifies the approach:
1. Make bio_may_need_split() use bi_iter instead of bi_vcnt. This handles
both cloned and non-cloned bios in a consistent way. Also move bi_io_vec
adjacent to bi_iter in struct bio since they're commonly accessed
together.
2. Stop copying iov_iter.nr_segs to bi_vcnt in bio_iov_bvec_set(), since
cloned bios should rely on bi_iter, not bi_vcnt.
3. Remove the nr_segs recalculation in io_uring, which was only needed
to provide an accurate bi_vcnt value.
Nitesh verified no performance regression on NVMe 512-byte fio/t/io_uring
workloads.
V2:
- improve bio layout by putting bi_iter and bi_io_vec together
- improve commit log
Ming Lei (3):
block: use bvec iterator helper for bio_may_need_split()
block: don't initialize bi_vcnt for cloned bio in bio_iov_bvec_set()
io_uring: remove nr_segs recalculation in io_import_kbuf()
block/bio.c | 5 ++++-
block/blk.h | 12 +++++++++---
include/linux/blk_types.h | 4 ++--
io_uring/rsrc.c | 11 -----------
4 files changed, 15 insertions(+), 17 deletions(-)
--
2.47.0
next reply other threads:[~2025-12-31 3:01 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-12-31 3:00 Ming Lei [this message]
2025-12-31 3:00 ` [PATCH V2 1/3] block: use bvec iterator helper for bio_may_need_split() Ming Lei
2026-01-07 10:38 ` Nitesh Shetty
2025-12-31 3:00 ` [PATCH V2 2/3] block: don't initialize bi_vcnt for cloned bio in bio_iov_bvec_set() Ming Lei
2026-01-07 10:39 ` Nitesh Shetty
2025-12-31 3:00 ` [PATCH V2 3/3] io_uring: remove nr_segs recalculation in io_import_kbuf() Ming Lei
2026-01-07 10:40 ` Nitesh Shetty
2026-01-07 4:11 ` [PATCH V2 0/3] block: avoid to use bi_vcnt in bio_may_need_split() Ming Lei
2026-01-07 15:08 ` Jens Axboe
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=20251231030101.3093960-1-ming.lei@redhat.com \
--to=ming.lei@redhat.com \
--cc=axboe@kernel.dk \
--cc=csander@purestorage.com \
--cc=io-uring@vger.kernel.org \
--cc=linux-block@vger.kernel.org \
--cc=nj.shetty@samsung.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