From: no-reply@sashiko.virtuozzo.com
To: Konstantin Khorenko <khorenko@virtuozzo.com>,
Liu Kui <kui.liu@virtuozzo.com>
Cc: devel@openvz.org, azaitsev@virtuozzo.com
Subject: Re: [Devel] [PATCH hci-8.1 1/3] fs/fuse kio: teach pcs_umem about hugepage-backed mappings
Date: Wed, 30 Sep 2026 07:41:56 +0000 [thread overview]
Message-ID: <010701a0f1435502-816204c4-752e-4b1f-95ba-4a4f09b5312b-000000@eu-central-1.amazonses.com> (raw)
In-Reply-To: <20260930-sockio-refactor-8-1-v1-1-806348b9eb26@virtuozzo.com>
> pcs_umem pins every mapping as a flat array of 4k pages and the krpc
> bvec builder emits one bio_vec per page, so a 1M chunk costs 256 page
> pointers and 256 bio_vecs even when it sits in a single hugepage.
>
> Detect hugepage backing after pinning: if the range is 2M-aligned and
> each 2M chunk is backed by one folio (hugetlbfs or fully-mapped THP),
> keep only the folio head page per chunk and raise umem->page_shift to
> PMD_SHIFT. The bvec builder then caps bv_len at the backing unit
> instead of PAGE_SIZE, so a 1M chunk in a hugepage umem becomes a
> single bio_vec. Mixed or misaligned mappings stay at 4k.
>
> Also pin with pin_user_pages_fast(), which needs no mmap_lock, and
> fix the pcs_umem_get() error path, which leaked chunk page arrays and
> could deref a NULL page_chunk.
>
> https://virtuozzo.atlassian.net/browse/VSTOR-142166
>
> Signed-off-by: Liu Kui <kui.liu@virtuozzo.com>
Sashiko has reviewed this patch and found no issues. It looks great!
--
Sashiko AI review · https://sashiko.vzint.dev/#/patchset/20260930-sockio-refactor-8-1-v1-0-806348b9eb26@virtuozzo.com?part=1
_______________________________________________
Devel mailing list
Devel@openvz.org
https://lists.openvz.org/mailman/listinfo/devel
next prev parent reply other threads:[~2026-09-30 7:43 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-30 7:34 [Devel] [PATCH hci-8.1 0/3] fs/fuse kio: cut per-4k-page sendmsg/recvmsg calls in the sockio path Liu Kui
2026-09-30 7:34 ` [Devel] [PATCH hci-8.1 1/3] fs/fuse kio: teach pcs_umem about hugepage-backed mappings Liu Kui
2026-09-30 7:41 ` no-reply [this message]
2026-09-30 7:34 ` [Devel] [PATCH hci-8.1 2/3] fs/fuse kio: refactor transmit path in pcs_sockio_send() Liu Kui
2026-09-30 7:42 ` no-reply
2026-09-30 16:29 ` Konstantin Khorenko
2026-09-30 7:34 ` [Devel] [PATCH hci-8.1 3/3] fs/fuse kio: refactor receive path in pcs_sockio_recv() Liu Kui
2026-09-30 7:42 ` no-reply
2026-09-30 16:09 ` [Devel] [PATCH hci-8.1 0/3] fs/fuse kio: cut per-4k-page sendmsg/recvmsg calls in the sockio path Konstantin Khorenko
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=010701a0f1435502-816204c4-752e-4b1f-95ba-4a4f09b5312b-000000@eu-central-1.amazonses.com \
--to=no-reply@sashiko.virtuozzo.com \
--cc=azaitsev@virtuozzo.com \
--cc=devel@openvz.org \
--cc=khorenko@virtuozzo.com \
--cc=kui.liu@virtuozzo.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