Linux fsverity development list
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Andrey Albershteyn" <aalbersh@kernel.org>,
	fsverity <fsverity@lists.linux.dev>,
	linux-ext4 <linux-ext4@vger.kernel.org>,
	dsterba@suse.com, "Eric Biggers" <ebiggers@kernel.org>,
	"Christian Brauner" <brauner@kernel.org>
Cc: tytso@mit.edu, chao@kernel.org, jaegeuk@kernel.org,
	viro@zeniv.linux.org.uk, dsterba@suse.cz
Subject: Re: [PATCH v3 4/7] iomap: use fsverity_is_building() to check for fsverity IO
Date: Fri, 20 Nov 2026 16:21:51 +0000	[thread overview]
Message-ID: <sashiko-outbox-225558@kernel.org> (raw)
In-Reply-To: <20261120155032.187246-5-aalbersh@kernel.org>

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [Medium] iomap: inode-wide fsverity_is_building() check in
  iomap_writeback_handle_eof()
- [Low] iomap: missing Fixes trailer for broken fsverity check in
  writeback
--

commit f52ac6dc651e2b12c5452600fd41f0db8111577f
Author: Andrey Albershteyn <aalbersh@kernel.org>

> iomap: use fsverity_is_building() to check for fsverity IO
>
> Recently, inode gained new I_VERITY_BUILDING flag. This flag indicates
> that inode is undergoing fsverity metadata construction. Previously, it
> was indicated by stale iomap flag on the IO, now we can check the inode
> instead. As fsverity inodes are read-only, the only write IO happening
> during I_VERITY_BUILDING is fsverity metadata.
>
> Signed-off-by: Andrey Albershteyn <aalbersh@kernel.org>

[Severity: Low]
This fixes the check that read an uninitialized or stale wpc->iomap
before wpc->ops->writeback_range() mapped the folio.  Should it carry:

  Fixes: 63e242afa466 ("iomap: introduce IOMAP_F_FSVERITY and teach writeback to handle fsverity")

> diff --git a/fs/iomap/buffered-io.c b/fs/iomap/buffered-io.c
> index d6e2f1ded5ca6..4fb4136908ec5 100644
> --- a/fs/iomap/buffered-io.c
> +++ b/fs/iomap/buffered-io.c
> @@ -1930,7 +1930,7 @@ static bool iomap_writeback_handle_eof(struct folio *folio,
>  	struct inode *inode = wpc->inode;
>  	u64 isize = i_size_read(inode);
>  
> -	if (wpc->iomap.flags & IOMAP_F_FSVERITY) {
> +	if (fsverity_is_building(inode)) {
>  		WARN_ON_ONCE(folio_pos(folio) < isize);
>  		return true;
>  	}

[Severity: Medium]
Can writeback of a dirty data folio while I_VERITY_BUILDING is set hit
this WARN_ON_ONCE() and skip post-EOF tail zeroing?

enable_verity() sets I_VERITY_BUILDING without flushing pre-existing
dirty data folios in inode->i_mapping, and filesystems write back both
dirty data and verity metadata folios while I_VERITY_BUILDING is still
set (either during background writeback or in ->end_enable_verity()).
Because fsverity_is_building(inode) is an inode-wide flag checked before
comparing the folio offset against isize, any dirty regular data folio
written back during verity construction triggers WARN_ON_ONCE(folio_pos(folio) < isize),
and a dirty data folio straddling isize returns early without running
folio_zero_segment() or clamping *end_pos to isize.

Should the fsverity_is_building(inode) exemption apply only when the
folio lies entirely at or beyond isize?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261120155032.187246-1-aalbersh@kernel.org?part=4

  parent reply	other threads:[~2026-11-20 16:21 UTC|newest]

Thread overview: 14+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-20 15:50 [PATCH v3 0/7] fsverity: lift ioctl() input masking and runtime inode flag to VFS Andrey Albershteyn
2026-11-20 15:50 ` [PATCH v3 1/7] fs: constify inode argument of inode_state_read_once() Andrey Albershteyn
2026-11-20 15:50 ` [PATCH v3 2/7] fs: strip FS_VERITY_FL from FS_IOC_SETFLAGS Andrey Albershteyn
2026-11-20 22:20   ` Eric Biggers
2026-11-20 15:50 ` [PATCH v3 3/7] fs,fsverity: add I_VERITY_BUILDING inode state flag Andrey Albershteyn
2026-11-20 22:17   ` Eric Biggers
2026-11-20 15:50 ` [PATCH v3 4/7] iomap: use fsverity_is_building() to check for fsverity IO Andrey Albershteyn
2026-11-20 16:13   ` sashiko-bot
2026-11-20 16:21   ` sashiko-bot [this message]
2026-11-20 16:42     ` Andrey Albershteyn
2026-11-20 22:23   ` Eric Biggers
2026-11-20 15:50 ` [PATCH v3 5/7] btrfs: use I_VERITY_BUILDING instead of private flags Andrey Albershteyn
2026-11-20 15:50 ` [PATCH v3 6/7] ext4: " Andrey Albershteyn
2026-11-20 15:50 ` [PATCH v3 7/7] f2fs: " Andrey Albershteyn

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=sashiko-outbox-225558@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=aalbersh@kernel.org \
    --cc=brauner@kernel.org \
    --cc=chao@kernel.org \
    --cc=dsterba@suse.com \
    --cc=dsterba@suse.cz \
    --cc=ebiggers@kernel.org \
    --cc=fsverity@lists.linux.dev \
    --cc=jaegeuk@kernel.org \
    --cc=linux-ext4@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=tytso@mit.edu \
    --cc=viro@zeniv.linux.org.uk \
    /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