Linux EXT4 FS development
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Tomás Senart" <ts@perfloop.ai>
Cc: linux-ext4@vger.kernel.org, tytso@mit.edu
Subject: Re: [PATCH] jbd2: stop fast commit replay at the end of the fast commit area
Date: Sat, 21 Nov 2026 03:13:23 +0000	[thread overview]
Message-ID: <sashiko-outbox-226561@kernel.org> (raw)
In-Reply-To: <CAEEBa3XJt0XU+5pXNQz_tjZyPEeEnY67i_QGupOnHLnXB=BVDw@mail.gmail.com>

Thank you for your contribution! Sashiko AI review found 1 potential
issue to consider:
- [Critical] jbd2: skipped ext4_fc_set_bitmaps_and_counters() when fast
  commit area is full
--

commit c70236779e98a78f5adca06274fb30cb8d0a0f06
Author: Tomás Senart <ts@perfloop.ai>

jbd2: stop fast commit replay at the end of the fast commit area

This commit changes the loop bound in fc_do_one_pass() from
<= journal->j_fc_last to < journal->j_fc_last so fast commit recovery
stops before reading past the end of the fast commit area.

> diff --git a/fs/jbd2/recovery.c b/fs/jbd2/recovery.c
> index cac8c2cd4a92a..a59de33a2a4a9 100644
> --- a/fs/jbd2/recovery.c
> +++ b/fs/jbd2/recovery.c
> @@ -242,7 +242,7 @@ static int fc_do_one_pass(journal_t *journal,
>  	if (!journal->j_fc_replay_callback)
>  		return 0;
>  
> -	while (next_fc_block <= journal->j_fc_last) {
> +	while (next_fc_block < journal->j_fc_last) {
>  		jbd2_debug(3, "Fast commit replay: next block %ld\n",
>  			  next_fc_block);
>  		err = jread(&bh, journal, next_fc_block);

[Severity: Critical]
When the fast commit area is full, does exiting the loop at
next_fc_block == journal->j_fc_last skip
ext4_fc_set_bitmaps_and_counters()?

In ext4_fc_replay(), state->fc_replay_num_tags is checked for zero only
at the top of the function and before decrementing it for the current tag:

fs/ext4/fast_commit.c:ext4_fc_replay() {
    ...
	if (!sbi->s_fc_replay_state.fc_replay_num_tags) {
		ext4_debug("Replay stops\n");
		return ext4_fc_set_bitmaps_and_counters(sb);
	}
    ...
	for (cur = start; cur < end;
	     cur = cur + EXT4_FC_TAG_BASE_LEN + tl.fc_len) {
		ext4_fc_get_tl(&tl, cur);

		if (state->fc_replay_num_tags == 0) {
			ret = ext4_fc_set_bitmaps_and_counters(sb);
			break;
		}

		ext4_debug("Replay phase, tag:%s\n", tag2str(tl.fc_tag));
		state->fc_replay_num_tags--;
    ...
}

Because ext4_fc_write_tail() sets the length of EXT4_FC_TAG_TAIL to cover
the rest of the block, the final tag in block journal->j_fc_last - 1
decrements state->fc_replay_num_tags from 1 to 0 and immediately finishes
the block loop, returning JBD2_FC_REPLAY_CONTINUE.

When the area is not full, fc_do_one_pass() invokes j_fc_replay_callback
on the next block, hitting !sbi->s_fc_replay_state.fc_replay_num_tags at
the start of ext4_fc_replay().  When the last usable block
(journal->j_fc_last - 1) holds the final EXT4_FC_TAG_TAIL, the while loop
in fc_do_one_pass() now terminates before calling j_fc_replay_callback
again.

During PASS_REPLAY, ext4_fc_replay_inode() clears the on-disk block
bitmaps for modified inodes via ext4_ext_clear_bb() and relies on
ext4_fc_set_bitmaps_and_counters() to mark their data and extent blocks
allocated again, so skipping it leaves live blocks marked free on disk.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/CAEEBa3XJt0XU+5pXNQz_tjZyPEeEnY67i_QGupOnHLnXB=BVDw@mail.gmail.com?part=1

      reply	other threads:[~2026-11-21  3:13 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-21  3:03 [PATCH] jbd2: stop fast commit replay at the end of the fast commit area Tomás Senart
2026-11-21  3:13 ` sashiko-bot [this message]

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-226561@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=linux-ext4@vger.kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    --cc=ts@perfloop.ai \
    --cc=tytso@mit.edu \
    /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