From: Dave Chinner <david@fromorbit.com>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 1/2] xfs: hoist refcount record merge predicates
Date: Wed, 30 Nov 2022 09:35:31 +1100 [thread overview]
Message-ID: <20221129223531.GJ3600936@dread.disaster.area> (raw)
In-Reply-To: <166975929118.3768925.9568770405264708473.stgit@magnolia>
On Tue, Nov 29, 2022 at 02:01:31PM -0800, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
>
> Hoist these multiline conditionals into separate static inline helpers
> to improve readability and set the stage for corruption fixes that will
> be introduced in the next patch.
>
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---
> fs/xfs/libxfs/xfs_refcount.c | 126 +++++++++++++++++++++++++++++++++++++-----
> 1 file changed, 110 insertions(+), 16 deletions(-)
Looks OK. Minor nit below.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
>
>
> diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
> index 3f34bafe18dd..8c68994d09f3 100644
> --- a/fs/xfs/libxfs/xfs_refcount.c
> +++ b/fs/xfs/libxfs/xfs_refcount.c
> @@ -815,11 +815,116 @@ xfs_refcount_find_right_extents(
> /* Is this extent valid? */
> static inline bool
> xfs_refc_valid(
> - struct xfs_refcount_irec *rc)
> + const struct xfs_refcount_irec *rc)
> {
> return rc->rc_startblock != NULLAGBLOCK;
> }
>
> +static inline bool
> +xfs_refc_want_merge_center(
> + const struct xfs_refcount_irec *left,
> + const struct xfs_refcount_irec *cleft,
> + const struct xfs_refcount_irec *cright,
> + const struct xfs_refcount_irec *right,
> + bool cleft_is_cright,
> + enum xfs_refc_adjust_op adjust,
> + unsigned long long *ulenp)
> +{
> + unsigned long long ulen = left->rc_blockcount;
> +
> + /*
> + * To merge with a center record, both shoulder records must be
> + * adjacent to the record we want to adjust. This is only true if
> + * find_left and find_right made all four records valid.
> + */
> + if (!xfs_refc_valid(left) || !xfs_refc_valid(right) ||
> + !xfs_refc_valid(cleft) || !xfs_refc_valid(cright))
> + return false;
> +
> + /* There must only be one record for the entire range. */
> + if (!cleft_is_cright)
> + return false;
> +
> + /* The shoulder record refcounts must match the new refcount. */
> + if (left->rc_refcount != cleft->rc_refcount + adjust)
> + return false;
> + if (right->rc_refcount != cleft->rc_refcount + adjust)
> + return false;
> +
> + /*
> + * The new record cannot exceed the max length. The funny computation
> + * of ulen avoids casting.
> + */
> + ulen += cleft->rc_blockcount + right->rc_blockcount;
> + if (ulen >= MAXREFCEXTLEN)
> + return false;
The comment took me a bit of spelunking to decipher what the "funny
computation" was. Better to spell it out directly (catch u32
overflows) than just hint that there's somethign special about it.
Say:
/*
* The new record cannot exceed the max length. ulen is a
* ULL as the individual record block counts can be up to
* (u32 - 1) in length hence we need to catch u32 addition
* overflows here.
*/
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
next prev parent reply other threads:[~2022-11-29 22:35 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-11-29 22:01 [PATCHSET 0/2] xfs: fix broken MAXREFCOUNT handling Darrick J. Wong
2022-11-29 22:01 ` [PATCH 1/2] xfs: hoist refcount record merge predicates Darrick J. Wong
2022-11-29 22:35 ` Dave Chinner [this message]
2022-11-30 0:13 ` Darrick J. Wong
2022-11-30 9:24 ` Yang, Xiao/杨 晓
2022-11-29 22:01 ` [PATCH 2/2] xfs: estimate post-merge refcounts correctly Darrick J. Wong
2022-11-29 22:37 ` Dave Chinner
2022-11-30 9:32 ` Yang, Xiao/杨 晓
2022-11-30 18:49 ` Darrick J. Wong
2022-11-29 22:06 ` [RFC PATCH] xfs/179: modify test to trigger refcount update bugs Darrick J. Wong
2022-11-29 22:42 ` Dave Chinner
2022-11-30 0:19 ` Darrick J. Wong
2022-11-30 10:07 ` Yang, Xiao/杨 晓
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=20221129223531.GJ3600936@dread.disaster.area \
--to=david@fromorbit.com \
--cc=djwong@kernel.org \
--cc=linux-xfs@vger.kernel.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