From: Dave Chinner <david@fromorbit.com>
To: "Darrick J. Wong" <djwong@kernel.org>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 09/12] xfs: check record domain when accessing refcount records
Date: Fri, 28 Oct 2022 08:15:31 +1100 [thread overview]
Message-ID: <20221027211531.GW3600936@dread.disaster.area> (raw)
In-Reply-To: <166689089384.3788582.15595498616742667720.stgit@magnolia>
On Thu, Oct 27, 2022 at 10:14:53AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
>
> Now that we've separated the startblock and CoW/shared extent domain in
> the incore refcount record structure, check the domain whenever we
> retrieve a record to ensure that it's still in the domain that we want.
> Depending on the circumstances, a change in domain either means we're
> done processing or that we've found a corruption and need to fail out.
>
> The refcount check in xchk_xref_is_cow_staging is redundant since
> _get_rec has done that for a long time now, so we can get rid of it.
>
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---
> fs/xfs/libxfs/xfs_refcount.c | 53 ++++++++++++++++++++++++++++++++----------
> fs/xfs/scrub/refcount.c | 4 ++-
> 2 files changed, 43 insertions(+), 14 deletions(-)
>
>
> diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
> index 3b1cb0578770..608a122eef16 100644
> --- a/fs/xfs/libxfs/xfs_refcount.c
> +++ b/fs/xfs/libxfs/xfs_refcount.c
> @@ -386,6 +386,8 @@ xfs_refcount_split_extent(
> goto out_error;
> }
>
> + if (rcext.rc_domain != domain)
> + return 0;
> if (rcext.rc_startblock == agbno || xfs_refc_next(&rcext) <= agbno)
> return 0;
>
> @@ -434,6 +436,9 @@ xfs_refcount_merge_center_extents(
> int error;
> int found_rec;
>
> + ASSERT(left->rc_domain == center->rc_domain);
> + ASSERT(right->rc_domain == center->rc_domain);
> +
> trace_xfs_refcount_merge_center_extents(cur->bc_mp,
> cur->bc_ag.pag->pag_agno, left, center, right);
Can you move the asserts to after the trace points? That way we if
need to debug the assert, we'll have a tracepoint with the record
information that triggered the assert emitted immediately before it
fires...
>
> @@ -510,6 +515,8 @@ xfs_refcount_merge_left_extent(
> int error;
> int found_rec;
>
> + ASSERT(left->rc_domain == cleft->rc_domain);
> +
> trace_xfs_refcount_merge_left_extent(cur->bc_mp,
> cur->bc_ag.pag->pag_agno, left, cleft);
>
> @@ -571,6 +578,8 @@ xfs_refcount_merge_right_extent(
> int error;
> int found_rec;
>
> + ASSERT(right->rc_domain == cright->rc_domain);
> +
> trace_xfs_refcount_merge_right_extent(cur->bc_mp,
> cur->bc_ag.pag->pag_agno, cright, right);
>
> @@ -654,12 +663,10 @@ xfs_refcount_find_left_extents(
> goto out_error;
> }
>
> + if (tmp.rc_domain != domain)
> + return 0;
> if (xfs_refc_next(&tmp) != agbno)
> return 0;
> - if (domain == XFS_REFC_DOMAIN_SHARED && tmp.rc_refcount < 2)
> - return 0;
> - if (domain == XFS_REFC_DOMAIN_COW && tmp.rc_refcount > 1)
> - return 0;
Ah, as these go away, you can ignore my comment about them in the
previous patches... :)
Otherwise, looks ok.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
next prev parent reply other threads:[~2022-10-27 21:15 UTC|newest]
Thread overview: 30+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-10-27 17:14 [PATCHSET v2 00/12] xfs: improve runtime refcountbt corruption detection Darrick J. Wong
2022-10-27 17:14 ` [PATCH 01/12] xfs: make sure aglen never goes negative in xfs_refcount_adjust_extents Darrick J. Wong
2022-10-27 20:41 ` Dave Chinner
2022-10-27 17:14 ` [PATCH 02/12] xfs: check deferred refcount op continuation parameters Darrick J. Wong
2022-10-27 20:49 ` Dave Chinner
2022-10-27 21:32 ` Darrick J. Wong
2022-10-27 21:42 ` Darrick J. Wong
2022-10-27 22:24 ` Dave Chinner
2022-10-27 23:25 ` Darrick J. Wong
2022-10-27 21:54 ` [PATCH v2.1 " Darrick J. Wong
2022-10-27 17:14 ` [PATCH 03/12] xfs: move _irec structs to xfs_types.h Darrick J. Wong
2022-10-27 17:14 ` [PATCH 04/12] xfs: refactor refcount record usage in xchk_refcountbt_rec Darrick J. Wong
2022-10-27 17:14 ` [PATCH 05/12] xfs: track cow/shared record domains explicitly in xfs_refcount_irec Darrick J. Wong
2022-10-27 21:03 ` Dave Chinner
2022-10-27 21:10 ` Darrick J. Wong
2022-10-27 17:14 ` [PATCH 06/12] xfs: report refcount domain in tracepoints Darrick J. Wong
2022-10-27 21:05 ` Dave Chinner
2022-10-27 17:14 ` [PATCH 07/12] xfs: refactor domain and refcount checking Darrick J. Wong
2022-10-27 21:07 ` Dave Chinner
2022-10-27 17:14 ` [PATCH 08/12] xfs: remove XFS_FIND_RCEXT_SHARED and _COW Darrick J. Wong
2022-10-27 21:11 ` Dave Chinner
2022-10-27 17:14 ` [PATCH 09/12] xfs: check record domain when accessing refcount records Darrick J. Wong
2022-10-27 21:15 ` Dave Chinner [this message]
2022-10-27 21:33 ` Darrick J. Wong
2022-10-27 17:14 ` [PATCH 10/12] xfs: fix agblocks check in the cow leftover recovery function Darrick J. Wong
2022-10-27 21:22 ` Dave Chinner
2022-10-27 17:15 ` [PATCH 11/12] xfs: fix uninitialized list head in struct xfs_refcount_recovery Darrick J. Wong
2022-10-27 21:24 ` Dave Chinner
2022-10-27 17:15 ` [PATCH 12/12] xfs: rename XFS_REFC_COW_START to _COWFLAG Darrick J. Wong
2022-10-27 21:25 ` Dave Chinner
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=20221027211531.GW3600936@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