From: Eric Sandeen <sandeen@sandeen.net>
To: xfs@oss.sgi.com
Subject: Re: [PATCH 2/6] xfs: shadow agfl indexes in the per-ag structures
Date: Fri, 2 Sep 2016 08:25:12 -0500 [thread overview]
Message-ID: <d32c1b10-107c-a1fc-3864-76223661ab71@sandeen.net> (raw)
In-Reply-To: <1472783257-15941-3-git-send-email-david@fromorbit.com>
On 9/1/16 9:27 PM, Dave Chinner wrote:
> From: Dave Chinner <dchinner@redhat.com>
>
> To verify that the AGFL contents is sane, we need to have access to
> the indexes that tell us what part of the AGFL is active. We cannot
> access the AGF buffer from the AGFL verifier, so we need to shadow
> these values in the struct xfs_perag so we check them when required.
>
> Signed-off-by: Dave Chinner <dchinner@redhat.com>
> ---
> fs/xfs/libxfs/xfs_alloc.c | 4 ++++
> fs/xfs/xfs_mount.h | 2 ++
> 2 files changed, 6 insertions(+)
>
> diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
> index 23559b9..1aef556 100644
> --- a/fs/xfs/libxfs/xfs_alloc.c
> +++ b/fs/xfs/libxfs/xfs_alloc.c
> @@ -2255,6 +2255,7 @@ xfs_alloc_get_freelist(
> be32_add_cpu(&agf->agf_flcount, -1);
> xfs_trans_agflist_delta(tp, -1);
> pag->pagf_flcount--;
> + pag->pagf_flfirst = be32_to_cpu(agf->agf_flfirst);
> xfs_perag_put(pag);
>
> logflags = XFS_AGF_FLFIRST | XFS_AGF_FLCOUNT;
Still reviewing, but this kind of jumped out at me, seems like the get/put
functions are a bit jumbled up:
/*
* Get the block number and update the data structures.
*/
agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
bno = be32_to_cpu(agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
be32_add_cpu(&agf->agf_flfirst, 1);
xfs_trans_brelse(tp, agflbp);
if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp))
agf->agf_flfirst = 0;
pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
be32_add_cpu(&agf->agf_flcount, -1);
xfs_trans_agflist_delta(tp, -1);
pag->pagf_flcount--;
xfs_perag_put(pag);
why is there a trans_brelse in between two lines which handle proper setting
of flfirst? I was looking at this thinking about where the pag structures
get updated, then saw the kind of swiss-cheese placement of the first/count
manipulation...
If pagf_flcount/agf_flcount, pagf_flfirst/agf_flfirst etc need to stay in
sync, should there be a wrapper to encapsulate it all?
like:
xfs_agf_{advance/drop/remove}_first(mp, agf, pagf)
{
be32_add_cpu(&agf->agf_flfirst, 1);
if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp))
agf->agf_flfirst = 0;
pag->pagf_flfirst = be32_to_cpu(agf->agf_flfirst);
be32_add_cpu(&agf->agf_flcount, -1);
pag->pagf_flcount--;
}
or something similar?
-Eric
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
next prev parent reply other threads:[~2016-09-02 13:25 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-09-02 2:27 [RFC PATCH 0/6] xfs: sort out the AGFL size mess Dave Chinner
2016-09-02 2:27 ` [PATCH 1/6] xfs: clean up XFS_AGFL_SIZE Dave Chinner
2018-01-03 22:27 ` Darrick J. Wong
2016-09-02 2:27 ` [PATCH 2/6] xfs: shadow agfl indexes in the per-ag structures Dave Chinner
2016-09-02 13:25 ` Eric Sandeen [this message]
2016-09-02 23:06 ` Dave Chinner
2016-09-02 2:27 ` [PATCH 3/6] xfs: detect AGFL size mismatches Dave Chinner
2016-09-13 23:39 ` Eric Sandeen
2016-09-14 21:26 ` Dave Chinner
2016-09-28 17:39 ` Eric Sandeen
2016-09-02 2:27 ` [PATCH 4/6] xfs: automatically fix up AGFL size issues Dave Chinner
2016-09-14 18:20 ` Darrick J. Wong
2016-09-14 21:50 ` Dave Chinner
2016-09-02 2:27 ` [PATCH 5/6] xfs: clean up AGFL index initialisation in growfs Dave Chinner
2016-09-02 2:27 ` [PATCH 6/6] xfs: use per-ag AGFL indexes everywhere 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=d32c1b10-107c-a1fc-3864-76223661ab71@sandeen.net \
--to=sandeen@sandeen.net \
--cc=xfs@oss.sgi.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.