linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Dave Chinner <david@fromorbit.com>
To: "Darrick J. Wong" <darrick.wong@oracle.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 4/6 v2] xfs: validate btree records on retreival
Date: Wed, 6 Jun 2018 09:02:46 +1000	[thread overview]
Message-ID: <20180605230246.GL10363@dastard> (raw)
In-Reply-To: <20180605174716.GM9437@magnolia>

On Tue, Jun 05, 2018 at 10:47:16AM -0700, Darrick J. Wong wrote:
> On Tue, Jun 05, 2018 at 04:40:43PM +1000, Dave Chinner wrote:
> > @@ -111,16 +111,51 @@ xfs_refcount_get_rec(
> >  	struct xfs_refcount_irec	*irec,
> >  	int				*stat)
> >  {
> > +	struct xfs_mount		*mp = cur->bc_mp;
> > +	xfs_agnumber_t			agno = cur->bc_private.a.agno;
> >  	union xfs_btree_rec		*rec;
> >  	int				error;
> > +	xfs_agblock_t			realstart;
> >  
> >  	error = xfs_btree_get_rec(cur, &rec, stat);
> > -	if (!error && *stat == 1) {
> > -		xfs_refcount_btrec_to_irec(rec, irec);
> > -		trace_xfs_refcount_get(cur->bc_mp, cur->bc_private.a.agno,
> > -				irec);
> > +	if (error || !*stat)
> > +		return error;
> > +
> > +	xfs_refcount_btrec_to_irec(rec, irec);
> > +
> > +	agno = cur->bc_private.a.agno;
> > +	if (irec->rc_blockcount == 0 || irec->rc_blockcount > MAXREFCEXTLEN)
> > +		goto out_bad_rec;
> > +
> > +	/* handle special COW-staging state */
> > +	realstart = irec->rc_startblock;
> > +	if (realstart & XFS_REFC_COW_START) {
> > +		if (irec->rc_refcount != 1)
> > +			goto out_bad_rec;
> > +		realstart &= ~XFS_REFC_COW_START;
> >  	}
> 
> If the record does not have the cow "flag" set then the refcount cannot
> be 1, so this needs an else clause:
> 
> } else {
> 	if (irec->rc_refcount == 1)
> 		goto out_bad_rec;
> }

Ah, yes, that is true, though it should be refcount < 2 because a
zero value is invalid, too.

> > +	if (irec->rm_owner == 0)
> > +		goto out_bad_rec;
> > +	if (irec->rm_owner > XFS_MAXINUMBER &&
> > +	    irec->rm_owner <= XFS_RMAP_OWN_MIN)
> 
> rm_owner should be (between XFS_RMAP_OWN_FS and XFS_RMAP_OWN_MIN) or
> (pass xfs_verify_fsino()) since we cannot have inodes between EOFS and
> MAXINUMBER.

Yes, that is better. I should have noticed xfs_verify_fsino() - my
bad.
>
> > +		goto out_bad_rec;
> > +
> 
> Should we check the rmap flags here?

They are checked in xfs_refcount_btrec_to_irec(), and it returns
-EFSCORRUPTED if invalid flags are set.

> > +	return 0;
> > +out_bad_rec:
> > +	xfs_warn(mp,
> > +		"RMAP BTree record corruption in AG %d detected!", agno);
> 
> "RMap" ?  Or "Reverse Mapping"?
> 
> (Consistent capitalization blah blah...)

I'll use the latter.

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

  reply	other threads:[~2018-06-05 23:03 UTC|newest]

Thread overview: 36+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-05  6:24 [PATCH 0/6 V2] xfs: more verifications! Dave Chinner
2018-06-05  6:24 ` [PATCH 1/6] xfs: catch bad stripe alignment configurations Dave Chinner
2018-06-05  9:27   ` Carlos Maiolino
2018-06-05  6:24 ` [PATCH 2/6] xfs: verify extent size hint is valid in inode verifier Dave Chinner
2018-06-05  9:53   ` Carlos Maiolino
2018-06-05 22:56     ` Dave Chinner
2018-06-05 17:10   ` Darrick J. Wong
2018-06-07 16:16     ` Darrick J. Wong
2018-06-08  1:10       ` Dave Chinner
2018-06-08  1:23         ` Darrick J. Wong
2018-06-08  2:23           ` Eric Sandeen
2018-07-24  6:39   ` Eric Sandeen
2018-07-24 16:43     ` Darrick J. Wong
2018-08-20 15:06       ` Brian Foster
2018-08-20 15:27         ` Eric Sandeen
2018-08-20 15:36           ` Darrick J. Wong
2018-08-20 15:59             ` Brian Foster
2018-08-20 22:15               ` Dave Chinner
2018-08-21 10:56                 ` Brian Foster
2018-08-22  0:41                   ` Dave Chinner
2018-06-05  6:24 ` [PATCH 3/6] xfs: verify COW " Dave Chinner
2018-06-05 10:00   ` Carlos Maiolino
2018-06-05 17:09   ` Darrick J. Wong
2018-06-05  6:24 ` [PATCH 4/6] xfs: validate btree records on retreival Dave Chinner
2018-06-05  6:40   ` [PATCH 4/6 v2] " Dave Chinner
2018-06-05 10:42     ` Carlos Maiolino
2018-06-05 23:00       ` Dave Chinner
2018-06-05 17:47     ` Darrick J. Wong
2018-06-05 23:02       ` Dave Chinner [this message]
2018-06-06  1:21     ` [PATCH 4/6 v3] " Dave Chinner
2018-06-05  6:24 ` [PATCH 5/6] xfs: verify root inode more thoroughly Dave Chinner
2018-06-05 10:50   ` Carlos Maiolino
2018-06-05 17:10   ` Darrick J. Wong
2018-06-05  6:24 ` [PATCH 6/6] xfs: push corruption -> ESTALE conversion to xfs_nfs_get_inode() Dave Chinner
2018-06-05 11:12   ` Carlos Maiolino
2018-06-05 17:11   ` Darrick J. Wong

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=20180605230246.GL10363@dastard \
    --to=david@fromorbit.com \
    --cc=darrick.wong@oracle.com \
    --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;
as well as URLs for NNTP newsgroup(s).