linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Allison Collins <allison.henderson@oracle.com>
Cc: sandeen@sandeen.net, linux-xfs@vger.kernel.org
Subject: Re: [PATCH 4/4] xfs_repair: don't corrupt a attr fork da3 node when clearing forw/back
Date: Tue, 4 Feb 2020 21:59:13 -0800	[thread overview]
Message-ID: <20200205055913.GD6870@magnolia> (raw)
In-Reply-To: <7baaea32-91dc-353d-6de2-ccae5bd79a52@oracle.com>

On Tue, Feb 04, 2020 at 10:29:26PM -0700, Allison Collins wrote:
> 
> 
> On 2/4/20 5:46 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > In process_longform_attr, we enforce that the root block of the
> > attribute index must have both forw or back pointers set to zero.
> > Unfortunately, the code that nulls out the pointers is not aware that
> > the root block could be in da3 node format.
> > 
> > This leads to corruption of da3 root node blocks because the functions
> > that convert attr3 leaf headers to and from the ondisk structures
> > perform some interpretation of firstused on what they think is an attr1
> > leaf block.
> > 
> > Found by using xfs/402 to fuzz hdr.info.hdr.forw.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> >   repair/attr_repair.c |  181 ++++++++++++++++++++++++++++++++------------------
> >   1 file changed, 117 insertions(+), 64 deletions(-)
> > 
> > 
> > diff --git a/repair/attr_repair.c b/repair/attr_repair.c
> > index 7b26df33..535fcfbb 100644
> > --- a/repair/attr_repair.c
> > +++ b/repair/attr_repair.c
> > @@ -952,6 +952,106 @@ _("wrong FS UUID, inode %" PRIu64 " attr block %" PRIu64 "\n"),
> >   	return 0;
> >   }
> > +static int
> > +process_longform_leaf_root(
> > +	struct xfs_mount	*mp,
> > +	xfs_ino_t		ino,
> > +	struct xfs_dinode	*dip,
> > +	struct blkmap		*blkmap,
> > +	int			*repair,
> > +	struct xfs_buf		*bp)
> > +{
> > +	struct xfs_attr3_icleaf_hdr leafhdr;
> > +	xfs_dahash_t		next_hashval;
> > +	int			badness;
> "badness" pretty much just looks like "error" here?  Or is it different in
> some way?

The return value for process_leaf_attr_block is 1 if the attr block is
bad and 0 for the attr block is ok.  It's not the usual -EFUBAR error
codes that we use in many other parts of xfs.

> > +	int			repairlinks = 0;
> > +
> > +	/*
> > +	 * check sibling pointers in leaf block or root block 0 before
> > +	 * we have to release the btree block
> > +	 */
> > +	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, bp->b_addr);
> > +	if (leafhdr.forw != 0 || leafhdr.back != 0)  {
> > +		if (!no_modify)  {
> > +			do_warn(
> > +_("clearing forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"),
> > +				ino);
> > +			repairlinks = 1;
> > +			leafhdr.forw = 0;
> > +			leafhdr.back = 0;
> > +			xfs_attr3_leaf_hdr_to_disk(mp->m_attr_geo, bp->b_addr,
> > +					&leafhdr);
> > +		} else  {
> > +			do_warn(
> > +_("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"), ino);
> > +		}
> > +	}
> > +
> > +	badness = process_leaf_attr_block(mp, bp->b_addr, 0, ino, blkmap, 0,
> > +			&next_hashval, repair);
> > +	if (badness) {
> > +		*repair = 0;
> > +		/* the block is bad.  lose the attribute fork. */
> > +		libxfs_putbuf(bp);
> > +		return 1;
> > +	}
> > +
> > +	*repair = *repair || repairlinks;
> > +
> > +	if (*repair && !no_modify)
> > +		libxfs_writebuf(bp, 0);
> > +	else
> > +		libxfs_putbuf(bp);
> > +
> > +	return 0;
> > +}
> > +
> > +static int
> > +process_longform_da_root(
> > +	struct xfs_mount	*mp,
> > +	xfs_ino_t		ino,
> > +	struct xfs_dinode	*dip,
> > +	struct blkmap		*blkmap,
> > +	int			*repair,
> > +	struct xfs_buf		*bp)
> > +{
> > +	struct xfs_da3_icnode_hdr	da3_hdr;
> > +	int			repairlinks = 0;
> > +	int			error;
> > +
> > +	libxfs_da3_node_hdr_from_disk(mp, &da3_hdr, bp->b_addr);
> > +	/*
> > +	 * check sibling pointers in leaf block or root block 0 before
> > +	 * we have to release the btree block
> > +	 */
> > +	if (da3_hdr.forw != 0 || da3_hdr.back != 0)  {
> > +		if (!no_modify)  {
> > +			do_warn(
> > +_("clearing forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"),
> > +				ino);
> > +
> > +			repairlinks = 1;
> > +			da3_hdr.forw = 0;
> > +			da3_hdr.back = 0;
> > +			xfs_da3_node_hdr_to_disk(mp, bp->b_addr, &da3_hdr);
> > +		} else  {
> > +			do_warn(
> > +_("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"), ino);
> > +		}
> > +	}
> > +
> > +	/* must do this now, to release block 0 before the traversal */
> Did you mean to reference *repair here without setting it?  I guess it was
> like that from the code it was taken from, but it makes it looks like repair
> is both an in and an out.  I wonder if it's really needed in the check
> below?

*repair is passed from a stack variable in process_inode_attr_fork ->
process_attributes -> process_longform_attr.  It's initialized properly,
but the code doesn't make it easy to figure that out since that's three
functions up in the call stack and anyone is allowed to mess with it.

One of the grottier bits of xfs_repair is that the return values for
those functions also have meaning...I think return 1 means "zap this
whole thing" vs. *repair ==1 means "we fixed it".

--D

> 
> Allison
> 
> > +	if ((*repair || repairlinks) && !no_modify) {
> > +		*repair = 1;
> > +		libxfs_writebuf(bp, 0);
> > +	} else
> > +		libxfs_putbuf(bp);
> > +	error = process_node_attr(mp, ino, dip, blkmap); /* + repair */
> > +	if (error)
> > +		*repair = 0;
> > +	return error;
> > +}
> > +
> >   /*
> >    * Start processing for a leaf or fuller btree.
> >    * A leaf directory is one where the attribute fork is too big for
> > @@ -963,19 +1063,15 @@ _("wrong FS UUID, inode %" PRIu64 " attr block %" PRIu64 "\n"),
> >    */
> >   static int
> >   process_longform_attr(
> > -	xfs_mount_t	*mp,
> > -	xfs_ino_t	ino,
> > -	xfs_dinode_t	*dip,
> > -	blkmap_t	*blkmap,
> > -	int		*repair)	/* out - 1 if something was fixed */
> > +	struct xfs_mount	*mp,
> > +	xfs_ino_t		ino,
> > +	struct xfs_dinode	*dip,
> > +	struct blkmap		*blkmap,
> > +	int			*repair) /* out - 1 if something was fixed */
> >   {
> > -	xfs_attr_leafblock_t	*leaf;
> > -	xfs_fsblock_t	bno;
> > -	xfs_buf_t	*bp;
> > -	xfs_dahash_t	next_hashval;
> > -	int		repairlinks = 0;
> > -	struct xfs_attr3_icleaf_hdr leafhdr;
> > -	int		error;
> > +	xfs_fsblock_t		bno;
> > +	struct xfs_buf		*bp;
> > +	struct xfs_da_blkinfo	*info;
> >   	*repair = 0;
> > @@ -1015,74 +1111,31 @@ process_longform_attr(
> >   		return 1;
> >   	}
> > -	/* verify leaf block */
> > -	leaf = bp->b_addr;
> > -	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
> > -
> > -	/* check sibling pointers in leaf block or root block 0 before
> > -	* we have to release the btree block
> > -	*/
> > -	if (leafhdr.forw != 0 || leafhdr.back != 0)  {
> > -		if (!no_modify)  {
> > -			do_warn(
> > -	_("clearing forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"),
> > -				ino);
> > -			repairlinks = 1;
> > -			leafhdr.forw = 0;
> > -			leafhdr.back = 0;
> > -			xfs_attr3_leaf_hdr_to_disk(mp->m_attr_geo,
> > -						   leaf, &leafhdr);
> > -		} else  {
> > -			do_warn(
> > -	_("would clear forw/back pointers in block 0 for attributes in inode %" PRIu64 "\n"), ino);
> > -		}
> > -	}
> > -
> >   	/*
> >   	 * use magic number to tell us what type of attribute this is.
> >   	 * it's possible to have a node or leaf attribute in either an
> >   	 * extent format or btree format attribute fork.
> >   	 */
> > -	switch (leafhdr.magic) {
> > +	info = bp->b_addr;
> > +	switch (be16_to_cpu(info->magic)) {
> >   	case XFS_ATTR_LEAF_MAGIC:	/* leaf-form attribute */
> >   	case XFS_ATTR3_LEAF_MAGIC:
> > -		if (process_leaf_attr_block(mp, leaf, 0, ino, blkmap,
> > -				0, &next_hashval, repair)) {
> > -			*repair = 0;
> > -			/* the block is bad.  lose the attribute fork. */
> > -			libxfs_putbuf(bp);
> > -			return(1);
> > -		}
> > -		*repair = *repair || repairlinks;
> > -		break;
> > -
> > +		return process_longform_leaf_root(mp, ino, dip, blkmap, repair,
> > +				bp);
> >   	case XFS_DA_NODE_MAGIC:		/* btree-form attribute */
> >   	case XFS_DA3_NODE_MAGIC:
> > -		/* must do this now, to release block 0 before the traversal */
> > -		if ((*repair || repairlinks) && !no_modify) {
> > -			*repair = 1;
> > -			libxfs_writebuf(bp, 0);
> > -		} else
> > -			libxfs_putbuf(bp);
> > -		error = process_node_attr(mp, ino, dip, blkmap); /* + repair */
> > -		if (error)
> > -			*repair = 0;
> > -		return error;
> > +		return process_longform_da_root(mp, ino, dip, blkmap, repair,
> > +				bp);
> >   	default:
> >   		do_warn(
> >   	_("bad attribute leaf magic # %#x for dir ino %" PRIu64 "\n"),
> > -			be16_to_cpu(leaf->hdr.info.magic), ino);
> > +			be16_to_cpu(info->magic), ino);
> >   		libxfs_putbuf(bp);
> >   		*repair = 0;
> > -		return(1);
> > +		return 1;
> >   	}
> > -	if (*repair && !no_modify)
> > -		libxfs_writebuf(bp, 0);
> > -	else
> > -		libxfs_putbuf(bp);
> > -
> > -	return(0);  /* repair may be set */
> > +	return 0; /* should never get here */
> >   }
> > 

  reply	other threads:[~2020-02-05  5:59 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-05  0:46 [PATCH 0/4] xfsprogs: random fixes Darrick J. Wong
2020-02-05  0:46 ` [PATCH 1/4] libxfs: re-sort libxfs_api_defs.h defines Darrick J. Wong
2020-02-05  2:11   ` Chaitanya Kulkarni
2020-02-05  5:28   ` Allison Collins
2020-02-17 13:45   ` Christoph Hellwig
2020-02-05  0:46 ` [PATCH 2/4] libfrog: remove libxfs.h dependencies in fsgeom.c and linux.c Darrick J. Wong
2020-02-05  5:28   ` Allison Collins
2020-02-17 13:46   ` Christoph Hellwig
2020-02-05  0:46 ` [PATCH 3/4] xfs_repair: refactor attr root block pointer check Darrick J. Wong
2020-02-05  5:28   ` Allison Collins
2020-02-13 23:14   ` Eric Sandeen
2020-02-14  4:24     ` Darrick J. Wong
2020-02-17 13:47   ` Christoph Hellwig
2020-02-05  0:46 ` [PATCH 4/4] xfs_repair: don't corrupt a attr fork da3 node when clearing forw/back Darrick J. Wong
2020-02-05  5:29   ` Allison Collins
2020-02-05  5:59     ` Darrick J. Wong [this message]
2020-02-17 13:48   ` Christoph Hellwig

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=20200205055913.GD6870@magnolia \
    --to=darrick.wong@oracle.com \
    --cc=allison.henderson@oracle.com \
    --cc=linux-xfs@vger.kernel.org \
    --cc=sandeen@sandeen.net \
    /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).