public inbox for linux-xfs@vger.kernel.org
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <darrick.wong@oracle.com>
To: Dave Chinner <david@fromorbit.com>
Cc: linux-xfs@vger.kernel.org
Subject: Re: [PATCH 04/13] xfs: refactor verifier callers to print address of failing check
Date: Thu, 14 Dec 2017 16:04:09 -0800	[thread overview]
Message-ID: <20171215000409.GM13436@magnolia> (raw)
In-Reply-To: <20171214220326.GI5858@dastard>

On Fri, Dec 15, 2017 at 09:03:26AM +1100, Dave Chinner wrote:
> On Wed, Dec 13, 2017 at 03:58:37PM -0800, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> > 
> > Refactor the callers of verifiers to print the instruction address of a
> > failing check.
> > 
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> 
> Just a quick comment about formatting as I browsed the patch...
> 
> > @@ -567,13 +568,14 @@ xfs_agfl_read_verify(
> >  	if (!xfs_sb_version_hascrc(&mp->m_sb))
> >  		return;
> >  
> > -	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF))
> > +	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) {
> > +		fa = __this_address;
> >  		xfs_buf_ioerror(bp, -EFSBADCRC);
> > -	else if (xfs_agfl_verify(bp))
> > +	} else if ((fa = xfs_agfl_verify(bp)))
> >  		xfs_buf_ioerror(bp, -EFSCORRUPTED);
> >  
> >  	if (bp->b_error)
> > -		xfs_verifier_error(bp);
> > +		xfs_verifier_error(bp, fa);
> 
> We are really trying to get rid of assignments in if() statements,
> so I'd prefer we don't add a bunch of new ones. While I understand
> there's a lot of mechanical change in this patch, I'd prefer to see
> these end up as something more like:
> 
> > -	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF))
> > +	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) {
> > +		fa = __this_address;
> >  		xfs_buf_ioerror(bp, -EFSBADCRC);
> > -	else if (xfs_agfl_verify(bp))
> > +	} else if ((fa = xfs_agfl_verify(bp)))
> >  		xfs_buf_ioerror(bp, -EFSCORRUPTED);
> 
> 	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) {
> 		fa = __this_address;
> 		error = -EFSBADCRC;
> 	} else {
> 		fa = xfs_agfl_verify(bp);
> 		if (fa)
> 			error = -EFSCORRUPTED;
> 	}
> 
> 	if (error) {
> 		xfs_buf_ioerror(bp, error);
> 		xfs_verifier_error(bp, fa);
> 	}
> 
> .....
> 
> > @@ -2459,16 +2462,18 @@ xfs_agf_read_verify(
> >  	struct xfs_buf	*bp)
> >  {
> >  	struct xfs_mount *mp = bp->b_target->bt_mount;
> > +	xfs_failaddr_t	fa;
> >  
> >  	if (xfs_sb_version_hascrc(&mp->m_sb) &&
> > -	    !xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF))
> > +	    !xfs_buf_verify_cksum(bp, XFS_AGF_CRC_OFF)) {
> > +		fa = __this_address;
> >  		xfs_buf_ioerror(bp, -EFSBADCRC);
> > -	else if (XFS_TEST_ERROR(xfs_agf_verify(mp, bp), mp,
> > -				XFS_ERRTAG_ALLOC_READ_AGF))
> > +	} else if (XFS_TEST_ERROR((fa = xfs_agf_verify(mp, bp)), mp,
> > +				  XFS_ERRTAG_ALLOC_READ_AGF))
> >  		xfs_buf_ioerror(bp, -EFSCORRUPTED);
> >  
> >  	if (bp->b_error)
> > -		xfs_verifier_error(bp);
> > +		xfs_verifier_error(bp, fa);
> >  }
> 
> Because this sort of thing is now getting towards being unreadable.
> With the way we keep adding to verifier checks, it's only going to
> get worse if we don't take steps to clean it up...

Ok.

> > diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
> > index 4c9f35d..0bbbf0b 100644
> > --- a/fs/xfs/xfs_error.c
> > +++ b/fs/xfs/xfs_error.c
> > @@ -347,13 +347,15 @@ xfs_corruption_error(
> >   */
> >  void
> >  xfs_verifier_error(
> > -	struct xfs_buf		*bp)
> > +	struct xfs_buf		*bp,
> > +	xfs_failaddr_t		fa)
> >  {
> >  	struct xfs_mount *mp = bp->b_target->bt_mount;
> >  
> >  	xfs_alert(mp, "Metadata %s detected at %pS, %s block 0x%llx",
> >  		  bp->b_error == -EFSBADCRC ? "CRC error" : "corruption",
> > -		  __return_address, bp->b_ops->name, bp->b_bn);
> > +		  fa ? fa : __return_address, bp->b_ops->name,
> > +		  bp->b_bn);
> >  
> >  	xfs_alert(mp, "Unmount and run xfs_repair");
> 
> I'm also wondering if we should move the xfs_buf_ioerror() call
> inside this function, too, rather than coding multiple calls in the
> verifiers to set bp->b_error in each branch of the verifier that
> has an error...

What, something like:

	if (!xfs_buf_verify_cksum(bp, XFS_AGFL_CRC_OFF)) {
		fa = __this_address;
		error = -EFSBADCRC;
	} else {
		fa = xfs_agfl_verify(bp);
		if (fa)
			error = -EFSCORRUPTED;
	}

	if (error)
		xfs_verifier_error(bp, fa, error);

???

--D

> 
> Cheers,
> 
> Dave.
> -- 
> Dave Chinner
> david@fromorbit.com

  reply	other threads:[~2017-12-15  0:04 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-12-13 23:58 [PATCH 00/13] xfs: more and better verifiers Darrick J. Wong
2017-12-13 23:58 ` [PATCH 01/13] xfs: refactor long-format btree header verification routines Darrick J. Wong
2017-12-14 22:06   ` Dave Chinner
2017-12-15  0:12     ` Darrick J. Wong
2017-12-13 23:58 ` [PATCH 02/13] xfs: remove XFS_WANT_CORRUPTED_RETURN from dir3 data verifiers Darrick J. Wong
2017-12-19  3:50   ` Dave Chinner
2017-12-13 23:58 ` [PATCH 03/13] xfs: have buffer verifier functions report failing address Darrick J. Wong
2017-12-19  4:12   ` Dave Chinner
2017-12-19 20:26     ` Darrick J. Wong
2017-12-13 23:58 ` [PATCH 04/13] xfs: refactor verifier callers to print address of failing check Darrick J. Wong
2017-12-14 22:03   ` Dave Chinner
2017-12-15  0:04     ` Darrick J. Wong [this message]
2017-12-15  3:09       ` Dave Chinner
2017-12-19 20:29         ` Darrick J. Wong
2017-12-13 23:58 ` [PATCH 05/13] xfs: verify dinode header first Darrick J. Wong
2017-12-19  4:13   ` Dave Chinner
2017-12-13 23:58 ` [PATCH 06/13] xfs: move inode fork verifiers to xfs_dinode_verify Darrick J. Wong
2017-12-19  5:16   ` Dave Chinner
2017-12-19 20:34     ` Darrick J. Wong
2017-12-19 20:48       ` Dave Chinner
2017-12-13 23:58 ` [PATCH 07/13] xfs: create structure verifier function for shortform xattrs Darrick J. Wong
2017-12-19  5:23   ` Dave Chinner
2017-12-19 20:41     ` Darrick J. Wong
2017-12-19 20:51       ` Dave Chinner
2017-12-19 21:04         ` Darrick J. Wong
2017-12-13 23:59 ` [PATCH 08/13] xfs: create structure verifier function for short form symlinks Darrick J. Wong
2017-12-19  5:27   ` Dave Chinner
2017-12-19 20:45     ` Darrick J. Wong
2017-12-13 23:59 ` [PATCH 09/13] xfs: refactor short form directory structure verifier function Darrick J. Wong
2017-12-19  5:45   ` Dave Chinner
2017-12-13 23:59 ` [PATCH 10/13] xfs: provide a centralized method for verifying inline fork data Darrick J. Wong
2017-12-19  6:06   ` Dave Chinner
2017-12-19 20:50     ` Darrick J. Wong
2017-12-13 23:59 ` [PATCH 11/13] xfs: fail out of xfs_attr3_leaf_lookup_int if it looks corrupt Darrick J. Wong
2017-12-19  6:13   ` Dave Chinner
2017-12-13 23:59 ` [PATCH 12/13] xfs: create a new buf_ops pointer to verify structure metadata Darrick J. Wong
2017-12-19  6:22   ` Dave Chinner
2017-12-19 18:15     ` Darrick J. Wong
2017-12-19 20:53       ` Dave Chinner
2017-12-13 23:59 ` [PATCH 13/13] xfs: scrub in-core metadata Darrick J. Wong
2017-12-19  6:23   ` 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=20171215000409.GM13436@magnolia \
    --to=darrick.wong@oracle.com \
    --cc=david@fromorbit.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