xfs
[Top] [All Lists]

Re: [PATCH 6/9] repair: report AG btree verifier errors

To: Brian Foster <bfoster@xxxxxxxxxx>
Subject: Re: [PATCH 6/9] repair: report AG btree verifier errors
From: Dave Chinner <david@xxxxxxxxxxxxx>
Date: Wed, 16 Apr 2014 07:53:59 +1000
Cc: xfs@xxxxxxxxxxx
Delivered-to: xfs@xxxxxxxxxxx
In-reply-to: <20140415194048.GD3470@xxxxxxxxxxxxxx>
References: <1397550301-31883-1-git-send-email-david@xxxxxxxxxxxxx> <1397550301-31883-7-git-send-email-david@xxxxxxxxxxxxx> <20140415194048.GD3470@xxxxxxxxxxxxxx>
User-agent: Mutt/1.5.21 (2010-09-15)
On Tue, Apr 15, 2014 at 03:40:48PM -0400, Brian Foster wrote:
> On Tue, Apr 15, 2014 at 06:24:58PM +1000, Dave Chinner wrote:
> > From: Dave Chinner <dchinner@xxxxxxxxxx>
> > 
> > When we scan the filesystem freespace and inode maps in phase 2, we
> > don't report CRC errors that are found. We don't really care from a
> > repair perspective, because the trees are completely rebuilt from
> > the ground up in phase 5, but froma checking perspective we need to
> > inform the user that we found inconsistencies.
> > 
> > Signed-off-by: Dave Chinner <dchinner@xxxxxxxxxx>
> > ---
> >  repair/scan.c | 27 ++++++++++++++++++++++++---
> >  1 file changed, 24 insertions(+), 3 deletions(-)
> > 
> > diff --git a/repair/scan.c b/repair/scan.c
> > index 6c43474..75b4b70 100644
> > --- a/repair/scan.c
> > +++ b/repair/scan.c
> > @@ -82,6 +82,12 @@ scan_sbtree(
> >             do_error(_("can't read btree block %d/%d\n"), agno, root);
> >             return;
> >     }
> > +   if (bp->b_error == EFSBADCRC || bp->b_error == EFSCORRUPTED) {
> > +           do_warn(_("btree block %d/%d is suspect, error %d\n"),
> > +                   agno, root, bp->b_error);
> > +           suspect = 1;
> > +   }
> > +
> >     (*func)(XFS_BUF_TO_BLOCK(bp), nlevels - 1, root, agno, suspect,
> >                                                     isroot, magic, priv);
> >     libxfs_putbuf(bp);
> > @@ -123,6 +129,7 @@ scan_lbtree(
> >     xfs_buf_t       *bp;
> >     int             err;
> >     int             dirty = 0;
> > +   bool            badcrc = false;
> >  
> >     bp = libxfs_readbuf(mp->m_dev, XFS_FSB_TO_DADDR(mp, root),
> >                   XFS_FSB_TO_BB(mp, 1), 0, ops);
> > @@ -132,14 +139,25 @@ scan_lbtree(
> >                     XFS_FSB_TO_AGBNO(mp, root));
> >             return(1);
> >     }
> > +
> > +   /*
> > +    * only check for bad CRC here - caller will determine if there
> > +    * is a corruption or not and whether it got corrected and so needs
> > +    * writing back. CRC errors always imply we need to write the block.
> > +    */
> > +   if (bp->b_error == EFSBADCRC) {
> > +           do_warn(_("btree block %d/%d is suspect, error %d\n"),
> > +                   XFS_FSB_TO_AGNO(mp, root),
> > +                   XFS_FSB_TO_AGBNO(mp, root), bp->b_error);
> > +           badcrc = true;
> > +   }
> > +
> >     err = (*func)(XFS_BUF_TO_BLOCK(bp), nlevels - 1,
> >                     type, whichfork, root, ino, tot, nex, blkmapp,
> >                     bm_cursor, isroot, check_dups, &dirty,
> >                     magic);
> >  
> > -   ASSERT(dirty == 0 || (dirty && !no_modify));
> > -
> 
> Here we remove the dirty assert, but don't change the semantics of dirty
> (we use badcrc).

Yeah, I went back and changed the mechanism for correcting pure
CRC errors not to use the dirty flag. I'll re-instate the assert...

Cheers,

Dave.
-- 
Dave Chinner
david@xxxxxxxxxxxxx

<Prev in Thread] Current Thread [Next in Thread>