ACK/Cmnt: [SRU oem-5.6] xfs: add agf freeblocks verify in xfs_agf_verify

[Thadeu Lima de Souza Cascardo]

On Thu, Sep 17, 2020 at 09:43:54AM +0200, Stefan Bader wrote:
> On 16.09.20 19:53, Thadeu Lima de Souza Cascardo wrote:
> > From: Zheng Bin <zhengbin13 at huawei.com>
> > 
> > We recently used fuzz(hydra) to test XFS and automatically generate
> > tmp.img(XFS v5 format, but some metadata is wrong)
> > 
> > xfs_repair information(just one AG):
> > agf_freeblks 0, counted 3224 in ag 0
> > agf_longest 536874136, counted 3224 in ag 0
> > sb_fdblocks 613, counted 3228
> > 
> > Test as follows:
> > mount tmp.img tmpdir
> > cp file1M tmpdir
> > sync
> > 
> > In 4.19-stable, sync will stuck, the reason is:
> > xfs_mountfs
> >   xfs_check_summary_counts
> >     if ((!xfs_sb_version_haslazysbcount(&mp->m_sb) ||
> >        XFS_LAST_UNMOUNT_WAS_CLEAN(mp)) &&
> >        !xfs_fs_has_sickness(mp, XFS_SICK_FS_COUNTERS))
> > 	return 0;  -->just return, incore sb_fdblocks still be 613
> >     xfs_initialize_perag_data
> > 
> > cp file1M tmpdir -->ok(write file to pagecache)
> > sync -->stuck(write pagecache to disk)
> > xfs_map_blocks
> >   xfs_iomap_write_allocate
> >     while (count_fsb != 0) {
> >       nimaps = 0;
> >       while (nimaps == 0) { --> endless loop
> >          nimaps = 1;
> >          xfs_bmapi_write(..., &nimaps) --> nimaps becomes 0 again
> > xfs_bmapi_write
> >   xfs_bmap_alloc
> >     xfs_bmap_btalloc
> >       xfs_alloc_vextent
> >         xfs_alloc_fix_freelist
> >           xfs_alloc_space_available -->fail(agf_freeblks is 0)
> > 
> > In linux-next, sync not stuck, cause commit c2b3164320b5 ("xfs:
> > use the latest extent at writeback delalloc conversion time") remove
> > the above while, dmesg is as follows:
> > [   55.250114] XFS (loop0): page discard on page ffffea0008bc7380, inode 0x1b0c, offset 0.
> > 
> > Users do not know why this page is discard, the better soultion is:
> > 1. Like xfs_repair, make sure sb_fdblocks is equal to counted
> > (xfs_initialize_perag_data did this, who is not called at this mount)
> > 2. Add agf verify, if fail, will tell users to repair
> > 
> > This patch use the second soultion.
> > 
> > Signed-off-by: Zheng Bin <zhengbin13 at huawei.com>
> > Signed-off-by: Ren Xudong <renxudong1 at huawei.com>
> > Reviewed-by: Darrick J. Wong <darrick.wong at oracle.com>
> > Signed-off-by: Darrick J. Wong <darrick.wong at oracle.com>
> > (cherry picked from commit d0c7feaf87678371c2c09b3709400be416b2dc62)
> > CVE-2020-12655
> > Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo at canonical.com>
> Acked-by: Stefan Bader <stefan.bader at canonical.com>
> > ---
> 
> Just as I noted it with this review, I would prefer to do the sob block like this:

Well, as far as I know, the tools don't care. And my order at least indicates
the CVE line was added after cherry-pick, and was not part of the original
commit message.

Cascardo.

> 
> ...
> 
> CVE-2020-12655
> (cherry picked from commit d0c7feaf87678371c2c09b3709400be416b2dc62)
> Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo at canonical.com>
> 
> >  fs/xfs/libxfs/xfs_alloc.c | 16 ++++++++++++++++
> >  1 file changed, 16 insertions(+)
> > 
> > diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
> > index 5a130409f173..0ebf4098a156 100644
> > --- a/fs/xfs/libxfs/xfs_alloc.c
> > +++ b/fs/xfs/libxfs/xfs_alloc.c
> > @@ -2858,6 +2858,13 @@ xfs_agf_verify(
> >  	      be32_to_cpu(agf->agf_flcount) <= xfs_agfl_size(mp)))
> >  		return __this_address;
> >  
> > +	if (be32_to_cpu(agf->agf_length) > mp->m_sb.sb_dblocks)
> > +		return __this_address;
> > +
> > +	if (be32_to_cpu(agf->agf_freeblks) < be32_to_cpu(agf->agf_longest) ||
> > +	    be32_to_cpu(agf->agf_freeblks) > be32_to_cpu(agf->agf_length))
> > +		return __this_address;
> > +
> >  	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 ||
> >  	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 ||
> >  	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS ||
> > @@ -2869,6 +2876,10 @@ xfs_agf_verify(
> >  	     be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS))
> >  		return __this_address;
> >  
> > +	if (xfs_sb_version_hasrmapbt(&mp->m_sb) &&
> > +	    be32_to_cpu(agf->agf_rmap_blocks) > be32_to_cpu(agf->agf_length))
> > +		return __this_address;
> > +
> >  	/*
> >  	 * during growfs operations, the perag is not fully initialised,
> >  	 * so we can't use it for any useful checking. growfs ensures we can't
> > @@ -2882,6 +2893,11 @@ xfs_agf_verify(
> >  	    be32_to_cpu(agf->agf_btreeblks) > be32_to_cpu(agf->agf_length))
> >  		return __this_address;
> >  
> > +	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
> > +	    be32_to_cpu(agf->agf_refcount_blocks) >
> > +	    be32_to_cpu(agf->agf_length))
> > +		return __this_address;
> > +
> >  	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
> >  	    (be32_to_cpu(agf->agf_refcount_level) < 1 ||
> >  	     be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS))
> > 
> 
>