Re: [PATCH 3/7] fs, xfs: convert xfs_buf_log_item.bli_refcount from atomic_t to refcount_t

[Brian Foster <bfoster@redhat.com>]

On Tue, Feb 21, 2017 at 09:06:20AM -0800, Darrick J. Wong wrote:
> On Tue, Feb 21, 2017 at 04:06:30PM +0000, Reshetova, Elena wrote:
> > > On Tue, Feb 21, 2017 at 05:49:03PM +0200, Elena Reshetova wrote:
> > > > refcount_t type and corresponding API should be
> > > > used instead of atomic_t when the variable is used as
> > > > a reference counter. This allows to avoid accidental
> > > > refcounter overflows that might lead to use-after-free
> > > > situations.
> > > 
> > > Changelog forgets to mention if this was runtime tested..
> > 
> > It was boot-tested in the whole refcount_t changes pile, which is not very useful for fs anyway. 
> > What's why we are sending this through maintainers to get through their tests. 
> > I am sure that testing would be better than what we can do. 
> 
> If you're going to go around making this many changes to XFS (or any
> other filesystem), please run the changes through xfstests first.
> Many fs projects (not just XFS) record their test cases there.
> 
> I think the kernel 0day build service is supposed to do that
> automatically...
> 

Be sure to use CONFIG_XFS_DEBUG and/or CONFIG_XFS_WARN to capture any
potential assert failures as well.

Brian

> --D
> 
> > 
> > > 
> > > 
> > > > @@ -371,7 +371,7 @@ xfs_trans_brelse(xfs_trans_t	*tp,
> > > >  	ASSERT(bip->bli_item.li_type == XFS_LI_BUF);
> > > >  	ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
> > > >  	ASSERT(!(bip->__bli_format.blf_flags & XFS_BLF_CANCEL));
> > > > -	ASSERT(atomic_read(&bip->bli_refcount) > 0);
> > > > +	ASSERT(refcount_read(&bip->bli_refcount) > 0);
> > > >
> > > >  	trace_xfs_trans_brelse(bip);
> > > >
> > > > @@ -419,7 +419,7 @@ xfs_trans_brelse(xfs_trans_t	*tp,
> > > >  	/*
> > > >  	 * Drop our reference to the buf log item.
> > > >  	 */
> > > > -	atomic_dec(&bip->bli_refcount);
> > > > +	refcount_dec(&bip->bli_refcount);
> > > >
> > > >  	/*
> > > >  	 * If the buf item is not tracking data in the log, then
> > > > @@ -432,7 +432,7 @@ xfs_trans_brelse(xfs_trans_t	*tp,
> > > >  /***
> > > >  		ASSERT(bp->b_pincount == 0);
> > > >  ***/
> > > > -		ASSERT(atomic_read(&bip->bli_refcount) == 0);
> > > > +		ASSERT(refcount_read(&bip->bli_refcount) == 0);
> > > >  		ASSERT(!(bip->bli_item.li_flags & XFS_LI_IN_AIL));
> > > >  		ASSERT(!(bip->bli_flags &
> > > XFS_BLI_INODE_ALLOC_BUF));
> > > >  		xfs_buf_item_relse(bp);
> > > 
> > > 
> > > This for example looks dodgy.
> > > 
> > > That seems to suggest the atomic_dec() there can actually hit 0, which
> > > _will_ generate a WARN.
> > 
> > True, but in some of this cases WARN might be ok, I think? As soon as functionality is not changed and object is not reused (by doing refcount_inc on it) anywhere later on. 
> > 
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> > the body of a message to majordomo@vger.kernel.org
> > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> --
> To unsubscribe from this list: send the line "unsubscribe linux-xfs" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html