Re: [PATCH v16 13/21] xfs: add XFS_BMAPI_UNWRITTEN to unmap unwritten extents in __xfs_bunmapi()

[Christoph Hellwig <hch@infradead.org>]

On Mon, Sep 21, 2026 at 09:34:02PM -0700, Darrick J. Wong wrote:
> > diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
> > index cc48f6e20e80..7ba76017ae56 100644
> > --- a/fs/xfs/libxfs/xfs_bmap.c
> > +++ b/fs/xfs/libxfs/xfs_bmap.c
> > @@ -5280,6 +5280,10 @@ __xfs_bunmapi(
> >  		if (del.br_startoff + del.br_blockcount > end + 1)
> >  			del.br_blockcount = end + 1 - del.br_startoff;
> >  
> > +		if ((flags & XFS_BMAPI_UNWRITTEN) &&
> > +				del.br_state != XFS_EXT_UNWRITTEN)
> 
> Odd indentation?
> 
> 		if ((flags & XFS_BMAPI_UNWRITTEN) &&
> 		    del.br_state != XFS_EXT_UNWRITTEN)

Yeah.

> /me again wonders if bunmapi ought to have its own flags instaed of
> reusing *some* of bmapi.  But I concede that ought to be a different
> cleanup, because figuring out which flags are for bunmapi.  I think it's
> REMAP, ATTR_FORK, NODISCARD, and UNWRITTEN?

I think that would create more confusion than it helps.  OTOH asserting
the valid flag would be helful.

Sashiko complains about interactions with the weird rtextent bunmap
behavior.  Should we disallow fsverity with a larger than fsb
rtextensive to not bother with this?  Btw, did you test this series
with RT classic and or zoned?