On Wed, Aug 30, 2017 at 12:53:21AM -0700, Christoph Hellwig wrote:

On Tue, Aug 29, 2017 at 09:41:02PM -0700, Darrick J. Wong wrote:

quoted

From: Darrick J. Wong <redacted>

Make sure that we update the rmapbt correctly when we collapse-range a
file and the extents on both sides of the hole can be merged.  We can
construct this pretty trivially with insert-range and write, so test
that too.

Does this break on the current tree?  From code inspection I suspect
the current code is doing the wrong thing there.

AFAICR the existing rmap code in bmse_shift_one /does/ work, though only
due to the subtlety that xfs_rmap_{,un}map_extent is smart enough to
automatically merge and unmerge rmap extents for you.

Hence if you have the following:

AAAAABBBAAAAAAAA

Where both A's could be merged after BBB goes away:

AAAAAAAAAAAAA---

The deferred rmap code will merge the two A regions into a single rmap
extent record automatically.

Hmm, so I guess that means that in the patch "xfs: use xfs_iext_*_extent
helpers in xfs_bmap_shift_extents" you can replace the three xfs_rmap
calls in xfs_bmse_merge with:

	/* update reverse mapping */
	error = xfs_rmap_unmap_extent(mp, dfops, ip, whichfork, got);
	if (error)
		return error;
	memcpy(&new, got, sizeof(new));
	new.br_startoff = left->br_startoff + left->br_blockcount;
	return xfs_rmap_map_extent(mp, dfops, ip, whichfork, &new);

...which saves us a deferred op.  Ok, will propose that in the actual
thread.

quoted

+./src/punch-alternating -o $((16 * blksz / file_blksz)) \

Shouldn't we always use $here/src/progname ?

Ok, I didn't realize that.  There are a lot of programs that still use ./src....

--D

--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html