Re: [PATCH] fstests: add btrfs fs-verity send/recv test

[Boris Burkov <boris@bur.io>]

On Wed, Aug 17, 2022 at 10:12:06PM -0700, Eric Biggers wrote:
> On Mon, Aug 15, 2022 at 01:57:56PM -0700, Boris Burkov wrote:
> > diff --git a/tests/btrfs/271 b/tests/btrfs/271
> > new file mode 100755
> > index 00000000..93b34540
> > --- /dev/null
> > +++ b/tests/btrfs/271
> 
> There is already a btrfs/271, so this patch doesn't apply anymore.  Best to use
> a higher number and let the maintainer renumber the test when applying.

I thought that, but now it looks like we have btrfs/290 and 291 from the
last tests I added :/

Will fix, though!

> 
> > @@ -0,0 +1,114 @@
> > +#! /bin/bash
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (c) 2022 YOUR NAME HERE.  All Rights Reserved.
> 
> YOUR NAME HERE is a great company to work for :-)
> 
> > +_require_test
> 
> I don't see where this uses the test filesystem; it seems to use scratch only.
> 
> > +	if [ $salt -eq 1 ]; then
> > +		extra_args+=" --salt=deadbeef"
> > +	fi
> 
> I like to use true and false for this sort of thing so you can just do:
> 
> 	if $salt; then
> 
> > +	echo "check received subvolume..."
> > +	echo 3 > /proc/sys/vm/drop_caches
> 
> A comment explaining why the drop_caches is needed would be helpful.
> And should there be a sync before it, and should it be _scratch_cycle_mount?
> 
> > +	_fsv_measure $fsv_file > $tmp.digest-after
> > +	$GETCAP_PROG $fsv_file > $tmp.cap-after
> > +	diff $tmp.digest-before $tmp.digest-after
> > +	diff $tmp.cap-before $tmp.cap-after
> > +	_scratch_unmount
> > +	echo OK
> 
> Should this compare the file's contents too?
> 
> - Eric