On Wed, May 23, 2018 at 10:01:29AM +0800, Eryu Guan wrote:
> On Tue, May 22, 2018 at 08:32:52AM +1000, Dave Chinner wrote:
> > On Fri, May 11, 2018 at 11:05:59AM +0800, Eryu Guan wrote:
> > > One of _scratch_mkfs_xfs_opts()'s job is to remove any metadata
> > > related mkfs options if mkfs.xfs binary doesn't support them, so
> > > that tests that pass metadata mkfs options to _scratch_mkfs, e.g.
> > > "_scratch_mkfs -m crc=0" won't break with old mkfs.xfs.
> > >
> > > But commit 596a068bf130 ("fstests: teach _scratch_mkfs to handle
> > > mkfs option conflicts") broke this, because it didn't pass any mkfs
> > > options to _scratch_mkfs_xfs_opts(). So fix it.
> > >
> > > Also declare a variable as 'local' and fix a typo.
> > >
> > > Reported-by: Dave Chinner <dchinner@xxxxxxxxxx>
> > > Signed-off-by: Eryu Guan <guaneryu@xxxxxxxxx>
> >
> > Looks fine.
> >
> > Reviewed-by: Dave Chinner <dchinner@xxxxxxxxxx>
> >
> > (sorry for missing this when you opsted it first time)
>
> Thanks for the review!
>
> Eryu
This turns out to be broken when specifying "-m crc=0/1" in MKFS_OPTIONS
explicitly and test has its own extra "-m crc=0/1" option (e.g. xfs/001,
mkfs fails and test uses SCRATCH_DEV created from previous test), as we
have "-m" option specified multiple times, and _scratch_do_mkfs() won't
try mkfs for a second time because $extra_mkfs_options is empty, which
is because I didn't pass "extra_mkfs_options" as extra options but
"mkfs_cmd".
I wrote a v2 patch and will send it out soon.
Thanks,
Eryu
--
To unsubscribe from this list: send the line "unsubscribe fstests" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
