Re: [PATCH][RESEND] generic/441: whitelist gfs2 for full test

[Eryu Guan <guaneryu@gmail.com>]

Hi Jeff,

On Fri, Apr 27, 2018 at 12:33:27PM -0400, Jeff Layton wrote:
> From: Jeff Layton <jlayton@redhat.com>
> 
> gfs2 passes the full-scale generic/441 test with the patch that
> converts it to use errseq_t reporting for fsync.
> 
> Signed-off-by: Jeff Layton <jlayton@redhat.com>

Sorry for letting you resend the same patch again and again, but I
didn't get any feedback on my previous replies to the patch.

Please see
https://patchwork.kernel.org/patch/9868465/
https://patchwork.kernel.org/patch/9952475/

> ---
>  tests/generic/441 | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/tests/generic/441 b/tests/generic/441
> index 5fbfececadc0..be903fbf94f8 100755
> --- a/tests/generic/441
> +++ b/tests/generic/441
> @@ -58,7 +58,7 @@ case $FSTYP in
>  	btrfs)
>  		_notrun "btrfs has a specialized test for this"
>  		;;
> -	ext3|ext4|xfs)
> +	ext3|ext4|gfs2|xfs)
>  		# Do the more thorough test if we have a logdev
>  		_has_logdev && sflag=''

So my concern remains, _scratch_mkfs doesn't create gfs2 with external
log device even if we set SCRATCH_LOGDEV and USE_EXTERNAL=yes, i.e.
_has_logdev returns true. So with this modification, gfs2 still runs
with internal log (and test still *fails* with v4.16 kernel, I hit the
same failure as in the second link above).

If running gfs2 test with internal log is expected, I think we can run
the test in full-scale mode unconditionally for gfs2, as suggested in
above two replies. i.e. something like

	ext3|ext4|xfs)
		# Do the more thorough test if we have a logdev
		_has_logdev && sflag=''
		;;
	gfs2)
		# <comments here>
		sflag=''
		;;
	*)
		;;

If gfs2 really requires external log to run in full-scale test, I think
we should also add external log device support to _scratch_mkfs().

Any comments?

Thanks,
Eryu