[fstests PATCH] ceph: don't run tests if we can't set a custom layout

[fstests PATCH] ceph: don't run tests if we can't set a custom layout

[Jeff Layton]

Some of the coming fscrypt patches prohibit non-default layout changes.
Skip running the tests that set custom layouts if setting the layout fails.

Cc: Luis Henriques <lhenriques@suse.de>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
---
 common/ceph | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/common/ceph b/common/ceph
index ca756dda8dd3..31b169af51b8 100644
--- a/common/ceph
+++ b/common/ceph
@@ -19,7 +19,7 @@ _ceph_create_file_layout()
 	touch $fname
 	$SETFATTR_PROG -n ceph.file.layout \
 		-v "stripe_unit=$objsz stripe_count=1 object_size=$objsz" \
-		$fname
+		$fname || _notrun "setting ceph.file.layout failed"
 }
 
 # this test requires to access file capabilities through vxattr 'ceph.caps'.
-- 
2.33.1

Re: [fstests PATCH] ceph: don't run tests if we can't set a custom layout

[Luís Henriques]

On Tue, Dec 07, 2021 at 11:04:37AM -0500, Jeff Layton wrote:
> Some of the coming fscrypt patches prohibit non-default layout changes.
> Skip running the tests that set custom layouts if setting the layout fails.
> 
> Cc: Luis Henriques <lhenriques@suse.de>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>  common/ceph | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/common/ceph b/common/ceph
> index ca756dda8dd3..31b169af51b8 100644
> --- a/common/ceph
> +++ b/common/ceph
> @@ -19,7 +19,7 @@ _ceph_create_file_layout()
>  	touch $fname
>  	$SETFATTR_PROG -n ceph.file.layout \
>  		-v "stripe_unit=$objsz stripe_count=1 object_size=$objsz" \
> -		$fname
> +		$fname || _notrun "setting ceph.file.layout failed"
>  }
>  
>  # this test requires to access file capabilities through vxattr 'ceph.caps'.
> -- 
> 2.33.1
> 

Yeah, this is a simple way to skip those tests.  Feel free to add my

Reviewed-by: Luis Henriques <lhenriques@suse.de>

Cheers,
--
Luís

Re: [fstests PATCH] ceph: don't run tests if we can't set a custom layout

[Eryu Guan]

On Tue, Dec 07, 2021 at 11:04:37AM -0500, Jeff Layton wrote:
> Some of the coming fscrypt patches prohibit non-default layout changes.
> Skip running the tests that set custom layouts if setting the layout fails.
> 
> Cc: Luis Henriques <lhenriques@suse.de>
> Signed-off-by: Jeff Layton <jlayton@kernel.org>
> ---
>  common/ceph | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/common/ceph b/common/ceph
> index ca756dda8dd3..31b169af51b8 100644
> --- a/common/ceph
> +++ b/common/ceph
> @@ -19,7 +19,7 @@ _ceph_create_file_layout()
>  	touch $fname
>  	$SETFATTR_PROG -n ceph.file.layout \
>  		-v "stripe_unit=$objsz stripe_count=1 object_size=$objsz" \
> -		$fname
> +		$fname || _notrun "setting ceph.file.layout failed"
>  }

IMO _ceph_create_file_layout() is a helper function that does the real
work, i.e. either prepares the test or does the test, not a function to
check if current env & setup meets what the test needs.

So I think it's better to check the ability to change layout explicitly
in a new _require rule. e.g. something like _require_ceph_change_layout?

Thanks,
Eryu

>  
>  # this test requires to access file capabilities through vxattr 'ceph.caps'.
> -- 
> 2.33.1

Re: [fstests PATCH] ceph: don't run tests if we can't set a custom layout

[Jeff Layton]

On Sun, 2021-12-12 at 22:02 +0800, Eryu Guan wrote:
> On Tue, Dec 07, 2021 at 11:04:37AM -0500, Jeff Layton wrote:
> > Some of the coming fscrypt patches prohibit non-default layout changes.
> > Skip running the tests that set custom layouts if setting the layout fails.
> > 
> > Cc: Luis Henriques <lhenriques@suse.de>
> > Signed-off-by: Jeff Layton <jlayton@kernel.org>
> > ---
> >  common/ceph | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/common/ceph b/common/ceph
> > index ca756dda8dd3..31b169af51b8 100644
> > --- a/common/ceph
> > +++ b/common/ceph
> > @@ -19,7 +19,7 @@ _ceph_create_file_layout()
> >  	touch $fname
> >  	$SETFATTR_PROG -n ceph.file.layout \
> >  		-v "stripe_unit=$objsz stripe_count=1 object_size=$objsz" \
> > -		$fname
> > +		$fname || _notrun "setting ceph.file.layout failed"
> >  }
> 
> IMO _ceph_create_file_layout() is a helper function that does the real
> work, i.e. either prepares the test or does the test, not a function to
> check if current env & setup meets what the test needs.
> 
> So I think it's better to check the ability to change layout explicitly
> in a new _require rule. e.g. something like _require_ceph_change_layout?
> 
> Thanks,
> Eryu


Good point. OTOH, relying on this program returning an error may be the
wrong approach. What might be best is to just have a
_require_not_encrypted check. That might be more universally useful
anyway.

Let me see what I can come up with. We can drop this patch for now.

Thanks!
-- 
Jeff Layton <jlayton@kernel.org>