Re: [PATCH OSSTEST] Toolstack: Do not pass -F to xm shutdown (Was: Re: [xen-4.2-testing test] 57630: regressions - FAIL)

[Ian Campbell <ian.campbell@citrix.com>]

On Mon, 2015-06-01 at 10:21 +0100, Jan Beulich wrote:
> >>> On 01.06.15 at 10:34, <ian.campbell@citrix.com> wrote:
> > Subject: [PATCH] Toolstack: Do not pass -F to xm shutdown
> > 
> > This is a feature of xl only.
> > 
> > Signed-off-by: Ian Campbell <ian.campbell@citrix.com>
> > ---
> >  Osstest/Toolstack/xl.pm | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/Osstest/Toolstack/xl.pm b/Osstest/Toolstack/xl.pm
> > index dd12ae1..23328d3 100644
> > --- a/Osstest/Toolstack/xl.pm
> > +++ b/Osstest/Toolstack/xl.pm
> > @@ -56,7 +56,7 @@ sub shutdown_wait ($$$) {
> >      my $ho = $self->{Host};
> >      my $gn = $gho->{Name};
> >      my $acpi_fallback = guest_var($gho,'acpi_shutdown','false') eq 'true'
> > -	? "F" : "";
> > +	&& $self->{Name} eq 'xl' ? "F" : "";
> >      target_cmd_root($ho,"$self->{_Command} shutdown -w${acpi_fallback} $gn", $timeout);
> 
> Well - to me, not knowing much about osstest's structure, catering
> for xm in a file called xl.pm seems odd. And then, taking into
> consideration a hypothetical 3rd toolstack, I wonder whether
> assuming only xl supports -F (instead of "only xm does not support
> it") is the more suitable check.

xend.pm inherits from xl.pm in a oo-ish way.

I did consider adding shutdown_wait to xend.pm to overload this one, but
that meant duplicating 95% unaltered.

A third toolstack is not hypothetical, libvirt's virsh is supported, but
virsh.pm sharing no historical legacy with xl/xm inherets from the
toplevel toolstack parent, not the xl one.

Ian.