Re: Stop testing SEDF almost at all (at least until Xen 4.2) [was: Re: [PATCH v2] OSSTest: stop testing SEDF, start testing RTDS]

[George Dunlap <george.dunlap@eu.citrix.com>]

On 06/26/2015 04:24 PM, Ian Jackson wrote:
> Dario Faggioli writes ("Stop testing SEDF almost at all (at least until Xen 4.2) [was: Re: [PATCH v2] OSSTest: stop testing SEDF, start testing RTDS]"):
>> However, I honestly think that testing SEDF for earlier versions than
>> xen-unstable (at least until 4.2) is also just a waste of test
>> resources.
> ...
>> Therefore, I'd argue for the attached patch, [...]
> 
> With my osstest maintainer hat on the aim of this patch is fine by me.
> But it seems that sedf is not coming back, so if we are going to do
> this to osstest we should probably rip the code out.  So I would
> welcome a followup which got rid of all the leftover traces in
> make-flight and allow.all.  (`git-grep -i sedf'.)
> 
> As to the substance, I would like to take advice from the relevant HV
> maintainers about this change.

I agree with Dario that there's no point in testing it, and with you
that we might as well rip the code out of osstest.

> With my tools maintainer hat on, we should consider removing the
> sedf-specific code from libxl.

With the obvious caveat that old code with SEDF macros and such still
needs to compile but return a runtime error if people try to use it.

 -George