[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] REST MAINTAINERS feedback requested Was:Re: [PATCH v5 01/28] HYPERCALL_version_op. New hypercall mirroring XENVER_ but sane.

Jan Beulich writes ("Re: REST MAINTAINERS feedback requested Was:Re: 
[Xen-devel] [PATCH v5 01/28] HYPERCALL_version_op. New hypercall mirroring 
XENVER_ but sane."):
> On 11.04.16 at 18:25, <Ian.Jackson@xxxxxxxxxxxxx> wrote:
> > But any improvement from an old API to a new one necessarily involves
> > providing a dual facility during a transition period.
> Except that, at least for most sub-ops, the new one doesn't really
> provide much advantage, and hence dealing with the lack of size
> for those sub-ops where it matters within the existing hypercall
> (perhaps by adding one or two new sub-ops) would limit duplication
> quite a bit.

ISTM that the lower duplication (which in principle is an advantage
which will be time limited if we are ever able to completely remove
teh old hypercall) comes with the cost of (in the long term) increased
mess in this particular subop.

Is that right ?  Obviously this cost is not very significant.  But
maybe the duplication isn't that significant either.  I was kind of
expecting to find stronger arguments on both sides in this

Andrew, can you please confirm what you think of Jan's suggestion to
instead add a new sub-op where the sub-op structure contains a
guest-provided length field ?  Are there other reasons to want to
introduce a new hypercall ?

> > I don't see an explicit deprecation in the patch that is in tree, but
> > it seems to me to be intended (and, perhaps, implied).  Certainly if
> > we are going to permit these strings etc. to be bigger than fits in
> > the old hypercall, the old hypercall needs to be deprecated on the
> > grounds that it can provide incomplete or inaccurate information.
> > 
> > Does this way of looking at it help ?
> If that means you approve of the introduction of the new hypercall,
> yes. After all the goal of this whole discussion is to determine
> whether another maintainer is willing to provide a replacement ack
> for the withdrawn one.

Err.  Well, I am still asking questions.  When I said "does that help"
I was meaning something like "does that help bring you closer to
agreement with Andrew".  But I don't want to focus on procedure here.

If we're going to have the argument over this bike shed I'd like to
figure out what colour I myself prefer :-).  But I need to know what
the reasons are behind people's preferences.


Xen-devel mailing list



Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.