Re: [Xen-devel] [PATCH 3 of 7] xen: allows more hypercalls from stubdoms

[Samuel Thibault <samuel.thibault@xxxxxxxxxxxx>]

Stefano Stabellini, le Tue 13 Oct 2009 13:18:50 +0100, a écrit :
> On Tue, 13 Oct 2009, Samuel Thibault wrote:
> > Stefano Stabellini, le Tue 13 Oct 2009 13:10:20 +0100, a écrit :
> > > On Tue, 13 Oct 2009, Samuel Thibault wrote:
> > > > Stefano Stabellini, le Tue 13 Oct 2009 13:00:09 +0100, a écrit :
> > > > > I agree with you that we need to redesign these hypercalls, but I am a
> > > > > fan of "doing one thing at a time" so I think we should decouple this
> > > > > goal from the other one of making passthrough work with stubdom for 
> > > > > the
> > > > > moment.
> > > > > This way we could first let people test it as it is, fix some bugs 
> > > > > that
> > > > > probably still affect the code, fix pci coldplug and add MSI-X 
> > > > > support,
> > > > > then redesign the hypercalls.
> > > > 
> > > > For that you could just provide testers with this patch along with a
> > > > strong warning that it's not safe (from a security point of view).
> > > 
> > > In theory yes, but in practice not so many people read the list
> > > compared to the people that actually chase xen-unstable.
> > 
> > Would the latter report bugs they encounter?
> 
> I hope so, at least on xen-users.

Well, I'd tend to think that if they're not following either, they're
probably not building from the hg repository.  Anyway, I let it to
Keir whether he's keen on opening a temporary security breach to allow
testing :)

Samuel

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel