Re: [Xen-devel] [Qemu-devel] [PATCH 2/4] 9pfs: introduce transport specific callbacks

[Stefano Stabellini <sstabellini@xxxxxxxxxx>]

On Thu, 24 Nov 2016, Greg Kurz wrote:
> On Thu, 24 Nov 2016 15:23:10 +0100
> Greg Kurz <groug@xxxxxxxx> wrote:
> 
> > On Thu, 24 Nov 2016 09:31:52 +0100
> > Greg Kurz <groug@xxxxxxxx> wrote:
> > 
> > > On Mon, 21 Nov 2016 13:39:30 -0800
> > > Stefano Stabellini <sstabellini@xxxxxxxxxx> wrote:
> > >   
> > > > Don't call virtio functions from 9pfs generic code, use generic function
> > > > callbacks instead.
> > > > 
> > > > Signed-off-by: Stefano Stabellini <sstabellini@xxxxxxxxxx>
> > > > ---    
> > > 
> > > Just a couple of indentation and line over 80 characters nits. I'll fix 
> > > them
> > > before pushing to 9p-next.
> > > 
> > > Reviewed-by: Greg Kurz <groug@xxxxxxxx>
> > >   
> > 
> > Hmm... second thought...
> > 
> 
> [...]
> 
> > > > +
> > > > +static inline int v9fs_register_transport(V9fsState *s, struct 
> > > > V9fsTransport *t)
> > > > +{
> > > > +    if (s->transport) {
> > > > +        return -EINVAL;
> > > > +    }
> 
> Calling v9fs_register_transport() several times for the same V9fsState looks
> more like a bug than an error...

Yes, you are right. I can turn this into an assert.


> also, is it possible to have several 9pfs
> shares with different transports ?

I think, at least theoretically, is possible. For example Xen HVM
guests, the ones most similar to KVM guests, already support virtio as a
transport. Additionally they will be able to support the Xen based
transport too. (Xen PV guests will only support
the Xen based transport as they don't support virtio in general.)

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel