On Thu, 22 Jan 2015 14:16:22 +1100
David Gibson wrote:
> On Wed, Jan 21, 2015 at 08:56:46AM +0100, Cornelia Huck wrote:
> > David: I saw you tested Stefan's endianness patch.
> >
> > Could you also give my alternative implementation a try? Thanks!
>
> Yes, sorry, I got a bit confused about whic
On Wed, Jan 21, 2015 at 08:56:46AM +0100, Cornelia Huck wrote:
> David: I saw you tested Stefan's endianness patch.
>
> Could you also give my alternative implementation a try? Thanks!
Yes, sorry, I got a bit confused about which series was which there.
I gave this a whirl on our P8 and with it
On Wed, 21 Jan 2015 13:30:02 +
Stefan Hajnoczi wrote:
> On Tue, Jan 20, 2015 at 4:27 PM, Cornelia Huck
> wrote:
> > @@ -154,7 +157,7 @@ bool vring_should_notify(VirtIODevice *vdev, Vring
> > *vring)
> > }
> >
> >
> > -static int get_desc(Vring *vring, VirtQueueElement *elem,
> > +static i
On Tue, Jan 20, 2015 at 4:27 PM, Cornelia Huck wrote:
> @@ -154,7 +157,7 @@ bool vring_should_notify(VirtIODevice *vdev, Vring *vring)
> }
>
>
> -static int get_desc(Vring *vring, VirtQueueElement *elem,
> +static int get_desc(VirtIODevice *vdev, Vring *vring, VirtQueueElement *elem,
>
David: I saw you tested Stefan's endianness patch.
Could you also give my alternative implementation a try? Thanks!
On Tue, 20 Jan 2015 17:27:50 +0100
Cornelia Huck wrote:
> The vring.c code currently assumes that guest and host endianness match,
> which is not true for a number of cases:
>
>
The vring.c code currently assumes that guest and host endianness match,
which is not true for a number of cases:
- emulating targets with a different endianness than the host
- bi-endian targets, where the correct endianness depends on the virtio
device
- upcoming support for the virtio-1 stand
