On Wed, 18 Nov 2015 22:46:55 +0100 Greg Kurz <gk...@linux.vnet.ibm.com> wrote:
> On Wed, 18 Nov 2015 22:48:06 +0200 > "Michael S. Tsirkin" <m...@redhat.com> wrote: > > > On Wed, Nov 18, 2015 at 05:23:00PM +0100, Greg Kurz wrote: > > > Since QEMU 2.4.0, vhost-net uses the cross-endian support of TAP devices > > > to > > > fix vnet headers. In fact, virtio-net can do the same instead of hackily > > > patching headers in virtio_net_hdr_swap(). > > > > > > This series moves the enablement of cross-endian support from vhost-net to > > > virtio-net: both vhost and full emulation can now benefit from it. Of > > > course > > > we keep the current hack to fall back on when the backend doesn't support > > > cross-endian. > > > > > > Thanks! > > This is an optimization rather than a bugfix, right? > > As such I'd rather defer this until after 2.5. > > > > Of course. I'll ping or repost later. > Hi Michael, 2.5 is now released and this series still applies cleanly. Should I repost ? -- Greg > > > --- > > > > > > Greg Kurz (3): > > > virtio-net: use the backend cross-endian capabilities > > > Revert "vhost-net: tell tap backend about the vnet endianness" > > > virtio: drop the virtio_needs_swap() helper > > > > > > > > > hw/net/vhost_net.c | 33 +------------------------------ > > > hw/net/virtio-net.c | 40 > > > +++++++++++++++++++++++++++++++++++-- > > > include/hw/virtio/virtio-access.h | 9 -------- > > > include/hw/virtio/virtio-net.h | 1 + > > > 4 files changed, 40 insertions(+), 43 deletions(-) > > > >