On 210807 1019, Philippe Mathieu-Daudé wrote:
> On 8/7/21 12:23 AM, Alexander Bulekov wrote:
> > Encapsulation offload (offload mode 1) is a valid mode present in the
> > kernel that isn't implemented in QEMU, yet.
> >
> > https://lore.kernel.org/lkml/20200528015426.8285-4-dos...@vmware.com/
> >
> > Add a stub for this mode, to avoid the guest-triggerable assertion.
> >
> > Resolves: https://gitlab.com/qemu-project/qemu/-/issues/517
> > Signed-off-by: Alexander Bulekov <alx...@bu.edu>
> > ---
> > hw/net/vmxnet3.c | 5 +++++
> > hw/net/vmxnet3.h | 1 +
> > 2 files changed, 6 insertions(+)
> >
> > diff --git a/hw/net/vmxnet3.c b/hw/net/vmxnet3.c
> > index 41f796a247..62e11d0e3e 100644
> > --- a/hw/net/vmxnet3.c
> > +++ b/hw/net/vmxnet3.c
> > @@ -443,6 +443,11 @@ vmxnet3_setup_tx_offloads(VMXNET3State *s)
> > net_tx_pkt_build_vheader(s->tx_pkt, false, false, 0);
> > break;
> >
> > + case VMXNET3_OM_ENCAP:
> > + VMW_PKPRN("Encapsulation offload requested, but not available\n");
> > + return false;
> > + break;
>
> No need to break if you returned (unreachable).
I included it in case the feature ends up being added, so there is one
less thing to worry about, but I can remove it. There are a couple of
places in QEMU where this occurs, so I wasn't sure about the best
practice.
-Alex
>
> > +
> > case VMXNET3_OM_CSUM:
> > net_tx_pkt_build_vheader(s->tx_pkt, false, true, 0);
> > VMW_PKPRN("L4 CSO requested\n");
> > diff --git a/hw/net/vmxnet3.h b/hw/net/vmxnet3.h
> > index 5b3b76ba7a..36a17b82aa 100644
> > --- a/hw/net/vmxnet3.h
> > +++ b/hw/net/vmxnet3.h
> > @@ -273,6 +273,7 @@ struct Vmxnet3_TxDesc {
> >
> > /* TxDesc.OM values */
> > #define VMXNET3_OM_NONE 0
> > +#define VMXNET3_OM_ENCAP 1
> > #define VMXNET3_OM_CSUM 2
> > #define VMXNET3_OM_TSO 3
> >
> >
>
