| |
|
 |
|
 |
|
|
|
| |
|
|
xen-devel
RE: [Xen-devel] [PATCH 4/6] xen/netback: Always pull throughPKT_PROT_LEN
On Wed, 2010-02-24 at 09:23 +0000, James Harper wrote:
> >
> > On Wed, 2010-02-24 at 08:28 +0000, Jan Beulich wrote:
> > > Could you point out what problem this addresses?
> >
> > It ensures that at least the TCP/IP headers will be pulled into the
> > linear part of the SKB. At least skb_checksum_setup relies on this and
> I
> > think it is a more generic assumption in at least some parts of the
> > network stack as well. The next patch increases PKT_PROT_LEN to
> include
> > the TCP options as well since we have observed cases where Windows
> > guests with PV drivers can generate a frame with a split at the point.
> >
>
> My PV drivers couldn't always be relied on to do this, because Windows
> couldn't be relied on to do it either. I just coalesced the first buffer
> to some minimum value depending on the header type. If you are doing TCP
> checksum offload then you have to tell Windows that you support IP
> checksum offload too - Linux doesn't support that so you have to lie to
> Windows about it and calculate the IP checksum in the PV drivers, and if
> you are doing that you have to have your own copy of the header anyway
> so it turns out not to be a big deal.
Yes, it's best if the guest takes care of this in the common case but
domain 0 can't trust the guest so we need a backstop in netback too. If
the guest is doing the sane thing the backstop shouldn't trigger.
Ian.
_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel
|
|
|
| |
|
Home