Re: [Xen-devel] [PATCH 5/7] xen-netfront: reduce gso_max_size to account for ethernet header

[Wei Liu <wei.liu2@xxxxxxxxxx>]

On Fri, Apr 12, 2013 at 09:57:15AM +0100, Ian Campbell wrote:
> On Fri, 2013-04-12 at 09:48 +0100, Wei Liu wrote:
> > On Fri, Apr 12, 2013 at 09:18:04AM +0100, Ian Campbell wrote:
> > > On Thu, 2013-04-11 at 21:04 +0100, Wei Liu wrote:
> > > > On Tue, Apr 09, 2013 at 12:07:33PM +0100, Wei Liu wrote:
> > > > > The maximum packet including ethernet header that can be handled by 
> > > > > netfront /
> > > > > netback wire format is 65535. Reduce gso_max_size accordingly.
> > > > > 
> > > > > Drop skb and print warning when skb->len > 65535. This can 1) save 
> > > > > the effort
> > > > > to send malformed packet to netback, 2) help spotting 
> > > > > misconfiguration of
> > > > > netfront in the future.
> > > > > 
> > > > 
> > > > Any opinion on how much space should be reserved? From a previous thread
> > > > Ben seemed to suggest 90 (Ethernet + VLAN tag + IPv6 + TCP + timestamp
> > > > option = 90 bytes).
> > > 
> > > I trust Ben and that seems as good as anything to me.
> > > 
> > > Is this the sort of limit others might be interested in, should we have
> > > a global #define?
> > > 
> > 
> > We shall have a global define in this case.
> > 
> > #define XEN_NETFRONT_MAX_HEADER ? I'm bad at naming things.
> 
> I meant an include/linux/skbuff.h (or some suitable header) #define
> SKB_MAX_FOO type thing...
> 

But we don't have handle on this. If I understand correctly the
discussion in other thread, 90 is empirical value, not something
documented.


Wei.

> Ian.
> 

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