On Sat, Nov 12, 2011 at 12:59 PM, Ben Pfaff <b...@nicira.com> wrote: > On Fri, Nov 11, 2011 at 05:21:52PM -0800, Jesse Gross wrote: >> On Fri, Nov 11, 2011 at 5:18 PM, Ben Pfaff <b...@nicira.com> wrote: >> > On Fri, Nov 11, 2011 at 03:50:56PM -0800, Ben Pfaff wrote: >> >> On Fri, Nov 11, 2011 at 03:48:25PM -0800, Jesse Gross wrote: >> >> > On Fri, Nov 11, 2011 at 3:43 PM, Ben Pfaff <b...@nicira.com> wrote: >> >> > > On Fri, Nov 11, 2011 at 03:38:51PM -0800, Jesse Gross wrote: >> >> > >> On Fri, Nov 11, 2011 at 3:13 PM, Ben Pfaff <b...@nicira.com> wrote: >> >> > >> > I guess that this will need reconsideration after the "encap" >> >> > >> > series, >> >> > >> > so I guess there's no point in a detailed response right now--let >> >> > >> > me >> >> > >> > know if you disagree. >> >> > >> >> >> > >> I agree that most of this patch has been overcome by events. ??I >> >> > >> think >> >> > >> there are a few useful things here that aren't addressed by the >> >> > >> latest >> >> > >> patch series though: >> >> > >> ??* Packets with invalid vlan headers shouldn't be dropped (maybe >> >> > >> this >> >> > >> is another rule, that invalid protocol headers should never drop the >> >> > >> packet?). >> >> > >> ??* I think that if the EtherType is 0x8100 then we should expect a >> >> > >> vlan attribute which may be an empty one, similar to other L3 >> >> > >> protocols. >> >> > > >> >> > > OK, do you want to address that or do you want me to do it? >> >> > >> >> > Could you do it while I review this series? >> >> >> >> OK. >> > >> > This is really preliminary. ??It compiles and passes "make check", I >> > won't vouch for anything else, hence no signed-off-by yet. ??It applies >> > after the "encap" series. >> > >> > Commit message is the same one you had, I haven't updated it at all. >> >> I haven't looked at the code at all but we should also add to the >> rules the principles that this patch implements (don't drop packets >> for broken tags; if the previous layer indicates that the attribute >> should be present then it should be, even if empty). > > How about this: > > Handling malformed packets
Looks good, thanks. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
