On Thu, May 17, 2012 at 12:17 PM, Ben Pfaff <b...@nicira.com> wrote: > On Thu, May 17, 2012 at 11:53:14AM -0700, Jesse Gross wrote: >> It's possible that packets that are sent on internal devices (from >> the OVS perspective) have already traversed the local IP stack. >> After they go through the internal device, they will again travel >> through the IP stack which may get confused by the presence of >> existing information in the skb. This clears out that information >> to avoid problems but deliberately leaves other metadata alone. >> This is to provide maximum flexibility in chaining together OVS >> and other Linux components. >> >> Bug #10995 >> >> Signed-off-by: Jesse Gross <je...@nicira.com> > > Acked-by: Ben Pfaff <b...@nicira.com> > > vport-internal_dev.c doesn't directly include <net/dst.h> for > skb_dst_drop(). I suppose it must include it indirectly, is that good > enough?
It's almost certain to always be OK because skbuff.h uses things from dst a fair amount but I added it as well. > Other paths that do similar cleaning also clear the rxhash. I guess > that's not necessary here, presumably because we're not actually > modifying the headers? That's right. > If there's a simple way to observe the user-visible change then it > would be kind to mention it in the commit log. I tried to add a little bit more detail. Thanks, I pushed this patch. _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
