[Cluster-devel] Re: [NFS] [RFC PATCH 1/3] NLM lock failover - lock release

[Neil Brown]

On Friday June 30, trond.myklebust at fys.uio.no wrote:
> On Fri, 2006-06-30 at 14:15 +1000, Neil Brown wrote:
> 
> > The socket can only be bound to INADDR_ANY for UDP, and in that case
> > we already set rq_daddr correctly.
> 
> As I understand it, Wendy is considering the case of a multi-homed
> server. She wants to record the IP address on which we received the
> datagram so that she knows which locks to invalidate in the case of a
> migration of that particular IP address onto another server.
> 
> My point is that she won't get that information if the socket is bound
> to INADDR_ANY, as would be the case for a UDP socket.

Yes.  But that code fragment covered the TCP case only.
The extra called to ->getname was placed in svc_recvfrom which is only
called from svc_tcp_recvfrom, not from svc_udp_recvfrom (Yes, I agree
there is room for confusion there).
svc_udp_recvfrom already has
	rqstp->rq_addr.sin_addr.s_addr = skb->nh.iph->saddr;
	rqstp->rq_daddr = skb->nh.iph->daddr;

so it sets rq_daddr correctly as each packet is received.
rq_daddr was never set for requests received via TCP because it was
never used for tcp (it was used only to set the source address for UDP
replies).

NeilBrown