Teach the loopback interface about checksum generation and validation

avoidance:

- Enable setting the RXCSUM and TXCSUM flags for loopback interfaces;
  set both by default.
- When RXCSUM is set, flag packets sent over the loopback interface as
  having checked and valid IP, UDP, TCP checksums so that higher
  protocol layers won't check them.
- Always clear CSUM_{IP,UDP_TCP} checksum required flags on transmit,
  as they will have gotten there as a result of TXCSUM being set.

This is done only for packets explicitly sent over the loopback, not
simulated loopback via if_simloop() due to !SIMPLEX interfaces, etc.

Note that enabling TXCSUM but not RXCSUM will lead to unhappiness, as
checksums won't be generated but will be validated.

Kris reports that this leads to significant performance improvements
in loopback benchmarking with TCP and UDP for throughput:

	RXCSUM 	RXCSUM+TXCSUM
TCP	15%	37%
UDP	10%	74%

Update man page.

Reviewed by:	sam
Tested by:	kris
MFC after:	1 week

1 files changed, 22 insertions, 1 deletions

diff --git a/sys/net/if_loop.c b/sys/net/if_loop.c
index d5548fcd935c..3fd003e3c524 100644
--- a/sys/net/if_loop.c
+++ b/sys/net/if_loop.c
@@ -138,6 +138,8 @@ lo_clone_create(struct if_clone *ifc, int unit, caddr_t params)
 	ifp->if_ioctl = loioctl;
 	ifp->if_output = looutput;
 	ifp->if_snd.ifq_maxlen = ifqmaxlen;
+	ifp->if_hwassist = ifp->if_capabilities = ifp->if_capenable =
+	    IFCAP_HWCSUM;
 	if_attach(ifp);
 	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
 	if (V_loif == NULL)
@@ -212,6 +214,13 @@ looutput(struct ifnet *ifp, struct mbuf *m, struct sockaddr *dst,
 #if 1	/* XXX */
 	switch (dst->sa_family) {
 	case AF_INET:
+		if (ifp->if_capenable & IFCAP_RXCSUM) {
+			m->m_pkthdr.csum_data = 0xffff;
+			m->m_pkthdr.csum_flags = CSUM_DATA_VALID |
+			    CSUM_PSEUDO_HDR | CSUM_IP_CHECKED |
+			    CSUM_IP_VALID | CSUM_SCTP_VALID;
+		}
+		m->m_pkthdr.csum_flags &= ~(CSUM_IP | CSUM_TCP | CSUM_UDP);
 	case AF_INET6:
 	case AF_IPX:
 	case AF_APPLETALK:
@@ -348,7 +357,7 @@ loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 {
 	struct ifaddr *ifa;
 	struct ifreq *ifr = (struct ifreq *)data;
-	int error = 0;
+	int error = 0, mask;
 
 	switch (cmd) {
 	case SIOCSIFADDR:
@@ -391,6 +400,18 @@ loioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 	case SIOCSIFFLAGS:
 		break;
 
+	case SIOCSIFCAP:
+		mask = ifp->if_capenable ^ ifr->ifr_reqcap;
+		if ((mask & IFCAP_RXCSUM) != 0)
+			ifp->if_capenable ^= IFCAP_RXCSUM;
+		if ((mask & IFCAP_TXCSUM) != 0)
+			ifp->if_capenable ^= IFCAP_TXCSUM;
+		if (ifp->if_capenable & IFCAP_TXCSUM)
+			ifp->if_hwassist = CSUM_IP | CSUM_TCP | CSUM_UDP;
+		else
+			ifp->if_hwassist = 0;
+		break;
+
 	default:
 		error = EINVAL;
 	}