Aren/NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

add support for skipping IPv6 checksum validation aka offloading -

Browse Source 
for xennet(4) both Rx and Tx, for xvif(4) only Tx for now
This commit is contained in:
jdolecek 2020-03-22 00:11:02 +00:00
parent 33e5bb572d
commit e02ab9f0fa
3 changed files with 108 additions and 39 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
sys/arch/xen/xen/if_xennet_xenbus.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: if_xennet_xenbus.c,v 1.92 2020/03/19 10:53:43 jdolecek Exp $ */
/* $NetBSD: if_xennet_xenbus.c,v 1.93 2020/03/22 00:11:02 jdolecek Exp $ */
/*
* Copyright (c) 2006 Manuel Bouyer.
@ -84,7 +84,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: if_xennet_xenbus.c,v 1.92 2020/03/19 10:53:43 jdolecek Exp $");
__KERNEL_RCSID(0, "$NetBSD: if_xennet_xenbus.c,v 1.93 2020/03/22 00:11:02 jdolecek Exp $");
#include "opt_xen.h"
#include "opt_nfs_boot.h"
@ -389,7 +389,14 @@ xennet_xenbus_attach(device_t parent, device_t self, void *aux)
ifp->if_capabilities =
IFCAP_CSUM_IPv4_Rx | IFCAP_CSUM_IPv4_Tx
| IFCAP_CSUM_UDPv4_Rx | IFCAP_CSUM_UDPv4_Tx
| IFCAP_CSUM_TCPv4_Rx | IFCAP_CSUM_TCPv4_Tx;
| IFCAP_CSUM_TCPv4_Rx | IFCAP_CSUM_TCPv4_Tx
| IFCAP_CSUM_UDPv6_Rx | IFCAP_CSUM_UDPv6_Tx
| IFCAP_CSUM_TCPv6_Rx | IFCAP_CSUM_TCPv6_Tx;
#define XN_M_CSUM_SUPPORTED ( \
M_CSUM_TCPv4 | M_CSUM_UDPv4 | M_CSUM_IPv4 \
| M_CSUM_TCPv6 | M_CSUM_UDPv6 \
)
IFQ_SET_READY(&ifp->if_snd);
if_attach(ifp);
ether_ifattach(ifp, sc->sc_enaddr);
@ -1229,8 +1236,7 @@ xennet_softstart(void *arg)
break;
}
if ((m->m_pkthdr.csum_flags &
(M_CSUM_TCPv4 | M_CSUM_UDPv4 | M_CSUM_IPv4)) != 0) {
if ((m->m_pkthdr.csum_flags & XN_M_CSUM_SUPPORTED) != 0) {
txflags = NETTXF_csum_blank;
} else {
txflags = NETTXF_data_validated;

115
sys/arch/xen/xen/xennet_checksum.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: xennet_checksum.c,v 1.8 2020/03/19 10:53:43 jdolecek Exp $ */
/* $NetBSD: xennet_checksum.c,v 1.9 2020/03/22 00:11:02 jdolecek Exp $ */
/*-
* Copyright (c)2006 YAMAMOTO Takashi,
@ -27,7 +27,11 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: xennet_checksum.c,v 1.8 2020/03/19 10:53:43 jdolecek Exp $");
__KERNEL_RCSID(0, "$NetBSD: xennet_checksum.c,v 1.9 2020/03/22 00:11:02 jdolecek Exp $");
#ifdef _KERNEL_OPT
#include "opt_inet.h"
#endif
#include <sys/types.h>
#include <sys/param.h>
@ -43,6 +47,8 @@ __KERNEL_RCSID(0, "$NetBSD: xennet_checksum.c,v 1.8 2020/03/19 10:53:43 jdolecek
#include <netinet/ip.h>
#include <netinet/tcp.h>
#include <netinet/udp.h>
#include <netinet/ip6.h>
#include <netinet6/in6_offload.h>
#include <xen/xennet_checksum.h>
@ -57,16 +63,15 @@ EVCNT_ATTACH_STATIC(xn_cksum_defer);
EVCNT_ATTACH_STATIC(xn_cksum_undefer);
EVCNT_ATTACH_STATIC(xn_cksum_valid);
#ifdef XENNET_DEBUG
/* ratecheck(9) for checksum validation failures */
static const struct timeval xn_cksum_errintvl = { 600, 0 }; /* 10 min, each */
#endif
static void *
m_extract(struct mbuf *m, int off, int len)
{
KASSERT(m->m_pkthdr.len >= off + len);
KASSERT(m->m_len >= off + len);
if (m->m_pkthdr.len >= off + len)
if (m->m_len >= off + len)
return mtod(m, char *) + off;
else
return NULL;
@ -80,7 +85,10 @@ int
xennet_checksum_fill(struct ifnet *ifp, struct mbuf *m, bool data_validated)
{
const struct ether_header *eh;
struct ip *iph;
struct ip *iph = NULL;
#ifdef INET6
struct ip6_hdr *ip6h = NULL;
#endif
int ehlen;
int iphlen;
int iplen;
@ -98,27 +106,49 @@ xennet_checksum_fill(struct ifnet *ifp, struct mbuf *m, bool data_validated)
return EINVAL;
}
etype = eh->ether_type;
if (etype == htobe16(ETHERTYPE_VLAN)) {
ehlen = ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN;
} else if (etype == htobe16(ETHERTYPE_IP)) {
ehlen = ETHER_HDR_LEN;
} else {
static struct timeval lasttime;
if (ratecheck(&lasttime, &xn_cksum_errintvl))
printf("%s: unknown etype %#x passed%s\n",
ifp->if_xname, ntohs(etype),
data_validated ? "" : " no checksum");
return EINVAL;
ehlen = ETHER_HDR_LEN;
if (__predict_false(etype == htons(ETHERTYPE_VLAN))) {
struct ether_vlan_header *evl = m_extract(m, 0, sizeof(*evl));
if (evl == NULL) {
/* Too short, packet will be dropped by upper layer */
return EINVAL;
}
ehlen += ETHER_VLAN_ENCAP_LEN;
etype = ntohs(evl->evl_proto);
}
iph = m_extract(m, ehlen, sizeof(*iph));
if (iph == NULL) {
/* Too short, packet will be dropped by upper layer */
return EINVAL;
switch (etype) {
case htons(ETHERTYPE_IP):
iph = m_extract(m, ehlen, sizeof(*iph));
if (iph == NULL) {
/* Too short, packet will be dropped by upper layer */
return EINVAL;
}
nxt = iph->ip_p;
iphlen = iph->ip_hl << 2;
iplen = ntohs(iph->ip_len);
break;
#ifdef INET6
case htons(ETHERTYPE_IPV6):
ip6h = m_extract(m, ehlen, sizeof(*ip6h));
if (ip6h == NULL) {
/* Too short, packet will be dropped by upper layer */
return EINVAL;
}
if ((ip6h->ip6_vfc & IPV6_VERSION_MASK) != IPV6_VERSION) {
/* Bad version */
return EINVAL;
}
nxt = ip6h->ip6_nxt;
iphlen = sizeof(*ip6h);
iplen = ntohs(ip6h->ip6_plen);
break;
#endif
default:
/* Not supported ethernet type */
return EOPNOTSUPP;
}
nxt = iph->ip_p;
iphlen = iph->ip_hl << 2;
iplen = ntohs(iph->ip_len);
if (ehlen + iplen > m->m_pkthdr.len) {
/* Too short, packet will be dropped by upper layer */
return EINVAL;
@ -126,27 +156,47 @@ xennet_checksum_fill(struct ifnet *ifp, struct mbuf *m, bool data_validated)
switch (nxt) {
case IPPROTO_UDP:
m->m_pkthdr.csum_flags = M_CSUM_UDPv4 | M_CSUM_IPv4;
if (iph)
m->m_pkthdr.csum_flags = M_CSUM_UDPv4 | M_CSUM_IPv4;
#ifdef INET6
else
m->m_pkthdr.csum_flags = M_CSUM_UDPv6;
#endif
m->m_pkthdr.csum_data = offsetof(struct udphdr, uh_sum);
m->m_pkthdr.csum_data |= iphlen << 16;
break;
case IPPROTO_TCP:
m->m_pkthdr.csum_flags = M_CSUM_TCPv4 | M_CSUM_IPv4;
if (iph)
m->m_pkthdr.csum_flags = M_CSUM_TCPv4 | M_CSUM_IPv4;
#ifdef INET6
else
m->m_pkthdr.csum_flags = M_CSUM_TCPv6;
#endif
m->m_pkthdr.csum_data = offsetof(struct tcphdr, th_sum);
m->m_pkthdr.csum_data |= iphlen << 16;
break;
case IPPROTO_ICMP:
case IPPROTO_IGMP:
m->m_pkthdr.csum_flags = M_CSUM_IPv4;
if (iph)
m->m_pkthdr.csum_flags = M_CSUM_IPv4;
m->m_pkthdr.csum_data = iphlen << 16;
break;
case IPPROTO_HOPOPTS:
case IPPROTO_ICMPV6:
case IPPROTO_FRAGMENT:
/* nothing to do */
error = 0;
goto out;
/* NOTREACHED */
default:
{
#ifdef XENNET_DEBUG
static struct timeval lasttime;
if (ratecheck(&lasttime, &xn_cksum_errintvl))
printf("%s: unknown proto %d passed%s\n",
ifp->if_xname, nxt,
data_validated ? "" : " no checksum");
#endif /* XENNET_DEBUG */
error = EINVAL;
goto out;
}
@ -163,7 +213,7 @@ xennet_checksum_fill(struct ifnet *ifp, struct mbuf *m, bool data_validated)
* checksumming requires this. in_undefer_cksum()
* also needs it to be zero.
*/
if (m->m_pkthdr.csum_flags & M_CSUM_IPv4)
if (iph != NULL && (m->m_pkthdr.csum_flags & M_CSUM_IPv4))
iph->ip_sum = 0;
if (sw_csum & (M_CSUM_IPv4|M_CSUM_UDPv4|M_CSUM_TCPv4)) {
@ -171,6 +221,13 @@ xennet_checksum_fill(struct ifnet *ifp, struct mbuf *m, bool data_validated)
sw_csum & (M_CSUM_IPv4|M_CSUM_UDPv4|M_CSUM_TCPv4));
}
#ifdef INET6
if (sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6)) {
in6_undefer_cksum(m, ehlen,
sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6));
}
#endif
if (m->m_pkthdr.csum_flags != 0) {
xn_cksum_defer.ev_count++;
#ifdef M_CSUM_BLANK

16
sys/arch/xen/xen/xennetback_xenbus.c
View File

@ -1,4 +1,4 @@
/* $NetBSD: xennetback_xenbus.c,v 1.83 2020/03/21 23:25:53 jdolecek Exp $ */
/* $NetBSD: xennetback_xenbus.c,v 1.84 2020/03/22 00:11:02 jdolecek Exp $ */
/*
* Copyright (c) 2006 Manuel Bouyer.
@ -25,7 +25,7 @@
*/
#include <sys/cdefs.h>
__KERNEL_RCSID(0, "$NetBSD: xennetback_xenbus.c,v 1.83 2020/03/21 23:25:53 jdolecek Exp $");
__KERNEL_RCSID(0, "$NetBSD: xennetback_xenbus.c,v 1.84 2020/03/22 00:11:02 jdolecek Exp $");
#include "opt_xen.h"
@ -306,7 +306,13 @@ xennetback_xenbus_create(struct xenbus_device *xbusd)
ifp->if_capabilities =
IFCAP_CSUM_IPv4_Tx
| IFCAP_CSUM_UDPv4_Tx
| IFCAP_CSUM_TCPv4_Tx;
| IFCAP_CSUM_TCPv4_Tx
| IFCAP_CSUM_UDPv6_Tx
| IFCAP_CSUM_TCPv6_Tx;
#define XN_M_CSUM_SUPPORTED ( \
M_CSUM_TCPv4 | M_CSUM_UDPv4 | M_CSUM_IPv4 \
| M_CSUM_TCPv6 | M_CSUM_UDPv6 \
)
ifp->if_ioctl = xennetback_ifioctl;
ifp->if_start = xennetback_ifstart;
ifp->if_watchdog = xennetback_ifwatchdog;
@ -1050,7 +1056,7 @@ xennetback_ifsoftstart_transfer(void *arg)
rxresp->offset = offset;
rxresp->status = m->m_pkthdr.len;
if ((m->m_pkthdr.csum_flags &
(M_CSUM_TCPv4 | M_CSUM_UDPv4 | M_CSUM_IPv4)) != 0) {
XN_M_CSUM_SUPPORTED) != 0) {
rxresp->flags = NETRXF_csum_blank;
} else {
rxresp->flags = NETRXF_data_validated;
@ -1366,7 +1372,7 @@ xennetback_ifsoftstart_copy(void *arg)
rxresp->offset = 0;
rxresp->status = m->m_pkthdr.len;
if ((m->m_pkthdr.csum_flags &
(M_CSUM_TCPv4 | M_CSUM_UDPv4 | M_CSUM_IPv4)) != 0) {
XN_M_CSUM_SUPPORTED) != 0) {
rxresp->flags = NETRXF_csum_blank;
} else {
rxresp->flags = NETRXF_data_validated;