Mon Apr 25 22:20:59 2011 UTC ()
undefer csum in looutput.
looutput is used by various code (ether_output, mcast) to loopback packets.


(yamt)
diff -r1.72 -r1.73 src/sys/net/if_loop.c
diff -r1.7 -r1.8 src/sys/netinet/in_offload.h
diff -r1.6 -r1.7 src/sys/netinet6/in6_offload.h
diff -r1.139 -r1.140 src/sys/netinet6/ip6_output.c

cvs diff -r1.72 -r1.73 src/sys/net/if_loop.c (switch to unified diff)

--- src/sys/net/if_loop.c 2010/04/05 07:22:23 1.72
+++ src/sys/net/if_loop.c 2011/04/25 22:20:59 1.73
@@ -1,457 +1,472 @@ @@ -1,457 +1,472 @@
1/* $NetBSD: if_loop.c,v 1.72 2010/04/05 07:22:23 joerg Exp $ */ 1/* $NetBSD: if_loop.c,v 1.73 2011/04/25 22:20:59 yamt Exp $ */
2 2
3/* 3/*
4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 4 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
5 * All rights reserved. 5 * All rights reserved.
6 * 6 *
7 * Redistribution and use in source and binary forms, with or without 7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions 8 * modification, are permitted provided that the following conditions
9 * are met: 9 * are met:
10 * 1. Redistributions of source code must retain the above copyright 10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer. 11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright 12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the 13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution. 14 * documentation and/or other materials provided with the distribution.
15 * 3. Neither the name of the project nor the names of its contributors 15 * 3. Neither the name of the project nor the names of its contributors
16 * may be used to endorse or promote products derived from this software 16 * may be used to endorse or promote products derived from this software
17 * without specific prior written permission. 17 * without specific prior written permission.
18 * 18 *
19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 19 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 22 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29 * SUCH DAMAGE. 29 * SUCH DAMAGE.
30 */ 30 */
31 31
32/* 32/*
33 * Copyright (c) 1982, 1986, 1993 33 * Copyright (c) 1982, 1986, 1993
34 * The Regents of the University of California. All rights reserved. 34 * The Regents of the University of California. All rights reserved.
35 * 35 *
36 * Redistribution and use in source and binary forms, with or without 36 * Redistribution and use in source and binary forms, with or without
37 * modification, are permitted provided that the following conditions 37 * modification, are permitted provided that the following conditions
38 * are met: 38 * are met:
39 * 1. Redistributions of source code must retain the above copyright 39 * 1. Redistributions of source code must retain the above copyright
40 * notice, this list of conditions and the following disclaimer. 40 * notice, this list of conditions and the following disclaimer.
41 * 2. Redistributions in binary form must reproduce the above copyright 41 * 2. Redistributions in binary form must reproduce the above copyright
42 * notice, this list of conditions and the following disclaimer in the 42 * notice, this list of conditions and the following disclaimer in the
43 * documentation and/or other materials provided with the distribution. 43 * documentation and/or other materials provided with the distribution.
44 * 3. Neither the name of the University nor the names of its contributors 44 * 3. Neither the name of the University nor the names of its contributors
45 * may be used to endorse or promote products derived from this software 45 * may be used to endorse or promote products derived from this software
46 * without specific prior written permission. 46 * without specific prior written permission.
47 * 47 *
48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 48 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 49 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 50 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 51 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 52 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 53 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 54 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 55 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 56 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 57 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
58 * SUCH DAMAGE. 58 * SUCH DAMAGE.
59 * 59 *
60 * @(#)if_loop.c 8.2 (Berkeley) 1/9/95 60 * @(#)if_loop.c 8.2 (Berkeley) 1/9/95
61 */ 61 */
62 62
63/* 63/*
64 * Loopback interface driver for protocol testing and timing. 64 * Loopback interface driver for protocol testing and timing.
65 */ 65 */
66 66
67#include <sys/cdefs.h> 67#include <sys/cdefs.h>
68__KERNEL_RCSID(0, "$NetBSD: if_loop.c,v 1.72 2010/04/05 07:22:23 joerg Exp $"); 68__KERNEL_RCSID(0, "$NetBSD: if_loop.c,v 1.73 2011/04/25 22:20:59 yamt Exp $");
69 69
70#include "opt_inet.h" 70#include "opt_inet.h"
71#include "opt_atalk.h" 71#include "opt_atalk.h"
72#include "opt_iso.h" 72#include "opt_iso.h"
73#include "opt_ipx.h" 73#include "opt_ipx.h"
74#include "opt_mbuftrace.h" 74#include "opt_mbuftrace.h"
75 75
76 76
77#include <sys/param.h> 77#include <sys/param.h>
78#include <sys/systm.h> 78#include <sys/systm.h>
79#include <sys/kernel.h> 79#include <sys/kernel.h>
80#include <sys/mbuf.h> 80#include <sys/mbuf.h>
81#include <sys/socket.h> 81#include <sys/socket.h>
82#include <sys/errno.h> 82#include <sys/errno.h>
83#include <sys/ioctl.h> 83#include <sys/ioctl.h>
84#include <sys/time.h> 84#include <sys/time.h>
85 85
86#include <sys/cpu.h> 86#include <sys/cpu.h>
87 87
88#include <net/if.h> 88#include <net/if.h>
89#include <net/if_types.h> 89#include <net/if_types.h>
90#include <net/netisr.h> 90#include <net/netisr.h>
91#include <net/route.h> 91#include <net/route.h>
92 92
93#ifdef INET 93#ifdef INET
94#include <netinet/in.h> 94#include <netinet/in.h>
95#include <netinet/in_systm.h> 95#include <netinet/in_systm.h>
96#include <netinet/in_var.h> 96#include <netinet/in_var.h>
 97#include <netinet/in_offload.h>
97#include <netinet/ip.h> 98#include <netinet/ip.h>
98#endif 99#endif
99 100
100#ifdef INET6 101#ifdef INET6
101#ifndef INET 102#ifndef INET
102#include <netinet/in.h> 103#include <netinet/in.h>
103#endif 104#endif
104#include <netinet6/in6_var.h> 105#include <netinet6/in6_var.h>
 106#include <netinet6/in6_offload.h>
105#include <netinet/ip6.h> 107#include <netinet/ip6.h>
106#endif 108#endif
107 109
108 
109#ifdef IPX 110#ifdef IPX
110#include <netipx/ipx.h> 111#include <netipx/ipx.h>
111#include <netipx/ipx_if.h> 112#include <netipx/ipx_if.h>
112#endif 113#endif
113 114
114#ifdef ISO 115#ifdef ISO
115#include <netiso/iso.h> 116#include <netiso/iso.h>
116#include <netiso/iso_var.h> 117#include <netiso/iso_var.h>
117#endif 118#endif
118 119
119#ifdef NETATALK 120#ifdef NETATALK
120#include <netatalk/at.h> 121#include <netatalk/at.h>
121#include <netatalk/at_var.h> 122#include <netatalk/at_var.h>
122#endif 123#endif
123 124
124#include <net/bpf.h> 125#include <net/bpf.h>
125 126
126#if defined(LARGE_LOMTU) 127#if defined(LARGE_LOMTU)
127#define LOMTU (131072 + MHLEN + MLEN) 128#define LOMTU (131072 + MHLEN + MLEN)
128#define LOMTU_MAX LOMTU 129#define LOMTU_MAX LOMTU
129#else 130#else
130#define LOMTU (32768 + MHLEN + MLEN) 131#define LOMTU (32768 + MHLEN + MLEN)
131#define LOMTU_MAX (65536 + MHLEN + MLEN) 132#define LOMTU_MAX (65536 + MHLEN + MLEN)
132#endif 133#endif
133 134
134#ifdef ALTQ 135#ifdef ALTQ
135static void lostart(struct ifnet *); 136static void lostart(struct ifnet *);
136#endif 137#endif
137 138
138static int loop_clone_create(struct if_clone *, int); 139static int loop_clone_create(struct if_clone *, int);
139static int loop_clone_destroy(struct ifnet *); 140static int loop_clone_destroy(struct ifnet *);
140 141
141static struct if_clone loop_cloner = 142static struct if_clone loop_cloner =
142 IF_CLONE_INITIALIZER("lo", loop_clone_create, loop_clone_destroy); 143 IF_CLONE_INITIALIZER("lo", loop_clone_create, loop_clone_destroy);
143 144
144void 145void
145loopattach(int n) 146loopattach(int n)
146{ 147{
147 148
148 (void)loop_clone_create(&loop_cloner, 0); /* lo0 always exists */ 149 (void)loop_clone_create(&loop_cloner, 0); /* lo0 always exists */
149 if_clone_attach(&loop_cloner); 150 if_clone_attach(&loop_cloner);
150} 151}
151 152
152static int 153static int
153loop_clone_create(struct if_clone *ifc, int unit) 154loop_clone_create(struct if_clone *ifc, int unit)
154{ 155{
155 struct ifnet *ifp; 156 struct ifnet *ifp;
156 157
157 ifp = if_alloc(IFT_LOOP); 158 ifp = if_alloc(IFT_LOOP);
158 159
159 if_initname(ifp, ifc->ifc_name, unit); 160 if_initname(ifp, ifc->ifc_name, unit);
160 161
161 ifp->if_mtu = LOMTU; 162 ifp->if_mtu = LOMTU;
162 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST | IFF_RUNNING; 163 ifp->if_flags = IFF_LOOPBACK | IFF_MULTICAST | IFF_RUNNING;
163 ifp->if_ioctl = loioctl; 164 ifp->if_ioctl = loioctl;
164 ifp->if_output = looutput; 165 ifp->if_output = looutput;
165#ifdef ALTQ 166#ifdef ALTQ
166 ifp->if_start = lostart; 167 ifp->if_start = lostart;
167#endif 168#endif
168 ifp->if_type = IFT_LOOP; 169 ifp->if_type = IFT_LOOP;
169 ifp->if_hdrlen = 0; 170 ifp->if_hdrlen = 0;
170 ifp->if_addrlen = 0; 171 ifp->if_addrlen = 0;
171 ifp->if_dlt = DLT_NULL; 172 ifp->if_dlt = DLT_NULL;
172 IFQ_SET_READY(&ifp->if_snd); 173 IFQ_SET_READY(&ifp->if_snd);
173 if (unit == 0) 174 if (unit == 0)
174 lo0ifp = ifp; 175 lo0ifp = ifp;
175 if_attach(ifp); 176 if_attach(ifp);
176 if_alloc_sadl(ifp); 177 if_alloc_sadl(ifp);
177 bpf_attach(ifp, DLT_NULL, sizeof(u_int)); 178 bpf_attach(ifp, DLT_NULL, sizeof(u_int));
178#ifdef MBUFTRACE 179#ifdef MBUFTRACE
179 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF, 180 ifp->if_mowner = malloc(sizeof(struct mowner), M_DEVBUF,
180 M_WAITOK | M_ZERO); 181 M_WAITOK | M_ZERO);
181 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname, 182 strlcpy(ifp->if_mowner->mo_name, ifp->if_xname,
182 sizeof(ifp->if_mowner->mo_name)); 183 sizeof(ifp->if_mowner->mo_name));
183 MOWNER_ATTACH(ifp->if_mowner); 184 MOWNER_ATTACH(ifp->if_mowner);
184#endif 185#endif
185 186
186 return (0); 187 return (0);
187} 188}
188 189
189static int 190static int
190loop_clone_destroy(struct ifnet *ifp) 191loop_clone_destroy(struct ifnet *ifp)
191{ 192{
192 193
193 if (ifp == lo0ifp) 194 if (ifp == lo0ifp)
194 return (EPERM); 195 return (EPERM);
195 196
196#ifdef MBUFTRACE 197#ifdef MBUFTRACE
197 MOWNER_DETACH(ifp->if_mowner); 198 MOWNER_DETACH(ifp->if_mowner);
198 free(ifp->if_mowner, M_DEVBUF); 199 free(ifp->if_mowner, M_DEVBUF);
199#endif 200#endif
200 201
201 bpf_detach(ifp); 202 bpf_detach(ifp);
202 if_detach(ifp); 203 if_detach(ifp);
203 204
204 free(ifp, M_DEVBUF); 205 free(ifp, M_DEVBUF);
205 206
206 return (0); 207 return (0);
207} 208}
208 209
209int 210int
210looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst, 211looutput(struct ifnet *ifp, struct mbuf *m, const struct sockaddr *dst,
211 struct rtentry *rt) 212 struct rtentry *rt)
212{ 213{
213 int s, isr; 214 int s, isr;
214 struct ifqueue *ifq = NULL; 215 struct ifqueue *ifq = NULL;
 216 int csum_flags;
215 217
216 MCLAIM(m, ifp->if_mowner); 218 MCLAIM(m, ifp->if_mowner);
217 if ((m->m_flags & M_PKTHDR) == 0) 219 if ((m->m_flags & M_PKTHDR) == 0)
218 panic("looutput: no header mbuf"); 220 panic("looutput: no header mbuf");
219 if (ifp->if_flags & IFF_LOOPBACK) 221 if (ifp->if_flags & IFF_LOOPBACK)
220 bpf_mtap_af(ifp, dst->sa_family, m); 222 bpf_mtap_af(ifp, dst->sa_family, m);
221 m->m_pkthdr.rcvif = ifp; 223 m->m_pkthdr.rcvif = ifp;
222 224
223 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) { 225 if (rt && rt->rt_flags & (RTF_REJECT|RTF_BLACKHOLE)) {
224 m_freem(m); 226 m_freem(m);
225 return (rt->rt_flags & RTF_BLACKHOLE ? 0 : 227 return (rt->rt_flags & RTF_BLACKHOLE ? 0 :
226 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH); 228 rt->rt_flags & RTF_HOST ? EHOSTUNREACH : ENETUNREACH);
227 } 229 }
228 230
229 ifp->if_opackets++; 231 ifp->if_opackets++;
230 ifp->if_obytes += m->m_pkthdr.len; 232 ifp->if_obytes += m->m_pkthdr.len;
231 233
232#ifdef ALTQ 234#ifdef ALTQ
233 /* 235 /*
234 * ALTQ on the loopback interface is just for debugging. It's 236 * ALTQ on the loopback interface is just for debugging. It's
235 * used only for loopback interfaces, not for a simplex interface. 237 * used only for loopback interfaces, not for a simplex interface.
236 */ 238 */
237 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) && 239 if ((ALTQ_IS_ENABLED(&ifp->if_snd) || TBR_IS_ENABLED(&ifp->if_snd)) &&
238 ifp->if_start == lostart) { 240 ifp->if_start == lostart) {
239 struct altq_pktattr pktattr; 241 struct altq_pktattr pktattr;
240 int error; 242 int error;
241 243
242 /* 244 /*
243 * If the queueing discipline needs packet classification, 245 * If the queueing discipline needs packet classification,
244 * do it before prepending the link headers. 246 * do it before prepending the link headers.
245 */ 247 */
246 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family, &pktattr); 248 IFQ_CLASSIFY(&ifp->if_snd, m, dst->sa_family, &pktattr);
247 249
248 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT); 250 M_PREPEND(m, sizeof(uint32_t), M_DONTWAIT);
249 if (m == NULL) 251 if (m == NULL)
250 return (ENOBUFS); 252 return (ENOBUFS);
251 *(mtod(m, uint32_t *)) = dst->sa_family; 253 *(mtod(m, uint32_t *)) = dst->sa_family;
252 254
253 s = splnet(); 255 s = splnet();
254 IFQ_ENQUEUE(&ifp->if_snd, m, &pktattr, error); 256 IFQ_ENQUEUE(&ifp->if_snd, m, &pktattr, error);
255 (*ifp->if_start)(ifp); 257 (*ifp->if_start)(ifp);
256 splx(s); 258 splx(s);
257 return (error); 259 return (error);
258 } 260 }
259#endif /* ALTQ */ 261#endif /* ALTQ */
260 262
261 m_tag_delete_nonpersistent(m); 263 m_tag_delete_nonpersistent(m);
262 264
263 switch (dst->sa_family) { 265 switch (dst->sa_family) {
264 266
265#ifdef INET 267#ifdef INET
266 case AF_INET: 268 case AF_INET:
 269 csum_flags = m->m_pkthdr.csum_flags;
 270 KASSERT((csum_flags & ~(M_CSUM_IPv4|M_CSUM_UDPv4)) == 0);
 271 if (csum_flags != 0 && IN_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
 272 ip_undefer_csum(m, 0, csum_flags);
 273 }
 274 m->m_pkthdr.csum_flags = 0;
267 ifq = &ipintrq; 275 ifq = &ipintrq;
268 isr = NETISR_IP; 276 isr = NETISR_IP;
269 break; 277 break;
270#endif 278#endif
271#ifdef INET6 279#ifdef INET6
272 case AF_INET6: 280 case AF_INET6:
 281 csum_flags = m->m_pkthdr.csum_flags;
 282 KASSERT((csum_flags & ~M_CSUM_UDPv6) == 0);
 283 if (csum_flags != 0 &&
 284 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)) {
 285 ip6_undefer_csum(m, 0, csum_flags);
 286 }
 287 m->m_pkthdr.csum_flags = 0;
273 m->m_flags |= M_LOOP; 288 m->m_flags |= M_LOOP;
274 ifq = &ip6intrq; 289 ifq = &ip6intrq;
275 isr = NETISR_IPV6; 290 isr = NETISR_IPV6;
276 break; 291 break;
277#endif 292#endif
278#ifdef ISO 293#ifdef ISO
279 case AF_ISO: 294 case AF_ISO:
280 ifq = &clnlintrq; 295 ifq = &clnlintrq;
281 isr = NETISR_ISO; 296 isr = NETISR_ISO;
282 break; 297 break;
283#endif 298#endif
284#ifdef IPX 299#ifdef IPX
285 case AF_IPX: 300 case AF_IPX:
286 ifq = &ipxintrq; 301 ifq = &ipxintrq;
287 isr = NETISR_IPX; 302 isr = NETISR_IPX;
288 break; 303 break;
289#endif 304#endif
290#ifdef NETATALK 305#ifdef NETATALK
291 case AF_APPLETALK: 306 case AF_APPLETALK:
292 ifq = &atintrq2; 307 ifq = &atintrq2;
293 isr = NETISR_ATALK; 308 isr = NETISR_ATALK;
294 break; 309 break;
295#endif 310#endif
296 default: 311 default:
297 printf("%s: can't handle af%d\n", ifp->if_xname, 312 printf("%s: can't handle af%d\n", ifp->if_xname,
298 dst->sa_family); 313 dst->sa_family);
299 m_freem(m); 314 m_freem(m);
300 return (EAFNOSUPPORT); 315 return (EAFNOSUPPORT);
301 } 316 }
302 s = splnet(); 317 s = splnet();
303 if (IF_QFULL(ifq)) { 318 if (IF_QFULL(ifq)) {
304 IF_DROP(ifq); 319 IF_DROP(ifq);
305 m_freem(m); 320 m_freem(m);
306 splx(s); 321 splx(s);
307 return (ENOBUFS); 322 return (ENOBUFS);
308 } 323 }
309 IF_ENQUEUE(ifq, m); 324 IF_ENQUEUE(ifq, m);
310 schednetisr(isr); 325 schednetisr(isr);
311 ifp->if_ipackets++; 326 ifp->if_ipackets++;
312 ifp->if_ibytes += m->m_pkthdr.len; 327 ifp->if_ibytes += m->m_pkthdr.len;
313 splx(s); 328 splx(s);
314 return (0); 329 return (0);
315} 330}
316 331
317#ifdef ALTQ 332#ifdef ALTQ
318static void 333static void
319lostart(struct ifnet *ifp) 334lostart(struct ifnet *ifp)
320{ 335{
321 struct ifqueue *ifq; 336 struct ifqueue *ifq;
322 struct mbuf *m; 337 struct mbuf *m;
323 uint32_t af; 338 uint32_t af;
324 int s, isr; 339 int s, isr;
325 340
326 for (;;) { 341 for (;;) {
327 IFQ_DEQUEUE(&ifp->if_snd, m); 342 IFQ_DEQUEUE(&ifp->if_snd, m);
328 if (m == NULL) 343 if (m == NULL)
329 return; 344 return;
330 345
331 af = *(mtod(m, uint32_t *)); 346 af = *(mtod(m, uint32_t *));
332 m_adj(m, sizeof(uint32_t)); 347 m_adj(m, sizeof(uint32_t));
333 348
334 switch (af) { 349 switch (af) {
335#ifdef INET 350#ifdef INET
336 case AF_INET: 351 case AF_INET:
337 ifq = &ipintrq; 352 ifq = &ipintrq;
338 isr = NETISR_IP; 353 isr = NETISR_IP;
339 break; 354 break;
340#endif 355#endif
341#ifdef INET6 356#ifdef INET6
342 case AF_INET6: 357 case AF_INET6:
343 m->m_flags |= M_LOOP; 358 m->m_flags |= M_LOOP;
344 ifq = &ip6intrq; 359 ifq = &ip6intrq;
345 isr = NETISR_IPV6; 360 isr = NETISR_IPV6;
346 break; 361 break;
347#endif 362#endif
348#ifdef IPX 363#ifdef IPX
349 case AF_IPX: 364 case AF_IPX:
350 ifq = &ipxintrq; 365 ifq = &ipxintrq;
351 isr = NETISR_IPX; 366 isr = NETISR_IPX;
352 break; 367 break;
353#endif 368#endif
354#ifdef ISO 369#ifdef ISO
355 case AF_ISO: 370 case AF_ISO:
356 ifq = &clnlintrq; 371 ifq = &clnlintrq;
357 isr = NETISR_ISO; 372 isr = NETISR_ISO;
358 break; 373 break;
359#endif 374#endif
360#ifdef NETATALK 375#ifdef NETATALK
361 case AF_APPLETALK: 376 case AF_APPLETALK:
362 ifq = &atintrq2; 377 ifq = &atintrq2;
363 isr = NETISR_ATALK; 378 isr = NETISR_ATALK;
364 break; 379 break;
365#endif 380#endif
366 default: 381 default:
367 printf("%s: can't handle af%d\n", ifp->if_xname, af); 382 printf("%s: can't handle af%d\n", ifp->if_xname, af);
368 m_freem(m); 383 m_freem(m);
369 return; 384 return;
370 } 385 }
371 386
372 s = splnet(); 387 s = splnet();
373 if (IF_QFULL(ifq)) { 388 if (IF_QFULL(ifq)) {
374 IF_DROP(ifq); 389 IF_DROP(ifq);
375 splx(s); 390 splx(s);
376 m_freem(m); 391 m_freem(m);
377 return; 392 return;
378 } 393 }
379 IF_ENQUEUE(ifq, m); 394 IF_ENQUEUE(ifq, m);
380 schednetisr(isr); 395 schednetisr(isr);
381 ifp->if_ipackets++; 396 ifp->if_ipackets++;
382 ifp->if_ibytes += m->m_pkthdr.len; 397 ifp->if_ibytes += m->m_pkthdr.len;
383 splx(s); 398 splx(s);
384 } 399 }
385} 400}
386#endif /* ALTQ */ 401#endif /* ALTQ */
387 402
388/* ARGSUSED */ 403/* ARGSUSED */
389void 404void
390lortrequest(int cmd, struct rtentry *rt, 405lortrequest(int cmd, struct rtentry *rt,
391 const struct rt_addrinfo *info) 406 const struct rt_addrinfo *info)
392{ 407{
393 408
394 if (rt) 409 if (rt)
395 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu; 410 rt->rt_rmx.rmx_mtu = lo0ifp->if_mtu;
396} 411}
397 412
398/* 413/*
399 * Process an ioctl request. 414 * Process an ioctl request.
400 */ 415 */
401/* ARGSUSED */ 416/* ARGSUSED */
402int 417int
403loioctl(struct ifnet *ifp, u_long cmd, void *data) 418loioctl(struct ifnet *ifp, u_long cmd, void *data)
404{ 419{
405 struct ifaddr *ifa; 420 struct ifaddr *ifa;
406 struct ifreq *ifr = data; 421 struct ifreq *ifr = data;
407 int error = 0; 422 int error = 0;
408 423
409 switch (cmd) { 424 switch (cmd) {
410 425
411 case SIOCINITIFADDR: 426 case SIOCINITIFADDR:
412 ifp->if_flags |= IFF_UP; 427 ifp->if_flags |= IFF_UP;
413 ifa = (struct ifaddr *)data; 428 ifa = (struct ifaddr *)data;
414 if (ifa != NULL /*&& ifa->ifa_addr->sa_family == AF_ISO*/) 429 if (ifa != NULL /*&& ifa->ifa_addr->sa_family == AF_ISO*/)
415 ifa->ifa_rtrequest = lortrequest; 430 ifa->ifa_rtrequest = lortrequest;
416 /* 431 /*
417 * Everything else is done at a higher level. 432 * Everything else is done at a higher level.
418 */ 433 */
419 break; 434 break;
420 435
421 case SIOCSIFMTU: 436 case SIOCSIFMTU:
422 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX) 437 if ((unsigned)ifr->ifr_mtu > LOMTU_MAX)
423 error = EINVAL; 438 error = EINVAL;
424 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){ 439 else if ((error = ifioctl_common(ifp, cmd, data)) == ENETRESET){
425 /* XXX update rt mtu for AF_ISO? */ 440 /* XXX update rt mtu for AF_ISO? */
426 error = 0; 441 error = 0;
427 } 442 }
428 break; 443 break;
429 444
430 case SIOCADDMULTI: 445 case SIOCADDMULTI:
431 case SIOCDELMULTI: 446 case SIOCDELMULTI:
432 if (ifr == NULL) { 447 if (ifr == NULL) {
433 error = EAFNOSUPPORT; /* XXX */ 448 error = EAFNOSUPPORT; /* XXX */
434 break; 449 break;
435 } 450 }
436 switch (ifreq_getaddr(cmd, ifr)->sa_family) { 451 switch (ifreq_getaddr(cmd, ifr)->sa_family) {
437 452
438#ifdef INET 453#ifdef INET
439 case AF_INET: 454 case AF_INET:
440 break; 455 break;
441#endif 456#endif
442#ifdef INET6 457#ifdef INET6
443 case AF_INET6: 458 case AF_INET6:
444 break; 459 break;
445#endif 460#endif
446 461
447 default: 462 default:
448 error = EAFNOSUPPORT; 463 error = EAFNOSUPPORT;
449 break; 464 break;
450 } 465 }
451 break; 466 break;
452 467
453 default: 468 default:
454 error = ifioctl_common(ifp, cmd, data); 469 error = ifioctl_common(ifp, cmd, data);
455 } 470 }
456 return (error); 471 return (error);
457} 472}

cvs diff -r1.7 -r1.8 src/sys/netinet/in_offload.h (switch to unified diff)

--- src/sys/netinet/in_offload.h 2010/12/11 22:37:46 1.7
+++ src/sys/netinet/in_offload.h 2011/04/25 22:20:59 1.8
@@ -1,57 +1,60 @@ @@ -1,57 +1,60 @@
1/* $NetBSD: in_offload.h,v 1.7 2010/12/11 22:37:46 matt Exp $ */ 1/* $NetBSD: in_offload.h,v 1.8 2011/04/25 22:20:59 yamt Exp $ */
2 2
3/*- 3/*-
4 * Copyright (c)2005, 2006 YAMAMOTO Takashi, 4 * Copyright (c)2005, 2006 YAMAMOTO Takashi,
5 * All rights reserved. 5 * All rights reserved.
6 * 6 *
7 * Redistribution and use in source and binary forms, with or without 7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions 8 * modification, are permitted provided that the following conditions
9 * are met: 9 * are met:
10 * 1. Redistributions of source code must retain the above copyright 10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer. 11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright 12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the 13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution. 14 * documentation and/or other materials provided with the distribution.
15 * 15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE. 26 * SUCH DAMAGE.
27 */ 27 */
28 28
29#ifndef _NETINET_IN_OFFLOAD_H_ 29#ifndef _NETINET_IN_OFFLOAD_H_
30#define _NETINET_IN_OFFLOAD_H_ 30#define _NETINET_IN_OFFLOAD_H_
31 31
32/* 32/*
33 * subroutines to do software-only equivalent of h/w offloading. 33 * subroutines to do software-only equivalent of h/w offloading.
34 */ 34 */
35 35
36int tcp4_segment(struct mbuf *, int (*)(void *, struct mbuf *), void *); 36int tcp4_segment(struct mbuf *, int (*)(void *, struct mbuf *), void *);
37int ip_tso_output(struct ifnet *, struct mbuf *, const struct sockaddr *, 37int ip_tso_output(struct ifnet *, struct mbuf *, const struct sockaddr *,
38 struct rtentry *); 38 struct rtentry *);
39void ip_undefer_csum(struct mbuf *, size_t, int);  39void ip_undefer_csum(struct mbuf *, size_t, int);
40 40
41/* 41/*
42 * offloading related sysctl variables. 42 * offloading related sysctl variables.
43 * 43 *
44 * they are here because it violates protocol layering in unusual way. 44 * they are here because it violates protocol layering in unusual way.
45 * ie. while they are TCP/UDP sysctls, they are used by IP layer. 45 * ie. while they are TCP/UDP sysctls, they are used by IP layer.
46 */ 46 */
47 47
48extern int tcp_do_loopback_cksum; /* do TCP checksum on loopback? */ 48extern int tcp_do_loopback_cksum; /* do TCP checksum on loopback? */
49extern int udp_do_loopback_cksum; /* do UDP checksum on loopback? */ 49extern int udp_do_loopback_cksum; /* do UDP checksum on loopback? */
50 50
 51#define IN_LOOPBACK_NEED_CHECKSUM(csum_flags) \
 52 ((((csum_flags) & M_CSUM_UDPv4) != 0 && udp_do_loopback_cksum) || \
 53 (((csum_flags) & M_CSUM_TCPv4) != 0 && tcp_do_loopback_cksum) || \
 54 (((csum_flags) & M_CSUM_IPv4) != 0 && ip_do_loopback_cksum))
 55
51#define IN_NEED_CHECKSUM(ifp, csum_flags) \ 56#define IN_NEED_CHECKSUM(ifp, csum_flags) \
52 (__predict_true(((ifp)->if_flags & IFF_LOOPBACK) == 0 || \ 57 (__predict_true(((ifp)->if_flags & IFF_LOOPBACK) == 0 || \
53 (((csum_flags) & M_CSUM_UDPv4) != 0 && udp_do_loopback_cksum) || \ 58 IN_LOOPBACK_NEED_CHECKSUM(csum_flags)))
54 (((csum_flags) & M_CSUM_TCPv4) != 0 && tcp_do_loopback_cksum) || \ 
55 (((csum_flags) & M_CSUM_IPv4) != 0 && ip_do_loopback_cksum))) 
56 59
57#endif /* !_NETINET_IN_OFFLOAD_H_ */ 60#endif /* !_NETINET_IN_OFFLOAD_H_ */

cvs diff -r1.6 -r1.7 src/sys/netinet6/in6_offload.h (switch to unified diff)

--- src/sys/netinet6/in6_offload.h 2010/12/11 22:37:47 1.6
+++ src/sys/netinet6/in6_offload.h 2011/04/25 22:20:59 1.7
@@ -1,41 +1,52 @@ @@ -1,41 +1,52 @@
1/* $NetBSD: in6_offload.h,v 1.6 2010/12/11 22:37:47 matt Exp $ */ 1/* $NetBSD: in6_offload.h,v 1.7 2011/04/25 22:20:59 yamt Exp $ */
2 2
3/*- 3/*-
4 * Copyright (c)2005, 2006 YAMAMOTO Takashi, 4 * Copyright (c)2005, 2006 YAMAMOTO Takashi,
5 * All rights reserved. 5 * All rights reserved.
6 * 6 *
7 * Redistribution and use in source and binary forms, with or without 7 * Redistribution and use in source and binary forms, with or without
8 * modification, are permitted provided that the following conditions 8 * modification, are permitted provided that the following conditions
9 * are met: 9 * are met:
10 * 1. Redistributions of source code must retain the above copyright 10 * 1. Redistributions of source code must retain the above copyright
11 * notice, this list of conditions and the following disclaimer. 11 * notice, this list of conditions and the following disclaimer.
12 * 2. Redistributions in binary form must reproduce the above copyright 12 * 2. Redistributions in binary form must reproduce the above copyright
13 * notice, this list of conditions and the following disclaimer in the 13 * notice, this list of conditions and the following disclaimer in the
14 * documentation and/or other materials provided with the distribution. 14 * documentation and/or other materials provided with the distribution.
15 * 15 *
16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26 * SUCH DAMAGE. 26 * SUCH DAMAGE.
27 */ 27 */
28 28
29/* 29/*
30 * subroutines to do software-only equivalent of h/w offloading. 30 * subroutines to do software-only equivalent of h/w offloading.
31 */ 31 */
32 32
33#if !defined(_NETINET6_IN6_OFFLOAD_H_) 33#if !defined(_NETINET6_IN6_OFFLOAD_H_)
34#define _NETINET6_IN6_OFFLOAD_H_ 34#define _NETINET6_IN6_OFFLOAD_H_
35 35
36int tcp6_segment(struct mbuf *, int (*)(void *, struct mbuf *), void *); 36int tcp6_segment(struct mbuf *, int (*)(void *, struct mbuf *), void *);
37int ip6_tso_output(struct ifnet *, struct ifnet *, struct mbuf *, 37int ip6_tso_output(struct ifnet *, struct ifnet *, struct mbuf *,
38 const struct sockaddr_in6 *, struct rtentry *); 38 const struct sockaddr_in6 *, struct rtentry *);
39void ip6_undefer_csum(struct mbuf *, size_t, int); 39void ip6_undefer_csum(struct mbuf *, size_t, int);
40 40
 41extern int tcp_do_loopback_cksum; /* do TCP checksum on loopback? */
 42extern int udp_do_loopback_cksum; /* do UDP checksum on loopback? */
 43
 44#define IN6_LOOPBACK_NEED_CHECKSUM(csum_flags) \
 45 ((((csum_flags) & M_CSUM_UDPv6) != 0 && udp_do_loopback_cksum) || \
 46 (((csum_flags) & M_CSUM_TCPv6) != 0 && tcp_do_loopback_cksum))
 47
 48#define IN6_NEED_CHECKSUM(ifp, csum_flags) \
 49 (__predict_true(((ifp)->if_flags & IFF_LOOPBACK) == 0 || \
 50 IN6_LOOPBACK_NEED_CHECKSUM(csum_flags)))
 51
41#endif /* !defined(_NETINET6_IN6_OFFLOAD_H_) */ 52#endif /* !defined(_NETINET6_IN6_OFFLOAD_H_) */

cvs diff -r1.139 -r1.140 src/sys/netinet6/ip6_output.c (switch to unified diff)

--- src/sys/netinet6/ip6_output.c 2009/05/07 21:51:47 1.139
+++ src/sys/netinet6/ip6_output.c 2011/04/25 22:20:59 1.140
@@ -1,1153 +1,1148 @@ @@ -1,1153 +1,1148 @@
1/* $NetBSD: ip6_output.c,v 1.139 2009/05/07 21:51:47 elad Exp $ */ 1/* $NetBSD: ip6_output.c,v 1.140 2011/04/25 22:20:59 yamt Exp $ */
2/* $KAME: ip6_output.c,v 1.172 2001/03/25 09:55:56 itojun Exp $ */ 2/* $KAME: ip6_output.c,v 1.172 2001/03/25 09:55:56 itojun Exp $ */
3 3
4/* 4/*
5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project. 5 * Copyright (C) 1995, 1996, 1997, and 1998 WIDE Project.
6 * All rights reserved. 6 * All rights reserved.
7 * 7 *
8 * Redistribution and use in source and binary forms, with or without 8 * Redistribution and use in source and binary forms, with or without
9 * modification, are permitted provided that the following conditions 9 * modification, are permitted provided that the following conditions
10 * are met: 10 * are met:
11 * 1. Redistributions of source code must retain the above copyright 11 * 1. Redistributions of source code must retain the above copyright
12 * notice, this list of conditions and the following disclaimer. 12 * notice, this list of conditions and the following disclaimer.
13 * 2. Redistributions in binary form must reproduce the above copyright 13 * 2. Redistributions in binary form must reproduce the above copyright
14 * notice, this list of conditions and the following disclaimer in the 14 * notice, this list of conditions and the following disclaimer in the
15 * documentation and/or other materials provided with the distribution. 15 * documentation and/or other materials provided with the distribution.
16 * 3. Neither the name of the project nor the names of its contributors 16 * 3. Neither the name of the project nor the names of its contributors
17 * may be used to endorse or promote products derived from this software 17 * may be used to endorse or promote products derived from this software
18 * without specific prior written permission. 18 * without specific prior written permission.
19 * 19 *
20 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND 20 * THIS SOFTWARE IS PROVIDED BY THE PROJECT AND CONTRIBUTORS ``AS IS'' AND
21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 21 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 22 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE 23 * ARE DISCLAIMED. IN NO EVENT SHALL THE PROJECT OR CONTRIBUTORS BE LIABLE
24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 24 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 25 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 26 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 27 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 28 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 29 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30 * SUCH DAMAGE. 30 * SUCH DAMAGE.
31 */ 31 */
32 32
33/* 33/*
34 * Copyright (c) 1982, 1986, 1988, 1990, 1993 34 * Copyright (c) 1982, 1986, 1988, 1990, 1993
35 * The Regents of the University of California. All rights reserved. 35 * The Regents of the University of California. All rights reserved.
36 * 36 *
37 * Redistribution and use in source and binary forms, with or without 37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions 38 * modification, are permitted provided that the following conditions
39 * are met: 39 * are met:
40 * 1. Redistributions of source code must retain the above copyright 40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer. 41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright 42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the 43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution. 44 * documentation and/or other materials provided with the distribution.
45 * 3. Neither the name of the University nor the names of its contributors 45 * 3. Neither the name of the University nor the names of its contributors
46 * may be used to endorse or promote products derived from this software 46 * may be used to endorse or promote products derived from this software
47 * without specific prior written permission. 47 * without specific prior written permission.
48 * 48 *
49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 49 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 50 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 51 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 52 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 53 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 54 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 55 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 56 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 57 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 58 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
59 * SUCH DAMAGE. 59 * SUCH DAMAGE.
60 * 60 *
61 * @(#)ip_output.c 8.3 (Berkeley) 1/21/94 61 * @(#)ip_output.c 8.3 (Berkeley) 1/21/94
62 */ 62 */
63 63
64#include <sys/cdefs.h> 64#include <sys/cdefs.h>
65__KERNEL_RCSID(0, "$NetBSD: ip6_output.c,v 1.139 2009/05/07 21:51:47 elad Exp $"); 65__KERNEL_RCSID(0, "$NetBSD: ip6_output.c,v 1.140 2011/04/25 22:20:59 yamt Exp $");
66 66
67#include "opt_inet.h" 67#include "opt_inet.h"
68#include "opt_inet6.h" 68#include "opt_inet6.h"
69#include "opt_ipsec.h" 69#include "opt_ipsec.h"
70#include "opt_pfil_hooks.h" 70#include "opt_pfil_hooks.h"
71 71
72#include <sys/param.h> 72#include <sys/param.h>
73#include <sys/malloc.h> 73#include <sys/malloc.h>
74#include <sys/mbuf.h> 74#include <sys/mbuf.h>
75#include <sys/errno.h> 75#include <sys/errno.h>
76#include <sys/protosw.h> 76#include <sys/protosw.h>
77#include <sys/socket.h> 77#include <sys/socket.h>
78#include <sys/socketvar.h> 78#include <sys/socketvar.h>
79#include <sys/systm.h> 79#include <sys/systm.h>
80#include <sys/proc.h> 80#include <sys/proc.h>
81#include <sys/kauth.h> 81#include <sys/kauth.h>
82 82
83#include <net/if.h> 83#include <net/if.h>
84#include <net/route.h> 84#include <net/route.h>
85#ifdef PFIL_HOOKS 85#ifdef PFIL_HOOKS
86#include <net/pfil.h> 86#include <net/pfil.h>
87#endif 87#endif
88 88
89#include <netinet/in.h> 89#include <netinet/in.h>
90#include <netinet/in_var.h> 90#include <netinet/in_var.h>
91#include <netinet/ip6.h> 91#include <netinet/ip6.h>
92#include <netinet/icmp6.h> 92#include <netinet/icmp6.h>
93#include <netinet/in_offload.h> 93#include <netinet/in_offload.h>
94#include <netinet6/in6_offload.h> 94#include <netinet6/in6_offload.h>
95#include <netinet6/ip6_var.h> 95#include <netinet6/ip6_var.h>
96#include <netinet6/ip6_private.h> 96#include <netinet6/ip6_private.h>
97#include <netinet6/in6_pcb.h> 97#include <netinet6/in6_pcb.h>
98#include <netinet6/nd6.h> 98#include <netinet6/nd6.h>
99#include <netinet6/ip6protosw.h> 99#include <netinet6/ip6protosw.h>
100#include <netinet6/scope6_var.h> 100#include <netinet6/scope6_var.h>
101 101
102#ifdef IPSEC 102#ifdef IPSEC
103#include <netinet6/ipsec.h> 103#include <netinet6/ipsec.h>
104#include <netinet6/ipsec_private.h> 104#include <netinet6/ipsec_private.h>
105#include <netkey/key.h> 105#include <netkey/key.h>
106#endif /* IPSEC */ 106#endif /* IPSEC */
107 107
108#ifdef FAST_IPSEC 108#ifdef FAST_IPSEC
109#include <netipsec/ipsec.h> 109#include <netipsec/ipsec.h>
110#include <netipsec/ipsec6.h> 110#include <netipsec/ipsec6.h>
111#include <netipsec/key.h> 111#include <netipsec/key.h>
112#include <netipsec/xform.h> 112#include <netipsec/xform.h>
113#endif 113#endif
114 114
115 115
116#include <net/net_osdep.h> 116#include <net/net_osdep.h>
117 117
118#ifdef PFIL_HOOKS 118#ifdef PFIL_HOOKS
119extern struct pfil_head inet6_pfil_hook; /* XXX */ 119extern struct pfil_head inet6_pfil_hook; /* XXX */
120#endif 120#endif
121 121
122struct ip6_exthdrs { 122struct ip6_exthdrs {
123 struct mbuf *ip6e_ip6; 123 struct mbuf *ip6e_ip6;
124 struct mbuf *ip6e_hbh; 124 struct mbuf *ip6e_hbh;
125 struct mbuf *ip6e_dest1; 125 struct mbuf *ip6e_dest1;
126 struct mbuf *ip6e_rthdr; 126 struct mbuf *ip6e_rthdr;
127 struct mbuf *ip6e_dest2; 127 struct mbuf *ip6e_dest2;
128}; 128};
129 129
130static int ip6_pcbopt(int, u_char *, int, struct ip6_pktopts **, 130static int ip6_pcbopt(int, u_char *, int, struct ip6_pktopts **,
131 kauth_cred_t, int); 131 kauth_cred_t, int);
132static int ip6_getpcbopt(struct ip6_pktopts *, int, struct sockopt *); 132static int ip6_getpcbopt(struct ip6_pktopts *, int, struct sockopt *);
133static int ip6_setpktopt(int, u_char *, int, struct ip6_pktopts *, kauth_cred_t, 133static int ip6_setpktopt(int, u_char *, int, struct ip6_pktopts *, kauth_cred_t,
134 int, int, int); 134 int, int, int);
135static int ip6_setmoptions(const struct sockopt *, struct ip6_moptions **); 135static int ip6_setmoptions(const struct sockopt *, struct ip6_moptions **);
136static int ip6_getmoptions(struct sockopt *, struct ip6_moptions *); 136static int ip6_getmoptions(struct sockopt *, struct ip6_moptions *);
137static int ip6_copyexthdr(struct mbuf **, void *, int); 137static int ip6_copyexthdr(struct mbuf **, void *, int);
138static int ip6_insertfraghdr(struct mbuf *, struct mbuf *, int, 138static int ip6_insertfraghdr(struct mbuf *, struct mbuf *, int,
139 struct ip6_frag **); 139 struct ip6_frag **);
140static int ip6_insert_jumboopt(struct ip6_exthdrs *, u_int32_t); 140static int ip6_insert_jumboopt(struct ip6_exthdrs *, u_int32_t);
141static int ip6_splithdr(struct mbuf *, struct ip6_exthdrs *); 141static int ip6_splithdr(struct mbuf *, struct ip6_exthdrs *);
142static int ip6_getpmtu(struct route *, struct route *, struct ifnet *, 142static int ip6_getpmtu(struct route *, struct route *, struct ifnet *,
143 const struct in6_addr *, u_long *, int *); 143 const struct in6_addr *, u_long *, int *);
144static int copypktopts(struct ip6_pktopts *, struct ip6_pktopts *, int); 144static int copypktopts(struct ip6_pktopts *, struct ip6_pktopts *, int);
145 145
146#ifdef RFC2292 146#ifdef RFC2292
147static int ip6_pcbopts(struct ip6_pktopts **, struct socket *, struct sockopt *); 147static int ip6_pcbopts(struct ip6_pktopts **, struct socket *, struct sockopt *);
148#endif 148#endif
149 149
150#define IN6_NEED_CHECKSUM(ifp, csum_flags) \ 
151 (__predict_true(((ifp)->if_flags & IFF_LOOPBACK) == 0 || \ 
152 (((csum_flags) & M_CSUM_UDPv6) != 0 && udp_do_loopback_cksum) || \ 
153 (((csum_flags) & M_CSUM_TCPv6) != 0 && tcp_do_loopback_cksum))) 
154 
155/* 150/*
156 * IP6 output. The packet in mbuf chain m contains a skeletal IP6 151 * IP6 output. The packet in mbuf chain m contains a skeletal IP6
157 * header (with pri, len, nxt, hlim, src, dst). 152 * header (with pri, len, nxt, hlim, src, dst).
158 * This function may modify ver and hlim only. 153 * This function may modify ver and hlim only.
159 * The mbuf chain containing the packet will be freed. 154 * The mbuf chain containing the packet will be freed.
160 * The mbuf opt, if present, will not be freed. 155 * The mbuf opt, if present, will not be freed.
161 * 156 *
162 * type of "mtu": rt_rmx.rmx_mtu is u_long, ifnet.ifr_mtu is int, and 157 * type of "mtu": rt_rmx.rmx_mtu is u_long, ifnet.ifr_mtu is int, and
163 * nd_ifinfo.linkmtu is u_int32_t. so we use u_long to hold largest one, 158 * nd_ifinfo.linkmtu is u_int32_t. so we use u_long to hold largest one,
164 * which is rt_rmx.rmx_mtu. 159 * which is rt_rmx.rmx_mtu.
165 */ 160 */
166int 161int
167ip6_output( 162ip6_output(
168 struct mbuf *m0, 163 struct mbuf *m0,
169 struct ip6_pktopts *opt, 164 struct ip6_pktopts *opt,
170 struct route *ro, 165 struct route *ro,
171 int flags, 166 int flags,
172 struct ip6_moptions *im6o, 167 struct ip6_moptions *im6o,
173 struct socket *so, 168 struct socket *so,
174 struct ifnet **ifpp /* XXX: just for statistics */ 169 struct ifnet **ifpp /* XXX: just for statistics */
175) 170)
176{ 171{
177 struct ip6_hdr *ip6, *mhip6; 172 struct ip6_hdr *ip6, *mhip6;
178 struct ifnet *ifp, *origifp; 173 struct ifnet *ifp, *origifp;
179 struct mbuf *m = m0; 174 struct mbuf *m = m0;
180 int hlen, tlen, len, off; 175 int hlen, tlen, len, off;
181 bool tso; 176 bool tso;
182 struct route ip6route; 177 struct route ip6route;
183 struct rtentry *rt = NULL; 178 struct rtentry *rt = NULL;
184 const struct sockaddr_in6 *dst = NULL; 179 const struct sockaddr_in6 *dst = NULL;
185 struct sockaddr_in6 src_sa, dst_sa; 180 struct sockaddr_in6 src_sa, dst_sa;
186 int error = 0; 181 int error = 0;
187 struct in6_ifaddr *ia = NULL; 182 struct in6_ifaddr *ia = NULL;
188 u_long mtu; 183 u_long mtu;
189 int alwaysfrag, dontfrag; 184 int alwaysfrag, dontfrag;
190 u_int32_t optlen = 0, plen = 0, unfragpartlen = 0; 185 u_int32_t optlen = 0, plen = 0, unfragpartlen = 0;
191 struct ip6_exthdrs exthdrs; 186 struct ip6_exthdrs exthdrs;
192 struct in6_addr finaldst, src0, dst0; 187 struct in6_addr finaldst, src0, dst0;
193 u_int32_t zone; 188 u_int32_t zone;
194 struct route *ro_pmtu = NULL; 189 struct route *ro_pmtu = NULL;
195 int hdrsplit = 0; 190 int hdrsplit = 0;
196 int needipsec = 0; 191 int needipsec = 0;
197#ifdef IPSEC 192#ifdef IPSEC
198 int needipsectun = 0; 193 int needipsectun = 0;
199 struct secpolicy *sp = NULL; 194 struct secpolicy *sp = NULL;
200 195
201 ip6 = mtod(m, struct ip6_hdr *); 196 ip6 = mtod(m, struct ip6_hdr *);
202#endif /* IPSEC */ 197#endif /* IPSEC */
203#ifdef FAST_IPSEC 198#ifdef FAST_IPSEC
204 struct secpolicy *sp = NULL; 199 struct secpolicy *sp = NULL;
205 int s; 200 int s;
206#endif 201#endif
207 202
208 memset(&ip6route, 0, sizeof(ip6route)); 203 memset(&ip6route, 0, sizeof(ip6route));
209 204
210#ifdef DIAGNOSTIC 205#ifdef DIAGNOSTIC
211 if ((m->m_flags & M_PKTHDR) == 0) 206 if ((m->m_flags & M_PKTHDR) == 0)
212 panic("ip6_output: no HDR"); 207 panic("ip6_output: no HDR");
213 208
214 if ((m->m_pkthdr.csum_flags & 209 if ((m->m_pkthdr.csum_flags &
215 (M_CSUM_TCPv4|M_CSUM_UDPv4|M_CSUM_TSOv4)) != 0) { 210 (M_CSUM_TCPv4|M_CSUM_UDPv4|M_CSUM_TSOv4)) != 0) {
216 panic("ip6_output: IPv4 checksum offload flags: %d", 211 panic("ip6_output: IPv4 checksum offload flags: %d",
217 m->m_pkthdr.csum_flags); 212 m->m_pkthdr.csum_flags);
218 } 213 }
219 214
220 if ((m->m_pkthdr.csum_flags & (M_CSUM_TCPv6|M_CSUM_UDPv6)) == 215 if ((m->m_pkthdr.csum_flags & (M_CSUM_TCPv6|M_CSUM_UDPv6)) ==
221 (M_CSUM_TCPv6|M_CSUM_UDPv6)) { 216 (M_CSUM_TCPv6|M_CSUM_UDPv6)) {
222 panic("ip6_output: conflicting checksum offload flags: %d", 217 panic("ip6_output: conflicting checksum offload flags: %d",
223 m->m_pkthdr.csum_flags); 218 m->m_pkthdr.csum_flags);
224 } 219 }
225#endif 220#endif
226 221
227 M_CSUM_DATA_IPv6_HL_SET(m->m_pkthdr.csum_data, sizeof(struct ip6_hdr)); 222 M_CSUM_DATA_IPv6_HL_SET(m->m_pkthdr.csum_data, sizeof(struct ip6_hdr));
228 223
229#define MAKE_EXTHDR(hp, mp) \ 224#define MAKE_EXTHDR(hp, mp) \
230 do { \ 225 do { \
231 if (hp) { \ 226 if (hp) { \
232 struct ip6_ext *eh = (struct ip6_ext *)(hp); \ 227 struct ip6_ext *eh = (struct ip6_ext *)(hp); \
233 error = ip6_copyexthdr((mp), (void *)(hp), \ 228 error = ip6_copyexthdr((mp), (void *)(hp), \
234 ((eh)->ip6e_len + 1) << 3); \ 229 ((eh)->ip6e_len + 1) << 3); \
235 if (error) \ 230 if (error) \
236 goto freehdrs; \ 231 goto freehdrs; \
237 } \ 232 } \
238 } while (/*CONSTCOND*/ 0) 233 } while (/*CONSTCOND*/ 0)
239 234
240 memset(&exthdrs, 0, sizeof(exthdrs)); 235 memset(&exthdrs, 0, sizeof(exthdrs));
241 if (opt) { 236 if (opt) {
242 /* Hop-by-Hop options header */ 237 /* Hop-by-Hop options header */
243 MAKE_EXTHDR(opt->ip6po_hbh, &exthdrs.ip6e_hbh); 238 MAKE_EXTHDR(opt->ip6po_hbh, &exthdrs.ip6e_hbh);
244 /* Destination options header(1st part) */ 239 /* Destination options header(1st part) */
245 MAKE_EXTHDR(opt->ip6po_dest1, &exthdrs.ip6e_dest1); 240 MAKE_EXTHDR(opt->ip6po_dest1, &exthdrs.ip6e_dest1);
246 /* Routing header */ 241 /* Routing header */
247 MAKE_EXTHDR(opt->ip6po_rthdr, &exthdrs.ip6e_rthdr); 242 MAKE_EXTHDR(opt->ip6po_rthdr, &exthdrs.ip6e_rthdr);
248 /* Destination options header(2nd part) */ 243 /* Destination options header(2nd part) */
249 MAKE_EXTHDR(opt->ip6po_dest2, &exthdrs.ip6e_dest2); 244 MAKE_EXTHDR(opt->ip6po_dest2, &exthdrs.ip6e_dest2);
250 } 245 }
251 246
252#ifdef IPSEC 247#ifdef IPSEC
253 if ((flags & IPV6_FORWARDING) != 0) { 248 if ((flags & IPV6_FORWARDING) != 0) {
254 needipsec = 0; 249 needipsec = 0;
255 goto skippolicycheck; 250 goto skippolicycheck;
256 } 251 }
257 252
258 /* get a security policy for this packet */ 253 /* get a security policy for this packet */
259 if (so == NULL) 254 if (so == NULL)
260 sp = ipsec6_getpolicybyaddr(m, IPSEC_DIR_OUTBOUND, 0, &error); 255 sp = ipsec6_getpolicybyaddr(m, IPSEC_DIR_OUTBOUND, 0, &error);
261 else { 256 else {
262 if (IPSEC_PCB_SKIP_IPSEC(sotoinpcb_hdr(so)->inph_sp, 257 if (IPSEC_PCB_SKIP_IPSEC(sotoinpcb_hdr(so)->inph_sp,
263 IPSEC_DIR_OUTBOUND)) { 258 IPSEC_DIR_OUTBOUND)) {
264 needipsec = 0; 259 needipsec = 0;
265 goto skippolicycheck; 260 goto skippolicycheck;
266 } 261 }
267 sp = ipsec6_getpolicybysock(m, IPSEC_DIR_OUTBOUND, so, &error); 262 sp = ipsec6_getpolicybysock(m, IPSEC_DIR_OUTBOUND, so, &error);
268 } 263 }
269 264
270 if (sp == NULL) { 265 if (sp == NULL) {
271 IPSEC6_STATINC(IPSEC_STAT_OUT_INVAL); 266 IPSEC6_STATINC(IPSEC_STAT_OUT_INVAL);
272 goto freehdrs; 267 goto freehdrs;
273 } 268 }
274 269
275 error = 0; 270 error = 0;
276 271
277 /* check policy */ 272 /* check policy */
278 switch (sp->policy) { 273 switch (sp->policy) {
279 case IPSEC_POLICY_DISCARD: 274 case IPSEC_POLICY_DISCARD:
280 /* 275 /*
281 * This packet is just discarded. 276 * This packet is just discarded.
282 */ 277 */
283 IPSEC6_STATINC(IPSEC_STAT_OUT_POLVIO); 278 IPSEC6_STATINC(IPSEC_STAT_OUT_POLVIO);
284 goto freehdrs; 279 goto freehdrs;
285 280
286 case IPSEC_POLICY_BYPASS: 281 case IPSEC_POLICY_BYPASS:
287 case IPSEC_POLICY_NONE: 282 case IPSEC_POLICY_NONE:
288 /* no need to do IPsec. */ 283 /* no need to do IPsec. */
289 needipsec = 0; 284 needipsec = 0;
290 break; 285 break;
291 286
292 case IPSEC_POLICY_IPSEC: 287 case IPSEC_POLICY_IPSEC:
293 if (sp->req == NULL) { 288 if (sp->req == NULL) {
294 /* XXX should be panic ? */ 289 /* XXX should be panic ? */
295 printf("ip6_output: No IPsec request specified.\n"); 290 printf("ip6_output: No IPsec request specified.\n");
296 error = EINVAL; 291 error = EINVAL;
297 goto freehdrs; 292 goto freehdrs;
298 } 293 }
299 needipsec = 1; 294 needipsec = 1;
300 break; 295 break;
301 296
302 case IPSEC_POLICY_ENTRUST: 297 case IPSEC_POLICY_ENTRUST:
303 default: 298 default:
304 printf("ip6_output: Invalid policy found. %d\n", sp->policy); 299 printf("ip6_output: Invalid policy found. %d\n", sp->policy);
305 } 300 }
306 301
307 skippolicycheck:; 302 skippolicycheck:;
308#endif /* IPSEC */ 303#endif /* IPSEC */
309 304
310 /* 305 /*
311 * Calculate the total length of the extension header chain. 306 * Calculate the total length of the extension header chain.
312 * Keep the length of the unfragmentable part for fragmentation. 307 * Keep the length of the unfragmentable part for fragmentation.
313 */ 308 */
314 optlen = 0; 309 optlen = 0;
315 if (exthdrs.ip6e_hbh) optlen += exthdrs.ip6e_hbh->m_len; 310 if (exthdrs.ip6e_hbh) optlen += exthdrs.ip6e_hbh->m_len;
316 if (exthdrs.ip6e_dest1) optlen += exthdrs.ip6e_dest1->m_len; 311 if (exthdrs.ip6e_dest1) optlen += exthdrs.ip6e_dest1->m_len;
317 if (exthdrs.ip6e_rthdr) optlen += exthdrs.ip6e_rthdr->m_len; 312 if (exthdrs.ip6e_rthdr) optlen += exthdrs.ip6e_rthdr->m_len;
318 unfragpartlen = optlen + sizeof(struct ip6_hdr); 313 unfragpartlen = optlen + sizeof(struct ip6_hdr);
319 /* NOTE: we don't add AH/ESP length here. do that later. */ 314 /* NOTE: we don't add AH/ESP length here. do that later. */
320 if (exthdrs.ip6e_dest2) optlen += exthdrs.ip6e_dest2->m_len; 315 if (exthdrs.ip6e_dest2) optlen += exthdrs.ip6e_dest2->m_len;
321 316
322#ifdef FAST_IPSEC 317#ifdef FAST_IPSEC
323 /* Check the security policy (SP) for the packet */ 318 /* Check the security policy (SP) for the packet */
324  319
325 /* XXX For moment, we doesn't support packet with extented action */ 320 /* XXX For moment, we doesn't support packet with extented action */
326 if (optlen !=0) 321 if (optlen !=0)
327 goto freehdrs; 322 goto freehdrs;
328 323
329 sp = ipsec6_check_policy(m,so,flags,&needipsec,&error); 324 sp = ipsec6_check_policy(m,so,flags,&needipsec,&error);
330 if (error != 0) { 325 if (error != 0) {
331 /* 326 /*
332 * Hack: -EINVAL is used to signal that a packet 327 * Hack: -EINVAL is used to signal that a packet
333 * should be silently discarded. This is typically 328 * should be silently discarded. This is typically
334 * because we asked key management for an SA and 329 * because we asked key management for an SA and
335 * it was delayed (e.g. kicked up to IKE). 330 * it was delayed (e.g. kicked up to IKE).
336 */ 331 */
337 if (error == -EINVAL)  332 if (error == -EINVAL)
338 error = 0; 333 error = 0;
339 goto freehdrs; 334 goto freehdrs;
340 } 335 }
341#endif /* FAST_IPSEC */ 336#endif /* FAST_IPSEC */
342 337
343 338
344 if (needipsec && 339 if (needipsec &&
345 (m->m_pkthdr.csum_flags & (M_CSUM_UDPv6|M_CSUM_TCPv6)) != 0) { 340 (m->m_pkthdr.csum_flags & (M_CSUM_UDPv6|M_CSUM_TCPv6)) != 0) {
346 in6_delayed_cksum(m); 341 in6_delayed_cksum(m);
347 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6); 342 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6);
348 } 343 }
349 344
350 345
351 /* 346 /*
352 * If we need IPsec, or there is at least one extension header, 347 * If we need IPsec, or there is at least one extension header,
353 * separate IP6 header from the payload. 348 * separate IP6 header from the payload.
354 */ 349 */
355 if ((needipsec || optlen) && !hdrsplit) { 350 if ((needipsec || optlen) && !hdrsplit) {
356 if ((error = ip6_splithdr(m, &exthdrs)) != 0) { 351 if ((error = ip6_splithdr(m, &exthdrs)) != 0) {
357 m = NULL; 352 m = NULL;
358 goto freehdrs; 353 goto freehdrs;
359 } 354 }
360 m = exthdrs.ip6e_ip6; 355 m = exthdrs.ip6e_ip6;
361 hdrsplit++; 356 hdrsplit++;
362 } 357 }
363 358
364 /* adjust pointer */ 359 /* adjust pointer */
365 ip6 = mtod(m, struct ip6_hdr *); 360 ip6 = mtod(m, struct ip6_hdr *);
366 361
367 /* adjust mbuf packet header length */ 362 /* adjust mbuf packet header length */
368 m->m_pkthdr.len += optlen; 363 m->m_pkthdr.len += optlen;
369 plen = m->m_pkthdr.len - sizeof(*ip6); 364 plen = m->m_pkthdr.len - sizeof(*ip6);
370 365
371 /* If this is a jumbo payload, insert a jumbo payload option. */ 366 /* If this is a jumbo payload, insert a jumbo payload option. */
372 if (plen > IPV6_MAXPACKET) { 367 if (plen > IPV6_MAXPACKET) {
373 if (!hdrsplit) { 368 if (!hdrsplit) {
374 if ((error = ip6_splithdr(m, &exthdrs)) != 0) { 369 if ((error = ip6_splithdr(m, &exthdrs)) != 0) {
375 m = NULL; 370 m = NULL;
376 goto freehdrs; 371 goto freehdrs;
377 } 372 }
378 m = exthdrs.ip6e_ip6; 373 m = exthdrs.ip6e_ip6;
379 hdrsplit++; 374 hdrsplit++;
380 } 375 }
381 /* adjust pointer */ 376 /* adjust pointer */
382 ip6 = mtod(m, struct ip6_hdr *); 377 ip6 = mtod(m, struct ip6_hdr *);
383 if ((error = ip6_insert_jumboopt(&exthdrs, plen)) != 0) 378 if ((error = ip6_insert_jumboopt(&exthdrs, plen)) != 0)
384 goto freehdrs; 379 goto freehdrs;
385 optlen += 8; /* XXX JUMBOOPTLEN */ 380 optlen += 8; /* XXX JUMBOOPTLEN */
386 ip6->ip6_plen = 0; 381 ip6->ip6_plen = 0;
387 } else 382 } else
388 ip6->ip6_plen = htons(plen); 383 ip6->ip6_plen = htons(plen);
389 384
390 /* 385 /*
391 * Concatenate headers and fill in next header fields. 386 * Concatenate headers and fill in next header fields.
392 * Here we have, on "m" 387 * Here we have, on "m"
393 * IPv6 payload 388 * IPv6 payload
394 * and we insert headers accordingly. Finally, we should be getting: 389 * and we insert headers accordingly. Finally, we should be getting:
395 * IPv6 hbh dest1 rthdr ah* [esp* dest2 payload] 390 * IPv6 hbh dest1 rthdr ah* [esp* dest2 payload]
396 * 391 *
397 * during the header composing process, "m" points to IPv6 header. 392 * during the header composing process, "m" points to IPv6 header.
398 * "mprev" points to an extension header prior to esp. 393 * "mprev" points to an extension header prior to esp.
399 */ 394 */
400 { 395 {
401 u_char *nexthdrp = &ip6->ip6_nxt; 396 u_char *nexthdrp = &ip6->ip6_nxt;
402 struct mbuf *mprev = m; 397 struct mbuf *mprev = m;
403 398
404 /* 399 /*
405 * we treat dest2 specially. this makes IPsec processing 400 * we treat dest2 specially. this makes IPsec processing
406 * much easier. the goal here is to make mprev point the 401 * much easier. the goal here is to make mprev point the
407 * mbuf prior to dest2. 402 * mbuf prior to dest2.
408 * 403 *
409 * result: IPv6 dest2 payload 404 * result: IPv6 dest2 payload
410 * m and mprev will point to IPv6 header. 405 * m and mprev will point to IPv6 header.
411 */ 406 */
412 if (exthdrs.ip6e_dest2) { 407 if (exthdrs.ip6e_dest2) {
413 if (!hdrsplit) 408 if (!hdrsplit)
414 panic("assumption failed: hdr not split"); 409 panic("assumption failed: hdr not split");
415 exthdrs.ip6e_dest2->m_next = m->m_next; 410 exthdrs.ip6e_dest2->m_next = m->m_next;
416 m->m_next = exthdrs.ip6e_dest2; 411 m->m_next = exthdrs.ip6e_dest2;
417 *mtod(exthdrs.ip6e_dest2, u_char *) = ip6->ip6_nxt; 412 *mtod(exthdrs.ip6e_dest2, u_char *) = ip6->ip6_nxt;
418 ip6->ip6_nxt = IPPROTO_DSTOPTS; 413 ip6->ip6_nxt = IPPROTO_DSTOPTS;
419 } 414 }
420 415
421#define MAKE_CHAIN(m, mp, p, i)\ 416#define MAKE_CHAIN(m, mp, p, i)\
422 do {\ 417 do {\
423 if (m) {\ 418 if (m) {\
424 if (!hdrsplit) \ 419 if (!hdrsplit) \
425 panic("assumption failed: hdr not split"); \ 420 panic("assumption failed: hdr not split"); \
426 *mtod((m), u_char *) = *(p);\ 421 *mtod((m), u_char *) = *(p);\
427 *(p) = (i);\ 422 *(p) = (i);\
428 p = mtod((m), u_char *);\ 423 p = mtod((m), u_char *);\
429 (m)->m_next = (mp)->m_next;\ 424 (m)->m_next = (mp)->m_next;\
430 (mp)->m_next = (m);\ 425 (mp)->m_next = (m);\
431 (mp) = (m);\ 426 (mp) = (m);\
432 }\ 427 }\
433 } while (/*CONSTCOND*/ 0) 428 } while (/*CONSTCOND*/ 0)
434 /* 429 /*
435 * result: IPv6 hbh dest1 rthdr dest2 payload 430 * result: IPv6 hbh dest1 rthdr dest2 payload
436 * m will point to IPv6 header. mprev will point to the 431 * m will point to IPv6 header. mprev will point to the
437 * extension header prior to dest2 (rthdr in the above case). 432 * extension header prior to dest2 (rthdr in the above case).
438 */ 433 */
439 MAKE_CHAIN(exthdrs.ip6e_hbh, mprev, nexthdrp, IPPROTO_HOPOPTS); 434 MAKE_CHAIN(exthdrs.ip6e_hbh, mprev, nexthdrp, IPPROTO_HOPOPTS);
440 MAKE_CHAIN(exthdrs.ip6e_dest1, mprev, nexthdrp, 435 MAKE_CHAIN(exthdrs.ip6e_dest1, mprev, nexthdrp,
441 IPPROTO_DSTOPTS); 436 IPPROTO_DSTOPTS);
442 MAKE_CHAIN(exthdrs.ip6e_rthdr, mprev, nexthdrp, 437 MAKE_CHAIN(exthdrs.ip6e_rthdr, mprev, nexthdrp,
443 IPPROTO_ROUTING); 438 IPPROTO_ROUTING);
444 439
445 M_CSUM_DATA_IPv6_HL_SET(m->m_pkthdr.csum_data, 440 M_CSUM_DATA_IPv6_HL_SET(m->m_pkthdr.csum_data,
446 sizeof(struct ip6_hdr) + optlen); 441 sizeof(struct ip6_hdr) + optlen);
447 442
448#ifdef IPSEC 443#ifdef IPSEC
449 if (!needipsec) 444 if (!needipsec)
450 goto skip_ipsec2; 445 goto skip_ipsec2;
451 446
452 /* 447 /*
453 * pointers after IPsec headers are not valid any more. 448 * pointers after IPsec headers are not valid any more.
454 * other pointers need a great care too. 449 * other pointers need a great care too.
455 * (IPsec routines should not mangle mbufs prior to AH/ESP) 450 * (IPsec routines should not mangle mbufs prior to AH/ESP)
456 */ 451 */
457 exthdrs.ip6e_dest2 = NULL; 452 exthdrs.ip6e_dest2 = NULL;
458 453
459 { 454 {
460 struct ip6_rthdr *rh = NULL; 455 struct ip6_rthdr *rh = NULL;
461 int segleft_org = 0; 456 int segleft_org = 0;
462 struct ipsec_output_state state; 457 struct ipsec_output_state state;
463 458
464 if (exthdrs.ip6e_rthdr) { 459 if (exthdrs.ip6e_rthdr) {
465 rh = mtod(exthdrs.ip6e_rthdr, struct ip6_rthdr *); 460 rh = mtod(exthdrs.ip6e_rthdr, struct ip6_rthdr *);
466 segleft_org = rh->ip6r_segleft; 461 segleft_org = rh->ip6r_segleft;
467 rh->ip6r_segleft = 0; 462 rh->ip6r_segleft = 0;
468 } 463 }
469 464
470 memset(&state, 0, sizeof(state)); 465 memset(&state, 0, sizeof(state));
471 state.m = m; 466 state.m = m;
472 error = ipsec6_output_trans(&state, nexthdrp, mprev, sp, flags, 467 error = ipsec6_output_trans(&state, nexthdrp, mprev, sp, flags,
473 &needipsectun); 468 &needipsectun);
474 m = state.m; 469 m = state.m;
475 if (error) { 470 if (error) {
476 rh = mtod(exthdrs.ip6e_rthdr, struct ip6_rthdr *); 471 rh = mtod(exthdrs.ip6e_rthdr, struct ip6_rthdr *);
477 /* mbuf is already reclaimed in ipsec6_output_trans. */ 472 /* mbuf is already reclaimed in ipsec6_output_trans. */
478 m = NULL; 473 m = NULL;
479 switch (error) { 474 switch (error) {
480 case EHOSTUNREACH: 475 case EHOSTUNREACH:
481 case ENETUNREACH: 476 case ENETUNREACH:
482 case EMSGSIZE: 477 case EMSGSIZE:
483 case ENOBUFS: 478 case ENOBUFS:
484 case ENOMEM: 479 case ENOMEM:
485 break; 480 break;
486 default: 481 default:
487 printf("ip6_output (ipsec): error code %d\n", error); 482 printf("ip6_output (ipsec): error code %d\n", error);
488 /* FALLTHROUGH */ 483 /* FALLTHROUGH */
489 case ENOENT: 484 case ENOENT:
490 /* don't show these error codes to the user */ 485 /* don't show these error codes to the user */
491 error = 0; 486 error = 0;
492 break; 487 break;
493 } 488 }
494 goto bad; 489 goto bad;
495 } 490 }
496 if (exthdrs.ip6e_rthdr) { 491 if (exthdrs.ip6e_rthdr) {
497 /* ah6_output doesn't modify mbuf chain */ 492 /* ah6_output doesn't modify mbuf chain */
498 rh->ip6r_segleft = segleft_org; 493 rh->ip6r_segleft = segleft_org;
499 } 494 }
500 } 495 }
501skip_ipsec2:; 496skip_ipsec2:;
502#endif 497#endif
503 } 498 }
504 499
505 /* 500 /*
506 * If there is a routing header, replace destination address field 501 * If there is a routing header, replace destination address field
507 * with the first hop of the routing header. 502 * with the first hop of the routing header.
508 */ 503 */
509 if (exthdrs.ip6e_rthdr) { 504 if (exthdrs.ip6e_rthdr) {
510 struct ip6_rthdr *rh; 505 struct ip6_rthdr *rh;
511 struct ip6_rthdr0 *rh0; 506 struct ip6_rthdr0 *rh0;
512 struct in6_addr *addr; 507 struct in6_addr *addr;
513 struct sockaddr_in6 sa; 508 struct sockaddr_in6 sa;
514 509
515 rh = (struct ip6_rthdr *)(mtod(exthdrs.ip6e_rthdr, 510 rh = (struct ip6_rthdr *)(mtod(exthdrs.ip6e_rthdr,
516 struct ip6_rthdr *)); 511 struct ip6_rthdr *));
517 finaldst = ip6->ip6_dst; 512 finaldst = ip6->ip6_dst;
518 switch (rh->ip6r_type) { 513 switch (rh->ip6r_type) {
519 case IPV6_RTHDR_TYPE_0: 514 case IPV6_RTHDR_TYPE_0:
520 rh0 = (struct ip6_rthdr0 *)rh; 515 rh0 = (struct ip6_rthdr0 *)rh;
521 addr = (struct in6_addr *)(rh0 + 1); 516 addr = (struct in6_addr *)(rh0 + 1);
522 517
523 /* 518 /*
524 * construct a sockaddr_in6 form of 519 * construct a sockaddr_in6 form of
525 * the first hop. 520 * the first hop.
526 * 521 *
527 * XXX: we may not have enough 522 * XXX: we may not have enough
528 * information about its scope zone; 523 * information about its scope zone;
529 * there is no standard API to pass 524 * there is no standard API to pass
530 * the information from the 525 * the information from the
531 * application. 526 * application.
532 */ 527 */
533 sockaddr_in6_init(&sa, addr, 0, 0, 0); 528 sockaddr_in6_init(&sa, addr, 0, 0, 0);
534 if ((error = sa6_embedscope(&sa, 529 if ((error = sa6_embedscope(&sa,
535 ip6_use_defzone)) != 0) { 530 ip6_use_defzone)) != 0) {
536 goto bad; 531 goto bad;
537 } 532 }
538 ip6->ip6_dst = sa.sin6_addr; 533 ip6->ip6_dst = sa.sin6_addr;
539 (void)memmove(&addr[0], &addr[1], 534 (void)memmove(&addr[0], &addr[1],
540 sizeof(struct in6_addr) * 535 sizeof(struct in6_addr) *
541 (rh0->ip6r0_segleft - 1)); 536 (rh0->ip6r0_segleft - 1));
542 addr[rh0->ip6r0_segleft - 1] = finaldst; 537 addr[rh0->ip6r0_segleft - 1] = finaldst;
543 /* XXX */ 538 /* XXX */
544 in6_clearscope(addr + rh0->ip6r0_segleft - 1); 539 in6_clearscope(addr + rh0->ip6r0_segleft - 1);
545 break; 540 break;
546 default: /* is it possible? */ 541 default: /* is it possible? */
547 error = EINVAL; 542 error = EINVAL;
548 goto bad; 543 goto bad;
549 } 544 }
550 } 545 }
551 546
552 /* Source address validation */ 547 /* Source address validation */
553 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) && 548 if (IN6_IS_ADDR_UNSPECIFIED(&ip6->ip6_src) &&
554 (flags & IPV6_UNSPECSRC) == 0) { 549 (flags & IPV6_UNSPECSRC) == 0) {
555 error = EOPNOTSUPP; 550 error = EOPNOTSUPP;
556 IP6_STATINC(IP6_STAT_BADSCOPE); 551 IP6_STATINC(IP6_STAT_BADSCOPE);
557 goto bad; 552 goto bad;
558 } 553 }
559 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src)) { 554 if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_src)) {
560 error = EOPNOTSUPP; 555 error = EOPNOTSUPP;
561 IP6_STATINC(IP6_STAT_BADSCOPE); 556 IP6_STATINC(IP6_STAT_BADSCOPE);
562 goto bad; 557 goto bad;
563 } 558 }
564 559
565 IP6_STATINC(IP6_STAT_LOCALOUT); 560 IP6_STATINC(IP6_STAT_LOCALOUT);
566 561
567 /* 562 /*
568 * Route packet. 563 * Route packet.
569 */ 564 */
570 /* initialize cached route */ 565 /* initialize cached route */
571 if (ro == NULL) { 566 if (ro == NULL) {
572 ro = &ip6route; 567 ro = &ip6route;
573 } 568 }
574 ro_pmtu = ro; 569 ro_pmtu = ro;
575 if (opt && opt->ip6po_rthdr) 570 if (opt && opt->ip6po_rthdr)
576 ro = &opt->ip6po_route; 571 ro = &opt->ip6po_route;
577 572
578 /* 573 /*
579 * if specified, try to fill in the traffic class field. 574 * if specified, try to fill in the traffic class field.
580 * do not override if a non-zero value is already set. 575 * do not override if a non-zero value is already set.
581 * we check the diffserv field and the ecn field separately. 576 * we check the diffserv field and the ecn field separately.
582 */ 577 */
583 if (opt && opt->ip6po_tclass >= 0) { 578 if (opt && opt->ip6po_tclass >= 0) {
584 int mask = 0; 579 int mask = 0;
585 580
586 if ((ip6->ip6_flow & htonl(0xfc << 20)) == 0) 581 if ((ip6->ip6_flow & htonl(0xfc << 20)) == 0)
587 mask |= 0xfc; 582 mask |= 0xfc;
588 if ((ip6->ip6_flow & htonl(0x03 << 20)) == 0) 583 if ((ip6->ip6_flow & htonl(0x03 << 20)) == 0)
589 mask |= 0x03; 584 mask |= 0x03;
590 if (mask != 0) 585 if (mask != 0)
591 ip6->ip6_flow |= htonl((opt->ip6po_tclass & mask) << 20); 586 ip6->ip6_flow |= htonl((opt->ip6po_tclass & mask) << 20);
592 } 587 }
593 588
594 /* fill in or override the hop limit field, if necessary. */ 589 /* fill in or override the hop limit field, if necessary. */
595 if (opt && opt->ip6po_hlim != -1) 590 if (opt && opt->ip6po_hlim != -1)
596 ip6->ip6_hlim = opt->ip6po_hlim & 0xff; 591 ip6->ip6_hlim = opt->ip6po_hlim & 0xff;
597 else if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { 592 else if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
598 if (im6o != NULL) 593 if (im6o != NULL)
599 ip6->ip6_hlim = im6o->im6o_multicast_hlim; 594 ip6->ip6_hlim = im6o->im6o_multicast_hlim;
600 else 595 else
601 ip6->ip6_hlim = ip6_defmcasthlim; 596 ip6->ip6_hlim = ip6_defmcasthlim;
602 } 597 }
603 598
604#ifdef IPSEC 599#ifdef IPSEC
605 if (needipsec && needipsectun) { 600 if (needipsec && needipsectun) {
606 struct ipsec_output_state state; 601 struct ipsec_output_state state;
607 602
608 /* 603 /*
609 * All the extension headers will become inaccessible 604 * All the extension headers will become inaccessible
610 * (since they can be encrypted). 605 * (since they can be encrypted).
611 * Don't panic, we need no more updates to extension headers 606 * Don't panic, we need no more updates to extension headers
612 * on inner IPv6 packet (since they are now encapsulated). 607 * on inner IPv6 packet (since they are now encapsulated).
613 * 608 *
614 * IPv6 [ESP|AH] IPv6 [extension headers] payload 609 * IPv6 [ESP|AH] IPv6 [extension headers] payload
615 */ 610 */
616 memset(&exthdrs, 0, sizeof(exthdrs)); 611 memset(&exthdrs, 0, sizeof(exthdrs));
617 exthdrs.ip6e_ip6 = m; 612 exthdrs.ip6e_ip6 = m;
618 613
619 memset(&state, 0, sizeof(state)); 614 memset(&state, 0, sizeof(state));
620 state.m = m; 615 state.m = m;
621 state.ro = ro; 616 state.ro = ro;
622 state.dst = rtcache_getdst(ro); 617 state.dst = rtcache_getdst(ro);
623 618
624 error = ipsec6_output_tunnel(&state, sp, flags); 619 error = ipsec6_output_tunnel(&state, sp, flags);
625 620
626 m = state.m; 621 m = state.m;
627 ro_pmtu = ro = state.ro; 622 ro_pmtu = ro = state.ro;
628 dst = satocsin6(state.dst); 623 dst = satocsin6(state.dst);
629 if (error) { 624 if (error) {
630 /* mbuf is already reclaimed in ipsec6_output_tunnel. */ 625 /* mbuf is already reclaimed in ipsec6_output_tunnel. */
631 m0 = m = NULL; 626 m0 = m = NULL;
632 m = NULL; 627 m = NULL;
633 switch (error) { 628 switch (error) {
634 case EHOSTUNREACH: 629 case EHOSTUNREACH:
635 case ENETUNREACH: 630 case ENETUNREACH:
636 case EMSGSIZE: 631 case EMSGSIZE:
637 case ENOBUFS: 632 case ENOBUFS:
638 case ENOMEM: 633 case ENOMEM:
639 break; 634 break;
640 default: 635 default:
641 printf("ip6_output (ipsec): error code %d\n", error); 636 printf("ip6_output (ipsec): error code %d\n", error);
642 /* FALLTHROUGH */ 637 /* FALLTHROUGH */
643 case ENOENT: 638 case ENOENT:
644 /* don't show these error codes to the user */ 639 /* don't show these error codes to the user */
645 error = 0; 640 error = 0;
646 break; 641 break;
647 } 642 }
648 goto bad; 643 goto bad;
649 } 644 }
650 645
651 exthdrs.ip6e_ip6 = m; 646 exthdrs.ip6e_ip6 = m;
652 } 647 }
653#endif /* IPSEC */ 648#endif /* IPSEC */
654#ifdef FAST_IPSEC 649#ifdef FAST_IPSEC
655 if (needipsec) { 650 if (needipsec) {
656 s = splsoftnet(); 651 s = splsoftnet();
657 error = ipsec6_process_packet(m,sp->req); 652 error = ipsec6_process_packet(m,sp->req);
658 653
659 /* 654 /*
660 * Preserve KAME behaviour: ENOENT can be returned 655 * Preserve KAME behaviour: ENOENT can be returned
661 * when an SA acquire is in progress. Don't propagate 656 * when an SA acquire is in progress. Don't propagate
662 * this to user-level; it confuses applications. 657 * this to user-level; it confuses applications.
663 * XXX this will go away when the SADB is redone. 658 * XXX this will go away when the SADB is redone.
664 */ 659 */
665 if (error == ENOENT) 660 if (error == ENOENT)
666 error = 0; 661 error = 0;
667 splx(s); 662 splx(s);
668 goto done; 663 goto done;
669 } 664 }
670#endif /* FAST_IPSEC */  665#endif /* FAST_IPSEC */
671 666
672 667
673 668
674 /* adjust pointer */ 669 /* adjust pointer */
675 ip6 = mtod(m, struct ip6_hdr *); 670 ip6 = mtod(m, struct ip6_hdr *);
676 671
677 sockaddr_in6_init(&dst_sa, &ip6->ip6_dst, 0, 0, 0); 672 sockaddr_in6_init(&dst_sa, &ip6->ip6_dst, 0, 0, 0);
678 if ((error = in6_selectroute(&dst_sa, opt, im6o, ro, 673 if ((error = in6_selectroute(&dst_sa, opt, im6o, ro,
679 &ifp, &rt, 0)) != 0) { 674 &ifp, &rt, 0)) != 0) {
680 if (ifp != NULL) 675 if (ifp != NULL)
681 in6_ifstat_inc(ifp, ifs6_out_discard); 676 in6_ifstat_inc(ifp, ifs6_out_discard);
682 goto bad; 677 goto bad;
683 } 678 }
684 if (rt == NULL) { 679 if (rt == NULL) {
685 /* 680 /*
686 * If in6_selectroute() does not return a route entry, 681 * If in6_selectroute() does not return a route entry,
687 * dst may not have been updated. 682 * dst may not have been updated.
688 */ 683 */
689 rtcache_setdst(ro, sin6tosa(&dst_sa)); 684 rtcache_setdst(ro, sin6tosa(&dst_sa));
690 } 685 }
691 686
692 /* 687 /*
693 * then rt (for unicast) and ifp must be non-NULL valid values. 688 * then rt (for unicast) and ifp must be non-NULL valid values.
694 */ 689 */
695 if ((flags & IPV6_FORWARDING) == 0) { 690 if ((flags & IPV6_FORWARDING) == 0) {
696 /* XXX: the FORWARDING flag can be set for mrouting. */ 691 /* XXX: the FORWARDING flag can be set for mrouting. */
697 in6_ifstat_inc(ifp, ifs6_out_request); 692 in6_ifstat_inc(ifp, ifs6_out_request);
698 } 693 }
699 if (rt != NULL) { 694 if (rt != NULL) {
700 ia = (struct in6_ifaddr *)(rt->rt_ifa); 695 ia = (struct in6_ifaddr *)(rt->rt_ifa);
701 rt->rt_use++; 696 rt->rt_use++;
702 } 697 }
703 698
704 /* 699 /*
705 * The outgoing interface must be in the zone of source and 700 * The outgoing interface must be in the zone of source and
706 * destination addresses. We should use ia_ifp to support the 701 * destination addresses. We should use ia_ifp to support the
707 * case of sending packets to an address of our own. 702 * case of sending packets to an address of our own.
708 */ 703 */
709 if (ia != NULL && ia->ia_ifp) 704 if (ia != NULL && ia->ia_ifp)
710 origifp = ia->ia_ifp; 705 origifp = ia->ia_ifp;
711 else 706 else
712 origifp = ifp; 707 origifp = ifp;
713 708
714 src0 = ip6->ip6_src; 709 src0 = ip6->ip6_src;
715 if (in6_setscope(&src0, origifp, &zone)) 710 if (in6_setscope(&src0, origifp, &zone))
716 goto badscope; 711 goto badscope;
717 sockaddr_in6_init(&src_sa, &ip6->ip6_src, 0, 0, 0); 712 sockaddr_in6_init(&src_sa, &ip6->ip6_src, 0, 0, 0);
718 if (sa6_recoverscope(&src_sa) || zone != src_sa.sin6_scope_id) 713 if (sa6_recoverscope(&src_sa) || zone != src_sa.sin6_scope_id)
719 goto badscope; 714 goto badscope;
720 715
721 dst0 = ip6->ip6_dst; 716 dst0 = ip6->ip6_dst;
722 if (in6_setscope(&dst0, origifp, &zone)) 717 if (in6_setscope(&dst0, origifp, &zone))
723 goto badscope; 718 goto badscope;
724 /* re-initialize to be sure */ 719 /* re-initialize to be sure */
725 sockaddr_in6_init(&dst_sa, &ip6->ip6_dst, 0, 0, 0); 720 sockaddr_in6_init(&dst_sa, &ip6->ip6_dst, 0, 0, 0);
726 if (sa6_recoverscope(&dst_sa) || zone != dst_sa.sin6_scope_id) 721 if (sa6_recoverscope(&dst_sa) || zone != dst_sa.sin6_scope_id)
727 goto badscope; 722 goto badscope;
728 723
729 /* scope check is done. */ 724 /* scope check is done. */
730 725
731 if (rt == NULL || IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) { 726 if (rt == NULL || IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) {
732 if (dst == NULL) 727 if (dst == NULL)
733 dst = satocsin6(rtcache_getdst(ro)); 728 dst = satocsin6(rtcache_getdst(ro));
734 KASSERT(dst != NULL); 729 KASSERT(dst != NULL);
735 } else if (opt && rtcache_validate(&opt->ip6po_nextroute) != NULL) { 730 } else if (opt && rtcache_validate(&opt->ip6po_nextroute) != NULL) {
736 /* 731 /*
737 * The nexthop is explicitly specified by the 732 * The nexthop is explicitly specified by the
738 * application. We assume the next hop is an IPv6 733 * application. We assume the next hop is an IPv6
739 * address. 734 * address.
740 */ 735 */
741 dst = (struct sockaddr_in6 *)opt->ip6po_nexthop; 736 dst = (struct sockaddr_in6 *)opt->ip6po_nexthop;
742 } else if ((rt->rt_flags & RTF_GATEWAY)) 737 } else if ((rt->rt_flags & RTF_GATEWAY))
743 dst = (struct sockaddr_in6 *)rt->rt_gateway; 738 dst = (struct sockaddr_in6 *)rt->rt_gateway;
744 else if (dst == NULL) 739 else if (dst == NULL)
745 dst = satocsin6(rtcache_getdst(ro)); 740 dst = satocsin6(rtcache_getdst(ro));
746 741
747 /* 742 /*
748 * XXXXXX: original code follows: 743 * XXXXXX: original code follows:
749 */ 744 */
750 if (!IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst)) 745 if (!IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst))
751 m->m_flags &= ~(M_BCAST | M_MCAST); /* just in case */ 746 m->m_flags &= ~(M_BCAST | M_MCAST); /* just in case */
752 else { 747 else {
753 struct in6_multi *in6m; 748 struct in6_multi *in6m;
754 749
755 m->m_flags = (m->m_flags & ~M_BCAST) | M_MCAST; 750 m->m_flags = (m->m_flags & ~M_BCAST) | M_MCAST;
756 751
757 in6_ifstat_inc(ifp, ifs6_out_mcast); 752 in6_ifstat_inc(ifp, ifs6_out_mcast);
758 753
759 /* 754 /*
760 * Confirm that the outgoing interface supports multicast. 755 * Confirm that the outgoing interface supports multicast.
761 */ 756 */
762 if (!(ifp->if_flags & IFF_MULTICAST)) { 757 if (!(ifp->if_flags & IFF_MULTICAST)) {
763 IP6_STATINC(IP6_STAT_NOROUTE); 758 IP6_STATINC(IP6_STAT_NOROUTE);
764 in6_ifstat_inc(ifp, ifs6_out_discard); 759 in6_ifstat_inc(ifp, ifs6_out_discard);
765 error = ENETUNREACH; 760 error = ENETUNREACH;
766 goto bad; 761 goto bad;
767 } 762 }
768 763
769 IN6_LOOKUP_MULTI(ip6->ip6_dst, ifp, in6m); 764 IN6_LOOKUP_MULTI(ip6->ip6_dst, ifp, in6m);
770 if (in6m != NULL && 765 if (in6m != NULL &&
771 (im6o == NULL || im6o->im6o_multicast_loop)) { 766 (im6o == NULL || im6o->im6o_multicast_loop)) {
772 /* 767 /*
773 * If we belong to the destination multicast group 768 * If we belong to the destination multicast group
774 * on the outgoing interface, and the caller did not 769 * on the outgoing interface, and the caller did not
775 * forbid loopback, loop back a copy. 770 * forbid loopback, loop back a copy.
776 */ 771 */
777 KASSERT(dst != NULL); 772 KASSERT(dst != NULL);
778 ip6_mloopback(ifp, m, dst); 773 ip6_mloopback(ifp, m, dst);
779 } else { 774 } else {
780 /* 775 /*
781 * If we are acting as a multicast router, perform 776 * If we are acting as a multicast router, perform
782 * multicast forwarding as if the packet had just 777 * multicast forwarding as if the packet had just
783 * arrived on the interface to which we are about 778 * arrived on the interface to which we are about
784 * to send. The multicast forwarding function 779 * to send. The multicast forwarding function
785 * recursively calls this function, using the 780 * recursively calls this function, using the
786 * IPV6_FORWARDING flag to prevent infinite recursion. 781 * IPV6_FORWARDING flag to prevent infinite recursion.
787 * 782 *
788 * Multicasts that are looped back by ip6_mloopback(), 783 * Multicasts that are looped back by ip6_mloopback(),
789 * above, will be forwarded by the ip6_input() routine, 784 * above, will be forwarded by the ip6_input() routine,
790 * if necessary. 785 * if necessary.
791 */ 786 */
792 if (ip6_mrouter && (flags & IPV6_FORWARDING) == 0) { 787 if (ip6_mrouter && (flags & IPV6_FORWARDING) == 0) {
793 if (ip6_mforward(ip6, ifp, m) != 0) { 788 if (ip6_mforward(ip6, ifp, m) != 0) {
794 m_freem(m); 789 m_freem(m);
795 goto done; 790 goto done;
796 } 791 }
797 } 792 }
798 } 793 }
799 /* 794 /*
800 * Multicasts with a hoplimit of zero may be looped back, 795 * Multicasts with a hoplimit of zero may be looped back,
801 * above, but must not be transmitted on a network. 796 * above, but must not be transmitted on a network.
802 * Also, multicasts addressed to the loopback interface 797 * Also, multicasts addressed to the loopback interface
803 * are not sent -- the above call to ip6_mloopback() will 798 * are not sent -- the above call to ip6_mloopback() will
804 * loop back a copy if this host actually belongs to the 799 * loop back a copy if this host actually belongs to the
805 * destination group on the loopback interface. 800 * destination group on the loopback interface.
806 */ 801 */
807 if (ip6->ip6_hlim == 0 || (ifp->if_flags & IFF_LOOPBACK) || 802 if (ip6->ip6_hlim == 0 || (ifp->if_flags & IFF_LOOPBACK) ||
808 IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst)) { 803 IN6_IS_ADDR_MC_INTFACELOCAL(&ip6->ip6_dst)) {
809 m_freem(m); 804 m_freem(m);
810 goto done; 805 goto done;
811 } 806 }
812 } 807 }
813 808
814 /* 809 /*
815 * Fill the outgoing inteface to tell the upper layer 810 * Fill the outgoing inteface to tell the upper layer
816 * to increment per-interface statistics. 811 * to increment per-interface statistics.
817 */ 812 */
818 if (ifpp) 813 if (ifpp)
819 *ifpp = ifp; 814 *ifpp = ifp;
820 815
821 /* Determine path MTU. */ 816 /* Determine path MTU. */
822 if ((error = ip6_getpmtu(ro_pmtu, ro, ifp, &finaldst, &mtu, 817 if ((error = ip6_getpmtu(ro_pmtu, ro, ifp, &finaldst, &mtu,
823 &alwaysfrag)) != 0) 818 &alwaysfrag)) != 0)
824 goto bad; 819 goto bad;
825#ifdef IPSEC 820#ifdef IPSEC
826 if (needipsectun) 821 if (needipsectun)
827 mtu = IPV6_MMTU; 822 mtu = IPV6_MMTU;
828#endif 823#endif
829 824
830 /* 825 /*
831 * The caller of this function may specify to use the minimum MTU 826 * The caller of this function may specify to use the minimum MTU
832 * in some cases. 827 * in some cases.
833 * An advanced API option (IPV6_USE_MIN_MTU) can also override MTU 828 * An advanced API option (IPV6_USE_MIN_MTU) can also override MTU
834 * setting. The logic is a bit complicated; by default, unicast 829 * setting. The logic is a bit complicated; by default, unicast
835 * packets will follow path MTU while multicast packets will be sent at 830 * packets will follow path MTU while multicast packets will be sent at
836 * the minimum MTU. If IP6PO_MINMTU_ALL is specified, all packets 831 * the minimum MTU. If IP6PO_MINMTU_ALL is specified, all packets
837 * including unicast ones will be sent at the minimum MTU. Multicast 832 * including unicast ones will be sent at the minimum MTU. Multicast
838 * packets will always be sent at the minimum MTU unless 833 * packets will always be sent at the minimum MTU unless
839 * IP6PO_MINMTU_DISABLE is explicitly specified. 834 * IP6PO_MINMTU_DISABLE is explicitly specified.
840 * See RFC 3542 for more details. 835 * See RFC 3542 for more details.
841 */ 836 */
842 if (mtu > IPV6_MMTU) { 837 if (mtu > IPV6_MMTU) {
843 if ((flags & IPV6_MINMTU)) 838 if ((flags & IPV6_MINMTU))
844 mtu = IPV6_MMTU; 839 mtu = IPV6_MMTU;
845 else if (opt && opt->ip6po_minmtu == IP6PO_MINMTU_ALL) 840 else if (opt && opt->ip6po_minmtu == IP6PO_MINMTU_ALL)
846 mtu = IPV6_MMTU; 841 mtu = IPV6_MMTU;
847 else if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) && 842 else if (IN6_IS_ADDR_MULTICAST(&ip6->ip6_dst) &&
848 (opt == NULL || 843 (opt == NULL ||
849 opt->ip6po_minmtu != IP6PO_MINMTU_DISABLE)) { 844 opt->ip6po_minmtu != IP6PO_MINMTU_DISABLE)) {
850 mtu = IPV6_MMTU; 845 mtu = IPV6_MMTU;
851 } 846 }
852 } 847 }
853 848
854 /* 849 /*
855 * clear embedded scope identifiers if necessary. 850 * clear embedded scope identifiers if necessary.
856 * in6_clearscope will touch the addresses only when necessary. 851 * in6_clearscope will touch the addresses only when necessary.
857 */ 852 */
858 in6_clearscope(&ip6->ip6_src); 853 in6_clearscope(&ip6->ip6_src);
859 in6_clearscope(&ip6->ip6_dst); 854 in6_clearscope(&ip6->ip6_dst);
860 855
861 /* 856 /*
862 * If the outgoing packet contains a hop-by-hop options header, 857 * If the outgoing packet contains a hop-by-hop options header,
863 * it must be examined and processed even by the source node. 858 * it must be examined and processed even by the source node.
864 * (RFC 2460, section 4.) 859 * (RFC 2460, section 4.)
865 */ 860 */
866 if (exthdrs.ip6e_hbh) { 861 if (exthdrs.ip6e_hbh) {
867 struct ip6_hbh *hbh = mtod(exthdrs.ip6e_hbh, struct ip6_hbh *); 862 struct ip6_hbh *hbh = mtod(exthdrs.ip6e_hbh, struct ip6_hbh *);
868 u_int32_t dummy1; /* XXX unused */ 863 u_int32_t dummy1; /* XXX unused */
869 u_int32_t dummy2; /* XXX unused */ 864 u_int32_t dummy2; /* XXX unused */
870 865
871 /* 866 /*
872 * XXX: if we have to send an ICMPv6 error to the sender, 867 * XXX: if we have to send an ICMPv6 error to the sender,
873 * we need the M_LOOP flag since icmp6_error() expects 868 * we need the M_LOOP flag since icmp6_error() expects
874 * the IPv6 and the hop-by-hop options header are 869 * the IPv6 and the hop-by-hop options header are
875 * continuous unless the flag is set. 870 * continuous unless the flag is set.
876 */ 871 */
877 m->m_flags |= M_LOOP; 872 m->m_flags |= M_LOOP;
878 m->m_pkthdr.rcvif = ifp; 873 m->m_pkthdr.rcvif = ifp;
879 if (ip6_process_hopopts(m, (u_int8_t *)(hbh + 1), 874 if (ip6_process_hopopts(m, (u_int8_t *)(hbh + 1),
880 ((hbh->ip6h_len + 1) << 3) - sizeof(struct ip6_hbh), 875 ((hbh->ip6h_len + 1) << 3) - sizeof(struct ip6_hbh),
881 &dummy1, &dummy2) < 0) { 876 &dummy1, &dummy2) < 0) {
882 /* m was already freed at this point */ 877 /* m was already freed at this point */
883 error = EINVAL;/* better error? */ 878 error = EINVAL;/* better error? */
884 goto done; 879 goto done;
885 } 880 }
886 m->m_flags &= ~M_LOOP; /* XXX */ 881 m->m_flags &= ~M_LOOP; /* XXX */
887 m->m_pkthdr.rcvif = NULL; 882 m->m_pkthdr.rcvif = NULL;
888 } 883 }
889 884
890#ifdef PFIL_HOOKS 885#ifdef PFIL_HOOKS
891 /* 886 /*
892 * Run through list of hooks for output packets. 887 * Run through list of hooks for output packets.
893 */ 888 */
894 if ((error = pfil_run_hooks(&inet6_pfil_hook, &m, ifp, PFIL_OUT)) != 0) 889 if ((error = pfil_run_hooks(&inet6_pfil_hook, &m, ifp, PFIL_OUT)) != 0)
895 goto done; 890 goto done;
896 if (m == NULL) 891 if (m == NULL)
897 goto done; 892 goto done;
898 ip6 = mtod(m, struct ip6_hdr *); 893 ip6 = mtod(m, struct ip6_hdr *);
899#endif /* PFIL_HOOKS */ 894#endif /* PFIL_HOOKS */
900 /* 895 /*
901 * Send the packet to the outgoing interface. 896 * Send the packet to the outgoing interface.
902 * If necessary, do IPv6 fragmentation before sending. 897 * If necessary, do IPv6 fragmentation before sending.
903 * 898 *
904 * the logic here is rather complex: 899 * the logic here is rather complex:
905 * 1: normal case (dontfrag == 0, alwaysfrag == 0) 900 * 1: normal case (dontfrag == 0, alwaysfrag == 0)
906 * 1-a: send as is if tlen <= path mtu 901 * 1-a: send as is if tlen <= path mtu
907 * 1-b: fragment if tlen > path mtu 902 * 1-b: fragment if tlen > path mtu
908 * 903 *
909 * 2: if user asks us not to fragment (dontfrag == 1) 904 * 2: if user asks us not to fragment (dontfrag == 1)
910 * 2-a: send as is if tlen <= interface mtu 905 * 2-a: send as is if tlen <= interface mtu
911 * 2-b: error if tlen > interface mtu 906 * 2-b: error if tlen > interface mtu
912 * 907 *
913 * 3: if we always need to attach fragment header (alwaysfrag == 1) 908 * 3: if we always need to attach fragment header (alwaysfrag == 1)
914 * always fragment 909 * always fragment
915 * 910 *
916 * 4: if dontfrag == 1 && alwaysfrag == 1 911 * 4: if dontfrag == 1 && alwaysfrag == 1
917 * error, as we cannot handle this conflicting request 912 * error, as we cannot handle this conflicting request
918 */ 913 */
919 tlen = m->m_pkthdr.len; 914 tlen = m->m_pkthdr.len;
920 tso = (m->m_pkthdr.csum_flags & M_CSUM_TSOv6) != 0; 915 tso = (m->m_pkthdr.csum_flags & M_CSUM_TSOv6) != 0;
921 if (opt && (opt->ip6po_flags & IP6PO_DONTFRAG)) 916 if (opt && (opt->ip6po_flags & IP6PO_DONTFRAG))
922 dontfrag = 1; 917 dontfrag = 1;
923 else 918 else
924 dontfrag = 0; 919 dontfrag = 0;
925 920
926 if (dontfrag && alwaysfrag) { /* case 4 */ 921 if (dontfrag && alwaysfrag) { /* case 4 */
927 /* conflicting request - can't transmit */ 922 /* conflicting request - can't transmit */
928 error = EMSGSIZE; 923 error = EMSGSIZE;
929 goto bad; 924 goto bad;
930 } 925 }
931 if (dontfrag && (!tso && tlen > IN6_LINKMTU(ifp))) { /* case 2-b */ 926 if (dontfrag && (!tso && tlen > IN6_LINKMTU(ifp))) { /* case 2-b */
932 /* 927 /*
933 * Even if the DONTFRAG option is specified, we cannot send the 928 * Even if the DONTFRAG option is specified, we cannot send the
934 * packet when the data length is larger than the MTU of the 929 * packet when the data length is larger than the MTU of the
935 * outgoing interface. 930 * outgoing interface.
936 * Notify the error by sending IPV6_PATHMTU ancillary data as 931 * Notify the error by sending IPV6_PATHMTU ancillary data as
937 * well as returning an error code (the latter is not described 932 * well as returning an error code (the latter is not described
938 * in the API spec.) 933 * in the API spec.)
939 */ 934 */
940 u_int32_t mtu32; 935 u_int32_t mtu32;
941 struct ip6ctlparam ip6cp; 936 struct ip6ctlparam ip6cp;
942 937
943 mtu32 = (u_int32_t)mtu; 938 mtu32 = (u_int32_t)mtu;
944 memset(&ip6cp, 0, sizeof(ip6cp)); 939 memset(&ip6cp, 0, sizeof(ip6cp));
945 ip6cp.ip6c_cmdarg = (void *)&mtu32; 940 ip6cp.ip6c_cmdarg = (void *)&mtu32;
946 pfctlinput2(PRC_MSGSIZE, 941 pfctlinput2(PRC_MSGSIZE,
947 rtcache_getdst(ro_pmtu), &ip6cp); 942 rtcache_getdst(ro_pmtu), &ip6cp);
948 943
949 error = EMSGSIZE; 944 error = EMSGSIZE;
950 goto bad; 945 goto bad;
951 } 946 }
952 947
953 /* 948 /*
954 * transmit packet without fragmentation 949 * transmit packet without fragmentation
955 */ 950 */
956 if (dontfrag || (!alwaysfrag && (tlen <= mtu || tso))) { 951 if (dontfrag || (!alwaysfrag && (tlen <= mtu || tso))) {
957 /* case 1-a and 2-a */ 952 /* case 1-a and 2-a */
958 struct in6_ifaddr *ia6; 953 struct in6_ifaddr *ia6;
959 int sw_csum; 954 int sw_csum;
960 955
961 ip6 = mtod(m, struct ip6_hdr *); 956 ip6 = mtod(m, struct ip6_hdr *);
962 ia6 = in6_ifawithifp(ifp, &ip6->ip6_src); 957 ia6 = in6_ifawithifp(ifp, &ip6->ip6_src);
963 if (ia6) { 958 if (ia6) {
964 /* Record statistics for this interface address. */ 959 /* Record statistics for this interface address. */
965 ia6->ia_ifa.ifa_data.ifad_outbytes += m->m_pkthdr.len; 960 ia6->ia_ifa.ifa_data.ifad_outbytes += m->m_pkthdr.len;
966 } 961 }
967#ifdef IPSEC 962#ifdef IPSEC
968 /* clean ipsec history once it goes out of the node */ 963 /* clean ipsec history once it goes out of the node */
969 ipsec_delaux(m); 964 ipsec_delaux(m);
970#endif 965#endif
971 966
972 sw_csum = m->m_pkthdr.csum_flags & ~ifp->if_csum_flags_tx; 967 sw_csum = m->m_pkthdr.csum_flags & ~ifp->if_csum_flags_tx;
973 if ((sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6)) != 0) { 968 if ((sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6)) != 0) {
974 if (IN6_NEED_CHECKSUM(ifp, 969 if (IN6_NEED_CHECKSUM(ifp,
975 sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6))) { 970 sw_csum & (M_CSUM_UDPv6|M_CSUM_TCPv6))) {
976 in6_delayed_cksum(m); 971 in6_delayed_cksum(m);
977 } 972 }
978 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6); 973 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6);
979 } 974 }
980 975
981 KASSERT(dst != NULL); 976 KASSERT(dst != NULL);
982 if (__predict_true(!tso || 977 if (__predict_true(!tso ||
983 (ifp->if_capenable & IFCAP_TSOv6) != 0)) { 978 (ifp->if_capenable & IFCAP_TSOv6) != 0)) {
984 error = nd6_output(ifp, origifp, m, dst, rt); 979 error = nd6_output(ifp, origifp, m, dst, rt);
985 } else { 980 } else {
986 error = ip6_tso_output(ifp, origifp, m, dst, rt); 981 error = ip6_tso_output(ifp, origifp, m, dst, rt);
987 } 982 }
988 goto done; 983 goto done;
989 } 984 }
990 985
991 if (tso) { 986 if (tso) {
992 error = EINVAL; /* XXX */ 987 error = EINVAL; /* XXX */
993 goto bad; 988 goto bad;
994 } 989 }
995 990
996 /* 991 /*
997 * try to fragment the packet. case 1-b and 3 992 * try to fragment the packet. case 1-b and 3
998 */ 993 */
999 if (mtu < IPV6_MMTU) { 994 if (mtu < IPV6_MMTU) {
1000 /* path MTU cannot be less than IPV6_MMTU */ 995 /* path MTU cannot be less than IPV6_MMTU */
1001 error = EMSGSIZE; 996 error = EMSGSIZE;
1002 in6_ifstat_inc(ifp, ifs6_out_fragfail); 997 in6_ifstat_inc(ifp, ifs6_out_fragfail);
1003 goto bad; 998 goto bad;
1004 } else if (ip6->ip6_plen == 0) { 999 } else if (ip6->ip6_plen == 0) {
1005 /* jumbo payload cannot be fragmented */ 1000 /* jumbo payload cannot be fragmented */
1006 error = EMSGSIZE; 1001 error = EMSGSIZE;
1007 in6_ifstat_inc(ifp, ifs6_out_fragfail); 1002 in6_ifstat_inc(ifp, ifs6_out_fragfail);
1008 goto bad; 1003 goto bad;
1009 } else { 1004 } else {
1010 struct mbuf **mnext, *m_frgpart; 1005 struct mbuf **mnext, *m_frgpart;
1011 struct ip6_frag *ip6f; 1006 struct ip6_frag *ip6f;
1012 u_int32_t id = htonl(ip6_randomid()); 1007 u_int32_t id = htonl(ip6_randomid());
1013 u_char nextproto; 1008 u_char nextproto;
1014#if 0 /* see below */ 1009#if 0 /* see below */
1015 struct ip6ctlparam ip6cp; 1010 struct ip6ctlparam ip6cp;
1016 u_int32_t mtu32; 1011 u_int32_t mtu32;
1017#endif 1012#endif
1018 1013
1019 /* 1014 /*
1020 * Too large for the destination or interface; 1015 * Too large for the destination or interface;
1021 * fragment if possible. 1016 * fragment if possible.
1022 * Must be able to put at least 8 bytes per fragment. 1017 * Must be able to put at least 8 bytes per fragment.
1023 */ 1018 */
1024 hlen = unfragpartlen; 1019 hlen = unfragpartlen;
1025 if (mtu > IPV6_MAXPACKET) 1020 if (mtu > IPV6_MAXPACKET)
1026 mtu = IPV6_MAXPACKET; 1021 mtu = IPV6_MAXPACKET;
1027 1022
1028#if 0 1023#if 0
1029 /* 1024 /*
1030 * It is believed this code is a leftover from the 1025 * It is believed this code is a leftover from the
1031 * development of the IPV6_RECVPATHMTU sockopt and 1026 * development of the IPV6_RECVPATHMTU sockopt and
1032 * associated work to implement RFC3542. 1027 * associated work to implement RFC3542.
1033 * It's not entirely clear what the intent of the API 1028 * It's not entirely clear what the intent of the API
1034 * is at this point, so disable this code for now. 1029 * is at this point, so disable this code for now.
1035 * The IPV6_RECVPATHMTU sockopt and/or IPV6_DONTFRAG 1030 * The IPV6_RECVPATHMTU sockopt and/or IPV6_DONTFRAG
1036 * will send notifications if the application requests. 1031 * will send notifications if the application requests.
1037 */ 1032 */
1038 1033
1039 /* Notify a proper path MTU to applications. */ 1034 /* Notify a proper path MTU to applications. */
1040 mtu32 = (u_int32_t)mtu; 1035 mtu32 = (u_int32_t)mtu;
1041 memset(&ip6cp, 0, sizeof(ip6cp)); 1036 memset(&ip6cp, 0, sizeof(ip6cp));
1042 ip6cp.ip6c_cmdarg = (void *)&mtu32; 1037 ip6cp.ip6c_cmdarg = (void *)&mtu32;
1043 pfctlinput2(PRC_MSGSIZE, 1038 pfctlinput2(PRC_MSGSIZE,
1044 rtcache_getdst(ro_pmtu), &ip6cp); 1039 rtcache_getdst(ro_pmtu), &ip6cp);
1045#endif 1040#endif
1046 1041
1047 len = (mtu - hlen - sizeof(struct ip6_frag)) & ~7; 1042 len = (mtu - hlen - sizeof(struct ip6_frag)) & ~7;
1048 if (len < 8) { 1043 if (len < 8) {
1049 error = EMSGSIZE; 1044 error = EMSGSIZE;
1050 in6_ifstat_inc(ifp, ifs6_out_fragfail); 1045 in6_ifstat_inc(ifp, ifs6_out_fragfail);
1051 goto bad; 1046 goto bad;
1052 } 1047 }
1053 1048
1054 mnext = &m->m_nextpkt; 1049 mnext = &m->m_nextpkt;
1055 1050
1056 /* 1051 /*
1057 * Change the next header field of the last header in the 1052 * Change the next header field of the last header in the
1058 * unfragmentable part. 1053 * unfragmentable part.
1059 */ 1054 */
1060 if (exthdrs.ip6e_rthdr) { 1055 if (exthdrs.ip6e_rthdr) {
1061 nextproto = *mtod(exthdrs.ip6e_rthdr, u_char *); 1056 nextproto = *mtod(exthdrs.ip6e_rthdr, u_char *);
1062 *mtod(exthdrs.ip6e_rthdr, u_char *) = IPPROTO_FRAGMENT; 1057 *mtod(exthdrs.ip6e_rthdr, u_char *) = IPPROTO_FRAGMENT;
1063 } else if (exthdrs.ip6e_dest1) { 1058 } else if (exthdrs.ip6e_dest1) {
1064 nextproto = *mtod(exthdrs.ip6e_dest1, u_char *); 1059 nextproto = *mtod(exthdrs.ip6e_dest1, u_char *);
1065 *mtod(exthdrs.ip6e_dest1, u_char *) = IPPROTO_FRAGMENT; 1060 *mtod(exthdrs.ip6e_dest1, u_char *) = IPPROTO_FRAGMENT;
1066 } else if (exthdrs.ip6e_hbh) { 1061 } else if (exthdrs.ip6e_hbh) {
1067 nextproto = *mtod(exthdrs.ip6e_hbh, u_char *); 1062 nextproto = *mtod(exthdrs.ip6e_hbh, u_char *);
1068 *mtod(exthdrs.ip6e_hbh, u_char *) = IPPROTO_FRAGMENT; 1063 *mtod(exthdrs.ip6e_hbh, u_char *) = IPPROTO_FRAGMENT;
1069 } else { 1064 } else {
1070 nextproto = ip6->ip6_nxt; 1065 nextproto = ip6->ip6_nxt;
1071 ip6->ip6_nxt = IPPROTO_FRAGMENT; 1066 ip6->ip6_nxt = IPPROTO_FRAGMENT;
1072 } 1067 }
1073 1068
1074 if ((m->m_pkthdr.csum_flags & (M_CSUM_UDPv6|M_CSUM_TCPv6)) 1069 if ((m->m_pkthdr.csum_flags & (M_CSUM_UDPv6|M_CSUM_TCPv6))
1075 != 0) { 1070 != 0) {
1076 if (IN6_NEED_CHECKSUM(ifp, 1071 if (IN6_NEED_CHECKSUM(ifp,
1077 m->m_pkthdr.csum_flags & 1072 m->m_pkthdr.csum_flags &
1078 (M_CSUM_UDPv6|M_CSUM_TCPv6))) { 1073 (M_CSUM_UDPv6|M_CSUM_TCPv6))) {
1079 in6_delayed_cksum(m); 1074 in6_delayed_cksum(m);
1080 } 1075 }
1081 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6); 1076 m->m_pkthdr.csum_flags &= ~(M_CSUM_UDPv6|M_CSUM_TCPv6);
1082 } 1077 }
1083 1078
1084 /* 1079 /*
1085 * Loop through length of segment after first fragment, 1080 * Loop through length of segment after first fragment,
1086 * make new header and copy data of each part and link onto 1081 * make new header and copy data of each part and link onto
1087 * chain. 1082 * chain.
1088 */ 1083 */
1089 m0 = m; 1084 m0 = m;
1090 for (off = hlen; off < tlen; off += len) { 1085 for (off = hlen; off < tlen; off += len) {
1091 struct mbuf *mlast; 1086 struct mbuf *mlast;
1092 1087
1093 MGETHDR(m, M_DONTWAIT, MT_HEADER); 1088 MGETHDR(m, M_DONTWAIT, MT_HEADER);
1094 if (!m) { 1089 if (!m) {
1095 error = ENOBUFS; 1090 error = ENOBUFS;
1096 IP6_STATINC(IP6_STAT_ODROPPED); 1091 IP6_STATINC(IP6_STAT_ODROPPED);
1097 goto sendorfree; 1092 goto sendorfree;
1098 } 1093 }
1099 m->m_pkthdr.rcvif = NULL; 1094 m->m_pkthdr.rcvif = NULL;
1100 m->m_flags = m0->m_flags & M_COPYFLAGS; 1095 m->m_flags = m0->m_flags & M_COPYFLAGS;
1101 *mnext = m; 1096 *mnext = m;
1102 mnext = &m->m_nextpkt; 1097 mnext = &m->m_nextpkt;
1103 m->m_data += max_linkhdr; 1098 m->m_data += max_linkhdr;
1104 mhip6 = mtod(m, struct ip6_hdr *); 1099 mhip6 = mtod(m, struct ip6_hdr *);
1105 *mhip6 = *ip6; 1100 *mhip6 = *ip6;
1106 m->m_len = sizeof(*mhip6); 1101 m->m_len = sizeof(*mhip6);
1107 error = ip6_insertfraghdr(m0, m, hlen, &ip6f); 1102 error = ip6_insertfraghdr(m0, m, hlen, &ip6f);
1108 if (error) { 1103 if (error) {
1109 IP6_STATINC(IP6_STAT_ODROPPED); 1104 IP6_STATINC(IP6_STAT_ODROPPED);
1110 goto sendorfree; 1105 goto sendorfree;
1111 } 1106 }
1112 ip6f->ip6f_offlg = htons((u_int16_t)((off - hlen) & ~7)); 1107 ip6f->ip6f_offlg = htons((u_int16_t)((off - hlen) & ~7));
1113 if (off + len >= tlen) 1108 if (off + len >= tlen)
1114 len = tlen - off; 1109 len = tlen - off;
1115 else 1110 else
1116 ip6f->ip6f_offlg |= IP6F_MORE_FRAG; 1111 ip6f->ip6f_offlg |= IP6F_MORE_FRAG;
1117 mhip6->ip6_plen = htons((u_int16_t)(len + hlen + 1112 mhip6->ip6_plen = htons((u_int16_t)(len + hlen +
1118 sizeof(*ip6f) - sizeof(struct ip6_hdr))); 1113 sizeof(*ip6f) - sizeof(struct ip6_hdr)));
1119 if ((m_frgpart = m_copy(m0, off, len)) == 0) { 1114 if ((m_frgpart = m_copy(m0, off, len)) == 0) {
1120 error = ENOBUFS; 1115 error = ENOBUFS;
1121 IP6_STATINC(IP6_STAT_ODROPPED); 1116 IP6_STATINC(IP6_STAT_ODROPPED);
1122 goto sendorfree; 1117 goto sendorfree;
1123 } 1118 }
1124 for (mlast = m; mlast->m_next; mlast = mlast->m_next) 1119 for (mlast = m; mlast->m_next; mlast = mlast->m_next)
1125 ; 1120 ;
1126 mlast->m_next = m_frgpart; 1121 mlast->m_next = m_frgpart;
1127 m->m_pkthdr.len = len + hlen + sizeof(*ip6f); 1122 m->m_pkthdr.len = len + hlen + sizeof(*ip6f);
1128 m->m_pkthdr.rcvif = (struct ifnet *)0; 1123 m->m_pkthdr.rcvif = (struct ifnet *)0;
1129 ip6f->ip6f_reserved = 0; 1124 ip6f->ip6f_reserved = 0;
1130 ip6f->ip6f_ident = id; 1125 ip6f->ip6f_ident = id;
1131 ip6f->ip6f_nxt = nextproto; 1126 ip6f->ip6f_nxt = nextproto;
1132 IP6_STATINC(IP6_STAT_OFRAGMENTS); 1127 IP6_STATINC(IP6_STAT_OFRAGMENTS);
1133 in6_ifstat_inc(ifp, ifs6_out_fragcreat); 1128 in6_ifstat_inc(ifp, ifs6_out_fragcreat);
1134 } 1129 }
1135 1130
1136 in6_ifstat_inc(ifp, ifs6_out_fragok); 1131 in6_ifstat_inc(ifp, ifs6_out_fragok);
1137 } 1132 }
1138 1133
1139 /* 1134 /*
1140 * Remove leading garbages. 1135 * Remove leading garbages.
1141 */ 1136 */
1142sendorfree: 1137sendorfree:
1143 m = m0->m_nextpkt; 1138 m = m0->m_nextpkt;
1144 m0->m_nextpkt = 0; 1139 m0->m_nextpkt = 0;
1145 m_freem(m0); 1140 m_freem(m0);
1146 for (m0 = m; m; m = m0) { 1141 for (m0 = m; m; m = m0) {
1147 m0 = m->m_nextpkt; 1142 m0 = m->m_nextpkt;
1148 m->m_nextpkt = 0; 1143 m->m_nextpkt = 0;
1149 if (error == 0) { 1144 if (error == 0) {
1150 struct in6_ifaddr *ia6; 1145 struct in6_ifaddr *ia6;
1151 ip6 = mtod(m, struct ip6_hdr *); 1146 ip6 = mtod(m, struct ip6_hdr *);
1152 ia6 = in6_ifawithifp(ifp, &ip6->ip6_src); 1147 ia6 = in6_ifawithifp(ifp, &ip6->ip6_src);
1153 if (ia6) { 1148 if (ia6) {