Merge projects/ipsec into head/.

 Small summary
 -------------

o Almost all IPsec releated code was moved into sys/netipsec.
o New kernel modules added: ipsec.ko and tcpmd5.ko. New kernel
  option IPSEC_SUPPORT added. It enables support for loading
  and unloading of ipsec.ko and tcpmd5.ko kernel modules.
o IPSEC_NAT_T option was removed. Now NAT-T support is enabled by
  default. The UDP_ENCAP_ESPINUDP_NON_IKE encapsulation type
  support was removed. Added TCP/UDP checksum handling for
  inbound packets that were decapsulated by transport mode SAs.
  setkey(8) modified to show run-time NAT-T configuration of SA.
o New network pseudo interface if_ipsec(4) added. For now it is
  build as part of ipsec.ko module (or with IPSEC kernel).
  It implements IPsec virtual tunnels to create route-based VPNs.
o The network stack now invokes IPsec functions using special
  methods. The only one header file <netipsec/ipsec_support.h>
  should be included to declare all the needed things to work
  with IPsec.
o All IPsec protocols handlers (ESP/AH/IPCOMP protosw) were removed.
  Now these protocols are handled directly via IPsec methods.
o TCP_SIGNATURE support was reworked to be more close to RFC.
o PF_KEY SADB was reworked:
  - now all security associations stored in the single SPI namespace,
    and all SAs MUST have unique SPI.
  - several hash tables added to speed up lookups in SADB.
  - SADB now uses rmlock to protect access, and concurrent threads
    can do SA lookups in the same time.
  - many PF_KEY message handlers were reworked to reflect changes
    in SADB.
  - SADB_UPDATE message was extended to support new PF_KEY headers:
    SADB_X_EXT_NEW_ADDRESS_SRC and SADB_X_EXT_NEW_ADDRESS_DST. They
    can be used by IKE daemon to change SA addresses.
o ipsecrequest and secpolicy structures were cardinally changed to
  avoid locking protection for ipsecrequest. Now we support
  only limited number (4) of bundled SAs, but they are supported
  for both INET and INET6.
o INPCB security policy cache was introduced. Each PCB now caches
  used security policies to avoid SP lookup for each packet.
o For inbound security policies added the mode, when the kernel does
  check for full history of applied IPsec transforms.
o References counting rules for security policies and security
  associations were changed. The proper SA locking added into xform
  code.
o xform code was also changed. Now it is possible to unregister xforms.
  tdb_xxx structures were changed and renamed to reflect changes in
  SADB/SPDB, and changed rules for locking and refcounting.

Reviewed by:	gnn, wblock
Obtained from:	Yandex LLC
Relnotes:	yes
Sponsored by:	Yandex LLC
Differential Revision:	https://reviews.freebsd.org/D9352

1 files changed, 115 insertions, 130 deletions

diff --git a/sys/netipsec/xform_ipcomp.c b/sys/netipsec/xform_ipcomp.c
index 066f048ba40d..eb68f2aa444e 100644
--- a/sys/netipsec/xform_ipcomp.c
+++ b/sys/netipsec/xform_ipcomp.c
@@ -37,7 +37,6 @@
 #include <sys/mbuf.h>
 #include <sys/lock.h>
 #include <sys/mutex.h>
-#include <sys/rwlock.h>
 #include <sys/socket.h>
 #include <sys/kernel.h>
 #include <sys/protosw.h>
@@ -65,6 +64,7 @@
 #include <netipsec/ipcomp_var.h>
 
 #include <netipsec/key.h>
+#include <netipsec/key_debug.h>
 
 #include <opencrypto/cryptodev.h>
 #include <opencrypto/deflate.h>
@@ -88,18 +88,6 @@ SYSCTL_VNET_PCPUSTAT(_net_inet_ipcomp, IPSECCTL_STATS, stats,
 static int ipcomp_input_cb(struct cryptop *crp);
 static int ipcomp_output_cb(struct cryptop *crp);
 
-struct comp_algo *
-ipcomp_algorithm_lookup(int alg)
-{
-	if (alg >= IPCOMP_ALG_MAX)
-		return NULL;
-	switch (alg) {
-	case SADB_X_CALG_DEFLATE:
-		return &comp_algo_deflate;
-	}
-	return NULL;
-}
-
 /*
  * RFC 3173 p 2.2. Non-Expansion Policy:
  * If the total size of a compressed payload and the IPComp header, as
@@ -116,10 +104,10 @@ ipcomp_encapcheck(union sockaddr_union *src, union sockaddr_union *dst)
 {
 	struct secasvar *sav;
 
-	sav = KEY_ALLOCSA_TUNNEL(src, dst, IPPROTO_IPCOMP);
+	sav = key_allocsa_tunnel(src, dst, IPPROTO_IPCOMP);
 	if (sav == NULL)
 		return (0);
-	KEY_FREESAV(&sav);
+	key_freesav(&sav);
 
 	if (src->sa.sa_family == AF_INET)
 		return (sizeof(struct in_addr) << 4);
@@ -161,11 +149,11 @@ ipcomp_nonexp_input(struct mbuf **mp, int *offp, int proto)
 static int
 ipcomp_init(struct secasvar *sav, struct xformsw *xsp)
 {
-	struct comp_algo *tcomp;
+	const struct comp_algo *tcomp;
 	struct cryptoini cric;
 
 	/* NB: algorithm really comes in alg_enc and not alg_comp! */
-	tcomp = ipcomp_algorithm_lookup(sav->alg_enc);
+	tcomp = comp_algorithm_lookup(sav->alg_enc);
 	if (tcomp == NULL) {
 		DPRINTF(("%s: unsupported compression algorithm %d\n", __func__,
 			 sav->alg_comp));
@@ -201,7 +189,7 @@ ipcomp_zeroize(struct secasvar *sav)
 static int
 ipcomp_input(struct mbuf *m, struct secasvar *sav, int skip, int protoff)
 {
-	struct tdb_crypto *tc;
+	struct xform_data *xd;
 	struct cryptodesc *crdc;
 	struct cryptop *crp;
 	struct ipcomp *ipcomp;
@@ -236,12 +224,12 @@ ipcomp_input(struct mbuf *m, struct secasvar *sav, int skip, int protoff)
 		return ENOBUFS;
 	}
 	/* Get IPsec-specific opaque pointer */
-	tc = (struct tdb_crypto *) malloc(sizeof (*tc), M_XDATA, M_NOWAIT|M_ZERO);
-	if (tc == NULL) {
-		m_freem(m);
-		crypto_freereq(crp);
-		DPRINTF(("%s: cannot allocate tdb_crypto\n", __func__));
+	xd = malloc(sizeof(*xd), M_XDATA, M_NOWAIT | M_ZERO);
+	if (xd == NULL) {
+		DPRINTF(("%s: cannot allocate xform_data\n", __func__));
 		IPCOMPSTAT_INC(ipcomps_crypto);
+		crypto_freereq(crp);
+		m_freem(m);
 		return ENOBUFS;
 	}
 	crdc = crp->crp_desc;
@@ -250,27 +238,25 @@ ipcomp_input(struct mbuf *m, struct secasvar *sav, int skip, int protoff)
 	crdc->crd_len = m->m_pkthdr.len - (skip + hlen);
 	crdc->crd_inject = skip;
 
-	tc->tc_ptr = 0;
-
 	/* Decompression operation */
 	crdc->crd_alg = sav->tdb_compalgxform->type;
 
+
 	/* Crypto operation descriptor */
 	crp->crp_ilen = m->m_pkthdr.len - (skip + hlen);
 	crp->crp_flags = CRYPTO_F_IMBUF | CRYPTO_F_CBIFSYNC;
 	crp->crp_buf = (caddr_t) m;
 	crp->crp_callback = ipcomp_input_cb;
-	crp->crp_sid = sav->tdb_cryptoid;
-	crp->crp_opaque = (caddr_t) tc;
+	crp->crp_opaque = (caddr_t) xd;
 
 	/* These are passed as-is to the callback */
-	tc->tc_spi = sav->spi;
-	tc->tc_dst = sav->sah->saidx.dst;
-	tc->tc_proto = sav->sah->saidx.proto;
-	tc->tc_protoff = protoff;
-	tc->tc_skip = skip;
-	KEY_ADDREFSA(sav);
-	tc->tc_sav = sav;
+	xd->sav = sav;
+	xd->protoff = protoff;
+	xd->skip = skip;
+
+	SECASVAR_LOCK(sav);
+	crp->crp_sid = xd->cryptoid = sav->tdb_cryptoid;
+	SECASVAR_UNLOCK(sav);
 
 	return crypto_dispatch(crp);
 }
@@ -281,28 +267,26 @@ ipcomp_input(struct mbuf *m, struct secasvar *sav, int skip, int protoff)
 static int
 ipcomp_input_cb(struct cryptop *crp)
 {
-	char buf[INET6_ADDRSTRLEN];
+	char buf[IPSEC_ADDRSTRLEN];
 	struct cryptodesc *crd;
-	struct tdb_crypto *tc;
-	int skip, protoff;
+	struct xform_data *xd;
 	struct mbuf *m;
 	struct secasvar *sav;
 	struct secasindex *saidx;
-	int hlen = IPCOMP_HLENGTH, error, clen;
-	u_int8_t nproto;
 	caddr_t addr;
+	uint64_t cryptoid;
+	int hlen = IPCOMP_HLENGTH, error, clen;
+	int skip, protoff;
+	uint8_t nproto;
 
 	crd = crp->crp_desc;
 
-	tc = (struct tdb_crypto *) crp->crp_opaque;
-	IPSEC_ASSERT(tc != NULL, ("null opaque crypto data area!"));
-	skip = tc->tc_skip;
-	protoff = tc->tc_protoff;
 	m = (struct mbuf *) crp->crp_buf;
-
-	sav = tc->tc_sav;
-	IPSEC_ASSERT(sav != NULL, ("null SA!"));
-
+	xd = (struct xform_data *) crp->crp_opaque;
+	sav = xd->sav;
+	skip = xd->skip;
+	protoff = xd->protoff;
+	cryptoid = xd->cryptoid;
 	saidx = &sav->sah->saidx;
 	IPSEC_ASSERT(saidx->dst.sa.sa_family == AF_INET ||
 		saidx->dst.sa.sa_family == AF_INET6,
@@ -310,12 +294,12 @@ ipcomp_input_cb(struct cryptop *crp)
 
 	/* Check for crypto errors */
 	if (crp->crp_etype) {
-		/* Reset the session ID */
-		if (sav->tdb_cryptoid != 0)
-			sav->tdb_cryptoid = crp->crp_sid;
-
 		if (crp->crp_etype == EAGAIN) {
-			return crypto_dispatch(crp);
+			/* Reset the session ID */
+			if (ipsec_updateid(sav, &crp->crp_sid, &cryptoid) != 0)
+				crypto_freesession(cryptoid);
+			xd->cryptoid = crp->crp_sid;
+			return (crypto_dispatch(crp));
 		}
 		IPCOMPSTAT_INC(ipcomps_noxform);
 		DPRINTF(("%s: crypto error %d\n", __func__, crp->crp_etype));
@@ -334,7 +318,7 @@ ipcomp_input_cb(struct cryptop *crp)
 	clen = crp->crp_olen;		/* Length of data after processing */
 
 	/* Release the crypto descriptors */
-	free(tc, M_XDATA), tc = NULL;
+	free(xd, M_XDATA), xd = NULL;
 	crypto_freereq(crp), crp = NULL;
 
 	/* In case it's not done already, adjust the size of the mbuf chain */
@@ -379,37 +363,33 @@ ipcomp_input_cb(struct cryptop *crp)
 		panic("%s: Unexpected address family: %d saidx=%p", __func__,
 		    saidx->dst.sa.sa_family, saidx);
 	}
-
-	KEY_FREESAV(&sav);
 	return error;
 bad:
-	if (sav)
-		KEY_FREESAV(&sav);
-	if (m)
+	if (sav != NULL)
+		key_freesav(&sav);
+	if (m != NULL)
 		m_freem(m);
-	if (tc != NULL)
-		free(tc, M_XDATA);
-	if (crp)
+	if (xd != NULL)
+		free(xd, M_XDATA);
+	if (crp != NULL)
 		crypto_freereq(crp);
 	return error;
 }
 
 /*
- * IPComp output routine, called by ipsec[46]_process_packet()
+ * IPComp output routine, called by ipsec[46]_perform_request()
  */
 static int
-ipcomp_output(struct mbuf *m, struct ipsecrequest *isr, struct mbuf **mp,
-    int skip, int protoff)
+ipcomp_output(struct mbuf *m, struct secpolicy *sp, struct secasvar *sav,
+    u_int idx, int skip, int protoff)
 {
-	char buf[INET6_ADDRSTRLEN];
-	struct secasvar *sav;
-	struct comp_algo *ipcompx;
-	int error, ralen, maxpacketsize;
+	char buf[IPSEC_ADDRSTRLEN];
+	const struct comp_algo *ipcompx;
 	struct cryptodesc *crdc;
 	struct cryptop *crp;
-	struct tdb_crypto *tc;
+	struct xform_data *xd;
+	int error, ralen, maxpacketsize;
 
-	sav = isr->sav;
 	IPSEC_ASSERT(sav != NULL, ("null SA"));
 	ipcompx = sav->tdb_compalgxform;
 	IPSEC_ASSERT(ipcompx != NULL, ("null compression xform"));
@@ -422,7 +402,7 @@ ipcomp_output(struct mbuf *m, struct ipsecrequest *isr, struct mbuf **mp,
 	 */
 	if (m->m_pkthdr.len <= ipcompx->minlen) {
 		IPCOMPSTAT_INC(ipcomps_threshold);
-		return ipsec_process_done(m, isr);
+		return ipsec_process_done(m, sp, sav, idx);
 	}
 
 	ralen = m->m_pkthdr.len - skip;	/* Raw payload length before comp. */
@@ -496,33 +476,31 @@ ipcomp_output(struct mbuf *m, struct ipsecrequest *isr, struct mbuf **mp,
 	crdc->crd_alg = ipcompx->type;
 
 	/* IPsec-specific opaque crypto info */
-	tc = (struct tdb_crypto *) malloc(sizeof(struct tdb_crypto),
-		M_XDATA, M_NOWAIT|M_ZERO);
-	if (tc == NULL) {
+	xd =  malloc(sizeof(struct xform_data), M_XDATA, M_NOWAIT | M_ZERO);
+	if (xd == NULL) {
 		IPCOMPSTAT_INC(ipcomps_crypto);
-		DPRINTF(("%s: failed to allocate tdb_crypto\n", __func__));
+		DPRINTF(("%s: failed to allocate xform_data\n", __func__));
 		crypto_freereq(crp);
 		error = ENOBUFS;
 		goto bad;
 	}
 
-	key_addref(isr->sp);
-	tc->tc_isr = isr;
-	KEY_ADDREFSA(sav);
-	tc->tc_sav = sav;
-	tc->tc_spi = sav->spi;
-	tc->tc_dst = sav->sah->saidx.dst;
-	tc->tc_proto = sav->sah->saidx.proto;
-	tc->tc_protoff = protoff;
-	tc->tc_skip = skip;
+	xd->sp = sp;
+	xd->sav = sav;
+	xd->idx = idx;
+	xd->skip = skip;
+	xd->protoff = protoff;
 
 	/* Crypto operation descriptor */
 	crp->crp_ilen = m->m_pkthdr.len;	/* Total input length */
 	crp->crp_flags = CRYPTO_F_IMBUF | CRYPTO_F_CBIFSYNC;
 	crp->crp_buf = (caddr_t) m;
 	crp->crp_callback = ipcomp_output_cb;
-	crp->crp_opaque = (caddr_t) tc;
-	crp->crp_sid = sav->tdb_cryptoid;
+	crp->crp_opaque = (caddr_t) xd;
+
+	SECASVAR_LOCK(sav);
+	crp->crp_sid = xd->cryptoid = sav->tdb_cryptoid;
+	SECASVAR_UNLOCK(sav);
 
 	return crypto_dispatch(crp);
 bad:
@@ -537,39 +515,32 @@ bad:
 static int
 ipcomp_output_cb(struct cryptop *crp)
 {
-	char buf[INET6_ADDRSTRLEN];
-	struct tdb_crypto *tc;
-	struct ipsecrequest *isr;
+	char buf[IPSEC_ADDRSTRLEN];
+	struct xform_data *xd;
+	struct secpolicy *sp;
 	struct secasvar *sav;
 	struct mbuf *m;
-	int error, skip;
+	uint64_t cryptoid;
+	u_int idx;
+	int error, skip, protoff;
 
-	tc = (struct tdb_crypto *) crp->crp_opaque;
-	IPSEC_ASSERT(tc != NULL, ("null opaque data area!"));
 	m = (struct mbuf *) crp->crp_buf;
-	skip = tc->tc_skip;
-
-	isr = tc->tc_isr;
-	IPSEC_ASSERT(isr->sp != NULL, ("NULL isr->sp"));
-	IPSECREQUEST_LOCK(isr);
-	sav = tc->tc_sav;
-	/* With the isr lock released SA pointer can be updated. */
-	if (sav != isr->sav) {
-		IPCOMPSTAT_INC(ipcomps_notdb);
-		DPRINTF(("%s: SA expired while in crypto\n", __func__));
-		error = ENOBUFS;		/*XXX*/
-		goto bad;
-	}
+	xd = (struct xform_data *) crp->crp_opaque;
+	idx = xd->idx;
+	sp = xd->sp;
+	sav = xd->sav;
+	skip = xd->skip;
+	protoff = xd->protoff;
+	cryptoid = xd->cryptoid;
 
 	/* Check for crypto errors */
 	if (crp->crp_etype) {
-		/* Reset the session ID */
-		if (sav->tdb_cryptoid != 0)
-			sav->tdb_cryptoid = crp->crp_sid;
-
 		if (crp->crp_etype == EAGAIN) {
-			IPSECREQUEST_UNLOCK(isr);
-			return crypto_dispatch(crp);
+			/* Reset the session ID */
+			if (ipsec_updateid(sav, &crp->crp_sid, &cryptoid) != 0)
+				crypto_freesession(cryptoid);
+			xd->cryptoid = crp->crp_sid;
+			return (crypto_dispatch(crp));
 		}
 		IPCOMPSTAT_INC(ipcomps_noxform);
 		DPRINTF(("%s: crypto error %d\n", __func__, crp->crp_etype));
@@ -595,7 +566,8 @@ ipcomp_output_cb(struct cryptop *crp)
 		mo = m_makespace(m, skip, IPCOMP_HLENGTH, &roff);
 		if (mo == NULL) {
 			IPCOMPSTAT_INC(ipcomps_wrap);
-			DPRINTF(("%s: IPCOMP header inject failed for IPCA %s/%08lx\n",
+			DPRINTF(("%s: IPCOMP header inject failed "
+			    "for IPCA %s/%08lx\n",
 			    __func__, ipsec_address(&sav->sah->saidx.dst, buf,
 			    sizeof(buf)), (u_long) ntohl(sav->spi)));
 			error = ENOBUFS;
@@ -622,7 +594,7 @@ ipcomp_output_cb(struct cryptop *crp)
 
 		/* Fix Next Protocol in IPv4/IPv6 header */
 		prot = IPPROTO_IPCOMP;
-		m_copyback(m, tc->tc_protoff, sizeof(u_int8_t),
+		m_copyback(m, protoff, sizeof(u_int8_t),
 		    (u_char *)&prot);
 
 		/* Adjust the length in the IP header */
@@ -658,33 +630,22 @@ ipcomp_output_cb(struct cryptop *crp)
 	}
 
 	/* Release the crypto descriptor */
-	free(tc, M_XDATA);
+	free(xd, M_XDATA);
 	crypto_freereq(crp);
 
 	/* NB: m is reclaimed by ipsec_process_done. */
-	error = ipsec_process_done(m, isr);
-	KEY_FREESAV(&sav);
-	IPSECREQUEST_UNLOCK(isr);
-	KEY_FREESP(&isr->sp);
+	error = ipsec_process_done(m, sp, sav, idx);
 	return (error);
 bad:
-	if (sav)
-		KEY_FREESAV(&sav);
-	IPSECREQUEST_UNLOCK(isr);
-	KEY_FREESP(&isr->sp);
 	if (m)
 		m_freem(m);
-	free(tc, M_XDATA);
+	free(xd, M_XDATA);
 	crypto_freereq(crp);
+	key_freesav(&sav);
+	key_freesp(&sp);
 	return (error);
 }
 
-static struct xformsw ipcomp_xformsw = {
-	XF_IPCOMP,		XFT_COMP,		"IPcomp",
-	ipcomp_init,		ipcomp_zeroize,		ipcomp_input,
-	ipcomp_output
-};
-
 #ifdef INET
 static const struct encaptab *ipe4_cookie = NULL;
 extern struct domain inetdomain;
@@ -768,6 +729,15 @@ ipcomp6_nonexp_encapcheck(const struct mbuf *m, int off, int proto,
 }
 #endif
 
+static struct xformsw ipcomp_xformsw = {
+	.xf_type =	XF_IPCOMP,
+	.xf_name =	"IPcomp",
+	.xf_init =	ipcomp_init,
+	.xf_zeroize =	ipcomp_zeroize,
+	.xf_input =	ipcomp_input,
+	.xf_output =	ipcomp_output,
+};
+
 static void
 ipcomp_attach(void)
 {
@@ -780,8 +750,23 @@ ipcomp_attach(void)
 	ipe6_cookie = encap_attach_func(AF_INET6, -1,
 	    ipcomp6_nonexp_encapcheck, &ipcomp6_protosw, NULL);
 #endif
-	xform_register(&ipcomp_xformsw);
+	xform_attach(&ipcomp_xformsw);
+}
+
+static void
+ipcomp_detach(void)
+{
+
+#ifdef INET
+	encap_detach(ipe4_cookie);
+#endif
+#ifdef INET6
+	encap_detach(ipe6_cookie);
+#endif
+	xform_detach(&ipcomp_xformsw);
 }
 
 SYSINIT(ipcomp_xform_init, SI_SUB_PROTO_DOMAIN, SI_ORDER_MIDDLE,
     ipcomp_attach, NULL);
+SYSUNINIT(ipcomp_xform_uninit, SI_SUB_PROTO_DOMAIN, SI_ORDER_MIDDLE,
+    ipcomp_detach, NULL);