net/smc: build and send V2 CLC proposal

The new format of an SMCD V2 CLC proposal is introduced, and
building and checking of SMCD V2 CLC proposals is adapted
accordingly.
Signed-off-by: Ursula Braun <ubraun@linux.ibm.com>
Signed-off-by: Karsten Graul <kgraul@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/smc/af_smc.c

@@ -1301,7 +1301,7 @@ static void smc_find_ism_device_serv(struct smc_sock *new_smc,
 	if (!smcd_indicated(pclc->hdr.typev1))
 		goto not_found;
 	ini->is_smcd = true; /* prepare ISM check */
-	ini->ism_peer_gid[0] = pclc_smcd->gid;
+	ini->ism_peer_gid[0] = ntohll(pclc_smcd->ism.gid);
 	if (smc_find_ism_device(new_smc, ini))
 		goto not_found;
 	if (!smc_listen_ism_init(new_smc, ini))

net/smc/smc.h

@@ -20,6 +20,7 @@
 
 #define SMC_V1		1		/* SMC version V1 */
 #define SMC_V2		2		/* SMC version V2 */
+#define SMC_RELEASE	0
 
 #define SMCPROTO_SMC		0	/* SMC protocol, IPv4 */
 #define SMCPROTO_SMC6		1	/* SMC protocol, IPv6 */
@@ -28,6 +29,8 @@
 					 * devices
 					 */
 
+#define SMC_MAX_EID_LEN		32
+
 extern struct proto smc_proto;
 extern struct proto smc_proto6;
 
@@ -251,6 +254,9 @@ extern struct workqueue_struct	*smc_close_wq;	/* wq for close work */
 
 extern u8	local_systemid[SMC_SYSTEMID_LEN]; /* unique system identifier */
 
+#define ntohll(x) be64_to_cpu(x)
+#define htonll(x) cpu_to_be64(x)
+
 /* convert an u32 value into network byte order, store it into a 3 byte field */
 static inline void hton24(u8 *net, u32 host)
 {

net/smc/smc_clc.c

@@ -34,12 +34,52 @@ static const char SMC_EYECATCHER[4] = {'\xe2', '\xd4', '\xc3', '\xd9'};
 /* eye catcher "SMCD" EBCDIC for CLC messages */
 static const char SMCD_EYECATCHER[4] = {'\xe2', '\xd4', '\xc3', '\xc4'};
 
+/* check arriving CLC proposal */
+static bool smc_clc_msg_prop_valid(struct smc_clc_msg_proposal *pclc)
+{
+	struct smc_clc_msg_proposal_prefix *pclc_prfx;
+	struct smc_clc_smcd_v2_extension *smcd_v2_ext;
+	struct smc_clc_msg_hdr *hdr = &pclc->hdr;
+	struct smc_clc_v2_extension *v2_ext;
+
+	v2_ext = smc_get_clc_v2_ext(pclc);
+	pclc_prfx = smc_clc_proposal_get_prefix(pclc);
+	if (hdr->version == SMC_V1) {
+		if (hdr->typev1 == SMC_TYPE_N)
+			return false;
+		if (ntohs(hdr->length) !=
+			sizeof(*pclc) + ntohs(pclc->iparea_offset) +
+			sizeof(*pclc_prfx) +
+			pclc_prfx->ipv6_prefixes_cnt *
+				sizeof(struct smc_clc_ipv6_prefix) +
+			sizeof(struct smc_clc_msg_trail))
+			return false;
+	} else {
+		if (ntohs(hdr->length) !=
+			sizeof(*pclc) +
+			sizeof(struct smc_clc_msg_smcd) +
+			(hdr->typev1 != SMC_TYPE_N ?
+				sizeof(*pclc_prfx) +
+				pclc_prfx->ipv6_prefixes_cnt *
+				sizeof(struct smc_clc_ipv6_prefix) : 0) +
+			(hdr->typev2 != SMC_TYPE_N ?
+				sizeof(*v2_ext) +
+				v2_ext->hdr.eid_cnt * SMC_MAX_EID_LEN : 0) +
+			(smcd_indicated(hdr->typev2) ?
+				sizeof(*smcd_v2_ext) + v2_ext->hdr.ism_gid_cnt *
+					sizeof(struct smc_clc_smcd_gid_chid) :
+				0) +
+			sizeof(struct smc_clc_msg_trail))
+			return false;
+	}
+	return true;
+}
+
 /* check if received message has a correct header length and contains valid
  * heading and trailing eyecatchers
  */
 static bool smc_clc_msg_hdr_valid(struct smc_clc_msg_hdr *clcm, bool check_trl)
 {
-	struct smc_clc_msg_proposal_prefix *pclc_prfx;
 	struct smc_clc_msg_accept_confirm *clc;
 	struct smc_clc_msg_proposal *pclc;
 	struct smc_clc_msg_decline *dclc;
@@ -51,13 +91,7 @@ static bool smc_clc_msg_hdr_valid(struct smc_clc_msg_hdr *clcm, bool check_trl)
 	switch (clcm->type) {
 	case SMC_CLC_PROPOSAL:
 		pclc = (struct smc_clc_msg_proposal *)clcm;
-		pclc_prfx = smc_clc_proposal_get_prefix(pclc);
-		if (ntohs(pclc->hdr.length) <
-			sizeof(*pclc) + ntohs(pclc->iparea_offset) +
-			sizeof(*pclc_prfx) +
-			pclc_prfx->ipv6_prefixes_cnt *
-				sizeof(struct smc_clc_ipv6_prefix) +
-			sizeof(*trl))
+		if (!smc_clc_msg_prop_valid(pclc))
 			return false;
 		trl = (struct smc_clc_msg_trail *)
 			((u8 *)pclc + ntohs(pclc->hdr.length) - sizeof(*trl));
@@ -327,9 +361,6 @@ int smc_clc_wait_msg(struct smc_sock *smc, void *buf, int buflen,
 		goto out;
 	}
 
-	if (clcm->type == SMC_CLC_PROPOSAL && clcm->typev1 == SMC_TYPE_N)
-		reason_code = SMC_CLC_DECL_VERSMISMAT; /* just V2 offered */
-
 	/* receive the complete CLC message */
 	memset(&msg, 0, sizeof(struct msghdr));
 	if (datlen > buflen) {
@@ -412,15 +443,18 @@ int smc_clc_send_decline(struct smc_sock *smc, u32 peer_diag_info)
 /* send CLC PROPOSAL message across internal TCP socket */
 int smc_clc_send_proposal(struct smc_sock *smc, struct smc_init_info *ini)
 {
+	struct smc_clc_smcd_v2_extension *smcd_v2_ext;
 	struct smc_clc_msg_proposal_prefix *pclc_prfx;
 	struct smc_clc_msg_proposal *pclc_base;
+	struct smc_clc_smcd_gid_chid *gidchids;
 	struct smc_clc_msg_proposal_area *pclc;
 	struct smc_clc_ipv6_prefix *ipv6_prfx;
+	struct smc_clc_v2_extension *v2_ext;
 	struct smc_clc_msg_smcd *pclc_smcd;
 	struct smc_clc_msg_trail *trl;
 	int len, i, plen, rc;
 	int reason_code = 0;
-	struct kvec vec[5];
+	struct kvec vec[8];
 	struct msghdr msg;
 
 	pclc = kzalloc(sizeof(*pclc), GFP_KERNEL);
@@ -431,24 +465,37 @@ int smc_clc_send_proposal(struct smc_sock *smc, struct smc_init_info *ini)
 	pclc_smcd = &pclc->pclc_smcd;
 	pclc_prfx = &pclc->pclc_prfx;
 	ipv6_prfx = pclc->pclc_prfx_ipv6;
+	v2_ext = &pclc->pclc_v2_ext;
+	smcd_v2_ext = &pclc->pclc_smcd_v2_ext;
+	gidchids = pclc->pclc_gidchids;
 	trl = &pclc->pclc_trl;
 
+	pclc_base->hdr.version = SMC_V2;
+	pclc_base->hdr.typev1 = ini->smc_type_v1;
+	pclc_base->hdr.typev2 = ini->smc_type_v2;
+	plen = sizeof(*pclc_base) + sizeof(*pclc_smcd) + sizeof(*trl);
+
 	/* retrieve ip prefixes for CLC proposal msg */
+	if (ini->smc_type_v1 != SMC_TYPE_N) {
 		rc = smc_clc_prfx_set(smc->clcsock, pclc_prfx, ipv6_prfx);
 		if (rc) {
+			if (ini->smc_type_v2 == SMC_TYPE_N) {
 				kfree(pclc);
-		return SMC_CLC_DECL_CNFERR; /* configuration error */
+				return SMC_CLC_DECL_CNFERR;
+			}
+			pclc_base->hdr.typev1 = SMC_TYPE_N;
+		} else {
+			pclc_base->iparea_offset = htons(sizeof(*pclc_smcd));
+			plen += sizeof(*pclc_prfx) +
+					pclc_prfx->ipv6_prefixes_cnt *
+					sizeof(ipv6_prfx[0]);
+		}
 	}
 
-	/* send SMC Proposal CLC message */
-	plen = sizeof(*pclc_base) + sizeof(*pclc_prfx) +
-	       (pclc_prfx->ipv6_prefixes_cnt * sizeof(ipv6_prfx[0])) +
-	       sizeof(*trl);
+	/* build SMC Proposal CLC message */
 	memcpy(pclc_base->hdr.eyecatcher, SMC_EYECATCHER,
 	       sizeof(SMC_EYECATCHER));
 	pclc_base->hdr.type = SMC_CLC_PROPOSAL;
-	pclc_base->hdr.version = SMC_V1;		/* SMC version */
-	pclc_base->hdr.typev1 = ini->smc_type_v1;
 	if (smcr_indicated(ini->smc_type_v1)) {
 		/* add SMC-R specifics */
 		memcpy(pclc_base->lcl.id_for_peer, local_systemid,
@@ -456,25 +503,65 @@ int smc_clc_send_proposal(struct smc_sock *smc, struct smc_init_info *ini)
 		memcpy(pclc_base->lcl.gid, ini->ib_gid, SMC_GID_SIZE);
 		memcpy(pclc_base->lcl.mac, &ini->ib_dev->mac[ini->ib_port - 1],
 		       ETH_ALEN);
-		pclc_base->iparea_offset = htons(0);
 	}
 	if (smcd_indicated(ini->smc_type_v1)) {
 		/* add SMC-D specifics */
-		plen += sizeof(*pclc_smcd);
-		pclc_base->iparea_offset = htons(sizeof(*pclc_smcd));
-		pclc_smcd->gid = ini->ism_dev[0]->local_gid;
+		if (ini->ism_dev[0]) {
+			pclc_smcd->ism.gid = htonll(ini->ism_dev[0]->local_gid);
+			pclc_smcd->ism.chid =
+				htons(smc_ism_get_chid(ini->ism_dev[0]));
 		}
-	pclc_base->hdr.length = htons(plen);
+	}
+	if (ini->smc_type_v2 == SMC_TYPE_N) {
+		pclc_smcd->v2_ext_offset = 0;
+	} else {
+		u16 v2_ext_offset;
+		u8 *eid = NULL;
 
+		v2_ext_offset = sizeof(*pclc_smcd) -
+			offsetofend(struct smc_clc_msg_smcd, v2_ext_offset);
+		if (ini->smc_type_v1 != SMC_TYPE_N)
+			v2_ext_offset += sizeof(*pclc_prfx) +
+						pclc_prfx->ipv6_prefixes_cnt *
+						sizeof(ipv6_prfx[0]);
+		pclc_smcd->v2_ext_offset = htons(v2_ext_offset);
+		v2_ext->hdr.eid_cnt = 0;
+		v2_ext->hdr.ism_gid_cnt = ini->ism_offered_cnt;
+		v2_ext->hdr.flag.release = SMC_RELEASE;
+		v2_ext->hdr.flag.seid = 1;
+		v2_ext->hdr.smcd_v2_ext_offset = htons(sizeof(*v2_ext) -
+				offsetofend(struct smc_clnt_opts_area_hdr,
+					    smcd_v2_ext_offset) +
+				v2_ext->hdr.eid_cnt * SMC_MAX_EID_LEN);
+		if (ini->ism_dev[0])
+			smc_ism_get_system_eid(ini->ism_dev[0], &eid);
+		else
+			smc_ism_get_system_eid(ini->ism_dev[1], &eid);
+		if (eid)
+			memcpy(smcd_v2_ext->system_eid, eid, SMC_MAX_EID_LEN);
+		plen += sizeof(*v2_ext) + sizeof(*smcd_v2_ext);
+		if (ini->ism_offered_cnt) {
+			for (i = 1; i <= ini->ism_offered_cnt; i++) {
+				gidchids[i - 1].gid =
+					htonll(ini->ism_dev[i]->local_gid);
+				gidchids[i - 1].chid =
+					htons(smc_ism_get_chid(ini->ism_dev[i]));
+			}
+			plen += ini->ism_offered_cnt *
+				sizeof(struct smc_clc_smcd_gid_chid);
+		}
+	}
+	pclc_base->hdr.length = htons(plen);
 	memcpy(trl->eyecatcher, SMC_EYECATCHER, sizeof(SMC_EYECATCHER));
+
+	/* send SMC Proposal CLC message */
 	memset(&msg, 0, sizeof(msg));
 	i = 0;
 	vec[i].iov_base = pclc_base;
 	vec[i++].iov_len = sizeof(*pclc_base);
-	if (smcd_indicated(ini->smc_type_v1)) {
 	vec[i].iov_base = pclc_smcd;
 	vec[i++].iov_len = sizeof(*pclc_smcd);
-	}
+	if (ini->smc_type_v1 != SMC_TYPE_N) {
 		vec[i].iov_base = pclc_prfx;
 		vec[i++].iov_len = sizeof(*pclc_prfx);
 		if (pclc_prfx->ipv6_prefixes_cnt > 0) {
@@ -482,6 +569,18 @@ int smc_clc_send_proposal(struct smc_sock *smc, struct smc_init_info *ini)
 			vec[i++].iov_len = pclc_prfx->ipv6_prefixes_cnt *
 					   sizeof(ipv6_prfx[0]);
 		}
+	}
+	if (ini->smc_type_v2 != SMC_TYPE_N) {
+		vec[i].iov_base = v2_ext;
+		vec[i++].iov_len = sizeof(*v2_ext);
+		vec[i].iov_base = smcd_v2_ext;
+		vec[i++].iov_len = sizeof(*smcd_v2_ext);
+		if (ini->ism_offered_cnt) {
+			vec[i].iov_base = gidchids;
+			vec[i++].iov_len = ini->ism_offered_cnt *
+					sizeof(struct smc_clc_smcd_gid_chid);
+		}
+	}
 	vec[i].iov_base = trl;
 	vec[i++].iov_len = sizeof(*trl);
 	/* due to the few bytes needed for clc-handshake this cannot block */

net/smc/smc_clc.h

@@ -81,8 +81,6 @@ struct smc_clc_msg_local {	/* header2 of clc messages */
 	u8 mac[6];		/* mac of ib_device port */
 };
 
-#define SMC_CLC_MAX_V6_PREFIX	8
-
 /* Struct would be 4 byte aligned, but it is used in an array that is sent
  * to peers and must conform to RFC7609, hence we need to use packed here.
  */
@@ -91,6 +89,44 @@ struct smc_clc_ipv6_prefix {
 	u8 prefix_len;
 } __packed;			/* format defined in RFC7609 */
 
+#if defined(__BIG_ENDIAN_BITFIELD)
+struct smc_clc_v2_flag {
+	u8 release : 4,
+	   rsvd    : 3,
+	   seid    : 1;
+};
+#elif defined(__LITTLE_ENDIAN_BITFIELD)
+struct smc_clc_v2_flag {
+	u8 seid   : 1,
+	rsvd      : 3,
+	release   : 4;
+};
+#endif
+
+struct smc_clnt_opts_area_hdr {
+	u8 eid_cnt;		/* number of user defined EIDs */
+	u8 ism_gid_cnt;		/* number of ISMv2 GIDs */
+	u8 reserved1;
+	struct smc_clc_v2_flag flag;
+	u8 reserved2[2];
+	__be16 smcd_v2_ext_offset; /* SMC-Dv2 Extension Offset */
+};
+
+struct smc_clc_smcd_gid_chid {
+	__be64 gid;		/* ISM GID */
+	__be16 chid;		/* ISMv2 CHID */
+} __packed;		/* format defined in
+			 * IBM Shared Memory Communications Version 2
+			 * (https://www.ibm.com/support/pages/node/6326337)
+			 */
+
+struct smc_clc_v2_extension {
+	struct smc_clnt_opts_area_hdr hdr;
+	u8 roce[16];		/* RoCEv2 GID */
+	u8 reserved[16];
+	u8 user_eids[0][SMC_MAX_EID_LEN];
+};
+
 struct smc_clc_msg_proposal_prefix {	/* prefix part of clc proposal message*/
 	__be32 outgoing_subnet;	/* subnet mask */
 	u8 prefix_len;		/* number of significant bits in mask */
@@ -99,8 +135,15 @@ struct smc_clc_msg_proposal_prefix {	/* prefix part of clc proposal message*/
 } __aligned(4);
 
 struct smc_clc_msg_smcd {	/* SMC-D GID information */
-	u64 gid;		/* ISM GID of requestor */
-	u8 res[32];
+	struct smc_clc_smcd_gid_chid ism; /* ISM native GID+CHID of requestor */
+	__be16 v2_ext_offset;	/* SMC Version 2 Extension Offset */
+	u8 reserved[28];
+};
+
+struct smc_clc_smcd_v2_extension {
+	u8 system_eid[SMC_MAX_EID_LEN];
+	u8 reserved[16];
+	struct smc_clc_smcd_gid_chid gidchid[0];
 };
 
 struct smc_clc_msg_proposal {	/* clc proposal message sent by Linux */
@@ -109,11 +152,16 @@ struct smc_clc_msg_proposal {	/* clc proposal message sent by Linux */
 	__be16 iparea_offset;	/* offset to IP address information area */
 } __aligned(4);
 
+#define SMC_CLC_MAX_V6_PREFIX		8
+
 struct smc_clc_msg_proposal_area {
 	struct smc_clc_msg_proposal		pclc_base;
 	struct smc_clc_msg_smcd			pclc_smcd;
 	struct smc_clc_msg_proposal_prefix	pclc_prfx;
 	struct smc_clc_ipv6_prefix	pclc_prfx_ipv6[SMC_CLC_MAX_V6_PREFIX];
+	struct smc_clc_v2_extension		pclc_v2_ext;
+	struct smc_clc_smcd_v2_extension	pclc_smcd_v2_ext;
+	struct smc_clc_smcd_gid_chid		pclc_gidchids[SMC_MAX_ISM_DEVS];
 	struct smc_clc_msg_trail		pclc_trl;
 };
 
@@ -190,13 +238,28 @@ static inline bool smcd_indicated(int smc_type)
 static inline struct smc_clc_msg_smcd *
 smc_get_clc_msg_smcd(struct smc_clc_msg_proposal *prop)
 {
-	if (smcd_indicated(prop->hdr.type) &&
+	if (smcd_indicated(prop->hdr.typev1) &&
 	    ntohs(prop->iparea_offset) != sizeof(struct smc_clc_msg_smcd))
 		return NULL;
 
 	return (struct smc_clc_msg_smcd *)(prop + 1);
 }
 
+static inline struct smc_clc_v2_extension *
+smc_get_clc_v2_ext(struct smc_clc_msg_proposal *prop)
+{
+	struct smc_clc_msg_smcd *prop_smcd = smc_get_clc_msg_smcd(prop);
+
+	if (!prop_smcd || !ntohs(prop_smcd->v2_ext_offset))
+		return NULL;
+
+	return (struct smc_clc_v2_extension *)
+	       ((u8 *)prop_smcd +
+	       offsetof(struct smc_clc_msg_smcd, v2_ext_offset) +
+	       sizeof(prop_smcd->v2_ext_offset) +
+	       ntohs(prop_smcd->v2_ext_offset));
+}
+
 struct smcd_dev;
 struct smc_init_info;