DPDK  20.05.0
examples/ipsec-secgw/ipsec.c
/* SPDX-License-Identifier: BSD-3-Clause
* Copyright(c) 2016-2020 Intel Corporation
*/
#include <sys/types.h>
#include <netinet/in.h>
#include <netinet/ip.h>
#include <rte_log.h>
#include <rte_crypto.h>
#include <rte_security.h>
#include <rte_cryptodev.h>
#include <rte_ipsec.h>
#include <rte_ethdev.h>
#include <rte_mbuf.h>
#include <rte_hash.h>
#include "ipsec.h"
#include "esp.h"
static inline void
set_ipsec_conf(struct ipsec_sa *sa, struct rte_security_ipsec_xform *ipsec)
{
&ipsec->tunnel;
if (IS_IP4_TUNNEL(sa->flags)) {
tunnel->type =
tunnel->ipv4.ttl = IPDEFTTL;
memcpy((uint8_t *)&tunnel->ipv4.src_ip,
(uint8_t *)&sa->src.ip.ip4, 4);
memcpy((uint8_t *)&tunnel->ipv4.dst_ip,
(uint8_t *)&sa->dst.ip.ip4, 4);
} else if (IS_IP6_TUNNEL(sa->flags)) {
tunnel->type =
tunnel->ipv6.hlimit = IPDEFTTL;
tunnel->ipv6.dscp = 0;
tunnel->ipv6.flabel = 0;
memcpy((uint8_t *)&tunnel->ipv6.src_addr,
(uint8_t *)&sa->src.ip.ip6.ip6_b, 16);
memcpy((uint8_t *)&tunnel->ipv6.dst_addr,
(uint8_t *)&sa->dst.ip.ip6.ip6_b, 16);
}
/* TODO support for Transport */
}
ipsec->esn_soft_limit = IPSEC_OFFLOAD_ESN_SOFTLIMIT;
ipsec->replay_win_sz = app_sa_prm.window_size;
ipsec->options.esn = app_sa_prm.enable_esn;
}
int
create_lookaside_session(struct ipsec_ctx *ipsec_ctx, struct ipsec_sa *sa,
struct rte_ipsec_session *ips)
{
struct rte_cryptodev_info cdev_info;
unsigned long cdev_id_qp = 0;
int32_t ret = 0;
struct cdev_key key = { 0 };
key.lcore_id = (uint8_t)rte_lcore_id();
key.cipher_algo = (uint8_t)sa->cipher_algo;
key.auth_algo = (uint8_t)sa->auth_algo;
key.aead_algo = (uint8_t)sa->aead_algo;
ret = rte_hash_lookup_data(ipsec_ctx->cdev_map, &key,
(void **)&cdev_id_qp);
if (ret < 0) {
RTE_LOG(ERR, IPSEC,
"No cryptodev: core %u, cipher_algo %u, "
"auth_algo %u, aead_algo %u\n",
key.lcore_id,
key.cipher_algo,
key.auth_algo,
key.aead_algo);
return -1;
}
RTE_LOG_DP(DEBUG, IPSEC, "Create session for SA spi %u on cryptodev "
"%u qp %u\n", sa->spi,
ipsec_ctx->tbl[cdev_id_qp].id,
ipsec_ctx->tbl[cdev_id_qp].qp);
struct rte_security_session_conf sess_conf = {
.action_type = ips->type,
{.ipsec = {
.spi = sa->spi,
.salt = sa->salt,
.options = { 0 },
.replay_win_sz = 0,
.direction = sa->direction,
.mode = (IS_TUNNEL(sa->flags)) ?
} },
.crypto_xform = sa->xforms,
.userdata = NULL,
};
struct rte_security_ctx *ctx = (struct rte_security_ctx *)
rte_cryptodev_get_sec_ctx(
ipsec_ctx->tbl[cdev_id_qp].id);
/* Set IPsec parameters in conf */
set_ipsec_conf(sa, &(sess_conf.ipsec));
ips->security.ses = rte_security_session_create(ctx,
&sess_conf, ipsec_ctx->session_priv_pool);
if (ips->security.ses == NULL) {
RTE_LOG(ERR, IPSEC,
"SEC Session init failed: err: %d\n", ret);
return -1;
}
} else {
RTE_LOG(ERR, IPSEC, "Inline not supported\n");
return -1;
}
} else {
struct rte_cryptodev_info info;
uint16_t cdev_id;
cdev_id = ipsec_ctx->tbl[cdev_id_qp].id;
rte_cryptodev_info_get(cdev_id, &info);
if (!(info.feature_flags &
return -ENOTSUP;
ips->crypto.dev_id = cdev_id;
}
ipsec_ctx->session_pool);
rte_cryptodev_sym_session_init(ipsec_ctx->tbl[cdev_id_qp].id,
ips->crypto.ses, sa->xforms,
ipsec_ctx->session_priv_pool);
rte_cryptodev_info_get(ipsec_ctx->tbl[cdev_id_qp].id,
&cdev_info);
}
sa->cdev_id_qp = cdev_id_qp;
return 0;
}
int
create_inline_session(struct socket_ctx *skt_ctx, struct ipsec_sa *sa,
struct rte_ipsec_session *ips)
{
int32_t ret = 0;
struct rte_security_ctx *sec_ctx;
struct rte_security_session_conf sess_conf = {
.action_type = ips->type,
{.ipsec = {
.spi = sa->spi,
.salt = sa->salt,
.options = { 0 },
.replay_win_sz = 0,
.direction = sa->direction,
.mode = (sa->flags == IP4_TUNNEL ||
sa->flags == IP6_TUNNEL) ?
} },
.crypto_xform = sa->xforms,
.userdata = NULL,
};
RTE_LOG_DP(DEBUG, IPSEC, "Create session for SA spi %u on port %u\n",
sa->spi, sa->portid);
struct rte_flow_error err;
const struct rte_security_capability *sec_cap;
int ret = 0;
sec_ctx = (struct rte_security_ctx *)
sa->portid);
if (sec_ctx == NULL) {
RTE_LOG(ERR, IPSEC,
" rte_eth_dev_get_sec_ctx failed\n");
return -1;
}
ips->security.ses = rte_security_session_create(sec_ctx,
&sess_conf, skt_ctx->session_pool);
if (ips->security.ses == NULL) {
RTE_LOG(ERR, IPSEC,
"SEC Session init failed: err: %d\n", ret);
return -1;
}
sec_cap = rte_security_capabilities_get(sec_ctx);
/* iterate until ESP tunnel*/
while (sec_cap->action != RTE_SECURITY_ACTION_TYPE_NONE) {
if (sec_cap->action == ips->type &&
sec_cap->protocol ==
sec_cap->ipsec.mode ==
sec_cap->ipsec.direction == sa->direction)
break;
sec_cap++;
}
RTE_LOG(ERR, IPSEC,
"No suitable security capability found\n");
return -1;
}
ips->security.ol_flags = sec_cap->ol_flags;
ips->security.ctx = sec_ctx;
sa->pattern[0].type = RTE_FLOW_ITEM_TYPE_ETH;
if (IS_IP6(sa->flags)) {
sa->pattern[1].mask = &rte_flow_item_ipv6_mask;
sa->pattern[1].type = RTE_FLOW_ITEM_TYPE_IPV6;
sa->pattern[1].spec = &sa->ipv6_spec;
memcpy(sa->ipv6_spec.hdr.dst_addr,
sa->dst.ip.ip6.ip6_b, 16);
memcpy(sa->ipv6_spec.hdr.src_addr,
sa->src.ip.ip6.ip6_b, 16);
} else if (IS_IP4(sa->flags)) {
sa->pattern[1].mask = &rte_flow_item_ipv4_mask;
sa->pattern[1].type = RTE_FLOW_ITEM_TYPE_IPV4;
sa->pattern[1].spec = &sa->ipv4_spec;
sa->ipv4_spec.hdr.dst_addr = sa->dst.ip.ip4;
sa->ipv4_spec.hdr.src_addr = sa->src.ip.ip4;
}
sa->pattern[2].type = RTE_FLOW_ITEM_TYPE_ESP;
sa->pattern[2].spec = &sa->esp_spec;
sa->pattern[2].mask = &rte_flow_item_esp_mask;
sa->esp_spec.hdr.spi = rte_cpu_to_be_32(sa->spi);
sa->pattern[3].type = RTE_FLOW_ITEM_TYPE_END;
sa->action[0].type = RTE_FLOW_ACTION_TYPE_SECURITY;
sa->action[0].conf = ips->security.ses;
sa->action[1].type = RTE_FLOW_ACTION_TYPE_END;
sa->attr.egress = (sa->direction ==
sa->attr.ingress = (sa->direction ==
if (sa->attr.ingress) {
uint8_t rss_key[40];
struct rte_eth_rss_conf rss_conf = {
.rss_key_len = 40,
};
struct rte_eth_dev_info dev_info;
uint16_t queue[RTE_MAX_QUEUES_PER_PORT];
struct rte_flow_action_rss action_rss;
unsigned int i;
unsigned int j;
/* Don't create flow if default flow is created */
if (flow_info_tbl[sa->portid].rx_def_flow)
return 0;
ret = rte_eth_dev_info_get(sa->portid, &dev_info);
if (ret != 0) {
RTE_LOG(ERR, IPSEC,
"Error during getting device (port %u) info: %s\n",
sa->portid, strerror(-ret));
return ret;
}
sa->action[2].type = RTE_FLOW_ACTION_TYPE_END;
/* Try RSS. */
sa->action[1].type = RTE_FLOW_ACTION_TYPE_RSS;
sa->action[1].conf = &action_rss;
&rss_conf);
if (ret != 0) {
RTE_LOG(ERR, IPSEC,
"rte_eth_dev_rss_hash_conf_get:ret=%d\n",
ret);
return -1;
}
for (i = 0, j = 0; i < dev_info.nb_rx_queues; ++i)
queue[j++] = i;
action_rss = (struct rte_flow_action_rss){
.types = rss_conf.rss_hf,
.key_len = rss_conf.rss_key_len,
.queue_num = j,
.key = rss_key,
.queue = queue,
};
ret = rte_flow_validate(sa->portid, &sa->attr,
sa->pattern, sa->action,
&err);
if (!ret)
goto flow_create;
/* Try Queue. */
sa->action[1].type = RTE_FLOW_ACTION_TYPE_QUEUE;
sa->action[1].conf =
.index = 0,
};
ret = rte_flow_validate(sa->portid, &sa->attr,
sa->pattern, sa->action,
&err);
/* Try End. */
sa->action[1].type = RTE_FLOW_ACTION_TYPE_END;
sa->action[1].conf = NULL;
ret = rte_flow_validate(sa->portid, &sa->attr,
sa->pattern, sa->action,
&err);
if (ret)
goto flow_create_failure;
} else if (sa->attr.egress &&
(ips->security.ol_flags &
sa->action[1].type =
sa->action[2].type =
}
flow_create:
sa->flow = rte_flow_create(sa->portid,
&sa->attr, sa->pattern, sa->action, &err);
if (sa->flow == NULL) {
flow_create_failure:
RTE_LOG(ERR, IPSEC,
"Failed to create ipsec flow msg: %s\n",
err.message);
return -1;
}
const struct rte_security_capability *sec_cap;
sec_ctx = (struct rte_security_ctx *)
if (sec_ctx == NULL) {
RTE_LOG(ERR, IPSEC,
"Ethernet device doesn't have security features registered\n");
return -1;
}
/* Set IPsec parameters in conf */
set_ipsec_conf(sa, &(sess_conf.ipsec));
/* Save SA as userdata for the security session. When
* the packet is received, this userdata will be
* retrieved using the metadata from the packet.
*
* The PMD is expected to set similar metadata for other
* operations, like rte_eth_event, which are tied to
* security session. In such cases, the userdata could
* be obtained to uniquely identify the security
* parameters denoted.
*/
sess_conf.userdata = (void *) sa;
ips->security.ses = rte_security_session_create(sec_ctx,
&sess_conf, skt_ctx->session_pool);
if (ips->security.ses == NULL) {
RTE_LOG(ERR, IPSEC,
"SEC Session init failed: err: %d\n", ret);
return -1;
}
sec_cap = rte_security_capabilities_get(sec_ctx);
if (sec_cap == NULL) {
RTE_LOG(ERR, IPSEC,
"No capabilities registered\n");
return -1;
}
/* iterate until ESP tunnel*/
while (sec_cap->action !=
if (sec_cap->action == ips->type &&
sec_cap->protocol ==
sec_cap->ipsec.mode ==
sess_conf.ipsec.mode &&
sec_cap->ipsec.direction == sa->direction)
break;
sec_cap++;
}
RTE_LOG(ERR, IPSEC,
"No suitable security capability found\n");
return -1;
}
ips->security.ol_flags = sec_cap->ol_flags;
ips->security.ctx = sec_ctx;
}
return 0;
}
int
create_ipsec_esp_flow(struct ipsec_sa *sa)
{
int ret = 0;
struct rte_flow_error err;
if (sa->direction == RTE_SECURITY_IPSEC_SA_DIR_EGRESS) {
RTE_LOG(ERR, IPSEC,
"No Flow director rule for Egress traffic\n");
return -1;
}
if (sa->flags == TRANSPORT) {
RTE_LOG(ERR, IPSEC,
"No Flow director rule for transport mode\n");
return -1;
}
sa->action[0].type = RTE_FLOW_ACTION_TYPE_QUEUE;
sa->pattern[0].type = RTE_FLOW_ITEM_TYPE_ETH;
sa->action[0].conf = &(struct rte_flow_action_queue) {
.index = sa->fdir_qid,
};
sa->attr.egress = 0;
sa->attr.ingress = 1;
if (IS_IP6(sa->flags)) {
sa->pattern[1].mask = &rte_flow_item_ipv6_mask;
sa->pattern[1].type = RTE_FLOW_ITEM_TYPE_IPV6;
sa->pattern[1].spec = &sa->ipv6_spec;
memcpy(sa->ipv6_spec.hdr.dst_addr,
sa->dst.ip.ip6.ip6_b, sizeof(sa->dst.ip.ip6.ip6_b));
memcpy(sa->ipv6_spec.hdr.src_addr,
sa->src.ip.ip6.ip6_b, sizeof(sa->src.ip.ip6.ip6_b));
sa->pattern[2].type = RTE_FLOW_ITEM_TYPE_ESP;
sa->pattern[2].spec = &sa->esp_spec;
sa->pattern[2].mask = &rte_flow_item_esp_mask;
sa->esp_spec.hdr.spi = rte_cpu_to_be_32(sa->spi);
sa->pattern[3].type = RTE_FLOW_ITEM_TYPE_END;
} else if (IS_IP4(sa->flags)) {
sa->pattern[1].mask = &rte_flow_item_ipv4_mask;
sa->pattern[1].type = RTE_FLOW_ITEM_TYPE_IPV4;
sa->pattern[1].spec = &sa->ipv4_spec;
sa->ipv4_spec.hdr.dst_addr = sa->dst.ip.ip4;
sa->ipv4_spec.hdr.src_addr = sa->src.ip.ip4;
sa->pattern[2].type = RTE_FLOW_ITEM_TYPE_ESP;
sa->pattern[2].spec = &sa->esp_spec;
sa->pattern[2].mask = &rte_flow_item_esp_mask;
sa->esp_spec.hdr.spi = rte_cpu_to_be_32(sa->spi);
sa->pattern[3].type = RTE_FLOW_ITEM_TYPE_END;
}
sa->action[1].type = RTE_FLOW_ACTION_TYPE_END;
ret = rte_flow_validate(sa->portid, &sa->attr, sa->pattern, sa->action,
&err);
if (ret < 0) {
RTE_LOG(ERR, IPSEC, "Flow validation failed %s\n", err.message);
return ret;
}
sa->flow = rte_flow_create(sa->portid, &sa->attr, sa->pattern,
sa->action, &err);
if (!sa->flow) {
RTE_LOG(ERR, IPSEC, "Flow creation failed %s\n", err.message);
return -1;
}
return 0;
}
/*
* queue crypto-ops into PMD queue.
*/
void
enqueue_cop_burst(struct cdev_qp *cqp)
{
uint32_t i, len, ret;
len = cqp->len;
ret = rte_cryptodev_enqueue_burst(cqp->id, cqp->qp, cqp->buf, len);
if (ret < len) {
RTE_LOG_DP(DEBUG, IPSEC, "Cryptodev %u queue %u:"
" enqueued %u crypto ops out of %u\n",
cqp->id, cqp->qp, ret, len);
/* drop packets that we fail to enqueue */
for (i = ret; i < len; i++)
rte_pktmbuf_free(cqp->buf[i]->sym->m_src);
}
cqp->in_flight += ret;
cqp->len = 0;
}
static inline void
enqueue_cop(struct cdev_qp *cqp, struct rte_crypto_op *cop)
{
cqp->buf[cqp->len++] = cop;
if (cqp->len == MAX_PKT_BURST)
enqueue_cop_burst(cqp);
}
static inline void
ipsec_enqueue(ipsec_xform_fn xform_func, struct ipsec_ctx *ipsec_ctx,
struct rte_mbuf *pkts[], void *sas[],
uint16_t nb_pkts)
{
int32_t ret = 0, i;
struct ipsec_mbuf_metadata *priv;
struct rte_crypto_sym_op *sym_cop;
struct ipsec_sa *sa;
struct rte_ipsec_session *ips;
for (i = 0; i < nb_pkts; i++) {
if (unlikely(sas[i] == NULL)) {
rte_pktmbuf_free(pkts[i]);
continue;
}
rte_prefetch0(sas[i]);
rte_prefetch0(pkts[i]);
priv = get_priv(pkts[i]);
sa = ipsec_mask_saptr(sas[i]);
priv->sa = sa;
ips = ipsec_get_primary_session(sa);
switch (ips->type) {
priv->cop.type = RTE_CRYPTO_OP_TYPE_SYMMETRIC;
rte_prefetch0(&priv->sym_cop);
if ((unlikely(ips->security.ses == NULL)) &&
create_lookaside_session(ipsec_ctx, sa, ips)) {
rte_pktmbuf_free(pkts[i]);
continue;
}
sym_cop = get_sym_cop(&priv->cop);
sym_cop->m_src = pkts[i];
ips->security.ses);
break;
RTE_LOG(ERR, IPSEC, "CPU crypto is not supported by the"
" legacy mode.");
rte_pktmbuf_free(pkts[i]);
continue;
priv->cop.type = RTE_CRYPTO_OP_TYPE_SYMMETRIC;
rte_prefetch0(&priv->sym_cop);
if ((unlikely(ips->crypto.ses == NULL)) &&
create_lookaside_session(ipsec_ctx, sa, ips)) {
rte_pktmbuf_free(pkts[i]);
continue;
}
ips->crypto.ses);
ret = xform_func(pkts[i], sa, &priv->cop);
if (unlikely(ret)) {
rte_pktmbuf_free(pkts[i]);
continue;
}
break;
RTE_ASSERT(ips->security.ses != NULL);
ipsec_ctx->ol_pkts[ipsec_ctx->ol_pkts_cnt++] = pkts[i];
if (ips->security.ol_flags &
ips->security.ctx, ips->security.ses,
pkts[i], NULL);
continue;
RTE_ASSERT(ips->security.ses != NULL);
priv->cop.type = RTE_CRYPTO_OP_TYPE_SYMMETRIC;
rte_prefetch0(&priv->sym_cop);
ips->security.ses);
ret = xform_func(pkts[i], sa, &priv->cop);
if (unlikely(ret)) {
rte_pktmbuf_free(pkts[i]);
continue;
}
ipsec_ctx->ol_pkts[ipsec_ctx->ol_pkts_cnt++] = pkts[i];
if (ips->security.ol_flags &
ips->security.ctx, ips->security.ses,
pkts[i], NULL);
continue;
}
RTE_ASSERT(sa->cdev_id_qp < ipsec_ctx->nb_qps);
enqueue_cop(&ipsec_ctx->tbl[sa->cdev_id_qp], &priv->cop);
}
}
static inline int32_t
ipsec_inline_dequeue(ipsec_xform_fn xform_func, struct ipsec_ctx *ipsec_ctx,
struct rte_mbuf *pkts[], uint16_t max_pkts)
{
int32_t nb_pkts, ret;
struct ipsec_mbuf_metadata *priv;
struct ipsec_sa *sa;
struct rte_mbuf *pkt;
nb_pkts = 0;
while (ipsec_ctx->ol_pkts_cnt > 0 && nb_pkts < max_pkts) {
pkt = ipsec_ctx->ol_pkts[--ipsec_ctx->ol_pkts_cnt];
priv = get_priv(pkt);
sa = priv->sa;
ret = xform_func(pkt, sa, &priv->cop);
if (unlikely(ret)) {
continue;
}
pkts[nb_pkts++] = pkt;
}
return nb_pkts;
}
static inline int
ipsec_dequeue(ipsec_xform_fn xform_func, struct ipsec_ctx *ipsec_ctx,
struct rte_mbuf *pkts[], uint16_t max_pkts)
{
int32_t nb_pkts = 0, ret = 0, i, j, nb_cops;
struct ipsec_mbuf_metadata *priv;
struct rte_crypto_op *cops[max_pkts];
struct ipsec_sa *sa;
struct rte_mbuf *pkt;
for (i = 0; i < ipsec_ctx->nb_qps && nb_pkts < max_pkts; i++) {
struct cdev_qp *cqp;
cqp = &ipsec_ctx->tbl[ipsec_ctx->last_qp++];
if (ipsec_ctx->last_qp == ipsec_ctx->nb_qps)
ipsec_ctx->last_qp %= ipsec_ctx->nb_qps;
if (cqp->in_flight == 0)
continue;
nb_cops = rte_cryptodev_dequeue_burst(cqp->id, cqp->qp,
cops, max_pkts - nb_pkts);
cqp->in_flight -= nb_cops;
for (j = 0; j < nb_cops; j++) {
pkt = cops[j]->sym->m_src;
priv = get_priv(pkt);
sa = priv->sa;
RTE_ASSERT(sa != NULL);
if (ipsec_get_action_type(sa) ==
ret = xform_func(pkt, sa, cops[j]);
if (unlikely(ret)) {
continue;
}
} else if (ipsec_get_action_type(sa) ==
if (cops[j]->status) {
continue;
}
}
pkts[nb_pkts++] = pkt;
}
}
/* return packets */
return nb_pkts;
}
uint16_t
ipsec_inbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
uint16_t nb_pkts, uint16_t len)
{
void *sas[nb_pkts];
inbound_sa_lookup(ctx->sa_ctx, pkts, sas, nb_pkts);
ipsec_enqueue(esp_inbound, ctx, pkts, sas, nb_pkts);
return ipsec_inline_dequeue(esp_inbound_post, ctx, pkts, len);
}
uint16_t
ipsec_inbound_cqp_dequeue(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
uint16_t len)
{
return ipsec_dequeue(esp_inbound_post, ctx, pkts, len);
}
uint16_t
ipsec_outbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
uint32_t sa_idx[], uint16_t nb_pkts, uint16_t len)
{
void *sas[nb_pkts];
outbound_sa_lookup(ctx->sa_ctx, sa_idx, sas, nb_pkts);
ipsec_enqueue(esp_outbound, ctx, pkts, sas, nb_pkts);
return ipsec_inline_dequeue(esp_outbound_post, ctx, pkts, len);
}
uint16_t
ipsec_outbound_cqp_dequeue(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
uint16_t len)
{
return ipsec_dequeue(esp_outbound_post, ctx, pkts, len);
}