summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authordimitri staessens <dimitri.staessens@intec.ugent.be>2016-10-31 18:35:17 +0100
committerdimitri staessens <dimitri.staessens@intec.ugent.be>2016-10-31 18:43:04 +0100
commit86b6d0ca1b871e629a95c9cf0ddceeeb7e1bd345 (patch)
tree02f048b2b148ad9426194967cbc332a93742ac52
parentadd7d5f1e59c661b431ba86ba80d62badc14e762 (diff)
downloadouroboros-86b6d0ca1b871e629a95c9cf0ddceeeb7e1bd345.tar.gz
ouroboros-86b6d0ca1b871e629a95c9cf0ddceeeb7e1bd345.zip
lib, ipcpd: Improve flow allocation
All calls for opening rbuffs are now concentrated on the dev side. This allows some simplifications in the np1 calls. The ipcp_fini call will not destroy the mutex associated with the state, since the final state needs to be checked before shutting down an IPCP.
-rw-r--r--include/ouroboros/np1_flow.h3
-rw-r--r--src/ipcpd/ipcp.c16
-rw-r--r--src/ipcpd/local/main.c20
-rw-r--r--src/ipcpd/shim-eth-llc/main.c30
-rw-r--r--src/ipcpd/shim-udp/main.c58
-rw-r--r--src/lib/dev.c185
6 files changed, 169 insertions, 143 deletions
diff --git a/include/ouroboros/np1_flow.h b/include/ouroboros/np1_flow.h
index 5b063213..4e7e6070 100644
--- a/include/ouroboros/np1_flow.h
+++ b/include/ouroboros/np1_flow.h
@@ -28,8 +28,7 @@
int np1_flow_alloc(pid_t n_api,
int port_id);
-int np1_flow_resp(pid_t n_api,
- int port_id);
+int np1_flow_resp(int port_id);
int np1_flow_dealloc(int port_id);
diff --git a/src/ipcpd/ipcp.c b/src/ipcpd/ipcp.c
index b7eff3fb..4dff86f4 100644
--- a/src/ipcpd/ipcp.c
+++ b/src/ipcpd/ipcp.c
@@ -114,7 +114,6 @@ void ipcp_fini()
ipcp_data_destroy(ipcpi.data);
pthread_cond_destroy(&ipcpi.state_cond);
- pthread_mutex_destroy(&ipcpi.state_mtx);
pthread_rwlock_destroy(&ipcpi.state_lock);
}
@@ -349,7 +348,7 @@ void * ipcp_main_loop(void * o)
}
fd = np1_flow_alloc(msg->api, msg->port_id);
if (fd < 0) {
- LOG_ERR("Could not get fd for port_id. %d",
+ LOG_ERR("Failed allocating fd on port_id %d.",
msg->port_id);
ret_msg.has_result = true;
ret_msg.result = -1;
@@ -362,11 +361,6 @@ void * ipcp_main_loop(void * o)
msg->dst_name,
msg->src_ae_name,
msg->qos_cube);
- if (ret_msg.result < 0) {
- LOG_DBG("Deallocate failed on port_id %d.",
- msg->port_id);
- flow_dealloc(fd);
- }
break;
case IPCP_MSG_CODE__IPCP_FLOW_ALLOC_RESP:
if (ipcpi.ops->ipcp_flow_alloc_resp == NULL) {
@@ -375,10 +369,10 @@ void * ipcp_main_loop(void * o)
}
if (!msg->response) {
- fd = np1_flow_resp(msg->api, msg->port_id);
+ fd = np1_flow_resp(msg->port_id);
if (fd < 0) {
- LOG_ERR("Could not get fd for port_id %d.",
- msg->port_id);
+ LOG_WARN("Port_id %d is not known.",
+ msg->port_id);
ret_msg.has_result = true;
ret_msg.result = -1;
break;
@@ -397,7 +391,7 @@ void * ipcp_main_loop(void * o)
fd = np1_flow_dealloc(msg->port_id);
if (fd < 0) {
- LOG_ERR("Could not deallocate port_id %d.",
+ LOG_WARN("Could not deallocate port_id %d.",
msg->port_id);
ret_msg.has_result = true;
ret_msg.result = -1;
diff --git a/src/ipcpd/local/main.c b/src/ipcpd/local/main.c
index f95eeae4..f0c85084 100644
--- a/src/ipcpd/local/main.c
+++ b/src/ipcpd/local/main.c
@@ -88,14 +88,10 @@ static void * ipcp_local_sdu_loop(void * o)
(void) o;
- while (true) {
+ while (flow_event_wait(local_data.flows, local_data.fq, &timeout)) {
int fd;
ssize_t idx;
- if (flow_event_wait(local_data.flows, local_data.fq, &timeout)
- == -ETIMEDOUT)
- continue;
-
pthread_rwlock_rdlock(&ipcpi.state_lock);
if (ipcp_get_state() != IPCP_ENROLLED) {
@@ -233,7 +229,8 @@ static int ipcp_local_flow_alloc(int fd,
LOG_DBG("Allocating flow to %s on fd %d.", dst_name, fd);
- assert(dst_name || src_ae_name);
+ assert(dst_name);
+ assert(src_ae_name);
pthread_rwlock_rdlock(&ipcpi.state_lock);
@@ -298,17 +295,24 @@ static int ipcp_local_flow_dealloc(int fd)
ipcp_flow_fini(fd);
pthread_rwlock_rdlock(&ipcpi.state_lock);
+
+ if (ipcp_get_state() != IPCP_ENROLLED) {
+ pthread_rwlock_unlock(&ipcpi.state_lock);
+ LOG_DBG("Won't register with non-enrolled IPCP.");
+ return -1; /* -ENOTENROLLED */
+ }
+
pthread_rwlock_wrlock(&local_data.lock);
flow_set_del(local_data.flows, fd);
local_data.in_out[fd] = -1;
+ flow_dealloc(fd);
+
pthread_rwlock_unlock(&local_data.lock);
pthread_rwlock_unlock(&ipcpi.state_lock);
- flow_dealloc(fd);
-
LOG_INFO("Flow with fd %d deallocated.", fd);
return 0;
diff --git a/src/ipcpd/shim-eth-llc/main.c b/src/ipcpd/shim-eth-llc/main.c
index 0d4c3903..3f3c0e1e 100644
--- a/src/ipcpd/shim-eth-llc/main.c
+++ b/src/ipcpd/shim-eth-llc/main.c
@@ -624,18 +624,22 @@ static void * eth_llc_ipcp_sdu_writer(void * o)
(void) o;
- while (true) {
- if (flow_event_wait(eth_llc_data.np1_flows,
- eth_llc_data.fq,
- &timeout) == -ETIMEDOUT)
- continue;
+ while (flow_event_wait(eth_llc_data.np1_flows,
+ eth_llc_data.fq,
+ &timeout)) {
+ pthread_rwlock_rdlock(&ipcpi.state_lock);
+
+ if (ipcp_get_state() != IPCP_ENROLLED) {
+ pthread_rwlock_unlock(&ipcpi.state_lock);
+ return (void *) -1; /* -ENOTENROLLED */
+ }
while ((fd = fqueue_next(eth_llc_data.fq)) >= 0) {
if (ipcp_flow_read(fd, &sdb)) {
LOG_ERR("Bad read from fd %d.", fd);
continue;
}
- pthread_rwlock_rdlock(&ipcpi.state_lock);
+
pthread_rwlock_rdlock(&eth_llc_data.flows_lock);
ssap = reverse_bits(eth_llc_data.fd_to_ef[fd].sap);
@@ -645,7 +649,6 @@ static void * eth_llc_ipcp_sdu_writer(void * o)
MAC_SIZE);
pthread_rwlock_unlock(&eth_llc_data.flows_lock);
- pthread_rwlock_unlock(&ipcpi.state_lock);
eth_llc_ipcp_send_frame(r_addr, dsap, ssap,
shm_du_buff_head(sdb),
@@ -653,6 +656,8 @@ static void * eth_llc_ipcp_sdu_writer(void * o)
- shm_du_buff_head(sdb));
ipcp_flow_del(sdb);
}
+
+ pthread_rwlock_unlock(&ipcpi.state_lock);
}
return (void *) 1;
@@ -1045,6 +1050,13 @@ static int eth_llc_ipcp_flow_dealloc(int fd)
ipcp_flow_fini(fd);
pthread_rwlock_rdlock(&ipcpi.state_lock);
+
+ if (ipcp_get_state() != IPCP_ENROLLED) {
+ pthread_rwlock_unlock(&ipcpi.state_lock);
+ LOG_DBG("Won't register with non-enrolled IPCP.");
+ return -1; /* -ENOTENROLLED */
+ }
+
pthread_rwlock_wrlock(&eth_llc_data.flows_lock);
flow_set_del(eth_llc_data.np1_flows, fd);
@@ -1058,11 +1070,11 @@ static int eth_llc_ipcp_flow_dealloc(int fd)
eth_llc_data.ef_to_fd[sap] = -1;
+ flow_dealloc(fd);
+
pthread_rwlock_unlock(&eth_llc_data.flows_lock);
pthread_rwlock_unlock(&ipcpi.state_lock);
- flow_dealloc(fd);
-
LOG_DBG("Flow with fd %d deallocated.", fd);
return 0;
diff --git a/src/ipcpd/shim-udp/main.c b/src/ipcpd/shim-udp/main.c
index 503dbd0b..eff0bd94 100644
--- a/src/ipcpd/shim-udp/main.c
+++ b/src/ipcpd/shim-udp/main.c
@@ -452,7 +452,6 @@ static void * ipcp_udp_sdu_reader(void * o)
continue;
flags = fcntl(skfd, F_GETFL, 0);
fcntl(skfd, F_SETFL, flags | O_NONBLOCK);
- fd = udp_data.uf_to_fd[skfd];
n = sizeof(r_saddr);
if ((n = recvfrom(skfd,
&buf,
@@ -462,7 +461,14 @@ static void * ipcp_udp_sdu_reader(void * o)
(unsigned *) &n)) <= 0)
continue;
+ pthread_rwlock_rdlock(&ipcpi.state_lock);
+ pthread_rwlock_rdlock(&udp_data.flows_lock);
+
+ fd = udp_data.uf_to_fd[skfd];
flow_write(fd, buf, n);
+
+ pthread_rwlock_unlock(&udp_data.flows_lock);
+ pthread_rwlock_unlock(&ipcpi.state_lock);
}
}
@@ -477,11 +483,16 @@ static void * ipcp_udp_sdu_loop(void * o)
(void) o;
- while (true) {
- if (flow_event_wait(udp_data.np1_flows,
- udp_data.fq,
- &timeout) == -ETIMEDOUT)
- continue;
+ while (flow_event_wait(udp_data.np1_flows, udp_data.fq, &timeout)) {
+ pthread_rwlock_rdlock(&ipcpi.state_lock);
+
+ if (ipcp_get_state() != IPCP_ENROLLED) {
+ pthread_rwlock_unlock(&ipcpi.state_lock);
+ return (void *) -1; /* -ENOTENROLLED */
+ }
+
+
+ pthread_rwlock_rdlock(&udp_data.flows_lock);
while ((fd = fqueue_next(udp_data.fq)) >= 0) {
if (ipcp_flow_read(fd, &sdb)) {
@@ -489,19 +500,18 @@ static void * ipcp_udp_sdu_loop(void * o)
continue;
}
- pthread_rwlock_rdlock(&ipcpi.state_lock);
- pthread_rwlock_rdlock(&udp_data.flows_lock);
-
if (send(udp_data.fd_to_uf[fd].skfd,
shm_du_buff_head(sdb),
shm_du_buff_tail(sdb) - shm_du_buff_head(sdb),
0) < 0)
LOG_ERR("Failed to send SDU.");
- pthread_rwlock_unlock(&udp_data.flows_lock);
- pthread_rwlock_unlock(&ipcpi.state_lock);
+ ipcp_flow_del(sdb);
+ }
- ipcp_flow_del(sdb); }
+
+ pthread_rwlock_unlock(&udp_data.flows_lock);
+ pthread_rwlock_unlock(&ipcpi.state_lock);
}
return (void *) 1;
@@ -948,8 +958,8 @@ static int ipcp_udp_flow_alloc(int fd,
LOG_DBG("Allocating flow to %s.", dst_name);
- if (dst_name == NULL || src_ae_name == NULL)
- return -1;
+ assert(dst_name);
+ assert(src_ae_name);
if (strlen(dst_name) > 255
|| strlen(src_ae_name) > 255) {
@@ -1101,32 +1111,36 @@ static int ipcp_udp_flow_dealloc(int fd)
ipcp_flow_fini(fd);
- flow_set_del(udp_data.np1_flows, fd);
-
pthread_rwlock_rdlock(&ipcpi.state_lock);
+
+ if (ipcp_get_state() != IPCP_ENROLLED) {
+ pthread_rwlock_unlock(&ipcpi.state_lock);
+ LOG_DBG("Won't register with non-enrolled IPCP.");
+ return -1; /* -ENOTENROLLED */
+ }
+
pthread_rwlock_wrlock(&udp_data.flows_lock);
+ flow_set_del(udp_data.np1_flows, fd);
+
skfd = udp_data.fd_to_uf[fd].skfd;
udp_data.uf_to_fd[skfd] = -1;
udp_data.fd_to_uf[fd].udp = -1;
udp_data.fd_to_uf[fd].skfd = -1;
+ close(skfd);
+
pthread_rwlock_unlock(&udp_data.flows_lock);
pthread_rwlock_rdlock(&udp_data.flows_lock);
clr_fd(skfd);
- pthread_rwlock_unlock(&udp_data.flows_lock);
- pthread_rwlock_wrlock(&udp_data.flows_lock);
-
- close(skfd);
+ flow_dealloc(fd);
pthread_rwlock_unlock(&udp_data.flows_lock);
pthread_rwlock_unlock(&ipcpi.state_lock);
- flow_dealloc(fd);
-
LOG_DBG("Flow with fd %d deallocated.", fd);
return 0;
diff --git a/src/lib/dev.c b/src/lib/dev.c
index 4b97428e..fc8739a2 100644
--- a/src/lib/dev.c
+++ b/src/lib/dev.c
@@ -388,6 +388,14 @@ int flow_accept(char ** ae_name, struct qos_spec * qos)
return -1;
}
+ ai.flows[fd].tx_rb = shm_rbuff_open(recv_msg->api, recv_msg->port_id);
+ if (ai.flows[fd].tx_rb == NULL) {
+ reset_flow(fd);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
ai.flows[fd].set = shm_flow_set_open(recv_msg->api);
if (ai.flows[fd].set == NULL) {
reset_flow(fd);
@@ -398,7 +406,6 @@ int flow_accept(char ** ae_name, struct qos_spec * qos)
return -1;
}
-
if (ae_name != NULL) {
*ae_name = strdup(recv_msg->ae_name);
if (*ae_name == NULL) {
@@ -452,40 +459,34 @@ int flow_alloc_resp(int fd, int response)
msg.port_id = ai.flows[fd].port_id;
pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
msg.has_response = true;
msg.response = response;
recv_msg = send_recv_irm_msg(&msg);
- if (recv_msg == NULL) {
- pthread_rwlock_unlock(&ai.data_lock);
+ if (recv_msg == NULL)
return -1;
- }
if (!recv_msg->has_result) {
- pthread_rwlock_unlock(&ai.data_lock);
irm_msg__free_unpacked(recv_msg, NULL);
return -1;
}
ret = recv_msg->result;
- pthread_rwlock_wrlock(&ai.flows_lock);
+ irm_msg__free_unpacked(recv_msg, NULL);
+
+ if (response) {
+ pthread_rwlock_rdlock(&ai.data_lock);
+ pthread_rwlock_wrlock(&ai.flows_lock);
- ai.flows[fd].tx_rb = shm_rbuff_open(ai.flows[fd].api,
- ai.flows[fd].port_id);
- if (ai.flows[fd].tx_rb == NULL) {
reset_flow(fd);
+
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
- return -1;
}
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
-
- irm_msg__free_unpacked(recv_msg, NULL);
-
return ret;
}
@@ -535,9 +536,6 @@ int flow_alloc(char * dst_name, char * src_ae_name, struct qos_spec * qos)
return -1;
}
- ai.flows[fd].port_id = recv_msg->port_id;
- ai.flows[fd].oflags = FLOW_O_DEFAULT;
- ai.flows[fd].api = recv_msg->api;
ai.flows[fd].rx_rb = shm_rbuff_open(ai.api, recv_msg->port_id);
if (ai.flows[fd].rx_rb == NULL) {
reset_flow(fd);
@@ -548,6 +546,25 @@ int flow_alloc(char * dst_name, char * src_ae_name, struct qos_spec * qos)
return -1;
}
+ ai.flows[fd].tx_rb = shm_rbuff_open(recv_msg->api, recv_msg->port_id);
+ if (ai.flows[fd].tx_rb == NULL) {
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
+ ai.flows[fd].set = shm_flow_set_open(recv_msg->api);
+ if (ai.flows[fd].set == NULL) {
+ reset_flow(fd);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
+ ai.flows[fd].port_id = recv_msg->port_id;
+ ai.flows[fd].oflags = FLOW_O_DEFAULT;
+ ai.flows[fd].api = recv_msg->api;
+
ai.ports[recv_msg->port_id].fd = fd;
ai.ports[recv_msg->port_id].state = PORT_ID_ASSIGNED;
@@ -582,22 +599,6 @@ int flow_alloc_res(int fd)
msg.port_id = ai.flows[fd].port_id;
- ai.flows[fd].tx_rb = shm_rbuff_open(ai.flows[fd].api,
- ai.flows[fd].port_id);
- if (ai.flows[fd].tx_rb == NULL) {
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
-
- ai.flows[fd].set = shm_flow_set_open(ai.flows[fd].api);
- if (ai.flows[fd].set == NULL) {
- reset_flow(fd);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
-
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
@@ -764,10 +765,14 @@ ssize_t flow_write(int fd, void * buf, size_t count)
}
} else { /* blocking */
struct shm_rdrbuff * rdrb = ai.rdrb;
- pid_t api = ai.flows[fd].api;
+ struct shm_rbuff * tx_rb = ai.flows[fd].tx_rb;
+ pid_t api = ai.flows[fd].api;
+
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
+ assert(tx_rb);
+
idx = shm_rdrbuff_write_b(rdrb,
api,
DU_BUFF_HEADSPACE,
@@ -775,15 +780,13 @@ ssize_t flow_write(int fd, void * buf, size_t count)
buf,
count);
- pthread_rwlock_rdlock(&ai.data_lock);
- pthread_rwlock_rdlock(&ai.flows_lock);
-
- if (shm_rbuff_write(ai.flows[fd].tx_rb, idx) < 0) {
- shm_rdrbuff_remove(ai.rdrb, idx);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
+ if (shm_rbuff_write(tx_rb, idx) < 0) {
+ shm_rdrbuff_remove(rdrb, idx);
return -ENOTALLOC;
}
+
+ pthread_rwlock_rdlock(&ai.data_lock);
+ pthread_rwlock_rdlock(&ai.flows_lock);
}
shm_flow_set_notify(ai.flows[fd].set, ai.flows[fd].port_id);
@@ -993,7 +996,7 @@ int flow_event_wait(struct flow_set * set,
{
ssize_t ret;
- if (set == NULL)
+ if (set == NULL || fq == NULL)
return -EINVAL;
if (fq->fqsize > 0)
@@ -1038,6 +1041,22 @@ int np1_flow_alloc(pid_t n_api, int port_id)
return -1;
}
+ ai.flows[fd].tx_rb = shm_rbuff_open(n_api, port_id);
+ if (ai.flows[fd].tx_rb == NULL) {
+ reset_flow(fd);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
+ ai.flows[fd].set = shm_flow_set_open(n_api);
+ if (ai.flows[fd].set == NULL) {
+ reset_flow(fd);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
ai.flows[fd].port_id = port_id;
ai.flows[fd].oflags = FLOW_O_DEFAULT;
ai.flows[fd].api = n_api;
@@ -1066,8 +1085,7 @@ int np1_flow_dealloc(int port_id)
return fd;
}
-
-int np1_flow_resp(pid_t n_api, int port_id)
+int np1_flow_resp(int port_id)
{
int fd;
@@ -1077,28 +1095,6 @@ int np1_flow_resp(pid_t n_api, int port_id)
pthread_rwlock_wrlock(&ai.flows_lock);
fd = ai.ports[port_id].fd;
- if (fd < 0) {
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return fd;
- }
-
- ai.flows[fd].tx_rb = shm_rbuff_open(n_api, port_id);
- if (ai.flows[fd].tx_rb == NULL) {
- reset_flow(fd);
- port_destroy(&ai.ports[port_id]);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
-
- ai.flows[fd].set = shm_flow_set_open(n_api);
- if (ai.flows[fd].set == NULL) {
- reset_flow(fd);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
@@ -1166,21 +1162,47 @@ int ipcp_flow_req_arr(pid_t api, char * dst_name, char * src_ae_name)
if (recv_msg == NULL)
return -1;
- if (!recv_msg->has_port_id) {
+ if (!recv_msg->has_port_id || !recv_msg->has_api) {
irm_msg__free_unpacked(recv_msg, NULL);
return -1;
}
+ if (recv_msg->has_result && recv_msg->result) {
+ irm_msg__free_unpacked(recv_msg, NULL);
+ return -1;
+ }
+
port_id = recv_msg->port_id;
- irm_msg__free_unpacked(recv_msg, NULL);
- if (port_id < 0)
+ if (port_id < 0) {
+ irm_msg__free_unpacked(recv_msg, NULL);
return -1;
+ }
pthread_rwlock_rdlock(&ai.data_lock);
pthread_rwlock_wrlock(&ai.flows_lock);
ai.flows[fd].rx_rb = shm_rbuff_open(ai.api, port_id);
if (ai.flows[fd].rx_rb == NULL) {
+ irm_msg__free_unpacked(recv_msg, NULL);
+ reset_flow(fd);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
+ ai.flows[fd].tx_rb = shm_rbuff_open(recv_msg->api, port_id);
+ if (ai.flows[fd].tx_rb == NULL) {
+ irm_msg__free_unpacked(recv_msg, NULL);
+ reset_flow(fd);
+ port_destroy(&ai.ports[port_id]);
+ pthread_rwlock_unlock(&ai.flows_lock);
+ pthread_rwlock_unlock(&ai.data_lock);
+ return -1;
+ }
+
+ ai.flows[fd].set = shm_flow_set_open(recv_msg->api);
+ if (ai.flows[fd].set == NULL) {
+ irm_msg__free_unpacked(recv_msg, NULL);
reset_flow(fd);
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
@@ -1196,6 +1218,8 @@ int ipcp_flow_req_arr(pid_t api, char * dst_name, char * src_ae_name)
pthread_rwlock_unlock(&ai.flows_lock);
pthread_rwlock_unlock(&ai.data_lock);
+ irm_msg__free_unpacked(recv_msg, NULL);
+
return fd;
}
@@ -1228,27 +1252,6 @@ int ipcp_flow_alloc_reply(int fd, int response)
ret = recv_msg->result;
- pthread_rwlock_wrlock(&ai.flows_lock);
-
- ai.flows[fd].tx_rb = shm_rbuff_open(ai.flows[fd].api,
- ai.flows[fd].port_id);
- if (ai.flows[fd].tx_rb == NULL) {
- reset_flow(fd);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
-
- ai.flows[fd].set = shm_flow_set_open(ai.flows[fd].api);
- if (ai.flows[fd].set == NULL) {
- reset_flow(fd);
- pthread_rwlock_unlock(&ai.flows_lock);
- pthread_rwlock_unlock(&ai.data_lock);
- return -1;
- }
-
- pthread_rwlock_unlock(&ai.flows_lock);
-
irm_msg__free_unpacked(recv_msg, NULL);
return ret;