#include <fcntl.h>
-#include <event2/event.h>
#include <talloc.h>
#include <errno.h>
#include <unistd.h>
#include <gsscon.h>
#include <tr_rp.h>
-#include <tr_event.h>
-#include <tr_debug.h>
+#include <trust_router/tr_name.h>
#include <trp_internal.h>
+#include <trp_rtable.h>
+#include <tr_debug.h>
+static int trps_destructor(void *object)
+{
+ TRPS_INSTANCE *trps=talloc_get_type_abort(object, TRPS_INSTANCE);
+ if (trps->rtable!=NULL)
+ trp_rtable_free(trps->rtable);
+ return 0;
+}
+
TRPS_INSTANCE *trps_new (TALLOC_CTX *mem_ctx)
{
TRPS_INSTANCE *trps=talloc(mem_ctx, TRPS_INSTANCE);
trps->port=0;
trps->cookie=NULL;
trps->conn=NULL;
+ trps->trpc=NULL;
trps->mq=tr_mq_new(trps);
if (trps->mq==NULL) {
/* failed to allocate mq */
talloc_free(trps);
trps=NULL;
+ } else {
+ trps->rtable=trp_rtable_new();
+ if (trps->rtable==NULL) {
+ /* failed to allocate rtable */
+ talloc_free(trps);
+ trps=NULL;
+ } else
+ talloc_set_destructor((void *)trps, trps_destructor);
}
}
return trps;
talloc_steal(trps, new);
}
-int trps_send_msg (TRPS_INSTANCE *trps,
- int conn,
- gss_ctx_id_t gssctx,
- const char *msg_content)
+/* ok to call more than once; guarantees connection no longer in the list.
+ * Caller is responsible for freeing the removed element afterwards. */
+void trps_remove_connection(TRPS_INSTANCE *trps, TRP_CONNECTION *remove)
{
- int err=0;
- int rc=0;
+ trps->conn=trp_connection_remove(trps->conn, remove);
+}
- /* Send the request over the connection */
- if (err = gsscon_write_encrypted_token (conn,
- gssctx,
- msg_content,
- strlen(msg_content))) {
- tr_err( "trps_send_msg: Error sending message over connection.\n");
- rc = -1;
- }
+void trps_add_trpc(TRPS_INSTANCE *trps, TRPC_INSTANCE *trpc)
+{
+ if (trps->trpc==NULL)
+ trps->trpc=trpc;
+ else
+ trpc_append(trps->trpc, trpc);
+
+ talloc_steal(trps, trpc);
+}
+
+/* ok to call more than once; guarantees trpc no longer in the list.
+ * Caller is responsible for freeing the removed element afterwards. */
+void trps_remove_trpc(TRPS_INSTANCE *trps, TRPC_INSTANCE *remove)
+{
+ trps->trpc=trpc_remove(trps->trpc, remove);
+}
+TRP_RC trps_send_msg (TRPS_INSTANCE *trps, void *peer, const char *msg)
+{
+ TALLOC_CTX *tmp_ctx=talloc_new(NULL);
+ TR_MQ_MSG *mq_msg=NULL;
+ char *msg_dup=NULL;
+ TRP_RC rc=TRP_ERROR;
+
+ /* Currently ignore peer and just send to an open connection.
+ * In reality, need to identify the correct peer and send via that
+ * one. */
+ if (trps->trpc != NULL) {
+ if (trpc_get_status(trps->trpc)!=TRP_CONNECTION_UP)
+ tr_debug("trps_send_msg: skipping message sent while TRPC connection not up.");
+ else {
+ mq_msg=tr_mq_msg_new(tmp_ctx, "trpc_send");
+ msg_dup=talloc_strdup(mq_msg, msg); /* get local copy in mq_msg context */
+ tr_mq_msg_set_payload(mq_msg, msg_dup, NULL); /* no need for a free() func */
+ trpc_mq_append(trps->trpc, mq_msg);
+ rc=TRP_SUCCESS;
+ }
+ }
+ talloc_free(tmp_ctx);
return rc;
}
static int trps_listen (TRPS_INSTANCE *trps, int port)
{
- int rc = 0;
- int conn = -1;
- int optval = 1;
+ int rc = 0;
+ int conn = -1;
+ int optval = 1;
- union {
- struct sockaddr_storage storage;
- struct sockaddr_in in4;
- } addr;
+ union {
+ struct sockaddr_storage storage;
+ struct sockaddr_in in4;
+ } addr;
- struct sockaddr_in *saddr = (struct sockaddr_in *) &addr.in4;
+ struct sockaddr_in *saddr = (struct sockaddr_in *) &addr.in4;
- saddr->sin_port = htons (port);
- saddr->sin_family = AF_INET;
- saddr->sin_addr.s_addr = INADDR_ANY;
+ saddr->sin_port = htons (port);
+ saddr->sin_family = AF_INET;
+ saddr->sin_addr.s_addr = INADDR_ANY;
- if (0 > (conn = socket (AF_INET, SOCK_STREAM, 0)))
- return conn;
+ if (0 > (conn = socket (AF_INET, SOCK_STREAM, 0)))
+ return conn;
- setsockopt(conn, SOL_SOCKET, SO_REUSEADDR, &optval, sizeof(optval));
+ setsockopt(conn, SOL_SOCKET, SO_REUSEADDR, &optval, sizeof(optval));
- if (0 > (rc = bind (conn, (struct sockaddr *) saddr, sizeof(struct sockaddr_in))))
- return rc;
+ if (0 > (rc = bind (conn, (struct sockaddr *) saddr, sizeof(struct sockaddr_in))))
+ return rc;
- if (0 > (rc = listen(conn, 512)))
- return rc;
+ if (0 > (rc = listen(conn, 512)))
+ return rc;
- tr_debug("trps_listen: TRP Server listening on port %d", port);
- return conn;
+ tr_debug("trps_listen: TRP Server listening on port %d", port);
+ return conn;
}
+#if 0 /* remove this if I forget to do so */
/* returns EACCES if authorization is denied */
int trps_auth_cb(gss_name_t clientName, gss_buffer_t displayName, void *data)
{
- TRPS_INSTANCE *inst = (TRPS_INSTANCE *)data;
+ TRPS_INSTANCE *trps = (TRPS_INSTANCE *)data;
int result=0;
- if (0!=inst->auth_handler(clientName, displayName, inst->cookie)) {
+ if (0!=trps->auth_handler(clientName, displayName, trps->cookie)) {
tr_debug("trps_auth_cb: client '%.*s' denied authorization.", displayName->length, displayName->value);
result=EACCES; /* denied */
}
return result;
}
+#endif
-#if 0
-static int trps_read_message (TRPS_INSTANCE *trps, int conn, gss_ctx_id_t *gssctx, char **msg)
+/* get the currently selected route if available */
+TRP_RENTRY *trps_get_route(TRPS_INSTANCE *trps, TR_NAME *comm, TR_NAME *realm, TR_NAME *peer)
{
- int err;
- char *buf;
- size_t buflen = 0;
+ return trp_rtable_get_entry(trps->rtable, comm, realm, peer);
+}
- if (err = gsscon_read_encrypted_token(conn, *gssctx, &buf, &buflen)) {
+TRP_RENTRY *trps_get_selected_route(TRPS_INSTANCE *trps, TR_NAME *comm, TR_NAME *realm)
+{
+ return trp_rtable_get_selected_entry(trps->rtable, comm, realm);
+}
+
+/* copy the result if you want to keep it */
+TR_NAME *trps_get_next_hop(TRPS_INSTANCE *trps, TR_NAME *comm, TR_NAME *realm)
+{
+ TRP_RENTRY *route=trps_get_selected_route(trps, comm, realm);
+ if (route==NULL)
+ return NULL;
+
+ return trp_rentry_get_next_hop(route);
+}
+
+
+/* mark a route as retracted */
+static void trps_retract_route(TRPS_INSTANCE *trps, TRP_RENTRY *entry)
+{
+ trp_rentry_set_metric(entry, TRP_METRIC_INFINITY);
+}
+
+/* is this route retracted? */
+static int trps_route_retracted(TRPS_INSTANCE *trps, TRP_RENTRY *entry)
+{
+ return (trp_rentry_get_metric(entry)==TRP_METRIC_INFINITY);
+}
+
+static TRP_RC trps_read_message(TRPS_INSTANCE *trps, TRP_CONNECTION *conn, TR_MSG **msg)
+{
+ int err=0;
+ char *buf=NULL;
+ size_t buflen = 0;
+ TR_NAME *peer=NULL;
+
+ tr_debug("trps_read_message: started");
+ if (err = gsscon_read_encrypted_token(trp_connection_get_fd(conn),
+ *(trp_connection_get_gssctx(conn)),
+ &buf,
+ &buflen)) {
+ tr_debug("trps_read_message: error");
if (buf)
free(buf);
- return -1;
+ return TRP_ERROR;
}
- tr_debug("trps_read_request(): Request Received, %u bytes.", (unsigned) buflen);
- tr_debug("trps_read_request(): %.*s", buflen, buf);
+ tr_debug("trps_read_message(): Request Received, %u bytes.", (unsigned) buflen);
+ tr_debug("trps_read_message(): %.*s", buflen, buf);
- *msg=talloc_strndup(NULL, buf, buflen); /* no context owns this! */
+ *msg=tr_msg_decode(buf, buflen);
free(buf);
- return buflen;
+ if (*msg==NULL)
+ return TRP_NOPARSE;
+
+ peer=trp_connection_get_peer(conn);
+ /* verify we received a message we support, otherwise drop it now */
+ switch (tr_msg_get_msg_type(*msg)) {
+ case TRP_UPDATE:
+ trp_upd_set_peer(tr_msg_get_trp_upd(*msg), tr_dup_name(peer));
+ break;
+
+ case TRP_REQUEST:
+ trp_req_set_peer(tr_msg_get_trp_req(*msg), tr_dup_name(peer));
+ break;
+
+ default:
+ tr_debug("trps_read_message: received unsupported message from %.*s", peer->len, peer->buf);
+ tr_msg_free_decoded(*msg);
+ *msg=NULL;
+ return TRP_UNSUPPORTED;
+ }
+
+ return TRP_SUCCESS;
}
-#endif
int trps_get_listener(TRPS_INSTANCE *trps,
- TRP_REQ_FUNC req_handler,
+ TRPS_MSG_FUNC msg_handler,
TRP_AUTH_FUNC auth_handler,
const char *hostname,
unsigned int port,
if (listen > 0) {
/* store the caller's request handler & cookie */
- trps->req_handler = req_handler;
+ trps->msg_handler = msg_handler;
trps->auth_handler = auth_handler;
trps->hostname = talloc_strdup(trps, hostname);
trps->port = port;
return listen;
}
-/* old cruft */
-#if 0
-static gss_ctx_id_t trps_establish_gss_context (TRPS_INSTANCE *trps, int conn)
+void trps_handle_connection(TRPS_INSTANCE *trps, TRP_CONNECTION *conn)
{
TALLOC_CTX *tmp_ctx=talloc_new(NULL);
- gss_ctx_id_t gssctx = GSS_C_NO_CONTEXT;
- char *msg_rec=NULL;
- int msg_len = 0;
- int rc=0;
-
- if (trps_auth_connection(trps, conn, &gssctx))
- tr_notice("trps_establish_gss_context: Error authorizing TID Server connection.");
- else:
- tr_notice("trps_establish_gss_context: Connection authorized!");
- return gssctx;
-
- msg_len = trps_read_message(trps, conn, &gssctx, &msg_rec);
- talloc_steal(tmp_ctx, msg_rec); /* get this in our context */
- if (0 > msg_len) {
- tr_debug("trps_handle_connection: Error from trps_read_message()");
- goto cleanup;
+ TR_MSG *msg=NULL;
+ TRP_RC rc=TRP_ERROR;
+
+ /* try to establish a GSS context */
+ if (0!=trp_connection_auth(conn, trps->auth_handler, trps->cookie)) {
+ tr_notice("tr_trps_conn_thread: failed to authorize connection");
+ pthread_exit(NULL);
}
+ tr_notice("trps_handle_connection: authorized connection");
- tr_debug("trps_handle_connection: msg_len=%d", msg_len);
- reply=talloc_asprintf(tmp_ctx, "TRPS heard: %.*s", msg_len, msg_rec);
- if (0 > (rc = trps_send_msg(trps, conn, gssctx, reply))) {
- tr_debug("trps_handle_connection: Error from trps_send_message(), rc = %d.", rc);
+ /* loop as long as the connection exists */
+ while (trp_connection_get_status(conn)==TRP_CONNECTION_UP) {
+ rc=trps_read_message(trps, conn, &msg);
+ switch(rc) {
+ case TRP_SUCCESS:
+ trps->msg_handler(trps, conn, msg); /* send the TR_MSG off to the callback */
+ break;
+
+ case TRP_ERROR:
+ trp_connection_close(conn);
+ break;
+
+ default:
+ tr_debug("trps_handle_connection: trps_read_message failed (%d)", rc);
+ }
}
-cleanup:
+ tr_debug("trps_handle_connection: connection closed.");
talloc_free(tmp_ctx);
- return conn;
}
-#endif
+
+static TRP_RC trps_validate_update(TRPS_INSTANCE *trps, TRP_UPD *upd)
+{
+ if (trp_upd_get_inforec(upd)==NULL) {
+ tr_notice("trps_validate_update: received TRP update with no info records.");
+ return TRP_ERROR;
+ }
+
+ if (trp_upd_get_peer(upd)==NULL) {
+ tr_notice("trps_validate_update: received TRP update without origin peer information.");
+ return TRP_ERROR;
+ }
+
+ return TRP_SUCCESS;
+}
+
+/* ensure that the update could be accepted if feasible */
+static TRP_RC trps_validate_inforec(TRPS_INSTANCE *trps, TRP_INFOREC *rec)
+{
+ switch(trp_inforec_get_type(rec)) {
+ case TRP_INFOREC_TYPE_ROUTE:
+ if ((trp_inforec_get_comm(rec)==NULL)
+ || (trp_inforec_get_realm(rec)==NULL)
+ || (trp_inforec_get_trust_router(rec)==NULL)
+ || (trp_inforec_get_next_hop(rec)==NULL)) {
+ tr_debug("trps_validate_inforec: missing record info.");
+ return TRP_ERROR;
+ }
+
+ /* check for valid metric */
+ if ((trp_inforec_get_metric(rec)==TRP_METRIC_INVALID)
+ || (trp_inforec_get_metric(rec)>TRP_METRIC_INFINITY)) {
+ tr_debug("trps_validate_inforec: invalid metric.");
+ return TRP_ERROR;
+ }
+
+ /* check for valid interval */
+ if (trp_inforec_get_interval(rec)==TRP_INTERVAL_INVALID) {
+ tr_debug("trps_validate_inforec: invalid interval.");
+ return TRP_ERROR;
+ }
+ break;
+
+ default:
+ tr_notice("trps_validate_inforec: unsupported record type.");
+ return TRP_UNSUPPORTED;
+ }
+
+ return TRP_SUCCESS;
+}
+
+/* link cost to a peer */
+static unsigned int trps_cost(TRPS_INSTANCE *trps, TR_NAME *peer)
+{
+ return 1;
+}
+
+static unsigned int trps_advertised_metric(TRPS_INSTANCE *trps, TR_NAME *comm, TR_NAME *realm, TR_NAME *peer)
+{
+ TRP_RENTRY *entry=trp_rtable_get_entry(trps->rtable, comm, realm, peer);
+ if (entry==NULL)
+ return TRP_METRIC_INFINITY;
+ return trp_rentry_get_metric(entry) + trps_cost(trps, peer);
+}
+
+static int trps_check_feasibility(TRPS_INSTANCE *trps, TRP_INFOREC *rec)
+{
+ unsigned int rec_metric=trp_inforec_get_metric(rec);
+ unsigned int new_metric=0;
+ unsigned int current_metric=0;
+ TR_NAME *next_hop=NULL;
+
+ /* we check these in the validation stage, but just in case... */
+ if ((rec_metric==TRP_METRIC_INVALID) || (rec_metric>TRP_METRIC_INFINITY))
+ return 0;
+
+ /* retractions (aka infinite metrics) are always feasible */
+ if (rec_metric==TRP_METRIC_INFINITY)
+ return 1;
+
+ /* updates from our current next hop are always feasible*/
+ next_hop=trps_get_next_hop(trps,
+ trp_inforec_get_comm(rec),
+ trp_inforec_get_realm(rec));;
+ if ((next_hop!=NULL)
+ && (0==tr_name_cmp(next_hop,trp_inforec_get_next_hop(rec)))) {
+ return 1;
+ }
+
+
+ /* compare the existing metric we advertise to what we would advertise
+ * if we accept this update */
+ current_metric=trps_advertised_metric(trps,
+ trp_inforec_get_comm(rec),
+ trp_inforec_get_realm(rec),
+ trp_inforec_get_next_hop(rec));
+ new_metric=rec_metric + trps_cost(trps, trp_inforec_get_next_hop(rec));
+ if (new_metric <= current_metric)
+ return 1;
+ else
+ return 0;
+}
+
+/* uses memory pointed to by *ts, also returns that value. On error, its contents are {0,0} */
+static struct timespec *trps_compute_expiry(TRPS_INSTANCE *trps, unsigned int interval, struct timespec *ts)
+{
+ const unsigned int small_factor=3; /* how many intervals we wait before expiring */
+ if (0!=clock_gettime(CLOCK_REALTIME, ts)) {
+ tr_err("trps_compute_expiry: could not read realtime clock.");
+ ts->tv_sec=0;
+ ts->tv_nsec=0;
+ }
+ ts->tv_sec += small_factor*interval;
+ return ts;
+}
+
+static TRP_RC trps_accept_update(TRPS_INSTANCE *trps, TRP_INFOREC *rec)
+{
+ TRP_RENTRY *entry=NULL;
+
+ entry=trp_rtable_get_entry(trps->rtable,
+ trp_inforec_get_comm(rec),
+ trp_inforec_get_realm(rec),
+ trp_inforec_get_next_hop(rec));
+ if (entry==NULL) {
+ entry=trp_rentry_new(NULL);
+ if (entry==NULL) {
+ tr_err("trps_accept_update: unable to allocate new entry.");
+ return TRP_NOMEM;
+ }
+
+ trp_rentry_set_apc(entry, tr_dup_name(trp_inforec_get_comm(rec)));
+ trp_rentry_set_realm(entry, tr_dup_name(trp_inforec_get_realm(rec)));
+ trp_rentry_set_peer(entry, tr_dup_name(trp_inforec_get_next_hop(rec)));
+ trp_rentry_set_trust_router(entry, tr_dup_name(trp_inforec_get_trust_router(rec)));
+ trp_rentry_set_next_hop(entry, tr_dup_name(trp_inforec_get_next_hop(rec)));
+ if ((trp_rentry_get_apc(entry)==NULL)
+ ||(trp_rentry_get_realm(entry)==NULL)
+ ||(trp_rentry_get_peer(entry)==NULL)
+ ||(trp_rentry_get_trust_router(entry)==NULL)
+ ||(trp_rentry_get_next_hop(entry)==NULL)) {
+ /* at least one field could not be allocated */
+ tr_err("trps_accept_update: unable to allocate all fields for entry.");
+ trp_rentry_free(entry);
+ return TRP_NOMEM;
+ }
+ trp_rtable_add(trps->rtable, entry);
+ }
+
+ /* We now have an entry in the table, whether it's new or not. Update metric and expiry, unless
+ * the metric is infinity. An infinite metric can only occur here if we just retracted an existing
+ * route (we never accept retractions as new routes), so there is no risk of leaving the expiry
+ * time unset on a new route entry. */
+ tr_debug("trps_accept_update: accepting route update.");
+ trp_rentry_set_metric(entry, trp_inforec_get_metric(rec));
+ if (!trps_route_retracted(trps, entry)) {
+ tr_debug("trps_accept_update: route not retracted, setting expiry timer.");
+ trp_rentry_set_expiry(entry, trps_compute_expiry(trps,
+ trp_inforec_get_interval(rec),
+ trp_rentry_get_expiry(entry)));
+ }
+ return TRP_SUCCESS;
+}
+
+/* TODO: handle community updates */
+static TRP_RC trps_handle_update(TRPS_INSTANCE *trps, TRP_UPD *upd)
+{
+ unsigned int feas=0;
+ TRP_INFOREC *rec=NULL;
+ TRP_RENTRY *route=NULL;
+
+ if (trps_validate_update(trps, upd) != TRP_SUCCESS) {
+ tr_notice("trps_handle_update: received invalid TRP update.");
+ return TRP_ERROR;
+ }
+
+ rec=trp_upd_get_inforec(upd);
+ for (;rec!=NULL; rec=trp_inforec_get_next(rec)) {
+ /* validate/sanity check the record update */
+ if (trps_validate_inforec(trps, rec) != TRP_SUCCESS) {
+ tr_notice("trps_handle_update: invalid record in TRP update.");
+ continue;
+ }
+
+ /* determine feasibility */
+ feas=trps_check_feasibility(trps, rec);
+ tr_debug("trps_handle_update: record feasibility=%d", feas);
+
+ /* do we have an existing route? */
+ route=trps_get_route(trps, trp_inforec_get_comm(rec), trp_inforec_get_realm(rec), trp_inforec_get_next_hop(rec));
+ if (route!=NULL) {
+ /* there was a route table entry already */
+ tr_debug("trps_handle_updates: route entry already exists.");
+ if (feas) {
+ /* Update is feasible. Accept it. */
+ trps_accept_update(trps, rec);
+ } else {
+ /* Update is infeasible. Ignore it unless the trust router has changed. */
+ if (0!=tr_name_cmp(trp_rentry_get_trust_router(route),
+ trp_inforec_get_trust_router(rec))) {
+ /* the trust router associated with the route has changed, treat update as a retraction */
+ trps_retract_route(trps, route);
+ }
+ }
+ } else {
+ /* No existing route table entry. Ignore it unless it is feasible and not a retraction. */
+ tr_debug("trps_handle_update: no route entry exists yet.");
+ if (feas && (trp_inforec_get_metric(rec) != TRP_METRIC_INFINITY))
+ trps_accept_update(trps, rec);
+ }
+ }
+ return TRP_SUCCESS;
+}
+
+/* TODO: think this through more carefully. At least ought to add hysteresis
+ * to avoid flapping between routers or routes. */
+static TRP_RC trps_update_active_routes(TRPS_INSTANCE *trps)
+{
+ size_t n_apc=0, ii=0;
+ TR_NAME **apc=trp_rtable_get_apcs(trps->rtable, &n_apc);
+ size_t n_realm=0, jj=0;
+ TR_NAME **realm=NULL;
+ size_t n_entry=0, kk=0, kk_min=0;
+ TRP_RENTRY **entry=NULL, *cur_route=NULL;
+ unsigned int min_metric=0, cur_metric=0;
+
+ for (ii=0; ii<n_apc; ii++) {
+ realm=trp_rtable_get_apc_realms(trps->rtable, apc[ii], &n_realm);
+ for (jj=0; jj<n_realm; jj++) {
+ entry=trp_rtable_get_realm_entries(trps->rtable, apc[ii], realm[jj], &n_entry);
+ for (kk=0,min_metric=TRP_METRIC_INFINITY; kk<n_entry; kk++) {
+ if (trp_rentry_get_metric(entry[kk]) < min_metric) {
+ kk_min=kk;
+ min_metric=trp_rentry_get_metric(entry[kk]);
+ }
+ }
+
+ cur_route=trps_get_selected_route(trps, apc[ii], realm[jj]);
+ if (cur_route!=NULL) {
+ cur_metric=trp_rentry_get_metric(cur_route);
+ if (min_metric < cur_metric) {
+ trp_rentry_set_selected(cur_route, 0);
+ trp_rentry_set_selected(entry[kk_min], 1);
+ } else if (cur_metric==TRP_METRIC_INFINITY)
+ trp_rentry_set_selected(cur_route, 0);
+ } else if (min_metric<TRP_METRIC_INFINITY)
+ trp_rentry_set_selected(entry[kk_min], 1);
+
+ talloc_free(entry);
+ entry=NULL; n_entry=0;
+ }
+ talloc_free(realm);
+ realm=NULL; n_realm=0;
+ }
+ talloc_free(apc);
+ apc=NULL; n_apc=0;
+
+ return TRP_SUCCESS;
+}
+
+TRP_RC trps_handle_tr_msg(TRPS_INSTANCE *trps, TR_MSG *tr_msg)
+{
+ TRP_RC rc=TRP_ERROR;
+
+ switch (tr_msg_get_msg_type(tr_msg)) {
+ case TRP_UPDATE:
+ rc=trps_handle_update(trps, tr_msg_get_trp_upd(tr_msg));
+ if (rc==TRP_SUCCESS) {
+ rc=trps_update_active_routes(trps);
+ }
+ return rc;
+
+ case TRP_REQUEST:
+ return TRP_UNSUPPORTED;
+
+ default:
+ /* unknown error or one we don't care about (e.g., TID messages) */
+ return TRP_ERROR;
+ }
+}
projects / trust_router.git / blobdiff