Changeset 706:4ffbc9f1e922 in freeDiameter for libfdcore/p_dp.c

Timestamp:

Feb 9, 2011, 3:26:58 PM (13 years ago)

Author:

Sebastien Decugis <sdecugis@nict.go.jp>

Branch:

default

Phase:

public

Message:

Large UNTESTED commit with the following changes:

• Improved DiameterIdentity? handling (esp. interationalization issues), and improve efficiency of some string operations in peers, sessions, and dictionary modules (closes #7)
• Cleanup in the session module to free only unreferenced sessions (#16)
• Removed fd_cpu_flush_cache(), replaced by more robust alternatives.
• Improved peer state machine algorithm to counter SCTP multistream race condition.

File:

• libfdcore/p_dp.c (modified) (6 diffs)

libfdcore/p_dp.c

@@ -38 +38 @@
 /* This file contains code to handle Disconnect Peer messages (DPR and DPA) */
 
+/* Delay to use before next reconnect attempt */
+int fd_p_dp_newdelay(struct fd_peer * peer) 
+{
+        int delay = peer->p_hdr.info.config.pic_tctimer ?: fd_g_config->cnf_timer_tc;
+        
+        switch (peer->p_hdr.info.runtime.pir_lastDC) {
+                case ACV_DC_REBOOTING:
+                default:
+                        /* We use TcTimer to attempt reconnection */
+                        break;
+                case ACV_DC_BUSY:
+                        /* No need to hammer the overloaded peer */
+                        delay *= 10;
+                        break;
+                case ACV_DC_NOT_FRIEND:
+                        /* He does not want to speak to us... let's retry a *lot* later maybe */
+                        delay *= 200;
+                        break;
+        }
+        return delay;
+}
+
 /* Handle a received message */
 int fd_p_dp_handle(struct msg ** msg, int req, struct fd_peer * peer)
@@ -44 +66 @@
         
         if (req) {
-                /* We received a DPR, save the Disconnect-Cause and terminate the connection */
+                /* We received a DPR, save the Disconnect-Cause and go to CLOSING_GRACE or terminate the connection */
                 struct avp * dc;
-                int delay = peer->p_hdr.info.config.pic_tctimer ?: fd_g_config->cnf_timer_tc;
                 
                 CHECK_FCT( fd_msg_search_avp ( *msg, fd_dict_avp_DC, &dc ));
                 if (dc) {
-                        /* Check the value is consistent with the saved one */
                         struct avp_hdr * hdr;
                         CHECK_FCT(  fd_msg_avp_hdr( dc, &hdr )  );
@@ -60 +80 @@
                         }
 
+                        /* save the cause */
                         peer->p_hdr.info.runtime.pir_lastDC = hdr->avp_value->u32;
-                        
-                        switch (hdr->avp_value->u32) {
-                                case ACV_DC_REBOOTING:
-                                default:
-                                        /* We use TcTimer to attempt reconnection */
-                                        break;
-                                case ACV_DC_BUSY:
-                                        /* No need to hammer the overloaded peer */
-                                        delay *= 10;
-                                        break;
-                                case ACV_DC_NOT_FRIEND:
-                                        /* He does not want to speak to us... let's retry a lot later maybe */
-                                        delay *= 200;
-                                        break;
-                        }
                 }
                 if (TRACE_BOOL(INFO)) {
                         if (dc) {
-                                struct dict_object * dictobj = NULL;
+                                struct dict_object * dictobj;
                                 struct dict_enumval_request er;
                                 memset(&er, 0, sizeof(er));
+                                
+                                /* prepare the request */
                                 CHECK_FCT( fd_dict_search( fd_g_config->cnf_dict, DICT_TYPE, TYPE_OF_AVP, fd_dict_avp_DC, &er.type_obj, ENOENT )  );
                                 er.search.enum_value.u32 = peer->p_hdr.info.runtime.pir_lastDC;
+                                
+                                /* Search the enum value */
                                 CHECK_FCT( fd_dict_search( fd_g_config->cnf_dict, DICT_ENUMVAL, ENUMVAL_BY_STRUCT, &er, &dictobj, 0 )  );
                                 if (dictobj) {
@@ -100 +110 @@
                 CHECK_FCT( fd_msg_rescode_set( *msg, "DIAMETER_SUCCESS", NULL, NULL, 1 ) );
                 
-                /* Move to CLOSING state to failover outgoing messages (and avoid failing the DPA...) */
+                /* Move to CLOSING state to failover outgoing messages (and avoid failing over the DPA...) */
                 CHECK_FCT( fd_psm_change_state(peer, STATE_CLOSING) );
                 
@@ -106 +116 @@
                 CHECK_FCT( fd_out_send( msg, NULL, peer, FD_CNX_ORDERED) );
                 
-                /* Move to CLOSED state */
-                fd_psm_cleanup(peer, 0);
-                
-                /* Reset the timer for next connection attempt -- we'll retry sooner or later depending on the disconnection cause */
-                fd_psm_next_timeout(peer, 1, delay);
-                
+                if (fd_cnx_isMultichan(peer->p_cnxctx)) {
+                        /* There is a possibililty that messages are still in the pipe coming here, so let's grace for 1 second */
+                        CHECK_FCT( fd_psm_change_state(peer, STATE_CLOSING_GRACE) );
+                        fd_psm_next_timeout(peer, 0, 1);
+                        
+                } else {
+                        /* Move to CLOSED state */
+                        fd_psm_cleanup(peer, 0);
+
+                        /* Reset the timer for next connection attempt -- we'll retry sooner or later depending on the disconnection cause */
+                        fd_psm_next_timeout(peer, 1, fd_p_dp_newdelay(peer));
+                }
         } else {
                 /* We received a DPA */
-                fd_cpu_flush_cache();
-                if (peer->p_hdr.info.runtime.pir_state != STATE_CLOSING) {
-                        TRACE_DEBUG(INFO, "Ignoring DPA received in state %s", STATE_STR(peer->p_hdr.info.runtime.pir_state));
+                int curstate = fd_peer_getstate(peer);
+                if (curstate != STATE_CLOSING) {
+                        TRACE_DEBUG(INFO, "Ignoring DPA received in state %s", STATE_STR(curstate));
                 }
                         
                 /* In theory, we should control the Result-Code AVP. But since we will not go back to OPEN state here anyway, let's skip it */
+                
+                /* TODO("Control Result-Code in the DPA") */
                 CHECK_FCT_DO( fd_msg_free( *msg ), /* continue */ );
                 *msg = NULL;
                 
-                /* The calling function handles cleaning the PSM and terminating the peer since we return in CLOSING state */
+                if (fd_cnx_isMultichan(peer->p_cnxctx)) {
+                        CHECK_FCT( fd_psm_change_state(peer, STATE_CLOSING_GRACE) );
+                        fd_psm_next_timeout(peer, 0, 1);
+                        peer->p_flags.pf_localterm = 1;
+                }
+                /* otherwise, return in CLOSING state, the psm will handle it */
         }
         
@@ -153 +176 @@
         CHECK_FCT( fd_dict_search( fd_g_config->cnf_dict, DICT_TYPE, TYPE_OF_AVP, fd_dict_avp_DC, &er.type_obj, ENOENT )  );
         er.search.enum_name = reason ?: "REBOOTING";
-        CHECK_FCT( fd_dict_search( fd_g_config->cnf_dict, DICT_ENUMVAL, ENUMVAL_BY_STRUCT, &er, &dictobj, ENOENT )  );
+        CHECK_FCT_DO( fd_dict_search( fd_g_config->cnf_dict, DICT_ENUMVAL, ENUMVAL_BY_STRUCT, &er, &dictobj, ENOENT ), { ASSERT(0); /* internal error: unknown reason */ }  );
         CHECK_FCT( fd_dict_getval( dictobj, &er.search ) );