Add more output. Fix some typos, and some small cleanups.

This commit was SVN r14327.
2007-04-12 05:01:29 +00:00 · 2007-04-12 05:01:29 +00:00 · cad93a7693
--- a/orte/mca/oob/tcp/oob_tcp.c
+++ b/orte/mca/oob/tcp/oob_tcp.c
@ -2,7 +2,7 @@
 * Copyright (c) 2004-2007 The Trustees of Indiana University and Indiana
 *                         University Research and Technology
 *                         Corporation.  All rights reserved.
- * Copyright (c) 2004-2006 The University of Tennessee and The University
+ * Copyright (c) 2004-2007 The University of Tennessee and The University
 *                         of Tennessee Research Foundation.  All rights
 *                         reserved.
 * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
@ -18,9 +18,9 @@
 * $HEADER$
 *
 * In windows, many of the socket functions return an EWOULDBLOCK
- * instead of \ things like EAGAIN, EINPROGRESS, etc. It has been
- * verified that this will \ not conflict with other error codes that
- * are returned by these functions \ under UNIX/Linux environments 
+ * instead of things like EAGAIN, EINPROGRESS, etc. It has been
+ * verified that this will not conflict with other error codes that
+ * are returned by these functions under UNIX/Linux environments 
 */

 #include "orte_config.h"
@ -930,6 +930,7 @@ void mca_oob_tcp_registry_callback(
                    &mca_oob_tcp_component.tcp_peer_names, &addr->addr_name);
                if(NULL != existing) {
                    /* TSW - need to update existing entry */
+  		    opal_output( 0, "WHY ARE WE RECEIVING THE SAME INFORMATION SEVERAL TIMES ?!?!?!?" );
                    orte_hash_table_set_proc(&mca_oob_tcp_component.tcp_peer_names, &addr->addr_name, addr);
                    OBJ_RELEASE(addr);
                    continue;
--- a/orte/mca/oob/tcp/oob_tcp_msg.c
+++ b/orte/mca/oob/tcp/oob_tcp_msg.c
@ -269,14 +269,14 @@ bool mca_oob_tcp_msg_send_handler(mca_oob_tcp_msg_t* msg, struct mca_oob_tcp_pee
                    return true;
                }
            }
-        } while(msg->msg_rwnum);
+        } while(1);
    }
 }

 /*
 * Receives message data.
 * @param msg the message to be recieved into
- * @param peer the peer to recieve from
+ * @param peer the peer to receive from
 * @retval true if the whole message was received
 * @retval false if the whole message was not received
 */
@ -307,6 +307,12 @@ bool mca_oob_tcp_msg_recv_handler(mca_oob_tcp_msg_t* msg, struct mca_oob_tcp_pee
             msg->msg_rwiov[1].iov_len = 0;
             msg->msg_rwnum = 0;
        }
+        if (mca_oob_tcp_component.tcp_debug >= OOB_TCP_DEBUG_CONNECT) {
+            opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_recv_handler*: size %lu\n",
+                        ORTE_NAME_ARGS(orte_process_info.my_name),
+                        ORTE_NAME_ARGS(&(peer->peer_name)),
+                        (unsigned long)(msg->msg_hdr.msg_size) );
+        }
    }

    /* do the right thing based on the message type */
@ -343,16 +349,14 @@ static bool mca_oob_tcp_msg_recv(mca_oob_tcp_msg_t* msg, mca_oob_tcp_peer_t* pee
               under UNIX/Linux environments */
            else if (opal_socket_errno == EAGAIN || opal_socket_errno == EWOULDBLOCK)
                return false;
-            else {
-                opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_msg_recv: readv failed: %s (%d)", 
-                    ORTE_NAME_ARGS(orte_process_info.my_name),
-                    ORTE_NAME_ARGS(&(peer->peer_name)),
-                    strerror(opal_socket_errno),
-                    opal_socket_errno);
-                mca_oob_tcp_peer_close(peer);
-                mca_oob_call_exception_handlers(&peer->peer_name, MCA_OOB_PEER_DISCONNECTED);
-                return false;
-            }
+	    opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_msg_recv: readv failed: %s (%d)", 
+			ORTE_NAME_ARGS(orte_process_info.my_name),
+			ORTE_NAME_ARGS(&(peer->peer_name)),
+			strerror(opal_socket_errno),
+			opal_socket_errno);
+	    mca_oob_tcp_peer_close(peer);
+	    mca_oob_call_exception_handlers(&peer->peer_name, MCA_OOB_PEER_DISCONNECTED);
+	    return false;
        } else if (rc == 0)  {
            if(mca_oob_tcp_component.tcp_debug >= OOB_TCP_DEBUG_CONNECT_FAIL) {
                opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_msg_recv: peer closed connection", 
@ -374,10 +378,11 @@ static bool mca_oob_tcp_msg_recv(mca_oob_tcp_msg_t* msg, mca_oob_tcp_peer_t* pee
                (msg->msg_rwnum)--;
                (msg->msg_rwptr)++;
                if(0 == msg->msg_rwnum) {
+		    assert( 0 == rc );
                    return true;
                }
            }
-        } while(msg->msg_rwnum);
+        } while(1);
    }
    return true;
 }
--- a/orte/mca/oob/tcp/oob_tcp_recv.c
+++ b/orte/mca/oob/tcp/oob_tcp_recv.c
@ -64,6 +64,11 @@ int mca_oob_tcp_recv(
            return msg->msg_rc;
        }
 
+        opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_recv*unexpected*: tag %d size %lu\n",
+		    ORTE_NAME_ARGS(orte_process_info.my_name),
+		    ORTE_NAME_ARGS(peer),
+		    tag, (unsigned long)(msg->msg_hdr.msg_size) );
+
        /* if we are returning an allocated buffer - just take it from the message */
        if(flags & MCA_OOB_ALLOC) {

@ -110,6 +115,13 @@ int mca_oob_tcp_recv(
        size += iov[i].iov_len;
    }

+    if (mca_oob_tcp_component.tcp_debug >= OOB_TCP_DEBUG_CONNECT) {
+        opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_recv*expected*: tag %d size %lu\n",
+                    ORTE_NAME_ARGS(orte_process_info.my_name),
+                    ORTE_NAME_ARGS(peer),
+                    tag, (unsigned long)(size) );
+    }
+
    /* fill in the struct */
    msg->msg_hdr.msg_size = size;
    msg->msg_hdr.msg_tag = tag;
--- a/orte/mca/oob/tcp/oob_tcp_send.c
+++ b/orte/mca/oob/tcp/oob_tcp_send.c
@ -100,11 +100,17 @@ int mca_oob_tcp_send(
    if(NULL == peer)
        return ORTE_ERR_UNREACH;

+    /* calculate the size of the message */
+    size = 0;
+    for(rc = 0; rc < count; rc++) {
+        size += iov[rc].iov_len;
+    }
+   
    if(mca_oob_tcp_component.tcp_debug >= OOB_TCP_DEBUG_ALL) {
-        opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_send: tag %d\n",
+        opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_send: tag %d size %lu\n",
            ORTE_NAME_ARGS(orte_process_info.my_name),
            ORTE_NAME_ARGS(&(peer->peer_name)),
-            tag);
+            tag, (unsigned long)size );
    }

    MCA_OOB_TCP_MSG_ALLOC(msg, rc);
@ -112,12 +118,6 @@ int mca_oob_tcp_send(
        return rc;
    }

-    /* calculate the size of the message */
-    size = 0;
-    for(rc = 0; rc < count; rc++) {
-        size += iov[rc].iov_len;
-    }
-   
    /* turn the size to network byte order so there will be no problems */
    msg->msg_hdr.msg_type = MCA_OOB_TCP_DATA;
    msg->msg_hdr.msg_size = size;
@ -210,6 +210,14 @@ int mca_oob_tcp_send_nb(
    for(rc = 0; rc < count; rc++) {
        size += iov[rc].iov_len;
    }
+
+    if(mca_oob_tcp_component.tcp_debug >= OOB_TCP_DEBUG_ALL) {
+        opal_output(0, "[%lu,%lu,%lu]-[%lu,%lu,%lu] mca_oob_tcp_send_nb: tag %d size %lu\n",
+            ORTE_NAME_ARGS(orte_process_info.my_name),
+            ORTE_NAME_ARGS(&(peer->peer_name)),
+            tag, (unsigned long)size );
+    }
+
    /* turn the size to network byte order so there will be no problems */
    msg->msg_hdr.msg_type = MCA_OOB_TCP_DATA;
    msg->msg_hdr.msg_size = size;