fix error when flush arp/route run into deleted routing table entries

It is possible to get "no such process" (ESRCH) errors during flush
arp or flush route commands due to concurrent deletion of routing
table entries by another process.

This is no reason to stop trying to delete more routes, so do not
abort the loop in this case.

Also, make nsh print a more informative error message in this case,
rather than "no such process". Example of the new error message:

% No such route to delete: 100.64.1.2 fe:e1:ba:d0:8e:45 255.255.255.255 fe:e1:bb:d1:c4:b8 100.64.1.3

Testing + OK Tom

kroute.c

@@ -74,6 +74,7 @@ void	 print_rtmsg(struct rt_msghdr *);
 void	 print_getmsg(struct rt_msghdr *, int);
 void	 pmsg_common(struct rt_msghdr *);
 void	 pmsg_addrs(char *, int);
+void	 print_addrs(char *, int);
 void	 bprintf(FILE *, int, u_char *);
 
 /*
@@ -201,6 +202,16 @@ flushroutes(int af, int af2)
 		rtm->rtm_seq = seqno;
 		rlen = write(s, next, rtm->rtm_msglen);
 		if (rlen < (int)rtm->rtm_msglen) {
+			if (errno == ESRCH) {
+				printf("%% No such route to delete:");
+				print_addrs((char *)rtm + rtm->rtm_hdrlen,
+				   rtm->rtm_addrs);
+				putchar('\n');
+				fflush(stdout);
+
+				seqno++;
+				continue;
+			}
 			printf("%% Unable to write to routing socket: %s\n",
 			    strerror(errno));
 			break;
@@ -517,22 +528,31 @@ pmsg_common(struct rt_msghdr *rtm)
 void
 pmsg_addrs(char	*cp, int addrs)
 {
-	struct sockaddr *sa;
-	int i;
-
 	if (addrs == 0)
 		return;
 	(void) printf("%% sockaddrs: ");
 	bprintf(stdout, addrs, addrnames);
 	(void) putchar('\n');
+	print_addrs(cp, addrs);
+	(void) putchar('\n');
+	(void) fflush(stdout);
+}
+
+void
+print_addrs(char *cp, int addrs)
+{
+	struct sockaddr *sa;
+	int i;
+
+	if (addrs == 0)
+		return;
+
 	for (i = 1; i; i <<= 1)
 		if (i & addrs) {
 			sa = (struct sockaddr *)cp;
 			(void) printf(" %s", routename(sa));
 			ADVANCE(cp, sa);
 		}
-	(void) putchar('\n');
-	(void) fflush(stdout);
 }
 
 void