Re: nis problems

[Matthew Dillon <dillon@xxxxxxxxxxxxxxxxxxxx>]

:I think I understand the problem now.  The application is expecting the
:implicit bind on output to be exactly the same as an explicit bind.
:Can you see if this fixes it?  Thanks for tracking this down, Matt.
:
:							Jeffrey
:
    That is essentially what my little patch did, except there were
    two places where the bind was taking place... oh, I see, the
    second place removes it from the wildcard hash immediately.

    I guess your patch is cleaner :-).   I abstracted out a udp_pcbbind()
    but I guess that is overkill.

    I'll test your patch and commit it if it works the same as mine,
    and it should.

					-Matt
					Matthew Dillon 
					<dillon@xxxxxxxxxxxxx>

:--==_Exmh_-4515556020
:Content-Type: text/plain ; name="connect.diff"; charset=us-ascii
:Content-Description: connect.diff
:Content-Disposition: attachment; filename="connect.diff"
:
:Index: udp_usrreq.c
:===================================================================
:RCS file: /j/dragonfly/dcvs/src/sys/netinet/udp_usrreq.c,v
:retrieving revision 1.20
:diff -u -p -r1.20 udp_usrreq.c
:--- udp_usrreq.c	31 Mar 2004 07:21:38 -0000	1.20
:+++ udp_usrreq.c	4 Apr 2004 22:01:26 -0000
:@@ -730,6 +730,7 @@ udp_output(inp, m, dstaddr, control, td)
: 		error = in_pcbbind(inp, (struct sockaddr *)NULL, td);
: 		if (error)
: 			goto release;
:+		in_pcbinswildcardhash(inp);
: 	}
: 
: 	if (dstaddr != NULL) {		/* destination address specified */
:
:--==_Exmh_-4515556020--
:
: