>Number:         10606
>Category:       kern
>Synopsis:       IPv6 NFS over TCP does not work
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Jul 16 00:43:00 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     Jun-ichiro itojun Hagino
>Release:        latest
>Organization:
	itojun.org
>Environment:
System: NetBSD starfruit.itojun.org 1.5B NetBSD 1.5B (STARFRUIT) #54: Sat Jul 15 09:42:05 JST 2000 itojun@starfruit.itojun.org:/usr/home/itojun/NetBSD/src/sys/arch/i386/compile/STARFRUIT i386

>Description:
	TCP NFS over IPv6 does not work right.  I've tested it with:
	- NetBSD 1.5_ALPHA client, NetBSD-current server
	- Solaris8 client, NetBSD-current server
	- NetBSD-current client/server (loopback)
	and they all failed if we use TCP.

	UDP NFS over IPv6 works just fine for all cases.
>How-To-Repeat:
	1. delete 127.0.0.1 from /etc/hosts, to map "localhost" into ::1.
	2. export /usr.
	3. mount -T localhost:/usr /mnt.  see that the operation makes no
	   progress.
	4. if you try tcpdump on loopback interface, you will see that tthe
	   tcp stream does not get ACKed (is it a TCP issue, not nfs issue?)

	for comparison:
	a. if you use IPv4 (map "localhost" into 127.0.0.1), it works.
	b. if you use udp (no -T), it works.

	there seems to be some other issue in mountd(8), with -noprivport and
	-noprivmnt.
>Fix:
>Release-Note:
>Audit-Trail:
>Unformatted: