>Number:         18040
>Category:       port-sparc64
>Synopsis:       NFS broken due to hme bugs on sparc64
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-sparc64-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Aug 22 06:54:03 PDT 2002
>Closed-Date:
>Last-Modified:
>Originator:     Hubert Feyrer
>Release:        NetBSD 1.6_BETA5
>Organization:
Hubert Feyrer <hubertf@channel.regensburg.org>
>Environment:
System: NetBSD rfhs8036 1.6_BETA5 NetBSD 1.6_BETA5 (RFHS8036) #0: Fri Aug 9 05:34:59 CEST 2002 root@rfhs8036:/disk1/cvs/src-1.6/sys/arch/sparc64/compile/RFHS8036 sparc64
Architecture: sparc64
Machine: sparc64
>Description:
	Trying to do a build with sources on NFS fails with signal 11 in
	random places in the source tree. dmesg(8) shows many of these
	errors:

		hme0: status=8000<DTIMEXP>

	I've ensured that it's not broken RAM that's causing the signals
	(exchanged RAM, also set diag-switch? to true to get RAM test on
	bootup). 

	After discussion on port-sparc64 where Grant Beattie
	<grant@netbsd.org> suggested to move sources to local disk, and
	after that everything works fine. No more random sig11's and core
	dumps. 

	See http://mail-index.netbsd.org/port-sparc64/2002/08/22/0002.html
	for Grant's mail. There's a strong feeling that the hme driver needs
	some fixing.

	FWIW, I'm using NFS over IPv6:

		rfhs8036% mount -v
		...
		smaug.ipv6.fh-regensburg.de:/disk1 on /disk1 type nfs (writes: sync 0 async 0)
	where smaug.ipv6.fh-regensburg.de only has a AAAA record, no A
	record. In general, NFS over IPv6 works fine here for a very long
	time, I do i386 pkg bulk builds over it on a i386 machine with
	smaug as the server, so that's probably not a problem.


>How-To-Repeat:
	Try to rebuild system with sources on NFS.

>Fix:
	Fix hme driver.

>Release-Note:
>Audit-Trail:
>Unformatted: