>Number:         10905
>Category:       port-i386
>Synopsis:       vr0 underruns with 8K NFS mounts
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-i386-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Aug 28 11:19:00 PDT 2000
>Closed-Date:
>Last-Modified:
>Originator:     Michael Richardson <mcr@solidum.com>
>Release:        NetBSD-ALPHA2
>Organization:
Solidum Systems Corporation
>Environment:
	
System: NetBSD phobos.solidum.com 1.5_ALPHA2 NetBSD 1.5_ALPHA2 (XTERM) #4: Mon Aug 28 12:02:12 EDT 2000 rootmcr@phobos.solidum.com:/usr/src/sys/arch/i386/compile/XTERM i386


>Description:
	NFS mounts of a NetAPP 720 start to fail, and one gets underruns.

vr0 at pci0 dev 9 function 0: VIA VT3043 (Rhine) 10/100 Ethernet
vr0: interrupting at irq 10
vr0: Ethernet address: 00:00:21:6d:c1:f2
ukphy0 at vr0 phy 1: Generic IEEE 802.3u media interface
ukphy0: OUI 0x00c0b4, model 0x0000, rev. 8
ukphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
....
vr0: transmit underrun
vr0: transmit aborted
vr0: transmit underrun
vr0: transmit aborted
...
nfs server sirius:/users/mcr: not responding
vr0: transmit underrun
vr0: transmit aborted
...
	Okay, so Via Rhine is not your high performance NIC card, but it
	is a nice simple card to stick in desktops.

>How-To-Repeat:
	NFS mount with 8K blocks with a VR0
>Fix:
	Workaround: hack AMD to use 1K mounts by default.
	Wish that this was an option in amd.conf (it isn't that I can see).

	Better: why is it a transmit *underrun* ??? shouldn't it be an overrun?

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