The following reply was made to PR kern/34023; it has been noted by GNATS.

From: "Henry R. Bent" <Henry.Bent@oberlin.edu>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/34023
Date: Sun, 15 Oct 2006 01:02:51 -0400

 I am seeing this problem here running i386 on an Asus A8N32-SLI-Deluxe 
 board.  Under moderate load everything is fine; if I flood the adapter 
 with traffic (say, moving large files over NFS while streaming a FLAC to 
 another machine) it will stall out after a few seconds.  It will stall 
 for perhaps 5 or 10 seconds, transmit a small amount of traffic, and 
 then stall again.  I can alleviate this situation somewhat by stopping 
 one of the transfers, but it will still eventually stall.
 
 I do not see the "watchdog timeout" messages that Perry was seeing.
 
 uname -a:
 NetBSD aelfric 4.99.3 NetBSD 4.99.3 (GENERIC.MPACPI) #0: Mon Oct  2 
 09:05:10 EDT 2006 
 root@hbent3:/usr/obj/src-current/sys/arch/i386/compile/GENERIC.MPACPI i386
 
 dmesg excerpt:
 nfe0 at pci0 dev 19 function 0LMAC: Picked IRQ 21 with weight 1
 : ioapic0 pin 21 (irq 7), address 00:15:f2:xx:xx:xx
 makphy1 at nfe0 phy 1: Marvell 88E1111 Gigabit PHY, rev. 2
 makphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 
 1000baseT-FDX, auto
 
 lspci says:
 00:13.0 Bridge: nVidia Corporation CK804 Ethernet Controller (rev a3)
 
 There is also an onboard msk that is even worse.  I will file a separate 
 pr for it.  I have a PCI fxp card that I will try when downtime is 
 available.
 
 -Henry Bent
 Oberlin College CIT