The following reply was made to PR kern/34952; it has been noted by GNATS.

From: martijnb@stack.nl (Martijn van Buul)
To: gnats-bugs@NetBSD.org
Cc: 
Subject: Re: kern/34952: re0 fails to attach some of the time
Date: Wed, 1 Nov 2006 11:01:56 +0100

 * Perry E. Metzger:
 >>Number:         34952
 >>Category:       kern
 >>Synopsis:       re0 fails to attach some of the time
 >
 > Some percentage of the time, this happens to the realtek ethernet
 > driver (re) during boot:
 >
 > re0 at pci4 dev 8 function 0: Linksys EG1032 rev. 3 Gigabit Ethernet
 > APC1: Picked IRQ 16 with weight 0
 > re0: interrupting at ioapic0 pin 16 (irq 5)
 > re0: Ethernet address 00:14:bf:57:21:ae
 > re0: using 256 tx descriptors
 > rgephy0 at re0 phy 7: RTL8169S/8110S 1000BASE-T media interface, rev. 0
 > rgephy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseT, 1000baseT-FDX, auto
 > re0: diagnostic failed, failed to receive packet in loopback mode
 > re0: attach aborted due to hardware diag failure
 > rgephy0 detached
 >
 > A reboot fixes the problem. The issue happens about one in every five
 > reboots. Naturally when the device is the only way to get to a remote
 > box, this can be very disturbing.
 
 I have this problem too (In an Asus A8N-E mobo, Athlon64 3000+, both i386 as
 well as amd64 ports). I ended up using a horrible workaround by 
 checking for the presence of the 're0: attach aborted" line in the dmesg in
 /etc/rc.local and reboot if this is the case. Not a particulary nice solution,
 but at least it prevents the machine from remaining unreachable.
 
 -- 
 Martijn van Buul - pino@dohd.org