problem with sk? (was: ssh: "Connection reset by peer" or "Corrupted MAC on input")

Michael Gerhards HM-Gerhards at
Sat May 6 15:08:47 UTC 2006

Artyom Viklenko <artem at> wrote:
>> I have just installed FreeBSD/i386 6.1-RC2 on a AMD Athlon 64 X2 3800+,
>> Asus A8N-SLI Deluxe Motherboard and 2 GB RAM. The Motherboard has two
>> Ethernet-Ports for 1000BaseT which dmesg gives out as:
>>  skc0: <Marvell Gigabit Ethernet> port 0xa400-0xa4ff mem 0xd1004000-
>>  0xd1007fff irq 17 at device 12.0 on pci5
>>  skc0: Marvell Yukon Lite Gigabit Ethernet rev. (0x9)
>>  sk0: Ethernet address: 00:11:d8:ca:ed:43
>>  nve0: Ethernet address 00:11:d8:ca:e1:8b
>> I read about problems with the nve driver, so I only configured the skc0
>> device. It is connected to a 100MBit Hub, and all other computers in my

Sorry, it's a switch, of course.

>> What might be the reason for my problems? Is there a problem in the
>> support for my skc0 device? What could I do to solve my problems?
> I my case ther was bad memory in target system. Try to check/replace
> RAM in AMD machine.

I will try running memtest86.

But by browsing through this list I found the thread "sk0: watchdog timeout"
and the problem described there is quite similiar to what I get here. So
perhaps this bug in the sk driver is also the cause for my trouble here?!

Can I somehow use this patch for sk0 _without_ changing everything to 
-current? Acutually, I wanted to stay to 6.1-RELEASE... 


More information about the freebsd-stable mailing list