On 2001-05-13 01:20 -0400, Mike Tancsa <[EMAIL PROTECTED]> wrote:
>
> Any luck with the 815E issue ?
>
> fxp1: <Intel Pro/100 Ethernet> port 0xc400-0xc43f mem 0xd5101000-0xd5101fff
> irq 15 at device 8.0 on pci1
> fxp1: Ethernet address 00:01:80:07:1e:f4
> inphy1: <i82562EM 10/100 media interface> on miibus1
> inphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
>
> it takes about 3 seconds only to get
>
> fxp1: SCB timeout
> fxp1: SCB timeout
> fxp1: device timeout
>
> when connected to a 10baseT switch (Compaq in this case)
>
> The other strange thing about this is that if I force it to 10baseT/UTP the
> problem is not quite as bad as when letting it detect it via autoselect. I
> still get the SCB timeouts, but the card seems to recover from them
> slightly faster.
I have the same problem. Intel D815EEAAL motherboard, onboard fxp.
fxp0: <Intel Pro/100 Ethernet> port 0xdf00-0xdf3f mem 0xff8ef000-0xff8effff irq 11 at
device 8.0 on pci1
fxp0: Ethernet address 00:03:47:fe:ef:fb
inphy0: <i82562ET 10/100 media interface> on miibus0
inphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
With jlemon's advice, I added a dummynet pipe to limit the card to
3Mbit/sec, and at that speed, it seems to have time to recover
properly from whatever problem is occurring. It still occasionally
times out (SCB timeout), but will then device timeout and work
properly for a while.
ipfw pipe 001 config bw 3Mbit/s
ipfw add 200 pipe 001 ip from any to any
Greg
--
Gregory S. Sutter Only two things are infinite, the
mailto:[EMAIL PROTECTED] universe and human stupidity, and I'm
http://www.zer0.org/~gsutter/ not sure about the former.
hkp://wwwkeys.pgp.net/0x845DFEDD - Albert Einstein
PGP signature
