[prev in list] [next in list] [prev in thread] [next in thread] 

List:       freebsd-net
Subject:    RE: FW: Network lockups on fxp0?
From:       "Deepak Jain" <deepak () ai ! net>
Date:       2001-03-28 21:08:36
[Download RAW message or body]



-----Original Message-----
From: owner-freebsd-net@FreeBSD.ORG
[mailto:owner-freebsd-net@FreeBSD.ORG]On Behalf Of Remy Nonnenmacher
Sent: Wednesday, March 28, 2001 2:35 AM
To: deepak@ai.net
Cc: jlemon@flugsvamp.com; net@FreeBSD.ORG
Subject: RE: FW: Network lockups on fxp0?


On 27 Mar, Deepak Jain wrote:
>
> I know ifconfig up/down does not solve the problem. That was the first
thing
> tried.
>
> No pattern detected at the network traffic or application level, though
> machines have been known to go down simultaneously (same minute/second)
> though they service unrelated tasks.
>
> No flow control setting changes have been made. I didn't know fast
ethernet
> had reliable flow control.
>
> Thanks for the insight,
>
> Deepak Jain
> AiNET
>
> -----Original Message-----
> From: owner-freebsd-net@FreeBSD.ORG
> [mailto:owner-freebsd-net@FreeBSD.ORG]On Behalf Of Jonathan Lemon
> Sent: Tuesday, March 27, 2001 5:54 PM
> To: deepak@ai.net; net@FreeBSD.ORG
> Subject: Re: FW: Network lockups on fxp0?
>
>
> In article <MLEFIJIDEIDHLFKGNEILGEEHCFAA.deepak@ai.net> you write:
>>>fxp0: SCB timeout
>>>fxp0: DMA timeout
>>>(repeating)
>
> SCB timeout comes about because the chip is refusing to accept
> any more commands; in this case, it probably is wedged.  Is there
> any pattern to this?  Do you happen to have hardware flowcontrol
> enabled?  Does an ifconfig up/down fix the problem?
> --
> Jonathan
>

I have a pattern to submit:

Here is the pciconf -l of network cards:

fxp0@pci0:3:0:  class=0x020000 card=0x12298086 chip=0x12298086 rev=0x08
hdr=0x00
fxp1@pci2:4:0:  class=0x020000 card=0x10f08086 chip=0x12298086 rev=0x05
hdr=0x00

---
fxp0@pci0:4:0:  class=0x020000 card=0x000c8086 chip=0x12298086 rev=0x08
hdr=0x00
fxp1@pci0:5:0:  class=0x020000 card=0x000c8086 chip=0x12298086 rev=0x08
hdr=0x00

Our fxp's are both built onto the motherboards, but our servers are running
NO NFS, and none of the systems that are affected are SMP.

Deepak Jain
AiNET


---

fxp2@pci2:5:0:  class=0x020000 card=0x10f08086 chip=0x12298086 rev=0x05
hdr=0x00

fxp0 if the embeded 82559 of an STL2 MB. fxp1 and fxp2 are 82558 on a
dual pro/100B card. System is 4.2-STABLE or -RELEASE.

what works:

- UP kernel with or without NFS activity on all interfaces
- SMP kernel without NFS (client) activity on all interfaces
- SMP kernel _with_ NFS (client) activity on fxp1 or fxp2

What do not work:

- SMP kernel _with_ NFS (client) activity on fxp0

(Symptoms: message "SCB Timeout", NFS related activities locked hard.
All other non NFS related network activities keep running).

Hope this may help.

RN.
IhM



To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-net" in the body of the message


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-net" in the body of the message

[prev in list] [next in list] [prev in thread] [next in thread]