2.6.0-test[56] pcnet32 problems

2.6.0-test[56] pcnet32 problems

[Danny ter Haar]

Got an cyrix box with built in ethernet & 2 extra ethernet cards
as a firewall at home with a dsl line (connected to eth0)

pcnet32.c:v1.27b 01.10.2002 tsbogend@alpha.franken.de
PCI: Found IRQ 11 for device 0000:00:0f.0
IRQ routing conflict for 0000:00:0f.0, have irq 9, want irq 11
pcnet32: PCnet/FAST III 79C973 at 0xfca0, warning: CSR address invalid,
    using instead PROM address of 00 00 e2 24 41 1d assigned IRQ 9.
eth0: registered as PCnet/FAST III 79C973
pcnet32: 1 cards_found.

Sometimes (even during low traffic) eth0 simply locks up:
In dmesg i see:
kernel: eth0: Bus master arbitration failure, status 88f3.

rmmod pcnet32 results in kernel-panic.
Only reboot works.
tried:
acpi == disabled
pci=noacpi

Another weird thing is output of ifconfig eth0:

eth0      Link encap:Ethernet  HWaddr 00:00:E2:24:41:1D  
          inet addr:195.64.94.48  Bcast:195.64.94.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:30345 errors:0 dropped:0 overruns:0 frame:0
          TX packets:0 errors:16042 dropped:0 overruns:0 carrier:16042
          collisions:0 txqueuelen:1000 
          RX bytes:38404293 (36.6 MiB)  TX bytes:1298028 (1.2 MiB)
          Interrupt:9 Base address:0xfca0 

0 packets transmitted , all errors ans carrier faults ?
Still it works! (this could be counters that are wrong)

kernel config, lspci -v and dmesg output available at:

http://dth.net/kernel/

Any help/suggestions/hints appreciated.

Danny
-- 
 /"\                        | Dying is to be avoided because
 \ /  ASCII RIBBON CAMPAIGN | it can ruin your whole career 
  X   against HTML MAIL     | 
 / \  and POSTINGS          | - Bob Hope

Re: 2.6.0-test[56] pcnet32 problems

[Vojtech Pavlik]

On Tue, Sep 30, 2003 at 03:13:11AM +0000, Danny ter Haar wrote:
> Got an cyrix box with built in ethernet & 2 extra ethernet cards
> as a firewall at home with a dsl line (connected to eth0)
> 
> pcnet32.c:v1.27b 01.10.2002 tsbogend@alpha.franken.de
> PCI: Found IRQ 11 for device 0000:00:0f.0
> IRQ routing conflict for 0000:00:0f.0, have irq 9, want irq 11
> pcnet32: PCnet/FAST III 79C973 at 0xfca0, warning: CSR address invalid,
>     using instead PROM address of 00 00 e2 24 41 1d assigned IRQ 9.
> eth0: registered as PCnet/FAST III 79C973
> pcnet32: 1 cards_found.
> 
> Sometimes (even during low traffic) eth0 simply locks up:
> In dmesg i see:
> kernel: eth0: Bus master arbitration failure, status 88f3.
> 
> rmmod pcnet32 results in kernel-panic.
> Only reboot works.
> tried:
> acpi == disabled
> pci=noacpi
> 
> Another weird thing is output of ifconfig eth0:
> 
> eth0      Link encap:Ethernet  HWaddr 00:00:E2:24:41:1D  
>           inet addr:195.64.94.48  Bcast:195.64.94.255  Mask:255.255.255.0
>           UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
>           RX packets:30345 errors:0 dropped:0 overruns:0 frame:0
>           TX packets:0 errors:16042 dropped:0 overruns:0 carrier:16042
>           collisions:0 txqueuelen:1000 
>           RX bytes:38404293 (36.6 MiB)  TX bytes:1298028 (1.2 MiB)
>           Interrupt:9 Base address:0xfca0 
> 
> 0 packets transmitted , all errors ans carrier faults ?
> Still it works! (this could be counters that are wrong)
> 
> kernel config, lspci -v and dmesg output available at:
> 
> http://dth.net/kernel/
> 
> Any help/suggestions/hints appreciated.

Upgrade ifconfig?

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR

Re: 2.6.0-test[56] pcnet32 problems

[Barry K. Nathan]

On Tue, Sep 30, 2003 at 03:13:11AM +0000, Danny ter Haar wrote:
> Sometimes (even during low traffic) eth0 simply locks up:
> In dmesg i see:
> kernel: eth0: Bus master arbitration failure, status 88f3.

Hmmm... I'm not seeing this at all.

If going back to an older kernel gets rid of this problem, perhaps it
could be an IRQ routing problem or something like that. That's just a
guess, however. (If it doesn't, I would suspect hardware failure, perhaps
your motherboard; motherboard failures are the only time I've seen this
message happen with pcnet32 cards.)

> rmmod pcnet32 results in kernel-panic.

I never tried that because my primary pcnet32-using machine runs
monolithic kernels...

-Barry K. Nathan <barryn@pobox.com>

Re: 2.6.0-test[56] pcnet32 problems

[Danny ter Haar]

Vojtech Pavlik  <vojtech@suse.cz> wrote:
>Upgrade ifconfig?

Latest unstable debian.
The other 2 network cards (eepro/8139too) do work/show ok.

Danny
-- 
 /"\                        | Dying is to be avoided because
 \ /  ASCII RIBBON CAMPAIGN | it can ruin your whole career 
  X   against HTML MAIL     | 
 / \  and POSTINGS          | - Bob Hope

Re: 2.6.0-test[56] pcnet32 problems

[Danny ter Haar]

Barry K. Nathan <barryn@pobox.com> wrote:
>If going back to an older kernel gets rid of this problem, perhaps it
>could be an IRQ routing problem or something like that.

I will try first a monolithic kernel
If that doesn't work i'll change it for a 2.4.23-preXX

>That's just a
>guess, however. (If it doesn't, I would suspect hardware failure, perhaps
>your motherboard; motherboard failures are the only time I've seen this
>message happen with pcnet32 cards.)

Last weekend i exchanged my old (but stable) pentium 75Mhz to the
cyrix setup which happens to be a prototype of a siemens settop box.
(ir keyboard, tv-out and DVB card).
It could be faulty motherboard but it doesn't happen if transfer huge
amounts of data over the other two ethernet cards.

Thank for your answer(s)

Danny
-- 
 /"\                        | Dying is to be avoided because
 \ /  ASCII RIBBON CAMPAIGN | it can ruin your whole career 
  X   against HTML MAIL     | 
 / \  and POSTINGS          | - Bob Hope