[BUG] 2.5.70-mm8 and the 3com NIC driver

[BUG] 2.5.70-mm8 and the 3com NIC driver

[Marek Habersack]

[-- Attachment #1: Type: text/plain, Size: 1351 bytes --]

Hello,

  As mentioned in the subject, the kernel doesn't work with a 3Com 3c905
card. Here's what's shown on the screen and in the logs on both startup and
in intervals during normal system run:

NETDEV WATCHDOG: eth0: transmit timed out
eth0: transmit timed out, tx_status 00 status e601.
diagnostics: net 0cc0 media 8802 dma 0000003b fifo 0000
eth0: Interrupt posted but not delivered -- IRQ blocked by another device?
Flags; bus-master 1, dirty 240(0) current 240(0)
Transmit list 00000000 vs. efce2200.
0: @efce2200  length 8000002a status 0000002a
1: @efce22a0  length 80000050 status 00000050
2: @efce2340  length 8000002a status 0000002a
3: @efce23e0  length 8000002a status 0000002a
4: @efce2480  length 8000002a status 0000002a
5: @efce2520  length 80000036 status 00000036
6: @efce25c0  length 80000036 status 00000036
7: @efce2660  length 80000036 status 00000036
8: @efce2700  length 80000050 status 00000050
9: @efce27a0  length 80000050 status 00000050
10: @efce2840  length 80000050 status 00000050
11: @efce28e0  length 80000050 status 00000050
12: @efce2980  length 80000050 status 00000050
13: @efce2a20  length 80000050 status 00000050
14: @efce2ac0  length 80000050 status 80000050
15: @efce2b60  length 80000050 status 80000050
eth0: Resetting the Tx ring pointer.

All the earlier -mm kernels worked fine in this regard,

marek

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [BUG] 2.5.70-mm8 and the 3com NIC driver

[Thomas Schlichter]

[-- Attachment #1: Type: text/plain, Size: 1911 bytes --]

Hi Marek!

I had a similar problem and for me there were two solutions.

1. Disable ACPI (for example with the 'acpi=off' boot option)
  This was no major impact for me as ACPI never worked here well...

2. Revert the pci-init-ordering-fix from
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.70/2.5.70-mm8/broken-out/pci-init-ordering-fix.patch
  This may be only suboptimal, as this patch seems to fix some problems...

Best regards
   Thomas Schlichter

Marek Habersack wrote:
> Hello,
>
>   As mentioned in the subject, the kernel doesn't work with a 3Com 3c905
> card. Here's what's shown on the screen and in the logs on both startup and
> in intervals during normal system run:
>
> NETDEV WATCHDOG: eth0: transmit timed out
> eth0: transmit timed out, tx_status 00 status e601.
> diagnostics: net 0cc0 media 8802 dma 0000003b fifo 0000
> eth0: Interrupt posted but not delivered -- IRQ blocked by another device?
> Flags; bus-master 1, dirty 240(0) current 240(0)
> Transmit list 00000000 vs. efce2200.
> 0: @efce2200  length 8000002a status 0000002a
> 1: @efce22a0  length 80000050 status 00000050
> 2: @efce2340  length 8000002a status 0000002a
> 3: @efce23e0  length 8000002a status 0000002a
> 4: @efce2480  length 8000002a status 0000002a
> 5: @efce2520  length 80000036 status 00000036
> 6: @efce25c0  length 80000036 status 00000036
> 7: @efce2660  length 80000036 status 00000036
> 8: @efce2700  length 80000050 status 00000050
> 9: @efce27a0  length 80000050 status 00000050
> 10: @efce2840  length 80000050 status 00000050
> 11: @efce28e0  length 80000050 status 00000050
> 12: @efce2980  length 80000050 status 00000050
> 13: @efce2a20  length 80000050 status 00000050
> 14: @efce2ac0  length 80000050 status 80000050
> 15: @efce2b60  length 80000050 status 80000050
> eth0: Resetting the Tx ring pointer.
>
> All the earlier -mm kernels worked fine in this regard,
>
> marek

[-- Attachment #2: signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [BUG] 2.5.70-mm8 and the 3com NIC driver

[Marek Habersack]

[-- Attachment #1: Type: text/plain, Size: 592 bytes --]

On Fri, Jun 13, 2003 at 09:27:02AM +0200, Thomas Schlichter scribbled:
[snip]
> 1. Disable ACPI (for example with the 'acpi=off' boot option)
>   This was no major impact for me as ACPI never worked here well...
> 
> 2. Revert the pci-init-ordering-fix from
> ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.70/2.5.70-mm8/broken-out/pci-init-ordering-fix.patch
>   This may be only suboptimal, as this patch seems to fix some problems...
Thanks for the tips but, fortunately, with -mm9 they were no longer needed -
it works like a charm.

thanks again,

marek

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]