3c509 + sb16 bug

3c509 + sb16 bug

[Steve]

[-- Attachment #1: Type: text/plain, Size: 2042 bytes --]

The evidence really speaks for itself:

firstly, I have been running a 2.2.18 kernel system, with a 3c509b and a 
soundblaster 16 (and sundry other hardware).

The soundblaster 16 is on 0x220, irq 5. Its a soundblaster 16 (vibra 16b, '94)
The 3c509 is pnp and detects under 2.2.18 as the following:
eth0: 3c509 at 0x300 tag 1, 10baseT port, address  00 a0 24 75 b7 28, IRQ 10.

Both cards work perfectly, and autodetect without any arguments.

Now:

Here are the interesting bits of the boot of the 2.4.x kernel:

PCI: PCI BIOS revision 2.10 entry at 0xfb4f0,
last bus=1
PCI: Using configuration type 1
PCI: Probing PCI hardware
Unknown bridge resource 0: assuming transparent
Unknown bridge resource 1: assuming transparent
Unknown bridge resource 2: assuming transparent 
PCI: Using IRQ router ALI [10b9/1533] at 00:07.0
isapnp: Scanning for Pnp cards...
isapnp: Card '3Com 3C509B EtherLink III'
isapnp: 1 Plug & Play card detected total

eth0: 3c509 at 0x220, 10baseT port, address  00 20 24 75 b7 28, IRQ 5.
3c509.c:1.16 (2.2) 2/3/98 becker@cesdis.gsfc.nasa.gov.

Soundblaster audio driver Copyright (C) by Hannu Savolainen 1993-1996
sb: No ISAPnP cards found, trying standard ones... 
sb: I/O, IRQ, and DMA are mandatory

NB: PCI stuff was interesting, but I'm not sure if its connected to this
situation.

After bootup, at a random time interval between 10 seconds and 5 minutes the
following error spams the screen:
eth0: infinite loop in interrupt, status 2001

I can only conclude that the kernel has mistaken an ethernet card for a
sound card.

for convience, here is an lspci:
00:00.0 Host bridge: Acer Laboratories Inc. [ALi] M1541 (rev 04)
00:01.0 PCI bridge: Acer Laboratories Inc. [ALi] M5243 (rev 04)
00:07.0 ISA bridge: Acer Laboratories Inc. [ALi] M1533 PCI to ISA Bridge
[Aladdin IV] (rev c3)
00:08.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W [Millennium]
(rev 01)
00:0f.0 IDE interface: Acer Laboratories Inc. [ALi] M5229 IDE (rev c2)

Stephen Thorne

[-- Attachment #2: Type: application/pgp-signature, Size: 232 bytes --]

Re: 3c509 + sb16 bug

[J Sloan]

Perhaps it's cold comfort, but I found long ago that
3c509 and SB don't mix too well, at least in Linux.

ISA devices are somewhat dumb, switching one
of the cards for a PCI version does the trick here.

SB128, SBlive work fine, or you might want to go
to a 10/100 pci ethernet card.

Just my $.02

jjs

Steve wrote:

> The evidence really speaks for itself:
>
> firstly, I have been running a 2.2.18 kernel system, with a 3c509b and a
> soundblaster 16 (and sundry other hardware).
>
> The soundblaster 16 is on 0x220, irq 5. Its a soundblaster 16 (vibra 16b, '94)
> The 3c509 is pnp and detects under 2.2.18 as the following:
> eth0: 3c509 at 0x300 tag 1, 10baseT port, address  00 a0 24 75 b7 28, IRQ 10.
>
> Both cards work perfectly, and autodetect without any arguments.
>
> Now:
>
> Here are the interesting bits of the boot of the 2.4.x kernel:
>
> PCI: PCI BIOS revision 2.10 entry at 0xfb4f0,
> last bus=1
> PCI: Using configuration type 1
> PCI: Probing PCI hardware
> Unknown bridge resource 0: assuming transparent
> Unknown bridge resource 1: assuming transparent
> Unknown bridge resource 2: assuming transparent
> PCI: Using IRQ router ALI [10b9/1533] at 00:07.0
> isapnp: Scanning for Pnp cards...
> isapnp: Card '3Com 3C509B EtherLink III'
> isapnp: 1 Plug & Play card detected total
>
> eth0: 3c509 at 0x220, 10baseT port, address  00 20 24 75 b7 28, IRQ 5.
> 3c509.c:1.16 (2.2) 2/3/98 becker@cesdis.gsfc.nasa.gov.
>
> Soundblaster audio driver Copyright (C) by Hannu Savolainen 1993-1996
> sb: No ISAPnP cards found, trying standard ones...
> sb: I/O, IRQ, and DMA are mandatory
>
> NB: PCI stuff was interesting, but I'm not sure if its connected to this
> situation.
>
> After bootup, at a random time interval between 10 seconds and 5 minutes the
> following error spams the screen:
> eth0: infinite loop in interrupt, status 2001
>
> I can only conclude that the kernel has mistaken an ethernet card for a
> sound card.
>
> for convience, here is an lspci:
> 00:00.0 Host bridge: Acer Laboratories Inc. [ALi] M1541 (rev 04)
> 00:01.0 PCI bridge: Acer Laboratories Inc. [ALi] M5243 (rev 04)
> 00:07.0 ISA bridge: Acer Laboratories Inc. [ALi] M1533 PCI to ISA Bridge
> [Aladdin IV] (rev c3)
> 00:08.0 VGA compatible controller: Matrox Graphics, Inc. MGA 2064W [Millennium]
> (rev 01)
> 00:0f.0 IDE interface: Acer Laboratories Inc. [ALi] M5229 IDE (rev c2)
>
> Stephen Thorne
>
>   ------------------------------------------------------------------------
>    Part 1.2Type: application/pgp-signature

Re: 3c509 + sb16 bug

[Alan Cox]

> Perhaps it's cold comfort, but I found long ago that
> 3c509 and SB don't mix too well, at least in Linux.

I've had them mixed ok before

> ISA devices are somewhat dumb, switching one
> of the cards for a PCI version does the trick here.

I think the problem here thought isnt the 3c509 and SB card, its the kernel
plug and play code. You might want to try building kernels with no PnP support
at all and see how they behave

[FAQ] pls add migrate 2.2 -> 2.4 Re: 3c509 + sb16 bug

[Rob Cermak]

Alan,

You've mentioned this before.  I emailed Steve directly since I had the
same sort of problem with a trio of ethernet cards.  Hardware detection
goes wacky when mixing isapnp userspace tools with the CONFIG_ISAPNP
support in the kernel.  Steve told me: "I'm running debian sid, with
isapnptools package version: 1.21-2"

This should go into the FAQ at http://www.tux.org/lkml/

Here is Alan's quote: Fri, 16 Feb 2001 12:34:03 +0000 (GMT) 
"Dont mix isapnp tools with a 2.4 kernel unless you disable ISA PnP
support in the kernel. It needs to have one or the other do it, not both"

If you use isapnp tools, make sure your config looks like this:

#
# Plug and Play configuration
#
CONFIG_PNP=y
# CONFIG_ISAPNP is not set

Rob

On Fri, 23 Feb 2001, Alan Cox wrote:

> > Perhaps it's cold comfort, but I found long ago that
> > 3c509 and SB don't mix too well, at least in Linux.
> 
> I've had them mixed ok before
> 
> > ISA devices are somewhat dumb, switching one
> > of the cards for a PCI version does the trick here.
> 
> I think the problem here thought isnt the 3c509 and SB card, its the kernel
> plug and play code. You might want to try building kernels with no PnP support
> at all and see how they behave
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/

Re: 3c509 + sb16 bug

[Mark Hindley]

Alan Cox writes:

 > I think the problem here thought isnt the 3c509 and SB card, its the kernel
 > plug and play code. You might want to try building kernels with no PnP support
 > at all and see how they behave
 
I agree. I have been having problems with isapnp reliably finding my
ALS100 since 2.4.1 (I think). I have just modularised isapnp which
means you can at least reload it when it fails to id the card.

Since doing that, I've noticed that the use count for isapnp in lsmod
continues rising eachtime the sb driver is removed. I am hoping to get
some more time this weekend to look at it.


Mark