2.4.19-rc5 from 2.4.18

2.4.19-rc5 from 2.4.18

[Bruce]

I have a problem that has manifested itself since 2.4.19-rcx.

An ADSL driver (1483-2.4.16.o) for the DLINK dsl100d PCI ADSL modem
operational under 2.4.18 has been broken since 2.4.19,
specifically under rc3,rc4 and rc5. (Haven't tested rc2 or rc1)

It appears to get/interpret an incorrect hardware address from the card.

There is a constraint. The driver was originally compiled for ITEX/DLINK
under 2.4.16 and I don't have access to the source.

The only known location of this driver is

http://www.nzadsl.co.nz/software/other/Kernel-2.4.16.tar.gz

and it is not strictly a DLINK driver, more a creation for itex.

For reference I use
inmod -f ./1483-2.4.16.o vpi=[8][0] vci=35 framing=1

then

ifconfig eth1 up (yes there is an eth0 card) This fails with the message

SIOCSIFFLAGS: No such device.


Short of rolling back to 2.4.18 which isn't a problem, any advice anyone?

Cheers,
Bruce Stephens.
Melbourne, Australia.

Re: 2.4.19-rc5 from 2.4.18

[Alan Cox]

On Thu, 2002-08-01 at 11:21, Bruce wrote:
> I have a problem that has manifested itself since 2.4.19-rcx.
> 
> An ADSL driver (1483-2.4.16.o) for the DLINK dsl100d PCI ADSL modem
> operational under 2.4.18 has been broken since 2.4.19,
> specifically under rc3,rc4 and rc5. (Haven't tested rc2 or rc1)
> 
> It appears to get/interpret an incorrect hardware address from the card.
> 
> There is a constraint. The driver was originally compiled for ITEX/DLINK
> under 2.4.16 and I don't have access to the source.

You will need the source to deal with it. Binary only modules have to
match the exact kernel properties, so you effectively threw away all the
freedom you got from Linux by relying on it.