Hi,

I have a netgear PCMCIA card (FA511) which gives out a wrong device id upon connection to the socket.

Thus the correct driver (tulip_cb.o) is not loaded. I tried to compile and install the new PCMCIA-3.2.8 package and modified the tulip_cb.c accordingly to check the new id.

But this way it does not load the yenta_socket.o due to unresolved issues after the new pcmcia_core.o.

It confuses me that yenta_socket.o is dependent on pcmcia_core.o but yenta_socket is not included in the PCMCIA package.

Please guide me on how to resolve this unresolved issue.

If there is someother way to make RH9 recognise another id than the correct one please let me know.

The correct id isdevice id:vendor id) (x1317:x1985) cardmgr reads (xffff:1985) the sub vendor and device ids are correctly provided.

Thanks for u're help. Really appreciate it. I have been trying this for over two weeks now.

Thanks again!

RedHat 9
Kernel: 2.4.20-8
PCMCIA: 3.2.8
ThinkPad 600 (PentiumII)

yenta_socket is part of the kernel now, all of the driver pcmcia support has been in the kernel since 2.4.3.

pcmcia-cs still exists, but if you compile your kernel for in-kernel pcmcia then when you compile pcmcia-cs all you're getting out of that is the /etc/pcmcia/ tree and the userland binaries like cardmgr.

You've got a card that is newer then the ancient RedHat 9.0 kernel, that's all. If you can re-install the RH pcmcia package, (so that userland is straight) and then hand compile a newer 2.4.27-ish say thereabouts, then the module for tulip_cb should have the right pci id.

Cheers,

Finegan

Thanks Finegan,

I really appreciate u're reply. I am just curious about a few things.

1. When i put the card in the slot it does not give the (1317:1985) id to the hotplug.pci-agent instead it gives (ffff:1985)

This made me feel that probably the device id was wrong in my card. For everyone else on the net got (1317:1985) as
is documented in the tulip.o

If u still feel the contrary please let me know u're suggestions and i'll go ahead and do it.

2. There seems to be some kernel parameter like PCI_ANY. How does that impact the id?

Thanks again.