Re: i82365 PCI-PCMCIA bridge still not working in 2.4.0-test11

Michel Salim (mas118@demeter.cs.york.ac.uk)
Wed, 03 Jan 2001 09:39:09 +0000


Manually inputting i365_base=0x1030 when loading the module results in
the same freeze - nothing works at all, had to do a full hardware reset.
This is on the latest 2.4.0-prerelease...

David, any luck on your side? Sorry for the long absence, been a bit
busy and then I went on holiday.

Any chance it's due to IRQ sharing? Since I've been struggling to use my
PCMCIA modem under kernel 2.2.x too... pcmcia-cs detects it and minicom
can dial using it but nothing else... :( the PCMCIA controller is on IRQ
11, same as my video card. Haven't tried swapping it since the cables
are a mess to reattach :p

A belated Happy New Year,

Michel
David Woodhouse wrote:

> mas118@demeter.cs.york.ac.uk said:
>
>> Thanks for the patch... but it does not quite work. It applies
>> cleanly, but upon booting the patched kernel, the machine freezes
>> completely upon PCMCIA initialisation (it got to the point where the
>> init script said 'Loading modules' then nothing). CTRL+ALT+DEL does
>> not work, either.
>
>
> Hmmm.
>
> mas118@demeter.cs.york.ac.uk said:
>
>> Anyone got a clue? Would love to help debug this if someone would
>> just walk me through it.
>
>
> I think you may have to wait until I next get left alone in the house with
> the opportunity to gut my girlfriend's machine and stick a PCMCIA
> controller in it, unless you're willing to work on this yourself.
>
> What I've done so far is to refer to the current i82365 code from David
> Hinds' standalone package, and put back into Linus' version all the
> non-CardBus CONFIG_PCI code which deals with PCI-PCMCIA bridges. As some
> changes have been made to the pcmcia core in David's version since the
> merge, that's not trivial, although it's quite simple.
>
> I probably missed something - probably a change which isn't in CONFIG_PCI
> ifdefs. Did it print anything before dying? Does it die if you haven't got
> any cards present when it initialises? Does it die if you ensure it's
> polling the socket for status changes and doesn't register the interrupt?
>
> Does it work with the standard driver in -test11 if you specify the correct
> i365_base= parameter as obtained from lspci?
>
> If it does work with the current driver, I'm half inclined to leave it
> alone until 2.5 when we can re-sync the pcmcia core and hopefully then
> it'll 'just work' if we drop in the standalone driver, perhaps with the
> cardbus-specific parts stripped out.
>
> --
> dwmw2

-- 
Michèl Alexandre Salim

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/