> You have an interrupt starvation problem. The PCMCIA-HOWTO
> has some information about this situation. When the i82365
> module loads, it reports what interrupts appear to be
> available for PCMCIA devices. It must have decided that irq
> 11 is not usable: maybe it is committed for PCI.
This is where I was assuming the issue was. Why is the
only question remaining.
> To say more, I'd want to see the PCIC probe messages from
> the i82365 driver.
Great, here is what it looks like:
# probe -v
PCI bridge probe: not found.
Intel PCIC probe:
ident(0)=0x83 ident(1)=0x83
i82365sl B step found, 2 sockets.
> You should probably bag the cs_irq setting and instead
> maybe force polling for card insertions, with
> "poll_interval=100". That may free up another interrupt.
Tried this in one of my "itterations" but went back to
your suggestion. No joy.
> The latest PCMCIA release may also do a better job of
> configuring interrupts, than the drivers included in RH6.1.
This is why it took me so long to reply. Since I didn't
have a lot of time invested in RH6.1, I poked around a bit
and have seen mostly praise for RH6.2. So, I jumped in
with both feet... and then looked around.
Tried the same configs with native RH6.2 (pcmcia 3.1.8),
with the same lack of results. So, I then downloaded
the latest pcmcia 3.1.14 bits, config'd, compiled and
installed... no joy.
Either card gets recognized, but only whoever is first
recognized gets configured.
Could this have something to do with the 0x83 base address
being the same on both. One thing I didn't try was forcing
the base address. But then, I'm not sure exactly just
how to do this.
Thanx again for your pointers.
-fred
|