Re: Problems with TI 1225 PCI->Cardbus adapter....

On Wed, Jun 18, 2003 at 03:29:55PM -0400, Steven J. Hill wrote:
> David Hinds wrote:
> >
> >Probably, the problem is that the card is behind a PCI-to-PCI bridge
> >that is only forwarding limited ranges of IO and memory transactions.
> >You may be able to deduce what these ranges are from the output of
> >'lspci -v'.  Look for the PCI bridge device and see if there are lines
> >like:
> >
> >>I/O behind bridge: 00002000-00002fff
> >>Memory behind bridge: f0200000-f02fffff
> >
> >Then put these ranges in your /etc/pcmcia/config.opts file.
> >
> 02:04.0 CardBus bridge: Texas Instruments PCI1225 (rev 01)
> 	Subsystem: Actiontec Electronics Inc: Unknown device 0293
> 	Flags: bus master, medium devsel, latency 32, IRQ 16
> 	Memory at fc032000 (32-bit, non-prefetchable) [size=4K]
> 	Bus: primary=02, secondary=03, subordinate=03, sec-latency=32
> 	Memory window 1: fc025000-fc026000 (prefetchable)
> 	I/O window 1: 0000a800-0000ac03
> 	16-bit legacy interface ports at 0001

Err you did not follow my instructions.  What I said, was to look for
the *PCI* bridge device, not the CardBus bridge.  And to look for the
specific lines, "I/O behind bridge" and "Memory behind bridge".

The resources used by the CardBus bridge are not the ones you want to
add to the config file... because they are a mix-up of resources that
the bridge itself needs for its own registers etc, and things that
could be allocated for a PCMCIA card.

You are looking for the PCI-to-PCI bridge with secondary PCI bus
number 2, because that's the bus the CardBus bridge is on.

Based on what sort-of worked, I would put:

  include memory 0xfc000000-0xfcffffff
  include port 0xa000-0xbfff

in /etc/pcmcia/config.opts and comment out all the other ranges that
do not work.

> I am stuck at a 2.4.18 kernel as my production kernel. You are
> suggesting that I use the PCMCIA in the kernel tree and not the
> pcmcia-cs, correct? I am concerned that failure to get the interrupt
> may have to due with the PCI issue you mentioned and routing the
> interrupt properly. Thank you for your help so far. Any more ideas
> off hand? Thanks

I don't recall seeing any failure to get an interrupt in any of your
logs?  I don't think you have any interrupt problems.  In any case,
that has nothing to do with the issue of using the kernel subsystem
versus the pcmcia-cs modules.

-- Dave


-------------------------------------------------------
This SF.Net email is sponsored by: INetU
Attention Web Developers & Consultants: Become An INetU Hosting Partner.
Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php