[PATCHES] 11-pcmciaresource

Russell King rmk+pcmcia at arm.linux.org.uk
Fri Mar 26 20:40:05 GMT 2004


On Fri, Mar 26, 2004 at 10:15:50AM +0100, Dominik Brodowski wrote:
> On Fri, Mar 26, 2004 at 07:49:01AM +0000, Russell King wrote:
> > On Thu, Mar 25, 2004 at 08:23:36PM -0500, Pavel Roskin wrote:
> > > Is my understanding correct that PCMCIA drivers will use resources already
> > > allocated for the bridges (with the exceptions for pcnet_cs and
> > > serial_cs)?  That would be great.
> > 
> > No.  Not yet at least.
> 
> Once the ioresource_t length is resolved, and the handle argument of
> 
> int pcmcia_adjust_resource_info(client_handle_t handle, adjust_t *adj)
> 
> is removed, we can try out a couple of things to adjust the resources the
> PCMCIA subsystem offers to its "clients", e.g. PCMCIA card drivers. Of
> course these are only valid for non-static sockets:
> 
> a) use the iomem and ioports reserved for CardBus cards,

This is something I suspect may be problematical, since we'll effectively
have two bridge windows overlapping each other.  I suspect what happens
will depend on the Cardbus bridge implementation, but we'll see when we
get there.

> b) use the iomem and ioport "behind" the parent PCI bridge, if such a bridge
>    exists,

We do this already; we have to.

> e) trust /proc/io{mem,ports}

<cough> with stupid boot loaders like grub passing the mem= parameter to
the kernel, you'll excuse me if I completely ignore you on that point.
I've seen two many instances here where the Cardbus bridge registers
have been mapped over the top of RAM, and no surprises that it doesn't
work.

-- 
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:  2.6 PCMCIA      - http://pcmcia.arm.linux.org.uk/
                 2.6 Serial core



More information about the linux-pcmcia mailing list