Re: CardBus testers needed

[Markus W Kilbinger <mk%kilbi.de@localhost>]

>>>>> "David" == David Young <dyoung%pobox.com@localhost> writes:

First: Thanks for your explanations!

    >> pci3 at ppb2 bus 3 pci3: i/o space, memory space enabled cbb0
    >> at pci3 dev 5 function 0: ENE Technology CB1410 CardBus
    >> Controller (rev. 0x01)

    David> Finally, we have the PCI-CardBus bridge.

    David> I'm concerned that your CardBus bridge's memory-mapped
    David> registers are at [0xcffe0000, 0xcfffffff],
    David> [...]
    David> The PCI bridge's window, [0xfe300000, 0xfebfffff], should
    David> enclose the CardBus bridge's registers, [0xcffe0000,
    David> 0xcfffffff], but the regions do not even overlap!

    David> I believe that this is just the kind of misconfiguration
    David> that PCI_ADDR_FIXUP was meant to fix.

I re-played with this (single) option which showed no changes within
these pcictl-output. But: If I've activated all 3 PCI_*_FIXUP's I've
got the following diff:

  $ diff -u pcictl.pci3.dump.-d.5.-f.0..without-pci-addr-fix 
pcictl.pci3.dump.-d.5.-f.0..with-pci-addr-fix
  --- pcictl.pci3.dump.-d.5.-f.0..without-pci-addr-fix    2010-03-12 
08:46:55.000000000 +0100
  +++ pcictl.pci3.dump.-d.5.-f.0..with-pci-addr-fix       2010-03-12 
09:04:18.000000000 +0100
  @@ -38,14 +38,14 @@
       Cache Line Size: 0x10
  
     Type 2 (PCI-CardBus bridge) header:
  -    0x10: 0xcffe0000 0x020000a0 0x40040403 0xfffff000
  +    0x10: 0x80000000 0x020000a0 0x40040403 0xfffff000
       0x20: 0x00000000 0xfffff000 0x00000000 0xfffffffc
       0x30: 0x00000000 0xfffffffc 0x00000000 0x04e3010a
       0x40: 0x14121524 0x00000001
  
       Base address register at 0x10 (CardBus socket/ExCA registers)
         type: 32-bit nonprefetchable memory
  -      base: 0xcffe0000, not sized
  +      base: 0x80000000, not sized
       Capability list pointer: 0xa0
       Secondary status register: 0x0200
         66 MHz capable: off

With no functional changes (== still no card detected/working).

    David> It seems that accesses to [0xcffe0000, 0xcfffffff] could
    David> not possibly go to the CardBus bridge. It is possible that
    David> those accesses go to RAM.

It seems to work under Linux (slackware 12.1 detects and runs some
cards (cardbus and pcmcia) I have fine), so there should be a way to
get it working under NetBSD, too. But how?

Would be some linux kernel output of detecting/attaching these cards
informative?

Markus.