From: Jouni Malinen (jkmaline_at_cc.hut.fi)
Date: 2002-07-25 15:27:59 UTC
On Thu, Jul 25, 2002 at 03:24:03PM +0200, Joerg Dorchain wrote:
> I recently got a Pheecom WL-PCI card. Unfortunately it did not work with
> the hostap driver. The card shows up in lspci -n as:
> 00:0c.0 Class 0280: 15e8:0131 (rev 01)
Thanks for the patch. Could you please send me more detailed PCI resource info, i.e., output of 'lspci -vvd 15e8:0131'?
> As it the patch is still somewhat ugly but in a "works-for-me" state, I'd
> like others to test and improve it.
I do not have such card, so I will not be able to test this. However, I'm willing to add support for it into the Host AP driver assuming you can test the end result. It looks like at least cor_sreset needs some changes since COR is in different location as in PLX9052.
> + pccard_ioaddr = pci_resource_start(pdev, 2);
> + plx_ioaddr = pci_resource_start(pdev,1);
> + pccard_attr_mem = 0;
So this card card has the pci resources in a bit different order and no pccard attribute memory access..
> + printk(KERN_INFO "TMD adapter at 0x%x, pccard at 0x%x\n", plx_ioaddr, pccard_ioaddr);
> + outb(0x45, plx_ioaddr);
> + mdelay(1000);
> + reg = inb(plx_ioaddr);
That's horribly long busy wait.. Are you sure it is really needed? What is this 0x45 based on (0x41 is OK, but 0x04 part)?
> + /* FIXME */
> + cor_offset = 0; /* no idea what they mean */
> + cor_index = 0; /* no idea what they mean */
These are supposed to be parsed from CIS, but the patch you sent does not seem to indicate any CIS access which is a bit odd.. That's why I would like to see more detailed lspci output for the card. The COR seemed to be available from cards own i/o space (pci resource 1) so it can be in the driver cor_sreset (just need to change the data structures a bit to be ready for this). Data sheet of the card/chip would also be useful.
-- Jouni Malinen PGP id EFC895FA