Yet More Patches

[Pavel Roskin]

On Mon, 8 Mar 2004, Russell King wrote:

> For this to happen, it needs to be thoroughly tested, which means I
> need to hear success reports.

I have tested it on Linux 2.6.4-rc2-bk5 (the latest snapshot).

One problem is that 01-debug.diff couldn't be applied to sa11xx_core.c
cleanly.  Since I wasn't testing on ARM, I ignored the failure and applied
the rest.

After I booted the patched kernel, I tested the CVS Orinoco driver.  There
was one error message:

orinoco_cs: Unknown symbol pcmcia_report_error

This can be fixed by using cs_error instead.  In fact, some standalone
drivers already use cs_error (HostAP), while others use
pcmcia_report_error() (linux-wlan-ng, atmelwlandriver).

Of the drivers in the kernel sources, only orinoco_cs.c uses
pcmcia_report_error().  It will need to be fixed in the kernel if
06-cardservices.diff is applied.  Just remove orinoco_cs_error()
implementation and use cs_error() instead.

I'll fix orinoco_cs.c in CVS to use cs_error().  I hope it will be in the
kernel soon, but I cannot promise that it will be there before
06-cardservices.diff is applied.

Actually, I think that pcmcia_report_error() is a much better name for a
publicly exported function than cs_error(), but we probably cannot do much
about it at this point.

The .config file used during testing is available here:
http://www.red-bean.com/~proski/pcmcia/config-test

-- 
Regards,
Pavel Roskin