latest commit

[Oleg Gusev]

On Wednesday 14 July 2004 11:20 pm, Francois-Rene Rideau wrote:
> Looking back at our archives, I realized that
> oleg had solved the flicker by using a different pixclock,
> and that Matan had tried to restore the WinCE pixclock.
> I reverted that part of Matan's patch,
> and the flicker seems to be gone once again. Patch committed.

The error probably lies in the MECR setting, but i am not absolutely
sure. Otherwise, why wince is not flickering ?

>
> As of 2.6, it looks like we are doing something very wrong,
> of the likes of allocating twice the same resource, so that
> the kernel overwrite some userspace stuff. Or maybe not.

The resource management is in a very wrong state now.
There are 10 "sub"devices, and it is necessary to write the
probe/suspend/etc routines (can be just empty stubs),
that grab only the resources that they need. The next step
will be to do the proper interdriver communication, so
the usb driver asks about the int, and not just modifies the
INTPOL1.

> Can one of the gurus out there suggest what may have gone horribly wrong
> since the time when it "worked" without either PCMCIA or USB?

I am not a guru, but can suggest you to compile the sa1101 as a module.
Then you will be forced to compile the module-init-tools :)

 Oleg.