Re: FastATA 4000 driver

[Frank Wille <frank%phoenix.owl.de@localhost>]

Radoslaw Kujawa wrote:

> Driver compiles and attaches correctly, but it does not work (yet) -
> kernel hangs during hard drive detection (both on m68k and ppc). I
> think it may be something related to interrupt code. Unfortunately it
> hangs so badly that I'm unable to enter DDB.
>
> Maybe someone will be able to point out obvious mistke in this code?

I didn't analyze the whole code but maybe this is a problem:
    sc->sc_iot.absm = &amiga_bus_stride_4swap;

You are using a bus space with a stride of 4, so the register offsets you
are using would have to be divided by 4.

A bus space with a stride of 0x100 could be useful here, but it doesn't
exist yet.


> Very short description of FastATA 4000 registers is available on
> Elbox's web site in archive with Linux drivers (
> http://www.elbox.com/download/fastata.tar.gz in file README.fastata ).

And maybe this sentence from the Readme is another problem:

"Use polling for find interrupt for devices, see FastATA Int Status
Register."

Does it mean that the FastATA controller does not generate any hardware
interrupts?

-- 
Frank Wille