[LinuxBIOS] [PATCH] Improved Intel 82371EB (and others) support

[Uwe Hermann]

See patch.

This is required to make all 440BX based boards in v2 almost fully
supported. Without the patch at least the Super I/O init is never
performed, thus lots of stuff is broken.

Uwe.

[Stefan Reinauer]

* Uwe Hermann  [071130 02:39]:
> Improve support for the Intel 82371FB/SB/AB/EB/MB southbridge(s):
> Implement ISA related support:
> Initialize the RTC
> Enable access to all BIOS regions (but _not_ write access to ROM)
> Enable ISA (not EIO) support
> Without the *_isa.c file, the Super I/O init is never performed
> Improve IDE support:
> Add config option to enable Ultra DMA/33 for each disk
> Add config option to enable legacy IDE port access
> Implement hard reset support
> Implement USB controller support
> Various code cleanups and improvements
>
> The code partially supports southbridges other than the 82371EB (but
> which are very similar), more complete support will follow.
>
> Signed-off-by: Uwe Hermann

Acked-by: Stefan Reinauer

Not sure about the USB controller part. It does nothing, yet, does it?
Maybe it should stay disabled until it has to do something?

[Uwe Hermann]

On Fri, Nov 30, 2007 at 02:50:31AM +0100, Stefan Reinauer wrote:
> Acked-by: Stefan Reinauer

Thanks, r2994.
> Not sure about the USB controller part. It does nothing, yet, does it?

Good question. This needs some more testing, but I _think_ Linux just
enables the USB ports anyway, no matter what we configure here
(even setting it to 'off' in Config.lb seems to be ignored by Linux (?))

That's not a reason to not attempt any init, though. We not only want to
support Linux but also ADLO/Windows, FILO/GRUB (which might require
the USB devices to be setup (?)), custom payloads etc.

Uwe.

[Stefan Reinauer]

* Uwe Hermann  [071130 03:12]:
> > Not sure about the USB controller part. It does nothing, yet, does it?
>
> Good question. This needs some more testing, but I _think_ Linux just
> enables the USB ports anyway, no matter what we configure here
> (even setting it to 'off' in Config.lb seems to be ignored by Linux (?))
>
> That's not a reason to not attempt any init, though. We not only want to
> support Linux but also ADLO/Windows, FILO/GRUB (which might require
> the USB devices to be setup (?)), custom payloads etc.

I assumed it would be possible to disable compiling the object file in
for now since there does not happen any init anyways: usb_init() is
empty..

static void usb_init(struct device *dev)
{
        /* TODO: No special init needed? */
}

It's only a few bytes anyways...

Stefan

[Žiga Mlinar]

On Friday 30 November 2007 02:39:25 Uwe Hermann wrote:
> See patch.
>
> This is required to make all 440BX based boards in v2 almost fully
> supported. Without the patch at least the Super I/O init is never
> performed, thus lots of stuff is broken.
>
> Uwe.

Works here on Azza/PT-6IBD. With this patch ps2 mouse and ps2 keyboard work.  
They didn't before.

Thanks

[Uwe Hermann]

On Fri, Nov 30, 2007 at 04:01:16AM +0100, Žiga Mlinar wrote:
> > This is required to make all 440BX based boards in v2 almost fully
> > supported. Without the patch at least the Super I/O init is never
> > performed, thus lots of stuff is broken.
>
> Works here on Azza/PT-6IBD. With this patch ps2 mouse and ps2 keyboard work.  
> They didn't before.

Great, thanks! Wiki is updated now.

Uwe.