Re: Give MDA/VGA a try.

[Matan Ziv-Av <ma...@sv...>]

On Sun, 9 Jul 2000, James Simmons wrote:

> 
> > > > The mdacon driver detects a mda card only by checking if memory at
> > > > 0xb0000 is accessible (for write and read). If your vga card does decode
> > > > those addressess, I can think of the following options:
> > > > * enable the commented out tests in mda_detect.
> > >
> > > Short term solution.
> >
> > It might be a good solution. It seems to me that test_mda_b is wrong. It
> > does not restore the register value to what it was, which might be the
> > reason. Try replacing the function with
> 
> I'mgoing to give it a try now. Thanks. If this works really well we can
> see if it will be applied to the standard kernel.

The results I have with this (changing test_mda_b and uncommenting the 
rextra tests):
with a noname mda card in the system - still recognised.
only SiS 6326 (no mda) - not recognised both with extra tests and without.
only Banshee (no mda) mdacon finds an MDA card without the extra tests,
but does not find without it.

So it does seem to help

> 
> > Actually, I am not sure that there is such a test. the 6845 is very
> > simple, and if a vga card emulates it it might emulate it completely.
> 
> Yeah. This might be one of those oplease only compile in if you really
> have a MDA card thing. Thank for the info below.

It is better to have a boot command line option, I think. 

-- 
Matan Ziv-Av.                         ma...@sv...