Re: [apostle-developers] design

[Guido de J. <gu...@us...>]

On Monday 08 October 2001 04:19, mohan kumar wrote:
> --- Guido de Jong <gu...@us...>
>
> wrote:
> >I meant the 2nd address: 0x07fd0008?
>
> oops... i was wrong... 2nd address sure is 0x00b8007

OK, good we cleared that up.

> hi guido,
> I have sorted out things here, and init is running!!!
> But i sure still clearly dunno what was wrong, and
> weather what i did is correct.
> I embedded a few printf and tracked down that the gpf
> was when we try to enable port 4, and preciesly when
> we try to memset(line 53 in port.c) the port bit map,
> it gpf's.

That was to be expected. The memset line modifies the page containing the IO 
port bitmap.

> I commented (without still knowing the piece of code's
> importance :-P ) the memset and everything is working
> fine.

Now that you no longer try to modify the IO bitmap, you will not get the page 
fault. The problem is that you must change the IO port bitmap in order to get 
it working as it should.

> can u help me to understand why its still working
> after the memset was commented.

I think that the IO port bitmap is overlapping some other region of memory. 
Can you run:

	nm -Cgn gemini-nucleus

This will give you a symbol table. Please look up the values for tss_system 
and io_port_bitmap.

> Now that i have got a bit of hand on the code, i will
> start with the aposle code. (nucleus is ur headache as
> of now, i suppose ;-)

We need to fix this porblem with the nucleus before we can start doing 
something with the Apostle sources. Note that we probably require some 
redesign for that too.

Guido