[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MiNT 1.15.1.beta-1 problems



On Sun, Nov 01, 1998 at 08:48:00PM +0100, Jo Even Skarstein wrote:
> Relocating the reset-handler to ST-RAM would solve many problems on
> computers with fastRAM.

I think removing the reset handler completely would be even better.

I just had a look at the code, and I can't say that I understand why it
should be necessary at all.

The TOS re-initializes all those system variables that are restored from the
reset vector / rest_intr() on a warmboot anyway - except shell_p (which does
not seem to be restored), and cookie_p in case of TOS <2.0x (IIRC).

So, when running on TOS 2.0x and above, we need no reset vector at all. 

On TOS <2.0x, it should be sufficient to clear cookie_p *if* the cookie jar
was allocated by MiNT - otherwise, the guy who set up the cookie jar did
already install its own reset handler. In that case, there is no risk in
doing so since TOS <2.0x does not support Fast/Alternate RAM, and so MiNT
will always be loaded to ST-RAM and thus cause no problems.

Comments? Did I miss something?

cu
Michael
-- 
Michael Schwingen, Ahornstrasse 36, 52074 Aachen