On Wed, Sep 04, 2019 at 06:21:58PM +0200, Michał Pleban wrote: > Ullrich von Bassewitz wrote: > > > As far as I remember, the coprocessor routines are faulty anyway, and could be removed. > > I've been using them extensively for the past two years and I haven't > noticed anything faulty about them :P You're right. I've used them in the CBM-II support for cc65 myself and they're ok. Seems it's my memory that is faulty :) > > Something strange is that the kernal does not seem to run in bank 15. This is > > mentioned in both HSG issues. No idea why this might be the case. > > Maybe it uses self-modifying code or some other tricks that make it > unable to run in ROM? Or, if it actually calls bank 15 instead of using > LDA (zp),T, then of course it would fall into an endless loop when > placed in bank 15. There's another mentioning of this kernal in HSG1, page 39/40 and this one explicitly says it is not using bank switching to call the kernal in bank 15. So self-modifying code is a possibility. Another interesting question that comes to mind is if this kernal uses it's own zero page or the one in bank 15. Regards Uz -- Ullrich von Bassewitz uz_at_musoftware.de Encrypted email preferred PGP Key-Id: 29D93B10Received on 2020-05-29 22:51:20
Archive generated by hypermail 2.3.0.