Path: utzoo!attcan!uunet!snorkelwacker!ira.uka.de!fauern!fauern!csbrod From: csbrod@medusa.informatik.uni-erlangen.de (Claus Brod ) Newsgroups: comp.sys.amiga Subject: Re: Atari1 emulator Message-ID: <3168@medusa.informatik.uni-erlangen.de> Date: 15 Oct 90 17:23:15 GMT References: <9010019@hpfcso.HP.COM> <1990Oct14.203824.8248@ecst.csuchico.edu> <11358@life.ai.mit.edu> <38264@ut-emx.uucp> Organization: CSD, University of Erlangen, W-Germany Lines: 24 greg@ccwf.cc.utexas.edu (Greg Harp) writes: >From what I hear (from an Atari hacker) the Atari TOS uses the line-F >instructions (which are set aside for coprocessors) for some type of system >calls... Therefore anything higher than an 68000 thinks it's passing >instructions to a co-processor. Atari doesn't have a 32-bit bit machine >(except for the new TT which is incompatible for this reason) because of >this. I'm sure that this is because they never planned on _using_ anything >higher than a 68000. Remember that AmigaDOS 1.0 (and I believe 1.1) used >the MOVE SR, EA instruction, which also chokes 010's and higher. There is not a single program I know that breaks on the TT because of that line-F hack that is now missing on the TT. Line F has been used for internal OS purposes only in TOS 1.0 to TOS 1.4. There is no line F in there since TOS 1.6. The problem that the ST emulator faces on a 68010 Amiga is that 68000 and 68010 use different and incompatible exception stack formats. Programs relying on 68000 stack structure (including TOS 1.0, TOS 1.2 and TOS 1.4) will break on a 68010. ---------------------------------------------------------------------- Claus Brod, Am Felsenkeller 2, Things. Take. Time. D-8772 Marktheidenfeld, West Germany (Piet Hein) csbrod@medusa.informatik.uni-erlangen.de ----------------------------------------------------------------------