ST00482@auvm.auvm.edu (dan drake) (09/23/90)
I've just been informed that MINIX will not run on an Amiga 3000 (running a 68030). How has AMIGAMINIX been limited to 24 bit pointers and integers? Is it just a case of recompiling it with a compiler that uses 32 bit integers? If this is the case, could I recompile the source under AmigaDos, and have a valid program? I see no reason why not, if the output is just plain vanilla 68000 code or 68030 code. (I'm not even sure that my compiler supports any other processors than the 68000.) Does the problem rise from the caches of the '030? In that case, they can be turned off. How much of the distribution is in assembly? Does this section use instructions that are priviledged in the 68030? Any help would be greatly appreciated, dan drake. st00482@auvm.bitnet (yuck!!) chopin!dan@uunet.uu.net
ast@cs.vu.nl (Andy Tanenbaum) (09/24/90)
In article <90266.095514ST00482@auvm.auvm.edu> ST00482@auvm.auvm.edu (dan drake) writes: >I've just been informed that MINIX will not run on an Amiga 3000 I believe this is correct. A 68030 has an MMU that must be set up properly. None of the 68000 versions of MINIX even know such a gadget exists, let alone setting it up correctly. Even with that done, my intuition says that plenty of other changes will be needed. When Bruce Evans went from 8088 to 80286 protected mode for the IBM version, large number of changes were needed. I think the 68000 to 68030 transition will be equally massive. This gives you a nice project for the coming year. Andy Tanenbaum (ast@cs.vu.nl)
mlord@bwdls58.bnr.ca (Mark Lord) (09/24/90)
In article <90266.095514ST00482@auvm.auvm.edu> ST00482@auvm.auvm.edu (dan drake) writes: >I've just been informed that MINIX will not run on an Amiga 3000 (running >a 68030). How has AMIGAMINIX been limited to 24 bit pointers and integers? >Is it just a case of recompiling it with a compiler that uses 32 bit integers? >If this is the case, could I recompile the source under AmigaDos, and have >a valid program? I see no reason why not, if the output is just plain >vanilla 68000 code or 68030 code. (I'm not even sure that my compiler >supports any other processors than the 68000.) > Does the problem rise from the caches of the '030? In that case, they >can be turned off. How much of the distribution is in assembly? Does >this section use instructions that are priviledged in the 68030? There ought to be little or no difference between an '020 and and '030. HOWEVER, the stack frames for exception ahndling ARE different between the regular 68000 and the '020/030. Try looking at the interrupt handling stuff. -- ___Mark S. Lord__________________________________________ | ..uunet!bnrgate!mlord%bmerh724 | Climb Free Or Die (NH) | | MLORD@BNR.CA Ottawa, Ontario | Personal views only. | |________________________________|________________________|
cyliax@ecn.purdue.edu (Ingo Cyliax) (09/25/90)
In article <7717@star.cs.vu.nl>, ast@cs.vu.nl (Andy Tanenbaum) writes: |> In article <90266.095514ST00482@auvm.auvm.edu> ST00482@auvm.auvm.edu (dan drake) writes: |> >I've just been informed that MINIX will not run on an Amiga 3000 |> I believe this is correct. A 68030 has an MMU that must be set up properly. |> None of the 68000 versions of MINIX even know such a gadget exists, let alone |> setting it up correctly. Even with that done, my intuition says that plenty |> of other changes will be needed. When Bruce Evans went from 8088 to 80286 |> protected mode for the IBM version, large number of changes were needed. |> I think the 68000 to 68030 transition will be equally massive. This gives |> you a nice project for the coming year. I have a port of Minix-ST to a 68020 system with 68851 (PMMU). The 68030 MMU is a subset of the 68851 PMMU. I was careful in selecting the modes and instructions used for the 68851 code, so it shouldn't be that hard to use my code if you're interested. My kernel is Minix-ST user binary compat. and doesn't use shadowing. BTW, it took me 3 months to get the whole thing running. 2 months to get the devices drivers for my system, and 1 month to get the MMU to work. Thanks (Andy) for a robust and modular OS and (Frans) for a (mostly) nice port to the Atari-ST. I say "mostly", since I wasn't too happy about the vector incoded in the top 8 bits of the PC in stmpx.s, but otherwise it was a nice port. Let me know if you're interested. -ingo -- /* Ingo Cyliax ECN, Electrical Engineering Bldg. * * cyliax@ecn.purdue.edu Purdue University, W. Lafayette,IN 47907 * * ing@cc.purdue.edu Work: (317) 494-9523 * * cyliax@pur-ee.UUCP Home: (317) 474-0031 */
KPURCELL@liverpool.ac.uk (Kevin Purcell) (09/25/90)
On Mon, 24 Sep 90 14:34:52 GMT Andy Tanenbaum (ast@NL.VU.CS) said: >In article <90266.095514ST00482@auvm.auvm.edu> ST00482@auvm.auvm.edu (dan drake) > writes: >>I've just been informed that MINIX will not run on an Amiga 3000 >I believe this is correct. A 68030 has an MMU that must be set up properly. >None of the 68000 versions of MINIX even know such a gadget exists, let alone >setting it up correctly. > >Andy Tanenbaum (ast@cs.vu.nl) This is presumably not true of the mac version, as it runs on top of the Mac Toolbox/OS, which hides the true machine. I expect (remebered from an earlier post?) that Minix will run on all Macs from Plus to IIfx. Is this true? -kgp Kevin G. Purcell | kpurcell@liverpool.ac.uk Surface Science, | Liverpool University | Programming the Macintosh is easy if you understand Liverpool L69 3BX | how the Mac works and hard if you don't. -- Dan Allen
archetyp@uxh.cso.uiuc.edu (Joseph R Pickert) (09/25/90)
KPURCELL@liverpool.ac.uk (Kevin Purcell) writes: >>In article <90266.095514ST00482@auvm.auvm.edu> ST00482@auvm.auvm.edu (dan >drake) >> writes: >>>I've just been informed that MINIX will not run on an Amiga 3000 >>I believe this is correct. A 68030 has an MMU that must be set up properly. >>None of the 68000 versions of MINIX even know such a gadget exists, let alone >>setting it up correctly. >> >>Andy Tanenbaum (ast@cs.vu.nl) >This is presumably not true of the mac version, as it runs on top of the >Mac Toolbox/OS, which hides the true machine. I expect (remebered from an >earlier post?) that Minix will run on all Macs from Plus to IIfx. Is this >true? Yes, although there were no claims about the IIfx, since I didn't have access to that machine. IN THEORY, however, it should run on a IIfx. The same can be said about the laptop. Joe Pickert
michel%segin4.segin.fr@relay.prime.com (Michel Pollet) (09/26/90)
In message of 25 Sep 90 08:54:02 GMT, kpurcell said : > This is presumably not true of the mac version, as it runs on top of the > Mac Toolbox/OS, which hides the true machine. I expect (remebered from an > earlier post?) that Minix will run on all Macs from Plus to IIfx. Is this > true? You're right... Many thanks to Apple ingenieers ;-) I've ordered my MacMinix last week with a special note 'PLEASE ACKNONEDGE THIS ORDER' I'm still waiting... Note to 'archetyp' : Please don't switch off your mac please, mails I wrote you bounce ;-) Michel 'BusError' Pollet | "This place is reserved for future michel%segin4.segin.fr@relay.prime.com | use and is intentionaly left blank"
peter@ficc.ferranti.com (Peter da Silva) (09/29/90)
In article <31594@nigel.ee.udel.edu> michel%segin4.segin.fr@relay.prime.com (Michel Pollet) writes: > In message of 25 Sep 90 08:54:02 GMT, kpurcell said : > > This is presumably not true of the mac version, as it runs on top of the > > Mac Toolbox/OS, which hides the true machine. > You're right... Many thanks to Apple ingenieers ;-) No, to the Mac-MINIX implementors who implemented it on top of the toolbox instead of replacing it. They could have done a quick blow-away-the-OS version like the ST and Amiga ports just as easily, with the same restrictions. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
michel%segin4.segin.fr@relay.prime.com (Michel Pollet) (10/01/90)
In message of 28 Sep 90 20:20:12 GMT, peter said : [Stuff Deleted] > > You're right... Many thanks to Apple ingenieers ;-) > > No, to the Mac-MINIX implementors who implemented it on top of the toolbox > instead of replacing it. They could have done a quick blow-away-the-OS > version like the ST and Amiga ports just as easily, with the same > restrictions. If you know where to find necessary informations about macintosh firmware, you're very lucky !!! I NEVER heard about any documentation on the subject !! Apple provide an uniform system that mask differrent hardwares. They don't like low level hackers... :-( So, you can try to make a Mac + minix OR a mac SE minix OR a mac II minix OR .... good luck :-) Michel 'BusError' Pollet | "This place is reserved for future michel%segin4.segin.fr@relay.prime.com | use and is intentionaly left blank"
peter@ficc.ferranti.com (Peter da Silva) (10/02/90)
In article <32014@nigel.ee.udel.edu> michel%segin4.segin.fr@relay.prime.com (Michel Pollet) writes: > So, you can try to make a Mac + minix OR a mac SE minix OR a > mac II minix OR .... good luck :-) Which is basically what the people who did Amiga Minix did. They did a 500/2000 only version that runs on the (basically similar) 500 and 2000, but not on the 2500 or 3000. In terms of Mac generations that's like an SE-only version. -- Peter da Silva. `-_-' +1 713 274 5180. 'U` peter@ferranti.com
guineau@wjg.enet.dec.com (W. John Guineau) (10/02/90)
I recieved Minix last night for the Amiga. I have an A2500/30 so obvoiusly Minix is dead (hung solid after asking for the root disk). I managed toi get the loader converted to compile under SAS/C (Lattice) and am now wondering if I can make the proper changes to the loader so it will null out the mmu and load the kernel into a "safe" area. If this is possible, will Minix boot or will it choke on the 030 exception frames (which *should* be easy to fix)? Will it only choke if an exception occurs or are there other gotchas? -- W. John Guineau guineau@wjg.enet.dec.com Digital Equipment Corporation Marlboro MA. 01752
s4016@mips1.uqam.ca (Vuong*Daniel) (10/02/90)
this user don't exist anymore, so please stop sending mail... -thanks.
HBO043%DJUKFA11.BITNET@cunyvm.cuny.edu (Christoph van Wuellen) (10/03/90)
I do not think the different exception frames are a real problem. As long as you do not have neither an address nor a bus error, these exceptions do not get activated and the exception handler code is not executed. An error-free system should boot with incorrect handlers installed. The only exceptions that must work are the interrupt and syscall exceptions. Perhaps I am wrong since I am not an Amigean, but this is the point of view of an 680x0 user. C.v.W.