tilley@ukfca1 (Jonathan Tilley) (10/12/90)
Last night I was looking through an old copy of GNU's Bulletin (June 1989) and I read a piece about CMU's MACH kernel. The article said that once the MACH kernel was completely free of ATT stuff and was stable, it was going to be available free. I know that the kernel is now stable - NeXT has it as their operating system, so does anyone know if or when MACH will become free? Jonathan Tilley
chris@mimsy.umd.edu (Chris Torek) (10/13/90)
In article <450@dill.UUCP> tilley@ukfca1 (Jonathan Tilley) writes: >[An] article said that once the MACH kernel was completely free of >ATT stuff and was stable, it was going to be available free. This is the plan. As the old saying goes, `Don't count your chickens before they hatch.' (Things look good; CMU already gives away Mach for free. The problem is that you must first show your AT&T source license.) >I know that the kernel is now stable - NeXT has it as their operating >system, (a) CMU Mach != NeXT Mach; (b) just because NeXT ships it does not mean it is `stable'.% Neither of these is really important, though: >so does anyone know if or when MACH will become free? It already is freely available. The problem is that the Mach kernel is also useless *by itself*. The Mach kernel is a *kernel*, NOT an Operating System; it is more like what IBM sometimes calls a `nucleus'. It does not provide a file system, for instance. No file system == no files on disks == no place to keep data. You can use it with the AT&T and/or Berkeley file systems, of course, but you have to pay money for those. It also does not come with little things like a shell, login program, .... (Note that Berkeley CSRG are working on a free 4BSD as well. The same set of caveats above apply, except that the 4BSD `kernel' is pretty much the whole O/S, not just the bits that cannot be paged.) ----- % See the lyrics to `There Must Be 50 Ways to Crash Your NeXT Box' :-) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 405 2750) Domain: chris@cs.umd.edu Path: uunet!mimsy!chris
shore@mtxinu.COM (Melinda Shore) (10/14/90)
In article <26980@mimsy.umd.edu> chris@mimsy.umd.edu (Chris Torek) writes: >(Things look good; CMU already gives away Mach >for free. The problem is that you must first show your AT&T source >license.) And NFS and others. AFS pulls out pretty easily if you don't want it or can't get a license, but the ufs filesystem is Sun vnode-based and (alas) doesn't come out as cleanly. Drivers, math libraries, debuggers, etc. are also usually under somebody or other's license (different companies hold licenses for different platforms - representative companies include Sun, DEC, IBM, Prime). >(a) CMU Mach != NeXT Mach; Indeed. NeXT Mach is based on Mach 2.0 (no external pagers, for starters) - we're shipping Mach 2.5 with extensions, and CMU is currently working on 3.0. 3.0 is the microkernel Mach, and the one most likely to become "free." This question keeps coming up over and over. It's going to be awhile before there is an entirely license-free Mach-based OS. Even when the microkernel becomes available and FSF provides a complete user-level environment, somebody will still need to provide drivers, filesystems, etc. [Get to work :-)] -- Melinda Shore shore@mtxinu.com mt Xinu ..!uunet!mtxinu.com!shore
gwyn@smoke.BRL.MIL (Doug Gwyn) (10/14/90)
In article <450@dill.UUCP> tilley@ukfca1 (Jonathan Tilley) writes: >when MACH will become free? The Mach kernel does not look like UNIX. It needs another level of functionality wrapped around it first; typically that is hacked from BSD kernel source code, which of course still requires AT&T UNIX 32/V or later source licensing. And UNIX is much more than a UNIX kernel, for example there are hundreds of more or less standard utilities in the UNIX user environment. You would also need replacements for all those. There are certainly people working on independent implementations of all these, but only a handful are ready at present. I wouldn't hold my breath..
henry@zoo.toronto.edu (Henry Spencer) (10/16/90)
In article <450@dill.UUCP> tilley@ukfca1 (Jonathan Tilley) writes: >... said that once the MACH kernel was completely free of ATT stuff and >was stable, it was going to be available free. I know that the kernel is now >stable - NeXT has it as their operating system, so does anyone know if or >when MACH will become free? Note that there is an "and" in the first statement. Mach may be "stable" in some sense of the word -- CMU probably wouldn't agree -- but it is by no means de-AT&Tized yet. And incidentally, even when the kernel is free and clear, there are still just a few :-) utility programs that need work too. The kernel (even using "kernel" in the normal Unix sense) isn't a lot of good by itself. -- "...the i860 is a wonderful source | Henry Spencer at U of Toronto Zoology of thesis topics." --Preston Briggs | henry@zoo.toronto.edu utzoo!henry
greywolf@unisoft.UUCP (The Grey Wolf) (10/24/90)
In article <1369@mtxinu.UUCP> shore@mtxinu.com (Melinda Shore) writes: >In article <26980@mimsy.umd.edu> chris@mimsy.umd.edu (Chris Torek) writes: >>(Things look good; CMU already gives away Mach >>for free. The problem is that you must first show your AT&T source >>license.) > >And NFS and others. AFS pulls out pretty easily if you don't want >it or can't get a license, but the ufs filesystem is Sun vnode-based >and (alas) doesn't come out as cleanly. Drivers, math libraries, >debuggers, etc. are also usually under somebody or other's license >(different companies hold licenses for different platforms - >representative companies include Sun, DEC, IBM, Prime). I keep seeing references to the ufs; who technically owns the concepts behind the original Berkeley FFS? And couldn't some other schmo just as easily come up with a scheme for a "Virtual File System" that wouldn't infringe upon Sun's vnode filesystem? I know the original ufs (read: System V file system) is AT&T property, but why anyone would want to be stuck with 14-byte filenames and u_short inode numbers is beyond me. >-- >Melinda Shore shore@mtxinu.com >mt Xinu ..!uunet!mtxinu.com!shore -- "This is *not* going to work!" "Well, why didn't you say so before?" "I *did* say so before!" ...!{ucbvax,acad,uunet,amdahl,pyramid}!unisoft!greywolf
shore@mtxinu.COM (Melinda Shore) (10/24/90)
In article <3180@unisoft.UUCP> greywolf@unisoft.UUCP (The Grey Wolf) writes: >I keep seeing references to the ufs; who technically owns the concepts >behind the original Berkeley FFS? Berkeley does, but there are still snippets of AT&T code in there, some of which is non-trivial (cache management). I could be wrong, but I do understand that once the buffer management is thrown away and rebuilt using 4.x vm, it will be free. Someone from CSRG or who at least has stronger ties there could give you a clearer idea of what they're doing and when to expect it. >And couldn't some other schmo just as >easily come up with a scheme for a "Virtual File System" that wouldn't >infringe upon Sun's vnode filesystem? Done, at Guelph University working together with CSRG. Dunno about availability right now, but you can probably expect it as part of 4.4. It's a PD implementation of NFS. The vnode format is different, so don't expect it to slide right in. -- Melinda Shore shore@mtxinu.com mt Xinu ..!uunet!mtxinu.com!shore
arnold@audiofax.com (Arnold Robbins) (10/24/90)
|In article <3180@unisoft.UUCP> greywolf@unisoft.UUCP (The Grey Wolf) writes: ||And couldn't some other schmo just as ||easily come up with a scheme for a "Virtual File System" that wouldn't ||infringe upon Sun's vnode filesystem? In article <1379@mtxinu.UUCP> shore@mtxinu.com (Melinda Shore) writes: |Done, at Guelph University working together with CSRG. Dunno about |availability right now, but you can probably expect it as part of |4.4. It's a PD implementation of NFS. The vnode format is different, |so don't expect it to slide right in. Actually, it's available now, in 4.3-Reno. CSRG specifically states, though, that 4.3-Reno is not a "production" release. And you can definitely expect it as part of 4.4. -- Arnold Robbins AudioFAX, Inc. | Laundry increases 2000 Powers Ferry Road, #200 / Marietta, GA. 30067 | exponentially in the INTERNET: arnold@audiofax.com Phone: +1 404 933 7612 | number of children. UUCP: emory!audfax!arnold Fax-box: +1 404 618 4581 | -- Miriam Robbins