In article <1992Aug27.044210.2753@murdoch.acc.Virginia.EDU>, lfoard@Turing.ORG (Lawrence C. Foard) writes:
|> In article <ROLAND.92Aug24194541@churchy.gnu.ai.mit.edu> linux-ss@gnu.ai.mit.edu writes:
|> >The GNU Project would like to find a volunteer to work on a
|> >single-server for the Mach 3.0 microkernel based on Linux.
|
|> What advantage does Mach have over straight Linux?
|> Is it worth loosing the simplicity and smallness of the Linux
|> kernel for it?
The advantage is Mach 3.0 runs on a large, large number of platforms not just
the ubiquitous PeeCee. Linux running on Mach3.0 would therefore also run on a
wide variety of platforms with 'no' porting.
The port to Mach3.0 is quite simple, rip out all the Linux VM code, Linux drivers, in fact anything that knows about the hardware then refer to the BSD Mach3.0 server and pull out the code you need (emulator, fork/exec, high level driver layer etc.). Not exactly exciting which is possibly why Roland McGrath hasn't done it himself.
In article <1992Aug25.123854.26792@uwm.edu>, rick@ee.uwm.edu (Rick Miller) writes:
> Randall Dean at CMU is finishing up a free BSD-based Unix emulator for
> Mach.
Randy's BSD Single Server has been pulled by CMU - at least until the outcome of the AT&T lawsuit is known, hence the *real* FSF interest.
Roland McGrath is maybe being a bit hopeful with his plans. What makes him thing that if Linux develops into a threat to AT&T it won't also become the victim of a lawsuit ? S/W developers should remember that it isn't neccesarily important for AT&T to win, just by starting the BSD lawsuit they've had sorts of folks running for cover. Other recent spoilers to note are be Apple vs D.R. over GEM or Intel vs AMD over the i386.
