From mboxrd@z Thu Jan 1 00:00:00 1970 From: arno.griffioen@ieee.org (Arno Griffioen) Date: Wed, 13 Jul 2011 15:25:04 +0200 Subject: [TUHS] SVR4 x86 -- Sources In-Reply-To: References: <1310483478.7906.14.camel@localhost.localdomain> <20110712232657.GB31526@bitmover.com> Message-ID: <20110713132504.GY13454@attic.nerdnet.nl> On Tue, Jul 12, 2011 at 08:23:41PM -0400, Jason Stevens wrote: > And then there was that whole SYSV to the Commodore Amiga that SUN tried to > piggy back on.... There had to be a lot more to that then meets the eye. The SVR4 Amiga UNIX implementation was an interesting oddball in itself as CBM was of course 'cheap' and trying to save money on the project, so they licensed the code-base for the 3B2 instead of the original M68k codebase from AT&T.. The M68k codebase was much more expensive to license as I recall from my days working at CBM The result was that the 'port' was a real SVR4 and worked as such, but it lacked the SVR4 M68K ABI support in the kernel, which meant that nearly all available off-the-shelf applications for M68K SVR4 did NOT work on these. Which 'slightly' hampered the rollout and acceptance of these UNIX machine (understatement!). Pity they disbanded the CBM UNIX devel group before it really got started and an 68040 version was never officially released so the whole product fizzled out. I remember that the decision to axe the whole UNIX team inside CBM was really made without anyone knowing about it. Some of the guys were off on visits to CBM offices in other countries when they were told they were fired :( The previous (mostly un-released/internal) SVR3.2 port to the A2500UX'es for 68020+MMU or 68030 (of which I still have one, just no SVR3.2 media..) was AFAIK based on the real M68k codebase. > Not to mention Commodore not letting SUN OEM the Amiga 3000/UX was their > biggest mistake. CBM in their late days were very good at making bad decisions ;) Bye, Arno