Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ncar!tank!mimsy!chris From: chris@mimsy.UUCP (Chris Torek) Newsgroups: comp.unix.wizards Subject: Re: csh pgrp problem Message-ID: <19143@mimsy.UUCP> Date: 18 Aug 89 17:42:45 GMT References: <712@skye.ed.ac.uk> <920@legato.LEGATO.COM> <184@sunquest.UUCP> Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742 Lines: 23 In article <184@sunquest.UUCP> terry@sunquest.UUCP (Terry Friedrichsen) writes: >And WHY was there no copy-on-write? In "Design and Implementation of >4.3 BSD" (title paraphrased from memory), the authors write that copy-on-write >was considered and abandoned because a microcode bug in one model of VAX >made it questionable that copy-on-write could be reliably implemented. > >They don't identify the model, though, so it's hard to say whether it would >have been better to write off that VAX instead of writing vfork(). As I heard it, the model was the 750, and it had something to do with peculiar addressing modes and stack pages (maybe something like movc3 with an (sp)+ argument :-) ). Instruction restart after a fault did not always work right. If this is correct, it would explain why not `write off that VAX': until recently, monet.berkeley.edu (a VAX-11/750) was *the* BSD development machine. (It also had---and still has---only 2 MB of memory. The new development machine is about 6 VAX MIPS---a 750 is about .6 or .7 ---and has 16 MB. Somehow I suspect 4.4BSD will have a more cavalier attitude towards CPU and memory usage :-/ . . . .) -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7163) Domain: chris@mimsy.umd.edu Path: uunet!mimsy!chris