Return to BSD News archive
Xref: sserve comp.unix.solaris:627 comp.unix.bsd:8318
Newsgroups: comp.unix.solaris,comp.unix.bsd
Path: sserve!manuel.anu.edu.au!munnari.oz.au!uunet!think.com!ames!agate!doc.ic.ac.uk!syma!tauk0
From: tauk0@syma.sussex.ac.uk (Fredrik Staxang)
Subject: Re: Solaris 1.1 vs. Solaris 2.0 (BSD vs AT&T)
Message-ID: <1992Nov19.165146.22006@syma.sussex.ac.uk>
Organization: University of Sussex
References: <1992Nov13.232053.7061@sjsumcs.sjsu.edu> <1992Nov15.014513.28154@nobeltech.se> <1992Nov15.035135.15514@ra.msstate.edu>
Date: Thu, 19 Nov 1992 16:51:46 GMT
Lines: 23
In article <1992Nov15.035135.15514@ra.msstate.edu> fwp@CC.MsState.Edu (Frank Peters) writes:
>scales better to the larger systems sun is beginning to offer (there is
>a reason that just about every large scale multiprocessor system vendor
>from Sequent to Cray uses SYSV).
^^^^^^^
Sequent uses a BSD kernel, very thinly disguised. Eg. the following does not
work in setuid program.
real = getuid() ;
eff = geteuid() ;
setuid(real) ;
system("..") ;
setuid(eff) ; /* This fails, whereas in SysV it works */
So instead you use setreuid(), which is a BSD call. There is also som other
things that work the BSD way, but the real giveaway is when you start to look
at the VM system.
/Fredrik