Path: utzoo!utgpu!water!watmath!clyde!att!rutgers!njin!princeton!idacrd!uunet!mcvax!unido!tub!uwbln!ckl
From: ckl@uwbln.UUCP (Christoph Kuenkel)
Newsgroups: comp.unix.wizards
Subject: P/V using SysV semop(2)
Keywords: fifo or lifo ?
Message-ID: <1001@uwbull.uwbln.UUCP>
Date: 24 Aug 88 13:45:43 GMT
Distribution: comp
Organization: UniWare GmbH, Berlin
Lines: 14

I just tried to implement Disjkstra's P/V semaphore operations using
system V's semops.  what i found on three different SysV R.2 ports was
that mutual exclusion was ok but the order of entrance into the critical
section enclosed into P/V was ``by random'' (probably by order of kernel
process slot, which is roughly equivalent to ``by random'').

I implemented it using an array of 1 semaphore and a value of -1 for
sem_op.  

As far as i understand, the documentations does not specify anything at all.
i cant believe it. is it impossible to implement P/V without starvation?
-- 
Christoph Kuenkel/UniWare GmbH       Kantstr. 152, 1000 Berlin 12, West Germany
ck@tub.BITNET                ckl@uwbln             {unido,tmpmbx,tub}!uwbln!ckl