Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!purdue!haven!adm!xadmx!rbj@dsys.ncsl.nist.gov
From: rbj@dsys.ncsl.nist.gov (Root Boy Jim)
Newsgroups: comp.unix.wizards
Subject: What kinds of things would you want in the GNU OS?
Message-ID: <19835@adm.BRL.MIL>
Date: 1 Jun 89 16:01:11 GMT
Sender: news@adm.BRL.MIL
Lines: 22

? From: Peter da Silva <peter@ficc.uu.net>
? Certainly I don't think I've ever seen a problem with utmp.

There is no problem with utmp because noone tries to write in the
same place at the same time. It would seem that small writes within
a block *would* be atomic. The trick is to pick a small enuf number
on the right boundary and hope that this is true in all cases.

Directory writes are atomic because they are done by the kernel.

BTW, I agree with you that threads need not duplicate the stack.
It doesn't seem to make much sense what happens when a thread
returns/exits. However, I would rather see the kernel/library
push the address of thread_exit on the stack. Eliminates one argument.

? Peter da Silva, Xenix Support, Ferranti International Controls Corporation.

? Business: uunet.uu.net!ficc!peter, peter@ficc.uu.net, +1 713 274 5180.
? Personal: ...!texbell!sugar!peter, peter@sugar.hackercorp.com.

	Root Boy Jim is what I am
	Are you what you are or what?