Received: from nic.funet.fi (nic.funet.fi [128.214.248.6])
by odie.barnet.ac.uk (8.8.6/8.8.6) with ESMTP id SAA06840
for <willy@odie.fluff.org>; Mon, 12 Jan 1998 18:46:13 GMT
Received: from vger.rutgers.edu ([128.6.190.2]:36474 "EHLO vger.rutgers.edu" ident: "root") by nic.funet.fi with ESMTP id <6844-150>; Mon, 12 Jan 1998 20:26:12 +0200
Received: by vger.rutgers.edu id <971186-241>; Mon, 12 Jan 1998 13:11:23 -0500
Received: from mail.tin.it ([194.243.154.39]:35023 "EHLO fes3.cs.tin.it" ident: "NO-IDENT-SERVICE") by vger.rutgers.edu with ESMTP id <971296-241>; Mon, 12 Jan 1998 12:59:41 -0500
Received: from riscpc (root@Monza7-133.tin.it [195.103.249.132])
by fes3.cs.tin.it (8.8.4/8.8.4) with SMTP
id SAA14585; Mon, 12 Jan 1998 18:40:06 +0100 (MET)
Date: Mon, 12 Jan 1998 12:01:18 +0100
From: Sergio Monesi <msergio@tin.it>
To: "Neil A. Carson" <neil@causality.com>
Cc: linux-arm@vger.rutgers.edu
Subject: Re: ARMlinux vs. RiscBSD
Message-ID: <48073E4995%msergio@tin.it>
In-Reply-To: <34B9B3BB.41C67EA6@causality.com>
Organization: None, AFAIK...
X-Mailer: Messenger v1.11 for RISC OS
X-Editor: Zap, using ZapEmail =?ISO-8859-1?Q?0.21=DF?=
X-Posting-Agent: RISC OS Newsbase 0.59h
X-Orcpt: rfc822;linux-arm@vger.rutgers.edu
Sender: owner-linux-arm@vger.rutgers.edu
Precedence: bulk
Status: RO
I'm sorry if I seemed to criticize RiscBSD without knowing what I was
speaking about...
However, I think that most Acorn users prefers to have a UNIX tuned for a
low-level workstation, not for a top-level server which the RiscPC obviously
isn't, so Linux seems to be the right OS for us...
In message <34B9B3BB.41C67EA6@causality.com>
"Neil A. Carson" <neil@causality.com> wrote:
> Sergio Monesi wrote:
> > In fact, the second time I run, say, "ls -lR" (or even something more
> > complex with some pipe) on Linux I got an instant output and no disc
> > access while there is no noticeable difference between the first and
> > second run of that on RiscBSD.
> There wouldn't be. As I've said before, these things aren't clean+dry
> that simple---it's different philosophies behind the OSs. What happens,
> is that the FFS filesystem synchronously writes through critical
> meta-data to the filesystem in order to help guarantee a consistent
> state in case the FS crashes due to power failure. This makes the system
> more robust, but slower on certain operations (opening, closing,
> creating and deleting a file---reads and writes of data will be the
> same).
OK, I understand this, but where are the write/create/delete operations in
"ls -lR"? I can't see any risk in buffering the data used by "ls -lR",
whatever happens during its execution, the filesystem integrity cannot be
compromised.
Incidentally, the slowness of the RiscPC internal IDE impacts heavily on
RiscBSD, let alone the SCSI drivers that seems to be terribly slow (from what
I read on the mailing list)... How do you hope to use RiscBSD as a file
server?
> Changing allocated buffers won't help. What you really want, again, is
> FreeBSDs new IOVM system :)
Since you seem to appreciate almost every feature of FreeBSD, why didn't you
(actually the RiscBSD kernel team) port it instead of NetBSD? Would it be
hard to switch to FreeBSD now? What about porting to NetBSD all those nice
peculiarities of FreeBSD?
Cheers,
Sergio
--
Sergio Monesi... \ . . . \ Cracking RC5-64 with a StrongARM RiscPC
