Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 14 Nov 1997 08:32:40 -0500
From:      "Dan O'Brien" <dmobrien@lucent.com>
To:        hackers@freebsd.org
Subject:   [Fwd: Pentium bug: FreeBSD workaround ?]
Message-ID:  <346C52F8.ADF4D658@lucent.com>

next in thread | raw e-mail | index | archive | help
This is a multi-part message in MIME format.
--------------B65F27E85DB1D044A31810D1
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Just for the record.  Here is the posting.  His reasoning is
rational and clear.   I support his "wait and see" attitude.

Thanks,

Dan O'Brien  (dmobrien@lucent.com)
Lucent Technologies, Bell Labs Innovations, Columbus OH 43213
Internal URL:<http://www.netminder.lucent.com/~dob>;
External URL:<http://www.bazzle.com/~dob>;
Work: 614-860-2392   Home:  614-927-2178
Fax:  614-868-3810   Home2: 614-927-2955
------------------------------------------------------
--------------B65F27E85DB1D044A31810D1
Content-Type: message/rfc822
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

Path: nntphub.cb.lucent.com!uunet!in5.uu.net!nntprelay.mathworks.com!news1.best.com!nntp2.ba.best.com!not-for-mail
From: "Jordan K. Hubbard" <jkh@FreeBSD.org>
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: Pentium bug: FreeBSD workaround ?
Date: Thu, 13 Nov 1997 08:13:31 -0800
Organization: Walnut Creek CDROM
Message-ID: <346B272B.167EB0E7@FreeBSD.org>
References: <64ckdr$2gg_002@cris.com> <64diac$k7t$3@otis.netspace.net.au> <346ADA14.C167CFD7@domgen.com> <mo4btzpdkg4.fsf@pooh.tky.hut.fi>
NNTP-Posting-Host: time.cdrom.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Trace: 879437616 7114 jkh  206.86.0.12
X-Mailer: Mozilla 3.03 (X11; I; FreeBSD 2.2.5-STABLE i386)
To: tri@iki.fi
Xref: nntphub.cb.lucent.com comp.unix.bsd.freebsd.misc:54115

[posted and mailed]

Timo J Rinne wrote:
> You understood wrong.  The patch makes idt entries 0-6 to point to
> unmapped space.  Then pagefault handler resolves, whether it's really
> a pagefault or should it forward call to the corresponding exception
> handler.
> 
> The patch should go into machdep.c and locore.s or trap.c in
> /sys/i386/i386 and it of course should be optional.
> 
> Is someone working on this or should I do it?

Well, several things to note here:

1. BSDI would appear to have *withdrawn* the patch which they released
to great fanfare yesterday.  This is disturbing and merits closer
attention before we make any moves ourselves.

2. Our principal architect has reviewed the Linux patch that someone
forwarded to us and he considers it, to use his own words, "a totally
disgusting hack."

Let's make sure that in our haste to deal with this latest political
football, we don't come up with a cure that's worse than the disease. I
also hasten to note that I've yet to hear *any* reports of serious DoS
attacks stemming from this bug, this being in all likelyhood another
FDIV-type situation where people who wouldn't be affected in a million
years by the bug are nonetheless steaming about it as if Intel had shot
their dog and raped their mothers.

I think folks need to put this in perspective and stop this silly
scare-mongering over it - this is hardly a critical emergency and I
suspect that many of the folks who are racing to implement a solution
simply so that they can claim the dubious distinction of "solving it
first" are only going to end up going back over their fixes later,
perhaps to do as BSDI has done in throwing out the first attempt.

Let's at least wait for Intel to release full details of their proposed
work-around in a day or two as they've promised, eh?  Again, if I were
hearing anguished cries from the ISPs about this then it would be a
different matter, but I haven't heard so much as a squeak from them and
this would all appear to be simply another chicken little episode, with
various people running around flapping and squawking simply because it
makes them feel important to be given the opportunity to run around and
flap about something. :-)

-- 
- Jordan Hubbard
  FreeBSD core team / Walnut Creek CDROM.

--------------B65F27E85DB1D044A31810D1--




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?346C52F8.ADF4D658>