From owner-freebsd-hackers Thu Nov 13 23:09:07 1997 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id XAA04901 for hackers-outgoing; Thu, 13 Nov 1997 23:09:07 -0800 (PST) (envelope-from owner-freebsd-hackers) Received: from implode.root.com (implode.root.com [198.145.90.17]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id XAA04891 for ; Thu, 13 Nov 1997 23:09:00 -0800 (PST) (envelope-from root@implode.root.com) Received: from implode.root.com (localhost [127.0.0.1]) by implode.root.com (8.8.5/8.8.5) with ESMTP id XAA29918; Thu, 13 Nov 1997 23:10:23 -0800 (PST) Message-Id: <199711140710.XAA29918@implode.root.com> To: Charles Henrich cc: freebsd-hackers@FreeBSD.ORG Subject: Re: Pentium lockup fix in FreeBSD In-reply-to: Your message of "Fri, 14 Nov 1997 01:02:45 EST." <199711140602.BAA01225@crh.cl.msu.edu> From: David Greenman Reply-To: dg@root.com Date: Thu, 13 Nov 1997 23:10:23 -0800 Sender: owner-freebsd-hackers@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk >In lists.freebsd.hackers you write: > >>> [ Private email ] > >>Oh? Just between you, me and all of -hackers, eh? :-) If you're going >>to send something in private email then cc'ing hackers doesn't make a >>lot of sense (not that much of this discussion has so far). > >Okay Kids, Go back to your room. Sean, you need to learn some etiquette and >A) not repost private email, and B) jump to conclusions that are contrary to >what many years of FreeBSD operation has shown us. > >As with all security patches, one will be provided in due time, after it is >evaluated critically. In this case I can honestly say "Fooey", yes its a >really NASTY bug for those of us with Pentium systems that have normal users >with accounts on them. However, its not the most security hole to ever be >invented. > >Lets all act like adults here, and let this thread drop right now. Thanks, Charles. I guess it's time that I said something as well. First, it has not been my deliberate intention to ignore Sean's private email regarding his problems trying to come up with a workaround for FreeBSD. I've spent the entire day (and the last several as well) in meetings negotiating a $200K Internet services contract for Walnut Creek CDROM. This has left me stressed out to the point of shaking and quite unable to deal with any email that didn't have quick answers. I did take some time out to look at the proposed patch for Linux. I have little to say about it other than it's not directly applicable to FreeBSD (both in terms of source code structure and algorithmic nature) and I think it is barely okay even for Linux. It seems to seriously violate machine dependant/independant seperation of the kernel sources by putting Pentium specific stuff in the generic VM fault handler, etc, not to mention being incredibly esoteric. When I say esoteric, I mean that in terms of Intel's lack of disclosure about how this fixes the problem more than I do about Linux's guesses. For these reasons and others, I refered to it as "a totally disgusting hack". On the other hand, I'm impressed that the Linux people have been able to get as far as they did with so little information from Intel. As mentioned above, I don't have any time to work on this problem myself, but I do very much appreciate the efforts that Sean and others have made on trying to make a FreeBSD workaround...I just can't be involved myself right now. With above said, I need to make one more important point. I have absolutely no problem with people coming up with a workaround for this, no matter how disgusting, if it does the job and doesn't harm people further. I fully support them making such a workaround available to anyone who wants it. HOWEVER, I do strongly object to the notion that such a hack should be brought into the FreeBSD source code repository at the first sign of life. Doing so would be a poor mode of operation in general and we must avoid this if we stand any chance of maintaining the high quality of our source tree. This is a policy issue that I believe enjoys full support of the entire FreeBSD core team. -DG David Greenman Core-team/Principal Architect, The FreeBSD Project