Return-Path: Received: by vger.rutgers.edu id <153682-31090>; Thu, 17 Dec 1998 02:12:10 -0500 Received: from tiger.coe.missouri.edu ([128.206.158.20]:11972 "EHLO COE.Missouri.EDU" ident: "IDENT-NONSENSE") by vger.rutgers.edu with ESMTP id <153881-31090>; Thu, 17 Dec 1998 02:10:54 -0500 Date: Thu, 17 Dec 1998 01:49:53 -0600 From: Tymm Twillman To: carguin@iname.com cc: linux-kernel@vger.rutgers.edu Subject: Re: process checkpointing In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-kernel@vger.rutgers.edu Content-Length: 2005 Lines: 53 Have you looked into libckpt? http://www.cs.utk.edu/~plank/papers/USENIX-95W.html -Tymm On Wed, 16 Dec 1998 carguin@iname.com wrote: > On Tue, 15 Dec 1998, Guest wrote: > > > are there any patches to support process checkpointin on x86 ? > > I'm currently working on this, but I have virtually nothing written yet. > Since you mention it, I'll take this opporunity to seek input from the > rest of the community. > > Actually, I'm surprised I haven't seen anything from anyone else yet. The > basic problem (ignoring, for now, things like open files) appears to be > fairly straightforward. I'm fairly certain I can implementing the dumping > of the process in just userspace (using ptrace and /proc), and I half > suspect that the reloading can also be done in userspace. > > Mind you, this is a problem, since I am doing this for my final in a class > about the Linux kernel. I kinda have to modify the kernel, and I'm not > sure I really need to. > > Of course, to actually be useful gets more complicated. Even ignoring > things like IPC it's complicated. You really have to handle pipes, or else > you have no stdin/stdout. Those are usually considered of some importance > to programs. And security.... I haven't really thought too hard about it, > but it seems to me that there are some definite security issues to > consider here. > > -- > Chris Arguin | "...All we had were Zeros and Ones -- And > CArguin@iname.com | sometimes we didn't even have Ones." > +--------------+ - Dilbert, by Scott Adams > http://leonardo.sr.unh.edu/arguin/home.html | > > > > > - > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in > the body of a message to majordomo@vger.rutgers.edu > Please read the FAQ at http://www.tux.org/lkml/ > - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/