From owner-freebsd-hackers Fri Dec 6 9:29:28 2002 Delivered-To: freebsd-hackers@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 247DE37B401 for ; Fri, 6 Dec 2002 09:29:27 -0800 (PST) Received: from HAL9000.homeunix.com (12-232-220-15.client.attbi.com [12.232.220.15]) by mx1.FreeBSD.org (Postfix) with ESMTP id 668E843EA9 for ; Fri, 6 Dec 2002 09:29:26 -0800 (PST) (envelope-from dschultz@uclink.Berkeley.EDU) Received: from HAL9000.homeunix.com (localhost [127.0.0.1]) by HAL9000.homeunix.com (8.12.6/8.12.5) with ESMTP id gB6HTOJt018392; Fri, 6 Dec 2002 09:29:24 -0800 (PST) (envelope-from dschultz@uclink.Berkeley.EDU) Received: (from das@localhost) by HAL9000.homeunix.com (8.12.6/8.12.5/Submit) id gB6HTOsL018391; Fri, 6 Dec 2002 09:29:24 -0800 (PST) (envelope-from dschultz@uclink.Berkeley.EDU) Date: Fri, 6 Dec 2002 09:29:24 -0800 From: David Schultz To: Gary Thorpe Cc: freebsd-hackers@freebsd.org Subject: Re: maxusers and random system freezes Message-ID: <20021206172924.GA18216@HAL9000.homeunix.com> Mail-Followup-To: Gary Thorpe , freebsd-hackers@freebsd.org References: <20021206123440.GA16544@HAL9000.homeunix.com> <20021206162935.14339.qmail@web41213.mail.yahoo.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20021206162935.14339.qmail@web41213.mail.yahoo.com> Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Thus spake Gary Thorpe : > I have a question: does the entire KVA *have* to be mapped into the > each process's address space? How much of the KVA does a process need > to communicate with the kernel effectively? No, it doesn't have to be that way. An alternative organization is one in which the kernel has its own address space, and some operating systems have done that. Unfortunately, that approach requires special tricks to copy data between user processes and the kernel, and since such copies are a bottleneck anyway, it's undesirable to make them even slower. Some architectures provide special instructions that allow the kernel to move data between address spaces relatively efficiently, but there's still a penalty. On the other end of the spectrum, people are designing operating systems for 64-bit architectures where everything fits into one address space (e.g. `Mungi'), minimizing context switch overhead and simplifying sharing. The idea isn't really new, however. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message