From owner-freebsd-hackers Mon Feb 6 21:07:11 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id VAA08157 for hackers-outgoing; Mon, 6 Feb 1995 21:07:11 -0800 Received: from Root.COM (implode.Root.COM [198.145.90.1]) by freefall.cdrom.com (8.6.9/8.6.6) with ESMTP id VAA08149 for ; Mon, 6 Feb 1995 21:07:07 -0800 Received: from corbin.Root.COM (corbin.Root.COM [198.145.90.18]) by Root.COM (8.6.8/8.6.5) with ESMTP id VAA28104; Mon, 6 Feb 1995 21:06:59 -0800 Received: from localhost (localhost [127.0.0.1]) by corbin.Root.COM (8.6.9/8.6.5) with SMTP id VAA00830; Mon, 6 Feb 1995 21:06:58 -0800 Message-Id: <199502070506.VAA00830@corbin.Root.COM> X-Authentication-Warning: corbin.Root.COM: Host localhost didn't use HELO protocol To: terry@cs.weber.edu (Terry Lambert) cc: jmb@kryten.atinc.com, hackers@FreeBSD.org Subject: Re: MIT SHM X11 extensions? (fwd) In-reply-to: Your message of "Mon, 06 Feb 95 21:04:23 MST." <9502070404.AA10165@cs.weber.edu> From: David Greenman Reply-To: davidg@Root.COM Date: Mon, 06 Feb 1995 21:06:58 -0800 Sender: hackers-owner@FreeBSD.org Precedence: bulk >> To protect executing, pageable, binaries from being clobbered...which is >> exactly how it is used. The kernel is not a pageable binary, is not "executed" >> in the traditional sense. > >See the first set of quoted material above -- how are you not tagging the >vnode, yet you know to return ETXTBUSY? > >The only locking I can see is advisory. And it *looks* like the VTEXT *is* >being used. Sure it is, in the kernel for normal files that the kernel execs. >How does this jive with it being OK to clobber your binary? It doesn't. I never said it did in the general case. My reply was about Jonathan saying that you must use 'mv' to install a new kernel - that 'cp'ing a new kernel would cause the system to crash. You then made the assertion that 'cp' can't be used because of the ETXTBUSY error. I said (more or less) that this is simply not true for the kernel binary. >Or did you think I was talking about the VTEXT flag being set on the kernel >vnode? How could I have thought any different??? That **was** what we were talking about. > I *know* that doesn't happen: the kernel isn't opened through the >VFS, it's opened by the boot code. > >Is there something here I am missing? I guess so. -DG