From owner-freebsd-current Mon Mar 6 05:57:59 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.10/8.6.6) id FAA05703 for current-outgoing; Mon, 6 Mar 1995 05:57:59 -0800 Received: from sbstark.cs.sunysb.edu (sbstark.cs.sunysb.edu [130.245.1.47]) by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id FAA05691 for ; Mon, 6 Mar 1995 05:57:56 -0800 Received: from starkhome.UUCP (root@localhost) by sbstark.cs.sunysb.edu (8.6.9/8.6.9) with UUCP id IAA18648 for current@freebsd.org; Mon, 6 Mar 1995 08:57:49 -0500 Received: by starkhome.cs.sunysb.edu (8.6.10/1.34) id IAA03953; Mon, 6 Mar 1995 08:56:07 -0500 Date: Mon, 6 Mar 1995 08:56:07 -0500 From: starkhome!gene@sbstark.cs.sunysb.edu (Gene Stark) Message-Id: <199503061356.IAA03953@starkhome.cs.sunysb.edu> To: davidg@Root.COM CC: current@FreeBSD.org, dyson@Root.COM In-reply-to: David Greenman's message of Sun, 05 Mar 1995 18:45:07 -0800 <199503060245.SAA00283@corbin.Root.COM> Subject: Page fault panics during make world in -current Sender: current-owner@FreeBSD.org Precedence: bulk Actually, none of the calls to vm_page_lookup() in vfs_bio.c lock the VM object first. However, after some of them, there are even calls to VMWAIT while the pointer to the page is held (e.g. line 1046). Why are you sure that the page will still be valid when VMWAIT is over? The page fault handler goes to great pains to keep checking that a page lookup is still valid after a sleep might have occurred, so I don't think it can really be safe to lookup a page with the object unlocked and then do arbitrary stuff. - Gene