From owner-freebsd-alpha Wed Sep 13 12: 0:20 2000 Delivered-To: freebsd-alpha@freebsd.org Received: from pike.osd.bsdi.com (pike.osd.bsdi.com [204.216.28.222]) by hub.freebsd.org (Postfix) with ESMTP id 94FB537B440 for ; Wed, 13 Sep 2000 12:00:17 -0700 (PDT) Received: (from jhb@localhost) by pike.osd.bsdi.com (8.9.3/8.9.3) id MAA42839 for alpha@FreeBSD.org; Wed, 13 Sep 2000 12:00:10 -0700 (PDT) (envelope-from jhb) From: John Baldwin Message-Id: <200009131900.MAA42839@pike.osd.bsdi.com> Subject: We have a missing Giant mtx_enter To: alpha@FreeBSD.org Date: Wed, 13 Sep 2000 12:00:10 -0700 (PDT) X-Mailer: ELM [version 2.4ME+ PL68 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-alpha@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org Well, before my last round of commits, if you compiled a kernel with INVARIANTS it didn't make it through the device probe before it died when it tried to release the sched_lock when it didn't own it (probably due to the recursion count not being properly saved & restored.) With the stuff I just committed, it can now boot into single user with INVARIANTS, but it fails on an assertion entering multi-user. We basically try to release Giant when we don't actually have it. My guess is that we aren't obtaining Giant somewhere that we are supposed to be. With my interrupt threads stuff, it is still hanging trying to probe the SCSI busses. I'm currently trying to get that working at the moment, but the missing mtx_enter on Giant is something we will need to solve eventually. My guess is it might be related to forking new processes based on where it died (when dhclient forked a child process to presumably dink with ifconfig). -- John Baldwin -- http://www.FreeBSD.org/~jhb/ PGP Key: http://www.cslab.vt.edu/~jobaldwi/pgpkey.asc "Power Users Use the Power to Serve!" - http://www.FreeBSD.org/ To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-alpha" in the body of the message