From owner-freebsd-current Fri Jul 6 21: 1:18 2001 Delivered-To: freebsd-current@freebsd.org Received: from bunrab.catwhisker.org (adsl-63-193-123-122.dsl.snfc21.pacbell.net [63.193.123.122]) by hub.freebsd.org (Postfix) with ESMTP id 04C1837B403 for ; Fri, 6 Jul 2001 21:01:14 -0700 (PDT) (envelope-from david@catwhisker.org) Received: (from david@localhost) by bunrab.catwhisker.org (8.11.4/8.11.4) id f6741B922471; Fri, 6 Jul 2001 21:01:11 -0700 (PDT) Date: Fri, 6 Jul 2001 21:01:11 -0700 (PDT) From: David Wolfskill Message-Id: <200107070401.f6741B922471@bunrab.catwhisker.org> To: current@freebsd.org, imp@harmony.village.org Subject: Re: Interruptable hang starting init in today's -CURRENT In-Reply-To: <200107060635.f666ZkJ56636@harmony.village.org> Sender: owner-freebsd-current@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG >Date: Fri, 06 Jul 2001 00:35:46 -0600 >From: Warner Losh >I saw something similar at usenix with my old, hacked kernel config >file, but not with GENERIC that I just booted. Well, I appreciate that hint... so I looked for differences between GENERIC and my kernel config (LAPTOP_30W)... but nothing I tried worked. So I built a brand-new kernel from GENERIC (which is rev. 1.313). And it exhibited the same symptom (with one rather subtle difference, below). As N.Dudorov (nnd@mail.nsk.ru) pointed out, what is happening is that the "sysctl -a" in /etc/rc (as part of the "entropy harvesting") is failing to terminate. Whether with GENERIC or LAPTOP_30W, the last entry shown from "sysctl -a" is net.inet.accf.unloadable: 0 The subtle difference is with "sysctl -N -a". With LAPTOP_30W, what gets written after the above is a stream of net.inet.accf.373 while with GENERIC, it's net.inet.accf.372 Now, I confess a relative lack of knowledge about how sysctls are defined, but I rather suspect that the "net.inet.accf.###" is bogus. I did spend some time with a hand-compiled copy of sysctl (to turn on debugging) inside gdb, but I can't claim to be much the wiser for the experience. I'll be happy to poke around some more, but I'd appreciate a clue as to a source of additional information about sysctls. As a reminder, I've been re-syncing my local copy of the CVS repository daily, thus: CVSup begin from cvsup14.freebsd.org at Mon Jul 2 03:47:01 PDT 2001 CVSup ended from cvsup14.freebsd.org at Mon Jul 2 03:52:46 PDT 2001 CVSup begin from cvsup14.freebsd.org at Tue Jul 3 03:47:01 PDT 2001 CVSup ended from cvsup14.freebsd.org at Tue Jul 3 03:52:47 PDT 2001 CVSup begin from cvsup14.freebsd.org at Wed Jul 4 03:47:01 PDT 2001 CVSup ended from cvsup14.freebsd.org at Wed Jul 4 03:53:33 PDT 2001 CVSup begin from cvsup14.freebsd.org at Thu Jul 5 03:47:00 PDT 2001 CVSup ended from cvsup14.freebsd.org at Thu Jul 5 03:53:22 PDT 2001 CVSup begin from cvsup14.freebsd.org at Fri Jul 6 03:47:01 PDT 2001 CVSup ended from cvsup14.freebsd.org at Fri Jul 6 03:52:59 PDT 2001 and tracking -CURRENT (and -STABLE, for that matter) daily. I first noticed this problem with the -CURRENT build I did on 04 July. And booting from a -CURRENT kernel I stashed away (based on the same kernel config file that I had been using for the last few weeks) as of 16 June did not have the problem. Thanks, david -- David H. Wolfskill david@catwhisker.org As a computing professional, I believe it would be unethical for me to advise, recommend, or support the use (save possibly for personal amusement) of any product that is or depends on any Microsoft product. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-current" in the body of the message