From owner-freebsd-stable@FreeBSD.ORG Tue Aug 17 21:13:18 2004 Return-Path: Delivered-To: freebsd-stable@freebsd.org Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125]) by hub.freebsd.org (Postfix) with ESMTP id 1705C16A4CE for ; Tue, 17 Aug 2004 21:13:18 +0000 (GMT) Received: from post-24.mail.nl.demon.net (post-24.mail.nl.demon.net [194.159.73.194]) by mx1.FreeBSD.org (Postfix) with ESMTP id A7F3143D49 for ; Tue, 17 Aug 2004 21:13:17 +0000 (GMT) (envelope-from freebsd@schooljan.nl) Received: from schooljan.demon.nl ([82.161.49.211]:25269 helo=cartman) by post-24.mail.nl.demon.net with smtp (Exim 4.34) id 1BxBGe-000Glz-P5 for freebsd-stable@freebsd.org; Tue, 17 Aug 2004 21:13:16 +0000 Message-ID: <001a01c4849f$0f353bf0$6400a8c0@cartman> From: "Henno Schooljan" To: References: <20040817185537.GA6694@root.ucsc.edu> Date: Tue, 17 Aug 2004 23:13:36 +0200 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="iso-8859-1"; reply-type=original Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.2180 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.2180 Subject: Re: key_verify failed for server_host_key solution X-BeenThere: freebsd-stable@freebsd.org X-Mailman-Version: 2.1.1 Precedence: list List-Id: Production branch of FreeBSD source code List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 17 Aug 2004 21:13:18 -0000 > The culprit turns out to be a problem in libcrypto that is tickled > by too aggressive optimization settings. > > When building 4.10, my make.conf included: > > CFLAGS= -O -pipe -funroll-loops -ffast-math > COPTFLAGS= -O -pipe -funroll-loops -ffast-math > > By removing both the unroll-loops and fast-math optimizations, > and rebuilding libcrypto, the problem disappears. I had the same issue. I had this in my make.conf when it went wrong: CFLAGS= -O2 -pipe -funroll-loops and this solved the problem: CFLAGS= -O -pipe So I suspect the unroll loops option to be the cause. I guess I could rebuild with -O2 again without -funroll-loops to prove it... regards, Henno