From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (qmail 44461 invoked by alias); 26 May 2018 18:42:50 -0000 Mailing-List: contact libffi-discuss-help@sourceware.org; run by ezmlm Precedence: bulk List-Id: List-Subscribe: List-Archive: List-Post: List-Help: , Sender: libffi-discuss-owner@sourceware.org Received: (qmail 44425 invoked by uid 89); 26 May 2018 18:42:46 -0000 Authentication-Results: sourceware.org; auth=none X-Virus-Found: No X-Spam-SWARE-Status: No, score=1.8 required=5.0 tests=BAYES_50,KAM_LAZY_DOMAIN_SECURITY,RCVD_IN_DNSWL_NONE autolearn=no version=3.3.2 spammy=Shell, HTo:U*libffi-discuss, Celeron, celeron X-HELO: giraffe.birch.relay.mailchannels.net Received: from giraffe.birch.relay.mailchannels.net (HELO giraffe.birch.relay.mailchannels.net) (23.83.209.69) by sourceware.org (qpsmtpd/0.93/v0.84-503-g423c35a) with ESMTP; Sat, 26 May 2018 18:42:45 +0000 X-Sender-Id: a2hosting|x-authuser|list1@michaelshell.org Received: from relay.mailchannels.net (localhost [127.0.0.1]) by relay.mailchannels.net (Postfix) with ESMTP id F2FC7500ECB; Sat, 26 May 2018 18:42:42 +0000 (UTC) Received: from mi3-ss5.a2hosting.com (unknown [100.96.31.16]) (Authenticated sender: a2hosting) by relay.mailchannels.net (Postfix) with ESMTPA id 96FA3500EBB; Sat, 26 May 2018 18:42:42 +0000 (UTC) X-Sender-Id: a2hosting|x-authuser|list1@michaelshell.org Received: from mi3-ss5.a2hosting.com (mi3-ss5.a2hosting.com [172.19.61.140]) (using TLSv1.2 with cipher DHE-RSA-AES256-GCM-SHA384) by 0.0.0.0:2500 (trex/5.15.2); Sat, 26 May 2018 18:42:42 +0000 X-MC-Relay: Neutral X-MailChannels-SenderId: a2hosting|x-authuser|list1@michaelshell.org X-MailChannels-Auth-Id: a2hosting X-Hysterical-Decisive: 08b085c00502230c_1527360162861_3510402420 X-MC-Loop-Signature: 1527360162861:3946494819 X-MC-Ingress-Time: 1527360162861 Received: from [172.56.27.155] (port=63203 helo=Thor) by mi3-ss5.a2hosting.com with esmtpsa (TLSv1.2:ECDHE-RSA-AES256-GCM-SHA384:256) (Exim 4.89_1) (envelope-from ) id 1fMe9J-00DULH-Gi; Sat, 26 May 2018 14:42:41 -0400 Date: Sat, 26 May 2018 18:42:00 -0000 From: Michael Shell To: libffi-discuss@sourceware.org Cc: green@moxielogic.com Subject: libffi misidentifies certain older Intel CPU capabilites Message-Id: <20180526144420.d153aa01884dae20cbedefc4@michaelshell.org> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit X-AuthUser: list1@michaelshell.org X-SW-Source: 2018/txt/msg00033.txt.bz2 Anthony and other libffi developers/users, Over at the Linux From Scratch site, we've been seeing users get bitten by a bug in libffi where it thinks certain Intel processors have capabilites they actually don't which results in Illegal Instruction errors. Specifically, gcc, or at least the most recent version of it, is employing the vmovq instruction on CPUs that don't support it, e.g., Celeron G530 and G640T: http://lists.linuxfromscratch.org/pipermail/lfs-support/2018-May/051993.html http://lists.linuxfromscratch.org/pipermail/lfs-support/2018-May/052029.html Solved here: http://lists.linuxfromscratch.org/pipermail/lfs-support/2018-May/052008.html So, libffi needs to disable support for avx and aes instructions on Intel CPUs that don't support them. Just thought I'd send you a heads up on this issue. Cheers and thanks, Mike Shell