From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail5.wrs.com (unknown [64.129.254.146]) by mail.openembedded.org (Postfix) with ESMTP id 36D7875D4A for ; Fri, 30 Oct 2015 21:42:44 +0000 (UTC) Received: from ALA-HCA.corp.ad.wrs.com (ala-hca.corp.ad.wrs.com [147.11.189.40]) by mail5.wrs.com (8.15.2/8.15.2) with ESMTPS id t9ULgh7B006295 (version=TLSv1 cipher=AES128-SHA bits=128 verify=OK); Fri, 30 Oct 2015 14:42:44 -0700 Received: from Marks-MacBook-Pro.local (172.25.36.227) by ALA-HCA.corp.ad.wrs.com (147.11.189.50) with Microsoft SMTP Server id 14.3.248.2; Fri, 30 Oct 2015 14:42:43 -0700 To: Patches and discussions about the oe-core layer , Khem Raj From: Mark Hatle Organization: Wind River Systems Message-ID: <5633E452.9040500@windriver.com> Date: Fri, 30 Oct 2015 16:42:42 -0500 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.11; rv:38.0) Gecko/20100101 Thunderbird/38.3.0 MIME-Version: 1.0 Subject: PowerPC / gcc / prelink X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.12 Precedence: list List-Id: Patches and discussions about the oe-core layer List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 30 Oct 2015 21:42:49 -0000 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit I finally chased down the PowerPC prelink problem. It appears to be within GCC. On most systems GCC (for PPC) is configured with --enable-secureplt. However, the OE-Core version is not. The prelinker assumes that for certain types of relocations that the PPC_GOT table will exist, otherwise it gives up. (It doesn't know how to do the relocations in the old style .bss-got table.) I'm wondering if there is a reason that PPC is defaulting to the old style .bss-got (other then the secureplt was not enabled by default). Is this something we should do in all cases, or should this become a DISTRO_FEATURE like the mips 'mplt' setting? I'm still building systems and running tests, but adding --enable-secureplt to both the gcc-cross-initial, and gcc-cross appear to have resolved the prelinking issues. --Mark