From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mail.pbcl.net ([88.198.119.4] helo=hetzner.pbcl.net) by linuxtogo.org with esmtp (Exim 4.72) (envelope-from ) id 1U5yqJ-00075N-Tg for openembedded-core@lists.openembedded.org; Thu, 14 Feb 2013 14:27:21 +0100 Received: from cpc14-cmbg17-2-0-cust423.5-4.cable.virginmedia.com ([86.14.229.168] helo=[172.30.1.45]) by hetzner.pbcl.net with esmtpsa (TLS1.0:RSA_AES_256_CBC_SHA1:32) (Exim 4.72) (envelope-from ) id 1U5yan-0005wW-0n; Thu, 14 Feb 2013 14:11:13 +0100 Message-ID: <1360847436.4246.122.camel@phil-desktop.brightsign> From: Phil Blundell To: Gary Thomas Date: Thu, 14 Feb 2013 13:10:36 +0000 In-Reply-To: <511CDA28.5040504@mlbassoc.com> References: <511CDA28.5040504@mlbassoc.com> X-Mailer: Evolution 3.4.4-1 Mime-Version: 1.0 Cc: Patches and discussions about the oe-core layer Subject: Re: What do these warnings mean? X-BeenThere: openembedded-core@lists.openembedded.org X-Mailman-Version: 2.1.11 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: Thu, 14 Feb 2013 13:27:21 -0000 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit On Thu, 2013-02-14 at 05:35 -0700, Gary Thomas wrote: > WARNING: QA Issue: ELF binary > '/home/local/p82_soft/tmp/work/cortexa9-vfp-neon-amltd-linux-gnueabi/gst-plugins-bad/0.10.23-r3.ti1.6.4.3/packages-split/gst-plugins-bad-vp8/usr/lib/gstreamer-0.10/libgstvp8.so' > has relocations in .text > WARNING: QA Issue: ELF binary '/home/local/p82_soft/tmp/work/cortexa9-vfp-neon-amltd-linux-gnueabi/libav/0.8.3-r0/packages-split/libavcodec/usr/lib/libavcodec.so.53.35.0' has > relocations in .text > > Exactly what do these mean? What do I do to get rid of them? Generally they mean that some non-PIC code has leaked into a shared library. The most common way for that to happen is for a DSO final link to accidentally pull in some members from a static library. There are a few other possible causes as well, mostly toolchain bugs or limitations (you used to get this under some circumstances when enabling full unwind tables on MIPS for example), but non-PIC code is the most likely one. You can use readelf or objdump to inspect the relocation table for the files in question. In most cases there is a fairly obvious culprit, e.g. an R_ARM_PC24 reloc with an incriminating function name. Having relocations in .text is undesirable because it wastes memory and slows down program startup. They can occasionally cause more severe problems: for example, if were using SELinux then you might find that your security policy didn't allow those binaries to be run at all. (This was the default policy on Fedora for a while; I've no idea if it still is.) p.