From owner-freebsd-arch@FreeBSD.ORG Wed Nov 5 13:57:59 2014 Return-Path: Delivered-To: arch@FreeBSD.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:1900:2254:206a::19:1]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by hub.freebsd.org (Postfix) with ESMTPS id 4B29378B; Wed, 5 Nov 2014 13:57:59 +0000 (UTC) Received: from kib.kiev.ua (kib.kiev.ua [IPv6:2001:470:d5e7:1::1]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (Client did not present a certificate) by mx1.freebsd.org (Postfix) with ESMTPS id B0C57E29; Wed, 5 Nov 2014 13:57:58 +0000 (UTC) Received: from tom.home (kostik@localhost [127.0.0.1]) by kib.kiev.ua (8.14.9/8.14.9) with ESMTP id sA5DvrlJ085550 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 5 Nov 2014 15:57:53 +0200 (EET) (envelope-from kostikbel@gmail.com) DKIM-Filter: OpenDKIM Filter v2.9.2 kib.kiev.ua sA5DvrlJ085550 Received: (from kostik@localhost) by tom.home (8.14.9/8.14.9/Submit) id sA5Dvph1085549; Wed, 5 Nov 2014 15:57:51 +0200 (EET) (envelope-from kostikbel@gmail.com) X-Authentication-Warning: tom.home: kostik set sender to kostikbel@gmail.com using -f Date: Wed, 5 Nov 2014 15:57:51 +0200 From: Konstantin Belousov To: Baptiste Daroussin Subject: Re: Overlinking in base Message-ID: <20141105135751.GI53947@kib.kiev.ua> References: <20141105113839.GG10388@ivaldir.etoilebsd.net> <20141105125431.GD53947@kib.kiev.ua> <20141105125931.GJ10388@ivaldir.etoilebsd.net> <20141105133029.GH53947@kib.kiev.ua> <20141105134006.GL10388@ivaldir.etoilebsd.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20141105134006.GL10388@ivaldir.etoilebsd.net> User-Agent: Mutt/1.5.23 (2014-03-12) X-Spam-Status: No, score=-2.0 required=5.0 tests=ALL_TRUSTED,BAYES_00, DKIM_ADSP_CUSTOM_MED,FREEMAIL_FROM,NML_ADSP_CUSTOM_MED autolearn=no autolearn_force=no version=3.4.0 X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on tom.home Cc: arch@FreeBSD.org X-BeenThere: freebsd-arch@freebsd.org X-Mailman-Version: 2.1.18-1 Precedence: list List-Id: Discussion related to FreeBSD architecture List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 05 Nov 2014 13:57:59 -0000 On Wed, Nov 05, 2014 at 02:40:07PM +0100, Baptiste Daroussin wrote: > On Wed, Nov 05, 2014 at 03:30:29PM +0200, Konstantin Belousov wrote: > > On Wed, Nov 05, 2014 at 01:59:57PM +0100, Baptiste Daroussin wrote: > > > On Wed, Nov 05, 2014 at 02:54:31PM +0200, Konstantin Belousov wrote: > > > > Could you, please, share the script to see how the overlinking is > > > > checked ? > > > > > > Here you are: > > > https://people.freebsd.org/~bapt/check-links.sh > > > > > > Beware it is dirty :) > > > > > > Run it as check-links.sh nameofthebinary > > > > Ok. It is mostly fine, but you do not account for symbol versions of > > the looked up symbols. There were weird changes, e.g. isnanf story, > > which essentially migrated from libc to libm. I suspect such cases > > are not very important. > > > My proposal to fix this overlinking while still supporting static linkage > Is to change the way we are declaring those dependencies, imho the library > should declare what it needs in case of dynamic linking and what it needs in > case of static linking, the binary Makefile should only list what it requires as > a direct dependency and the framework should do all the magic. > > This can be done via .pc files (and calling pkgconf) or can be done via .mk > files in the library directory. > > In the first case we could have something like: > PCADD= liba libb libc > Which will result in the build system querying though pkgconf: > pkgconf --libs (--static if calling static linkage) > > In the second case we could do it via make(1) > LIBADD= liba libc libc > this will open something like a ${PATHTOTHELIB}/link.mk which will define > DYNAMIC_ADD > STATIC_ADD > > And this could be recursive. > > (note that pkgconf is also recursive as well). Well, there is third option, the linker option --as-needed (I think our old binutils are new enough). Personally, I would prefer make-based solution described above, but --as-needed may be used as the the check. Or, we should have an easy escape to --no-as-needed.