From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Wed, 29 Mar 2017 23:35:28 +0200 Subject: [Buildroot] [PATCH 4/5] package/mesa3d: explicitely disable asm In-Reply-To: <49fce2a1-49a7-317c-486f-b3b6c37604a9@gmail.com> References: <20170326162853.3022-1-romain.naour@gmail.com> <20170326162853.3022-4-romain.naour@gmail.com> <49fce2a1-49a7-317c-486f-b3b6c37604a9@gmail.com> Message-ID: <20170329233528.1eb13394@free-electrons.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, On Wed, 29 Mar 2017 23:28:56 +0200, Romain Naour wrote: > Le 29/03/2017 ? 06:41, Bernd Kuhls a ?crit : > > Am Sun, 26 Mar 2017 18:28:52 +0200 schrieb Romain Naour: > > > >> assembly are enabled by default on supported plaforms even while > >> crosscompiling if host == target: > >> > >> checking whether to enable assembly... yes, x86_64 > >> > >> It was also disabled in the mesa package in Fedora since 7.6 [1]. > >> > >> [1] http://pkgs.fedoraproject.org/cgit/rpms/mesa.git/commit/?id=629c8726 > > > > Hi, > > > > I am unsure about this patch, why shouldn't we use asm code when > > crosscompiling for host == target CPU? > > To be honest I don't really have an issue with assembly being enabled for my use > case. But there is an issue when you want a reproducible build with the same > defconfig build on a x86 and x86_64 host. For which architectures is the assembly code available? We should forcefully enable it for those architectures. I agree with Romain that we shouldn't depend on something that looks at the architecture of the host machine to decide if a feature for the target should be enabled or not. Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com