From mboxrd@z Thu Jan 1 00:00:00 1970 From: Arnout Vandecappelle Date: Sun, 6 Mar 2016 20:02:59 +0100 Subject: [Buildroot] [PATCH v3 1/2] erlang: use erlang's native atomic ops In-Reply-To: References: <1456082246-28181-1-git-send-email-fhunleth@troodon-software.com> <56D22060.4020108@mind.be> Message-ID: <56DC7EE3.50107@mind.be> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net On 03/02/16 21:47, Frank Hunleth wrote: > Hi Arnout, all, > > On Sat, Feb 27, 2016 at 5:17 PM, Arnout Vandecappelle wrote: >> On 02/21/16 20:17, Frank Hunleth wrote: >>> Now that we're using Erlang 18, the preferred atomic ops implementation >>> for Erlang is its own built-in implementation, so use it. It is still >>> possible to use libatomic_ops if the native implementation does not >>> work. BR2_PACKAGE_ERLANG_ARCH_SUPPORTS is added now that >>> BR2_PACKAGE_LIBATOMIC_OPS_ARCH_SUPPORTS is no longer appropriate for >>> checking whether Erlang can be built for a platform. >>> >>> This fixes an autobuilder failure when using libatomic_ops on aarch64. >>> Erlang's native atomic ops implementation works on this platform. >>> >>> Fixes: >>> http://autobuild.buildroot.net/results/0cd/0cd22eb74fa29e5a85bf897762e16ab3daf33962/ >>> >>> Signed-off-by: Frank Hunleth >>> --- >> [snip] >>> +config BR2_PACKAGE_ERLANG_ARCH_SUPPORTS >>> + bool >>> + default y if BR2_i386 || BR2_x86_64 || BR2_powerpc || BR2_sparc_v9 || BR2_arm || BR2_aarch64 # case (1) >> >> This line is too long, please split it. >> >>> + default y if BR2_PACKAGE_LIBATOMIC_OPS_ARCH_SUPPORTS # case (4) >>> >>> config BR2_PACKAGE_ERLANG >>> bool "erlang" >>> depends on BR2_USE_MMU # fork() >>> depends on !BR2_STATIC_LIBS >>> - depends on BR2_PACKAGE_LIBATOMIC_OPS_ARCH_SUPPORTS >>> - select BR2_PACKAGE_LIBATOMIC_OPS >>> + depends on BR2_PACKAGE_ERLANG_ARCH_SUPPORTS >>> help >>> Erlang is a programming language used to build massively scalable >>> soft real-time systems with requirements on high availability. >>> @@ -20,6 +36,14 @@ config BR2_PACKAGE_ERLANG >>> >>> if BR2_PACKAGE_ERLANG >>> >>> +config BR2_PACKAGE_ERLANG_LIBATOMIC_OPS >>> + bool "libatomic_ops" >> >> Does it make sense to let the user select this? I don't think so... >> >> I think we should have: >> >> config BR2_PACKAGE_ERLANG_ARCH_SUPPORTS_DIRECTLY # case (1) >> bool >> default y if BR2_i386 || BR2_x86_64 || BR2_powerpc || \ >> BR2_sparc_v9 || BR2_arm # aarch64 causes build failures >> >> config BR2_PACKAGE_ERLANG_ARCH_SUPPORTS >> bool >> default y if BR2_PACKAGE_ERLANG_ARCH_SUPPORTS_DIRECTLY >> default y if BR2_PACKAGE_LIBATOMIC_OPS_ARCH_SUPPORTS # case (4) >> >> and nothing more. > > The more I think about this patch, the more I want to drop all > references to libatomic_ops. The reason is that the native Erlang > atomic ops implementation supports almost every platform supported by > libatomic_ops. In fact, it looks like every _defconfig in Buildroot > that supports Erlang uses the native atomic ops implementation rather > than libatomic_ops. (I actually tried building all _defconfigs at one > point, but I'll double check if there are doubts.) > > Does it make sense to remove libatomic_ops altogether? I think the > original reason for including it doesn't apply now that we upgraded to > Erlang 18. Sounds OK to me. Regards, Arnout > > Thanks, > Frank > > >> >>> + depends on BR2_PACKAGE_LIBATOMIC_OPS_ARCH_SUPPORTS >>> + depends on !BR2_aarch64 # causes build failures >>> + select BR2_PACKAGE_LIBATOMIC_OPS >>> + help >>> + Use libatomic_ops instead of Erlang's native atomic ops support. >>> + [snip] -- Arnout Vandecappelle arnout at mind be Senior Embedded Software Architect +32-16-286500 Essensium/Mind http://www.mind.be G.Geenslaan 9, 3001 Leuven, Belgium BE 872 984 063 RPR Leuven LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle GPG fingerprint: 7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF