From mboxrd@z Thu Jan 1 00:00:00 1970 From: Luca Ceresoli Date: Sun, 28 Jun 2015 21:54:25 +0200 Subject: [Buildroot] [V4 1/2] A20-OLinuXino-Lime: new board (mainline) In-Reply-To: References: <1435313217-12871-1-git-send-email-francois.perrad@gadz.org> <1435313217-12871-2-git-send-email-francois.perrad@gadz.org> <558D7D84.1000605@lucaceresoli.net> Message-ID: <559050F1.1030101@lucaceresoli.net> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Dear Fran?ois, thanks for your patches, which are now on master. I'll base my A20-OLinuXino-MICRO patches on top of you good work. A comment below about minimal defconfigs. Fran?ois Perrad wrote: [...] >>> diff --git a/configs/olimex_a20_olinuxino_lime_defconfig >>> b/configs/olimex_a20_olinuxino_lime_defconfig >>> new file mode 100644 >>> index 0000000..62f0a4c >>> --- /dev/null >>> +++ b/configs/olimex_a20_olinuxino_lime_defconfig >>> @@ -0,0 +1,45 @@ >>> +# Architecture >>> +BR2_arm=y >>> +BR2_cortex_a7=y >>> +BR2_ARM_EABIHF=y >> >> >> This SoC has NEON and VFPv4, so you should add: >> BR2_ARM_FPU_NEON_VFPV4=y >> > > Dear Luca, > > From the reading of arch/Config.in.arm: > - BR2_cortex_a7 selects BR2_ARM_CPU_HAS_NEON and BR2_ARM_CPU_HAS_VFPV4 > - BR2_CPU_HAS_VFPV4 choices by default BR2_ARM_FPU_VFPV4D16 as > "Floating point strategy" > > BR2_ARM_FPU_NEON_VFPV4 could be another choice for this "Floating > point strategy". > > As far I see, the BR defconfigs try to stay minimalist. > For me, the default settings provided by BR are good enough. > > I added BR2_ARM_EABIHF because it is required by the proprietary blob > binary mali r3p0, not because it is a nice to have or a fine tuning > optimisation. I understand your argument. However, my understanding of the "minimal defconfig" choice is only that they should not include unneeded packages (i.e. everything but busybox) in the root filesystem. I don't see any advantage of configuring toolchain and target options in a sub-optimal way. In fact, we have several Cortex-A defconfigs that enable NEON (5 out of 26) and EABIHF (9 out of 26): $ git grep -l cortex_a -- configs/ |wc 26 26 878 $ git grep -l cortex_a -- configs/ | xargs grep -l EABIHF|wc 9 9 308 $ git grep -l cortex_a -- configs/ | xargs grep -l NEON|wc 5 5 158 $ But as I said, my understanding might be wrong. Any authoritative opinions on this? I'd be glad to update the docs with whatever is "official". -- Luca