From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Tue, 13 Sep 2016 17:42:56 +0200 Subject: [Buildroot] [autobuild.buildroot.net] Build results for 2016-09-04 In-Reply-To: References: <20160905063032.D7B78102B17@stock.ovh.net> Message-ID: <20160913174256.3c9eb965@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 Mon, 5 Sep 2016 16:15:47 -0300, Ezequiel Garcia wrote: > Tried to tackle this one, but it seems the problem is with the > external toolchain. Let me post the error here to make the > discussion easier: > > In file included from > /home/test/autobuild/run/instance-0/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/endian.h:60:0, > from > /home/test/autobuild/run/instance-0/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/bits/waitstatus.h:64, > from > /home/test/autobuild/run/instance-0/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/stdlib.h:42, > from t/../flist.h:4, > from t/btrace2fio.c:9: > /home/test/autobuild/run/instance-0/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/bits/byteswap.h: > In function 'main': > /home/test/autobuild/run/instance-0/output/host/usr/arm-buildroot-linux-gnueabihf/sysroot/usr/include/bits/byteswap.h:46:10: > fatal error: You must enable NEON instructions (e.g. > -mfloat-abi=softfp -mfpu=neon) to use these intrinsics. > return __builtin_bswap32 (__bsx); > ^~~~~~~~~~~~~~~~~~~~~~~~~ > compilation terminated. > make[1]: *** [t/btrace2fio.o] Error 1 > > AFAICS, the code doesn't directly use NEON intrinsics, > so it's not fio's fault. After some digging, I found > a gcc upstream report [1]. Quoting: > > "" > Note that `NEON` isn't enabled and there's no direct use > of NEON instructions/intrinsics in the code so the NEON > instructions must have been added by the optimizer. > Seemingly subtle change can make the error disappear. > "" > > Which is exaclty what I'm seeing here. The issue is fixed in gcc v6.2. > > [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71056 Thanks for the investigation! The problematic toolchain is indeed based on gcc 6.1, affected by the issue. We've updated to gcc 6.2 recently, so I'll try to re-build all the toolchains in the near future, which should help getting rid of this problem. Thanks! Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux and Kernel engineering http://free-electrons.com