Merge lp:~zhenqiang-chen/gcc-linaro/eabihf into lp:gcc-linaro/4.7

[Michael Hope]

Change the configury to recognise the eabihf tuple.

[Michael Hope (michaelh1)]

cbuild has taken a snapshot of this branch at r114985 and queued it for build.

The diff against the ancestor r114984 is available at:
 http://builds.linaro.org/toolchain/snapshots/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf.diff

and will be built on the following builders:
 a9-builder armv5-builder i686 x86_64

You can track the build queue at:
 http://ex.seabright.co.nz/helpers/scheduler

cbuild-snapshot: gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf
cbuild-ancestor: lp:gcc-linaro+bzr114984
cbuild-state: check

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on x86_64-natty-cbuild292-oort1-x86_64r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/x86_64-natty-cbuild292-oort1-x86_64r1

The testsuite results are the same as the branch point lp:gcc-linaro+bzr114984

The full diff is at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/x86_64-natty-cbuild292-oort1-x86_64r1/testsuite-diff.txt

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/x86_64-natty-cbuild292-oort1-x86_64r1/gcc-testsuite.txt

cbuild-checked: x86_64-natty-cbuild292-oort1-x86_64r1

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on i686-natty-cbuild292-oort2-i686r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/i686-natty-cbuild292-oort2-i686r1

The testsuite results are the same as the branch point lp:gcc-linaro+bzr114984

The full diff is at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/i686-natty-cbuild292-oort2-i686r1/testsuite-diff.txt

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/i686-natty-cbuild292-oort2-i686r1/gcc-testsuite.txt

cbuild-checked: i686-natty-cbuild292-oort2-i686r1

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on armv7l-natty-cbuild292-tcpanda04-cortexa9r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda04-cortexa9r1

The testsuite results are the same as the branch point lp:gcc-linaro+bzr114984

The full diff is at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda04-cortexa9r1/testsuite-diff.txt

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda04-cortexa9r1/gcc-testsuite.txt

cbuild-checked: armv7l-natty-cbuild292-tcpanda04-cortexa9r1

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on armv7l-natty-cbuild292-tcpanda02-armv5r2.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda02-armv5r2

The testsuite results are the same as the branch point lp:gcc-linaro+bzr114984

The full diff is at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda02-armv5r2/testsuite-diff.txt

The full testsuite results are at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-natty-cbuild292-tcpanda02-armv5r2/gcc-testsuite.txt

cbuild-checked: armv7l-natty-cbuild292-tcpanda02-armv5r2

[Michael Hope (michaelh1)]

Note that a armhf build has also completed:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7lhf-precise-cbuild292-ursa3-cortexa9hfr1/

[Ulrich Weigand (uweigand)]

So with this patch, GCC can be configured using any tuple matching arm-*-linux-*eabi* ... but they will all behave in exactly the same way. Is this the intention?

I would have expected that using the eabihf tuple would actually cause GCC to default to hard-float ...

[Linaro Toolchain Builder (cbuild)]

cbuild successfully built this on armv7l-precise-cbuild294-ursa3-cortexa9r1.

The build results are available at:
 http://ex.seabright.co.nz/build/gcc-linaro-4.7+bzr114985~zhenqiang-chen~eabihf/logs/armv7l-precise-cbuild294-ursa3-cortexa9r1

+PASS: g++.dg/graphite/id-1.C -std=c++11 (test for excess errors)
+PASS: g++.dg/graphite/id-1.C -std=c++98 (test for excess errors)
+PASS: g++.dg/graphite/pr39447.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr39447.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr41305.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr41305.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr41562.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr41562.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr42130.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr42130.C -std=gnu++11 execution test
+PASS: g++.dg/graphite/pr42130.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr42130.C -std=gnu++98 execution test
+PASS: g++.dg/graphite/pr42681.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr42681.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr42930.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr42930.C -std=gnu++98 (test for excess errors)
+PASS: g++.dg/graphite/pr43026.C -std=gnu++11 (test for excess errors)
+PASS: g++.dg/graphite/pr43026.C -std=gnu++98 (test for excess errors)
-PASS: g++.old-deja/g++.jason/enum6.C -std=gnu++11 (test for excess errors)
+FAIL: g++.old-deja/g++.jason/enum6.C -std=gnu++11 (test for excess errors)
-PASS: g++.old-deja/g++.jason/enum6.C -std=gnu++98 (test for excess errors)
+FAIL: g++.old-deja/g++.jason/enum6.C -std=gnu++98 (test for excess errors)
-PASS: g++.old-deja/g++.law/enum9.C -std=gnu++11 (test for excess errors)
+FAIL: g++.old-deja/g++.law/enum9.C -std=gnu++11 (test for excess errors)
-PASS: g++.old-deja/g++.law/enum9.C -std=gnu++98 (test for excess errors)
+FAIL: g++.old-deja/g++.law/enum9.C -std=gnu++98 (test for excess errors)
-PASS: g++.old-deja/g++.law/profile1.C -std=gnu++11 (test for excess errors)
-PASS: g++.old-deja/g++.law/profile1.C -std=gnu++11 execution test
-PASS: g++.old-deja/g++.law/profile1.C -std=gnu++98 (test for excess errors)
-PASS: g++.old-deja/g++.law/profile1.C -std=gnu++98 execution test
+FAIL: g++.old-deja/g++.law/profile1.C -std=gnu++11 (test for excess errors)
+UNRESOLVED: g++.old-deja/g++.law/profile1.C -std=gnu++11 compilation failed to produce executable
+FAIL: g++.old-deja/g++.law/profile1.C -std=gnu++98 (test for excess errors)
+UNRESOLVED: g++.old-deja/g++.law/profile1.C -std=gnu++98 compilation failed to produce executable
-PASS: g++.old-deja/g++.other/enum4.C -std=gnu++11 (test for excess errors)
+FAIL: g++.old-deja/g++.other/enum4.C -std=gnu++11 (test for excess errors)
-PASS: g++.old-deja/g++.other/enum4.C -std=gnu++98 (test for excess errors)
+FAIL: g++.old-deja/g++.other/enum4.C -std=gnu++98 (test for excess errors)
-PASS: gcc.dg/20021014-1.c (test for excess errors)
-PASS: gcc.dg/20021014-1.c execution test
+FAIL: gcc.dg/20021014-1.c ...

[Michael Hope (michaelh1)]

I had a look in gcc/config.gcc and there is a precedent in mips for setting an ABI feature based on the triplet.

Zhenqiang, could you update the patch so if with_float is not set and the triplet is arm-*-*hf then set with_float to hard?

Something like:

case ${target} in
arm*-*-*hf)
    with_float=${with_float:-hard}
esac

Also drop the uclibc changes. It's up to that community to define a hard float triplet if they want it.

review: Disapprove

[Zhenqiang Chen (zhenqiang-chen)]

On 8 May 2012 10:17, Michael Hope <email address hidden> wrote:
> Review: Disapprove
>
> I had a look in gcc/config.gcc and there is a precedent in mips for setting an ABI feature based on the triplet.
>
> Zhenqiang, could you update the patch so if with_float is not set and the triplet is arm-*-*hf then set with_float to hard?
>
> Something like:
>
> case ${target} in
> arm*-*-*hf)
>    with_float=${with_float:-hard}
> esac
>
> Also drop the uclibc changes.  It's up to that community to define a hard float triplet if they want it.

Thanks. I will update it.
-Zhenqiang

[Michael Hope (michaelh1)]

Superseded by https://code.launchpad.net/~zhenqiang-chen/gcc-linaro/gnueabihf/+merge/105602