On Thu, Mar 27, 2014 at 02:51:38PM -0400, Vladimir Makarov wrote:
> 2014-03-27 Vladimir Makarov <vmaka...@redhat.com>
>
> PR rtl-optimization/60650
> * gcc.target/arm/pr60650.c: New.
The test fails for me with:
spawn -ignore SIGHUP
/builddir/build/BUILD/gcc-4.9.0-20140328/obj-armv7hl-redhat-linux-gnueabi/gcc/xgcc
-B/builddir/build/BUILD/gcc-4.9.0-20140328/obj-armv7hl-redhat-linux-gnueabi/gcc/builddir/build/BUILD/gcc-4.9.0-20140328/gcc/testsuite/gcc.target/arm/pr60650.c
-fno-diagnostics-show-caret -fdiagnostics-color=never -O2
-fno-omit-frame-pointer -mabi=apcs-gnu -march=armv7-a -S -o pr60650.s
/builddir/build/BUILD/gcc-4.9.0-20140328/gcc/testsuite/gcc.target/arm/pr60650.c:1:0:
sorry, unimplemented: -mfloat-abi=hard and VFP
compiler exited with status 1
Does one have to pass -mfloat-abi= option too unconditionally?
Can somebody from the ARM/Linaro people handle this?
Also:
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/arm/pr60650.c
...
> + asm ("1\t.long ");
This asm looks quite bogus, and with
asm ("");
it ICEs the same way, can it be changed?
> + __builtin_unreachable ();
> + }
> +}
Jakub
