It looks like that wild_bctr on powerpc/fixes is still not compiling

Michael Ellerman mpe at ellerman.id.au
Wed Nov 14 11:58:53 AEDT 2018


Hi Gromero ;)

Gustavo Romero <gromero at linux.vnet.ibm.com> writes:
> Hi mpe,
>
> Even after the latest fix for the wild_bctr selftest I'm still getting the
> following compilation (actually, an assembling error) because UL is not
> understood by the assembler:
>
> BUILD_TARGET=/home/gromero/git/linux/tools/testing/selftests/powerpc/mm; mkdir -p $BUILD_TARGET; make OUTPUT=$BUILD_TARGET -k -C mm all
> make[1]: Entering directory '/home/gromero/git/linux/tools/testing/selftests/powerpc/mm'
> gcc -std=gnu99 -O2 -Wall -Werror -DGIT_VERSION='"v4.20-rc1-8-g2c7645b"' -I/home/gromero/git/linux/tools/testing/selftests/powerpc/include  -m64    wild_bctr.c ../harness.c  -o /home/gromero/git/linux/tools/testing/selftests/powerpc/mm/wild_bctr
> /tmp/cctUajlx.s: Assembler messages:
> /tmp/cctUajlx.s:270: Error: syntax error; found `U', expected `,'
> /tmp/cctUajlx.s:270: Error: junk at end of line: `UL'
...
> /tmp/cctUajlx.s:270: Error: syntax error; found `U', expected `,'
> /tmp/cctUajlx.s:270: Error: junk at end of line: `UL'
> ../../lib.mk:152: recipe for target '/home/gromero/git/linux/tools/testing/selftests/powerpc/mm/wild_bctr' failed
> make[1]: *** [/home/gromero/git/linux/tools/testing/selftests/powerpc/mm/wild_bctr] Error 1
> make[1]: Target 'all' not remade because of errors.
> make[1]: Leaving directory '/home/gromero/git/linux/tools/testing/selftests/powerpc/mm'
> Makefile:39: recipe for target 'mm' failed
> make: *** [mm] Error 2
>
> For:
> git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux.git (powerpc/fixes)
> $ git describe
> v4.20-rc1-8-g2c7645b
>
> This is gcc:
> $ gcc --version
> gcc (Ubuntu/IBM 5.4.0-6ubuntu1~16.04.10) 5.4.0 20160609
>
> but it should not make a difference, so  I'm wondering if anybody else is
> getting the same error too...

It seems to make a difference :)

I've been building with Ubuntu 7.3.0-27ubuntu1~18.04 and it builds just fine.

It looks like binutils 2.27 doesn't accept ULL but binutils 2.28 does.
 
Ah yep, here:

  https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commit;h=86b80085c889cd388fa677a5ae9053fd4be3776c

> The following trivial workaround can solve it by forcing a type promotion on
> the compiler side whilst leaving the macro taken into the asm code without
> the UL string:
>
> diff --git a/tools/testing/selftests/powerpc/mm/wild_bctr.c b/tools/testing/selftests/powerpc/mm/wild_bctr.c
> index 90469a9..d2772f4 100644
> --- a/tools/testing/selftests/powerpc/mm/wild_bctr.c
> +++ b/tools/testing/selftests/powerpc/mm/wild_bctr.c
> @@ -47,8 +47,9 @@ static int ok(void)
>          return 0;
>   }
>   
> -#define REG_POISON     0x5a5aUL
> -#define POISONED_REG(n)        ((REG_POISON << 48) | ((n) << 32) | (REG_POISON << 16) | (n))
> +#define REG_POISON     0x5a5a
> +#define POISONED_REG(n)        (((REG_POISON+0UL) << 48) | ((n) << 32) | ((REG_POISON+0UL) << 16) | (n))
>   
>   static inline void poison_regs(void)
>   {
>
>
> Should I contribute such a fix?

Yes thanks.

cheers


More information about the Linuxppc-dev mailing list