On Mon, 3 Dec 2018 at 20:22, Will Deacon <will.deacon@xxxxxxx> wrote:
>
> On Wed, Nov 28, 2018 at 09:09:00AM +0800, Jackie Liu wrote:
> > In a way similar to ARM commit 09096f6a0ee2 ("ARM: 7822/1: add workaround
> > for ambiguous C99 stdint.h types"), this patch redefines the macros that
> > are used in stdint.h so its definitions of uint64_t and int64_t are
> > compatible with those of the kernel.
> >
> > This patch comes from: https://patchwork.kernel.org/patch/3540001/
> > Wrote by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> >
> > We mark this file as a private file and don't have to override asm/types.h
> >
> > Reviewed-by: Ard Biesheuvel <ard.biesheuvel@xxxxxxxxxx>
> > Signed-off-by: Jackie Liu <liuyun01@xxxxxxxxxx>
> > ---
> > arch/arm64/include/asm/neon-intrinsics.h | 34 ++++++++++++++++++++++++++++++++
> > 1 file changed, 34 insertions(+)
> > create mode 100644 arch/arm64/include/asm/neon-intrinsics.h
> >
> > diff --git a/arch/arm64/include/asm/neon-intrinsics.h b/arch/arm64/include/asm/neon-intrinsics.h
> > new file mode 100644
> > index 0000000..e378766
> > --- /dev/null
> > +++ b/arch/arm64/include/asm/neon-intrinsics.h
> > @@ -0,0 +1,34 @@
> > +#ifndef _NEON_INTRINSICS_H
> > +#define _NEON_INTRINSICS_H
>
> We tend to name these with an __ASM_ prefix, so it should be:
>
> #ifndef __ASM_NEON_INTRINSICS_H
>
> That said, I notice that the commit you refer to for arch/arm/ actually
> places this stuff under uapi/. Is that needed?
>
No, it doesn't. It creates asm/types.h which has been moved into uap/
at a later date (which I guess means we're stuck with it). In
hindsight, it would have been better for ARM to create a neon
instrinsics header file such as this one, since the override is only
needed when you include <arm_neon.h>.
> > +#include <asm-generic/int-ll64.h>
> > +
> > +/*
> > + * For Aarch64, there is some ambiguity in the definition of the types below
> > + * between the kernel and GCC itself. This is usually not a big deal, but it
> > + * causes trouble when including GCC's version of 'stdint.h' (this is the file
> > + * that gets included when you #include <stdint.h> on a -ffreestanding build).
> > + * As this file also gets included implicitly when including 'arm_neon.h' (the
> > + * NEON intrinsics support header), we need the following to work around the
> > + * issue if we want to use NEON intrinsics in the kernel.
> > + */
>
> Could you elaborate on what the ambiguities / conflicts in the types are
> please? I think you can also remove the sentence about directly including
> stdint on a freestanding build, since it doesn't seem relevant to the
> kernel afaict (we only pull it in via arm_neon.h).
>
In the kernel, u64/s64 are [un]signed long long, not [un]signed long.
So by redefining these macros to the former, we can force gcc-stdint.h
to define uint64_t / in64_t in a compatible manner.
> > +
> > +#ifdef __INT64_TYPE__
> > +#undef __INT64_TYPE__
> > +#define __INT64_TYPE__ __signed__ long long
>
> Do we need this __signed__ part?
>
No that seems redundant to me.
