On Fri, 26 Apr 2013 15:51:05 +1000 Stephen Rothwell
wrote:
> Hi,
>
> On Fri, 26 Apr 2013 14:02:01 +0900 "Chanho Min" wrote:
> >
> >
> > @@ -0,0 +1,23 @@
> > +#include
> > +
> > +int __attribute__((weak)) __clzsi2(int val)
>
> We have __weak in
And lib/clz.c needs a few more includes, for
Hi,
On Fri, 26 Apr 2013 14:02:01 +0900 "Chanho Min" wrote:
>
>
> @@ -0,0 +1,23 @@
> +#include
> +
> +int __attribute__((weak)) __clzsi2(int val)
We have __weak in
--
Cheers,
Stephen Rothwells...@canb.auug.org.au
pgp21BCDTqop1.pgp
Description: PGP signature
>> gcc seems to define __builtin_clz as __clzsi2 in some architecture.
>> But, kernel doesn't link libgcc.a.
>> If kernel should use gcc's built-in function without libgcc.a,
>> do we need to port __clzsi2 to 'arch/*/lib/*'?
>
>This breaks alpha (gcc-4.4.4) as well. Can we please get this fixed
>
On Mon, 22 Apr 2013 18:22:18 +0900 "Chanho Min" wrote:
> >> +#define HTYPE const u8*
> >> +
> >> +#ifdef __BIG_ENDIAN
> >> +#define LZ4_NBCOMMONBYTES(val) (__builtin_clz(val) >> 3)
> >> +#else
> >> +#define LZ4_NBCOMMONBYTES(val) (__builtin_ctz(val) >> 3)
> >> +#endif
> >
> >It seems at least m68
On Mon, 22 Apr 2013 11:24:21 +0200 Geert Uytterhoeven
wrote:
> On Mon, Apr 22, 2013 at 11:22 AM, Chanho Min wrote:
> >>> +#define HTYPE const u8*
> >>> +
> >>> +#ifdef __BIG_ENDIAN
> >>> +#define LZ4_NBCOMMONBYTES(val) (__builtin_clz(val) >> 3)
> >>> +#else
> >>> +#define LZ4_NBCOMMONBYTES(val)
On Mon, Apr 22, 2013 at 11:22 AM, Chanho Min wrote:
>>> +#define HTYPE const u8*
>>> +
>>> +#ifdef __BIG_ENDIAN
>>> +#define LZ4_NBCOMMONBYTES(val) (__builtin_clz(val) >> 3)
>>> +#else
>>> +#define LZ4_NBCOMMONBYTES(val) (__builtin_ctz(val) >> 3)
>>> +#endif
>>
>>It seems at least m68k and sparc d
On Thu, Mar 14, 2013 at 10:48 AM, Chanho Min wrote:
> +#ifdef __BIG_ENDIAN
> +#define LZ4_NBCOMMONBYTES(val) (__builtin_clzll(val) >> 3)
> +#else
> +#define LZ4_NBCOMMONBYTES(val) (__builtin_ctzll(val) >> 3)
> +#endif
>
> #else /* 32-bit */
> #define STEPSIZE 4
> @@ -83,6 +130,14 @@ typedef str
