Re: Oops in 2.5.64

[Andrew Morton]

Denis Vlasenko <[e-mail]> wrote:
>
> > Eh?  How come the compiler didn't inline
> > __constant_c_and_count_memset? What compiler version are you using?
> 
> ...
> 
> +/* GCC 3 (and probably earlier, I'm not sure) can be told to always inline
> +   a function. */
> +#if __GNUC__ < 3
> +#define force_inline inline
> +#else
> +#define force_inline inline __attribute__ ((always_inline))
> +#endif

Well I'd consider this a workaround for a rampant compiler bug.  It's just
weird that it refuses to inline a function like that.  Having to make 10,000
edits to the kernel tree to work around this does not appeal.

Cannot we just stick:

	#define inline	__inline__ __attribute__((always_inline))

in kernel.h?