Re: [PATCH 1/7] posix-timers: tidy up clock interfaces andconsolidate dispatch logic

[Andrew Morton]

George Anzinger <george@xxxxxxxxxx> wrote:
>
>  >  /*
>   > + * Define this to initialize every k_clock function table so all its
>   > + * function pointers are non-null, and always do indirect calls through the
>   > + * table.  Leave it undefined to instead leave null function pointers and
>   > + * decide at the call sites between a direct call (maybe inlined) to the
>   > + * default function and an indirect call through the table when it's filled
>   > + * in.  Which style is preferable is whichever performs better in the
>   > + * common case of using the default functions.
> 
>   > +#define CLOCK_DISPATCH_DIRECT
> 
>  As I understand it modern machines, the indirect call does really bad things to 
>  the pipeline.  The default call, even preceeded by the if, will be much faster 
>  by this reasoning.  I would, therefor, prefer not defining CLOCK_DISPATCH_DIRECT.

We do need to do one or the other.  I assume the current indecision is
pending some benchmarking work?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/