Re: [PATCH 2/4] x86/pvops: target CREATE_TRACE_POINTS to particularsubsystems

[Steven Rostedt]

On Thu, 16 Apr 2009, Jeremy Fitzhardinge wrote:
> 
> diff --git a/include/trace/events/irq.h b/include/trace/events/irq.h
> index 75e3468..ddc62f5 100644
> --- a/include/trace/events/irq.h
> +++ b/include/trace/events/irq.h
> @@ -54,4 +54,9 @@ TRACE_FORMAT(softirq_exit,
>  #endif /*  _TRACE_IRQ_H */
>  
>  /* This part must be outside protection */
> +#ifdef CREATE_IRQ_TRACE_POINTS
> +#undef CREATE_IRQ_TRACE_POINTS	/* avoid infinite recursion */
> +#include <trace/instantiate_trace.h>
> +#else
>  #include <trace/define_trace.h>
> +#endif




> --- /dev/null
> +++ b/include/trace/instantiate_trace.h
> @@ -0,0 +1,7 @@
> +/* 
> + * trace/events/foo.h include this when their subsystem-specific
> + * CREATE_FOO_TRACE_POINTS is defined.
> + */
> +#define CREATE_TRACE_POINTS
> +#include <trace/define_trace.h>
> +#undef CREATE_TRACE_POINTS

Instead of doing it this way, what about not having this new header, and 
just do:


#ifdef CREATE_IRQ_TRACE_POINTS
#define CREATE_TRACE_POINTS
#endif
#include <trace/define_trace.h>

Heck, define_trace.h is only defined when CREATE_TRACE_POINTS is defined, 
why not just remove that and have:

#ifdef CREATE_IRQ_TRACE_POINTS
#define <trace/define_trace.h>
#endif

The whole trickery with the CREATE_TRACE_POINTS was to avoid the #if in 
the trace headers anyway. If we can't avoid it. We don't need to add more 
confusion to the mix.

-- Steve

--
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/