Re: [PATCH 2/3] ftrace: add tracepoint for hrtimer

[Thomas Gleixner]

On Fri, 22 May 2009, Xiao Guangrong wrote:
> +TRACE_EVENT(hrtimer_start,
> +
> +	TP_PROTO(struct hrtimer *timer),
> +
> +	TP_ARGS(timer),
> +
> +	TP_STRUCT__entry(
> +		__field( void *,	timer		)
> +		__field( void *,	function	)
> +		__field( s64,		expires		)
> +		__field( s64,		softexpires	)
> +	),
> +
> +	TP_fast_assign(
> +		__entry->timer		= timer;
> +		__entry->function	= timer->function;
> +		__entry->expires	= ktime_to_ns(hrtimer_get_expires(timer));
> +		__entry->softexpires	= ktime_to_ns(hrtimer_get_softexpires(timer));
> +	),

  Yuck, we really do not want expensive conversions in the fast
  path. This can be done by printk or user space tools.

> @@ -1119,6 +1122,7 @@ void hrtimer_init(struct hrtimer *timer, clockid_t clock_id,
>  {
>  	debug_hrtimer_init(timer);
>  	__hrtimer_init(timer, clock_id, mode);
> +	trace_hrtimer_init(timer, clock_id, mode);

  The comments I made about the timer.c tracepoints vs. debugobjects
  apply here as well.

Thanks,

	tglx
--
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/