__profile_frequency

[george anzinger]

Richard Henderson wrote:
> Um, so how's 
> 
> 45        tim.it_interval.tv_sec = 0;
> 46        tim.it_interval.tv_usec = 1;
> 47        tim.it_value.tv_sec = 0;
> 48        tim.it_value.tv_usec = 0;
> 49        __setitimer(ITIMER_REAL, &tim, 0);
> 50        __setitimer(ITIMER_REAL, 0, &tim);
> 51        if (tim.it_interval.tv_usec < 2)
> 52          return 0;
> 53        return (1000000 / tim.it_interval.tv_usec);
> 
> supposed to work?  AFAICS there's no reason for the kernel 
> (1) not to round down instead of up and 

The standard requires time to be rounded UP to the next resolution value.
(2) not do something
> else legal like schedule in between the two __setitimer calls.

Well, the above looks at interval, not the value, so a delay will not 
make a difference.  All you should see is the rounded interval.

I don't know about the setitimer, but the posix set_timer call with a 
zero value part, by the standard, resets the timer and would not pass 
the above test.

Time to check the standard...
> 
> On alpha-linux 2.5.73, tim is always zero after the second
> call, which results in divide-by-zero FPE in __sprofil.
> 
> Why wouldn't this be the same as __getclktck?
> 
> 
> r~
> 
> 

-- 
George Anzinger   george@mvista.com
High-res-timers:  http://sourceforge.net/projects/high-res-timers/
Preemption patch: http://www.kernel.org/pub/linux/kernel/people/rml