Re: gettimeofday problem

Pavel Machek (pavel@ucw.cz)
Mon, 1 Jul 2002 04:29:58 +0200


Hi!

> > In this piece of code I convert seconds and microseconds in
> > milliseconds. I think the problem is not in my code, in fact I wrote the
> > following piece of code in Java, and it does not work too. In the for
> > loop the 90% of times b > a while for 10% of times not.
> >
> ...
> > long a = System.currentTimeMillis();
> > long b = System.currentTimeMillis();
> > if (a > b) {
> > System.out.println("Wrong!!!!!!!!!!!!!");
> > }
>
>
> So in 10% of the cases, two successive calls yield time
> rolling BACK ?
>
> I used gettimeofday() call, and compared the original data
> from the code.
>
> At a modern uniprocessor machine I never get anything except
> monotonously increasing time (TSC is used in betwen timer ticks
> to supply time increase.) At a dual processor machine, on
> occasion I do get SAME value twice. I have never seen time
> rolling backwards.
>
> Uh.. correction: 216199245 0:-1 -- it did step backwards,
> but only once within about 216 million gettimeofday() calls.
> (I am running 2.4.19-pre8smp at the test box.)

Hmm, so it is buggy even for you. He probably has way crappier
hardware. Neptun chipsets and via chipsets have bugs in time
implementation.
Pavel

-- 
(about SSSCA) "I don't say this lightly.  However, I really think that the U.S.
no longer is classifiable as a democracy, but rather as a plutocracy." --hpa
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/