Am Dienstag, 10. September 2013, 17:14:54 schrieb Theodore Ts'o:
Hi Theodore,
>On Tue, Sep 10, 2013 at 10:48:00PM +0200, Geert Uytterhoeven wrote:
>> So the first importance for random_get_fast_cycles() is that it needs
>> to be fast. What's most important next: number of bits or
>> high-frequency?
>High-frequency. For example MIPS has a register which is bumped at
>every clock tick, modulo the number of lines in the TLB. That's what
>we're probably going to end up using for MIPS, on the assumption that
>the time between interrupts is not likely going to be related to the
>number of lines in the TLB. :-)
>
>Something like jiffies has lots of bits, but since it's updated at a
>much slower rate, it's not as useful if we are trying to measure
>uncertainity based on the interrupt time. (Worse yet, depending on
>how the architecture handles the clock, there mgiht be a very high
>correlation between when the jiffies counter gets incremented and the
>timer interrupt....)
>
>And yes, we will need to make sure this gets well documented in the
>sources when we introduce random_get_fast_cycles()....
As general hunch on the speed, I would say is a simple test that two
adjacent calls to obtain a time stamp show a delta. Eg.
__u64 tmp = random_get_fast_cycles() - random_get_fast_cycles();
if(0 == tmp)
return fail;
return pass;
>
>Cheers,
>
> - Ted
Ciao
Stephan
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
