Re: [PATCH 01/10] random: make 'add_interrupt_randomness()' dosomething sane

[Theodore Ts'o]

On Thu, Jul 05, 2012 at 07:59:00PM -0700, Linus Torvalds wrote:
> On Thu, Jul 5, 2012 at 4:21 PM, Theodore Ts'o <tytso@xxxxxxx> wrote:
> > On Thu, Jul 05, 2012 at 05:31:22PM -0500, Matt Mackall wrote:
> >>
> >> It's better to mix and not credit than to not mix at all. Instead just
> >> check the fast count against HZ before the credit.
> >
> > I'm not sure what you mean by this?
> 
> So what I think Matt meant was to check number of timer interrupts
> against the fast count.
> 
> IOW, always call "add_interrupt_randomness()", but then in that
> function, when you determine the amount of entropy, check if there
> were non-timer interrupts in the last HZ cycle. If there were purely
> timer interrupts, you can still mix in the cycle information, but it's
> likely to be fairly weak.

I'm going to have to plead ignorance to how the timer code works these
days.  With the advent of clockevents and NOHZ, it's gotten a lot more
complicated.  What in the world is "fast count"?  I've grepped for it,
and I can't find it.

I can simply not credit entropy of the timer is on the irq, but I
think you and Matt were suggesting more subtle.  I just have no idea
how to tell if there were non-timer interrupts during the last HZ
cycle.  Can you give me a hint?

Thanks,

						- ted
--
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/