[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[altq 470] Re: **HEADS UP** altq change 1st round
Francis Dupont wrote:
> In your previous mail you wrote:
> The performance impact is negligible. The tokenbucket regulator
> uses the CPU cycle counter on i386 and alpha as it is much cheaper
> than microtime() on these platforms.
> => don't forget this (use CPU cycle counter) doesn't work
> on a multiprocessor!
Ah, yes. A SMP box needs ALTQ_NOPCC (means don't use processor cycle
counter) in the kernel config. In that case, the token bucket
regulator falls back to microtime().
This should be addressed in the document. Thanks for the reminder.
By the way, I've already committed the changes last night so that the
new code will be in the next KAME snap.