Re: [PATCH] percpu: preemptless __per_cpu_counter_add

[Tejun Heo <tj@kernel.org>]

Hello,

On Wed, Apr 13, 2011 at 03:22:36PM -0500, Christoph Lameter wrote:
> +	do {
> +		count = this_cpu_read(*fbc->counters);
> +
> +		new = count + amount;
> +		/* In case of overflow fold it into the global counter instead */
> +		if (new >= batch || new <= -batch) {
> +			spin_lock(&fbc->lock);
> +			fbc->count += __this_cpu_read(*fbc->counters) + amount;
> +			spin_unlock(&fbc->lock);
> +			amount = 0;
> +			new = 0;
> +		}
> +
> +	} while (this_cpu_cmpxchg(*fbc->counters, count, new) != count);

Is this correct?  If the percpu count changes in the middle, doesn't
the count get added twice?  Can you please use the cmpxchg() only in
the fast path?  ie.

	do {
		count = this_cpu_read();
		if (overflow) {
			disable preemption and do the slow thing.
			return;
		}
	} while (this_cpu_cmpxchg());

Thanks.

-- 
tejun

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>