Re: [PATCH] another minor bit of cpumask cleanup

[William Lee Irwin III <wli@holomorphy.com>]

At some point in the past, akpm wrote:
>> Please, hang onto it until we get things synced up a bit more.

On Sun, Dec 21, 2003 at 11:19:18PM -0800, Paul Jackson wrote:
> Ok - good idea.  I'll resend later on.  There is no hurry on this one.

A rereading of this thread reveals the point of the thing was missed.
It's supposed to iterate over online cpus in an given bitmap. It was
meant to replace iterations like:

for_each_cpu(cpu, mask) {
	if (!cpu_online(cpu))
		continue;
	do_something(cpu);
}

with

for_each_online_cpu(cpu, mask)
	do_something(cpu);

Using any_online_cpu() as the starting point repairs it, since that
properly ands mask with cpu_online_map and hands back the first cpu,
though it's only a coincidence it hands back the first such cpu. There
isn't a a first_online_cpu() in the API, that's just effectively what
any_online_cpu() does at the moment.

On the other hand, I just don't care anymore, apart from clarifying
intent so as to counter the implication that all I did back then was
crap gibberish all over the tree. I personally have received zero
recognition or other return on my efforts in this area apart from the
mere fact it was merged. In fact, rather the opposite.


-- wli