Ahh, great.. thanks for clarifying.
Also, do you know when the developers (maybe Henning) are planning on making
the queue system 64bit so we can define queues greater in size than 4294Mbps?
Am I right in thinking that once we do move to a 64bit queue system, there will
also be better resolution for finer grained queue "quantization" steps?
We have 10gig OpenBSD firewalls and currently I have to restrict traffic to
4.2G. I can't just open the gates and use prio only, as our Committed Data Rate
with our WAN providers are not a full 10gig..
Cheers, Andy.
As an aside; I must say that with the advent of DDR4, and the potential of
mp_networking in OpenBSD in a couple years, I am really excited at the thought
of just how fast and powerful OpenBSD could easily become.
It really could be a threat to the big boys - An avid PF fan :)
On 26 Feb 2015, at 16:30, Stuart Henderson <s...@spacehopper.org> wrote:
> In gmane.os.openbsd.misc, Andy Lemin wrote:
>> Hopefully this is just a quick question and I'm missing something here, but
>> it
>> seems that we can no longer use percentages in our PF child queues.
>
> It hasn't been implemented for the queue rewrite yet.
>
> That said, there is a bug which makes it less obvious than it
> should be :-)
>
> Index: parse.y
> ===================================================================
> RCS file: /cvs/src/sbin/pfctl/parse.y,v
> retrieving revision 1.646
> diff -u -p -r1.646 parse.y
> --- parse.y 14 Feb 2015 23:32:41 -0000 1.646
> +++ parse.y 26 Feb 2015 16:29:22 -0000
> @@ -1364,7 +1364,7 @@ scspec : bandwidth
> {
> $$.m2 = $1;
> $$.d = 0;
> if ($$.m2.bw_percent) {
> - yyerror("no bandwidth in % yet");
> + yyerror("no bandwidth in %% yet");
> YYERROR;
> }
> }
> @@ -1383,7 +1383,7 @@ scspec : bandwidth
> {
> $$.m2 = $1;
>
> if ($$.m1.bw_percent || $$.m2.bw_percent) {
> - yyerror("no bandwidth in % yet");
> + yyerror("no bandwidth in %% yet");
> YYERROR;
> }
> }
