(no subject)

Z Wing zeddersuk at yahoo.co.uk
Sun Apr 11 04:32:55 UTC 2010

Hiya all,

Could someone clarify for me the purpose of the 
bandwidth parameter when used with hfsc? Please consider my queue (512Kb
 max upload through the ADSL line):

 altq on em1 bandwidth 500Kb 
hfsc queue { ack, dns, ssh, other }
  queue ack        bandwidth 95% 
priority 8 qlimit 500 hfsc (realtime   20%)
  queue dns        
bandwidth 95% priority 7 qlimit 500 hfsc (realtime    5%)
ssh        bandwidth 95% priority 6 qlimit 500 hfsc (realtime   20%) 
{ssh_login, ssh_bulk}
   queue ssh_login bandwidth 95% priority 6 
qlimit 500 hfsc
   queue ssh_bulk  bandwidth 95% priority 5 qlimit 
500 hfsc
  queue other      bandwidth 95% priority 5 qlimit 500 hfsc 
(realtime   20% default)

My understanding was that "bandwidth 
xx%" tells pf that the queue can only use xx% of the total parent queue 
bandwidth and the various guides on the Internet say that it cannot go 
above 100% which sort of makes sense. However what I want, for example, 
is the 'other' queue to get all the upload bandwidth as long as there is
 no acks or ssh traffic in the queue. If I set bandwidth 20% for other, 
won't that mean that no matter what, the 'other' queue will only get a 
maximum of 20% of the bandwidth (even if there is no ssh traffic 

I think I haven't quite understood this properly but 
perhaps someone could clarify it. I found a lot out from calomel.org's 
excellent page but that's the only bit that confused me.


More information about the freebsd-pf mailing list