From mboxrd@z Thu Jan 1 00:00:00 1970 From: Viktar Sakovich Date: Fri, 17 Nov 2006 13:53:04 +0000 Subject: Re: [LARTC] HTB prio: global or per class ? Message-Id: <06111715530400.00818@localhost> List-Id: References: <81c11a560611160926seb0c72ey4a6426a364274849@mail.gmail.com> In-Reply-To: <81c11a560611160926seb0c72ey4a6426a364274849@mail.gmail.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: lartc@vger.kernel.org On Friday 17 November 2006 09:57, you wrote: > Okay > > Which among you rightly? Who tests that ? Me for example. > If the the prio is global, it's strange and not very usefull and limit a > lot the HTB scheduler. I agree. But it's the way it works. > Normaly if HTB respecte the diffserv, the prio are per class basis. > Why is it global a > > 2006/11/16, Viktar Sakovich : > > On Thursday 16 November 2006 19:26, you wrote: > > > Hi all, > > > > > > Is the prio specification in the htb class global or is it on a per > > > class basis ? > > > > > > A simple example: > > > > > > class 1:10 parent 1: > > > class 1:100 parent 1:10 prio 3 > > > > > > class 1:200 parent 1:10 prio 7 > > > class 1:201 parent 1:200 prio 1 > > > class 1:202 parent 1:200 prio 2 > > > > > > Which class will get excessive bandwidth first? 100 or 201/202 ? > > > > Prio specification in the htb class is global. In this example 201/202 > > will > > get all bandwidth. _______________________________________________ LARTC mailing list LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc