Hey, fellow LART controlllers, I''d like to ask if someome has a real esfq experience with HTB class hierarchy and esfq qdiscs. Lets say that I want to unite the connections to 10 client PCs with u32 or fw filters (doesn''t matter which one in fact) into one HTB class with esfq qdisc at the end. Did anyone tried it, and more important, did it worked the way it is supposed to? As of my experience, I saw several connections matched right against the filters and put into the correct class, but the bandwidth management wasn''t even remotely fair. TIA, Nickola
On Sat, Apr 26, 2003 at 12:29:12PM +0300, Nickola Kolev wrote:> Did anyone tried it, and more important, did it worked the way it is supposed to? > As of my experience, I saw several connections matched right against the filters > and put into the correct class, but the bandwidth management wasn''t even remotely > fair.Did you use a kernel with #define PSCHED_CLOCK_SOURCE PSCHED_CPU in linux/include/net/sched/pkt_sched.h or just the standard kernel? Matthias _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/
On Sat, 26 Apr 2003 15:07:52 +0200 Matthias Weingart <lartc@pentax.boerde.de> wrote: : On Sat, Apr 26, 2003 at 12:29:12PM +0300, Nickola Kolev wrote: : : > Did anyone tried it, and more important, did it worked the way it is supposed to? : > As of my experience, I saw several connections matched right against the filters : > and put into the correct class, but the bandwidth management wasn''t even remotely : > fair. : : Did you use a kernel with : #define PSCHED_CLOCK_SOURCE PSCHED_CPU in linux/include/net/sched/pkt_sched.h : or just the standard kernel? : : Matthias Nope, The place for pkt_sched.h here is linux/include/net/pkt_sched.h. And there #define PSCHED_CLOCK_SOURCE PSCHED_JIFFIES Can you tell me what''s the difference? I mean, I see that in your case clock source for packet scheduling is #defined as PSCHED_CPU, and in mine as PSCHED_JIFFIES, but what''s the impact on esfq? TIA, Nickola
On Saturday 26 April 2003 16:47, Nickola Kolev wrote:> On Sat, 26 Apr 2003 15:07:52 +0200 > > Matthias Weingart <lartc@pentax.boerde.de> wrote: > : On Sat, Apr 26, 2003 at 12:29:12PM +0300, Nickola Kolev wrote: > : > Did anyone tried it, and more important, did it worked the way it is > : > supposed to? As of my experience, I saw several connections matched > : > right against the filters and put into the correct class, but the > : > bandwidth management wasn''t even remotely fair. > : > : Did you use a kernel with > : #define PSCHED_CLOCK_SOURCE PSCHED_CPU in > : linux/include/net/sched/pkt_sched.h or just the standard kernel? > : > : Matthias > > Nope, > > The place for pkt_sched.h here is linux/include/net/pkt_sched.h. > And there #define PSCHED_CLOCK_SOURCE PSCHED_JIFFIES > > Can you tell me what''s the difference? I mean, I see that in your > case clock source for packet scheduling is #defined as PSCHED_CPU, > and in mine as PSCHED_JIFFIES, but what''s the impact on esfq?About the difference : http://www.docum.org/stef.coene/qos/faq/cache/40.html The impact is increased precision. Stef -- stef.coene@docum.org "Using Linux as bandwidth manager" http://www.docum.org/ #lartc @ irc.oftc.net _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/