It seems that downgrading kernel to 2.6.32 from squeeze resolves the issue (no more warnings, and filtering/shaping work as expected). But, the problem might go deeper because when I've been digging some lists and forums if anyone faced similar problem and I've found someone got very much the same warning using HTB in kernel 2.6.26.6. This could be helpful, but I don't know if I can/should post this to netdev list, so if this is the right thing to do, please, somebody forward this :-)
This is the link (polish): http://forum.dug.net.pl/viewtopic.php?pid=112061#p112061 He uses imq (patched kernel by himself), esfq, l7 and ipmark. So this is pretty much not the same stuff as in my configuration. However symptoms are the same (he claims warnings apear once a second but it can be proportional to traffic passing through the machine). Also this is DUG forum (Debian Users Gang) so I conclude he uses self-compiled debian kernel, but man can't be sure. :-) Here's his kernel log in casse the link was dead: Feb 24 20:53:10 cerber kernel: ------------[ cut here ]------------ Feb 24 20:53:10 cerber kernel: WARNING: at net/sched/sch_htb.c:361 htb_dequeue+0x131/0x31d() Feb 24 20:53:10 cerber kernel: Modules linked in: xt_connmark nf_nat_h323 nf_conntrack_h323 nf_nat_sip nf_conntrack_sip xt_d$ Feb 24 20:53:10 cerber kernel: Pid: 0, comm: swapper Tainted: G W 2.6.26.6-cerber #4 Feb 24 20:53:10 cerber kernel: Feb 24 20:53:10 cerber kernel: Call Trace: Feb 24 20:53:10 cerber kernel: <IRQ> [<ffffffff802309f8>] warn_on_slowpath+0x51/0x8c Feb 24 20:53:10 cerber kernel: [<ffffffff803fabc7>] __qdisc_run+0x8c/0x205 Feb 24 20:53:10 cerber kernel: [<ffffffff804265d7>] ip_finish_output+0x0/0x2ab Feb 24 20:53:10 cerber kernel: [<ffffffff80211d7c>] read_tsc+0x9/0x1f Feb 24 20:53:10 cerber kernel: [<ffffffff80248b74>] getnstimeofday+0x40/0xad Feb 24 20:53:10 cerber kernel: [<ffffffff8024748e>] ktime_get_ts+0x28/0x56 Feb 24 20:53:10 cerber kernel: [<ffffffff802474c8>] ktime_get+0xc/0x64 Feb 24 20:53:10 cerber kernel: [<ffffffff80402784>] htb_dequeue+0x131/0x31d Feb 24 20:53:10 cerber kernel: [<ffffffff803facfe>] __qdisc_run+0x1c3/0x205 Feb 24 20:53:10 cerber kernel: [<ffffffff803ead15>] net_tx_action+0xe1/0x109 Feb 24 20:53:10 cerber kernel: [<ffffffff8023645a>] __do_softirq+0x69/0xd7 Feb 24 20:53:10 cerber kernel: [<ffffffff8020c66c>] call_softirq+0x1c/0x30 Feb 24 20:53:10 cerber kernel: [<ffffffff8020e220>] do_softirq+0x30/0x6b Feb 24 20:53:10 cerber kernel: [<ffffffff8020e453>] do_IRQ+0x73/0xd0 Feb 24 20:53:10 cerber kernel: [<ffffffff80212354>] mwait_idle+0x0/0x3f Feb 24 20:53:10 cerber kernel: [<ffffffff8020b9f1>] ret_from_intr+0x0/0xa Feb 24 20:53:10 cerber kernel: <EOI> [<ffffffff802093e0>] default_idle+0x0/0x3b Feb 24 20:53:10 cerber kernel: [<ffffffff8042b101>] tcp_poll+0x0/0x131 Feb 24 20:53:10 cerber kernel: [<ffffffff80212390>] mwait_idle+0x3c/0x3f Feb 24 20:53:10 cerber kernel: [<ffffffff8020a51a>] cpu_idle+0x5d/0xb8 Feb 24 20:53:10 cerber kernel: Feb 24 20:53:10 cerber kernel: ---[ end trace 43d0351cdc38f35a ]--- -- To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: http://lists.debian.org/3454ea9b.4ff10a15.4e12e76d.a3...@o2.pl