From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Borkmann Subject: Re: [PATCH 2/2] net: sched: convert qdisc linked list to hashtable Date: Fri, 12 Aug 2016 16:44:32 +0200 Message-ID: <57ADE0D0.2080500@iogearbox.net> References: <57ADC6AA.5070506@iogearbox.net> <57ADDC8B.9010700@iogearbox.net> Mime-Version: 1.0 Content-Type: text/plain; charset=windows-1252; format=flowed Content-Transfer-Encoding: 7bit Cc: Eric Dumazet , Jamal Hadi Salim , Phil Sutter , Cong Wang , David Miller , linux-kernel@vger.kernel.org, netdev@vger.kernel.org To: Jiri Kosina Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On 08/12/2016 04:36 PM, Jiri Kosina wrote: > On Fri, 12 Aug 2016, Daniel Borkmann wrote: > >>> I was thinking about something like the patch below (the reasong being >>> that ->dev would be NULL only in cases of singletonish qdiscs) ... >>> wouldn't that also fix the issue you're seeing? Have to think it >>> through a little bit more .. >> >> Ahh, so this has the same effect as previously observed with the other fix. > > Thanks a lot for confirming that this fixes the panic. I still have to > think a little bit more about this though. > >> Perhaps it's just a dumping issue, but to the below clsact, there shouldn't >> be pfifo_fast instances appearing. >> >> # tc qdisc show dev wlp2s0b1 >> qdisc mq 0: root >> qdisc pfifo_fast 0: parent :4 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :3 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :1 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> # tc qdisc add dev wlp2s0b1 clsact >> # tc qdisc show dev wlp2s0b1 >> qdisc mq 0: root >> qdisc pfifo_fast 0: parent :4 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :3 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :1 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc clsact ffff: parent ffff:fff1 >> qdisc pfifo_fast 0: parent :4 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :3 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :2 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 >> qdisc pfifo_fast 0: parent :1 bands 3 priomap 1 2 2 2 1 2 0 0 1 1 1 1 1 1 1 1 > > Hmm, no immediate idea where those are coming from, we'll have to figure > it out. The mq device used here has 4 queues, right? Yes, the first tc qdisc show is after boot and how it should normally look like, so 4 tx queues. # ls /sys/class/net/wlp2s0b1/queues/ rx-0 tx-0 tx-1 tx-2 tx-3 When adding clsact, only the 'qdisc clsact' line should be extra. Given the extra pfifo_fast ones look the same as above, I would suspect a htab dumping issue, perhaps. I can debug a bit later tonight on this.