From mboxrd@z Thu Jan 1 00:00:00 1970 From: nuclearcat@nuclearcat.com Subject: Re: 4.6.3 kernel panic on ppp interface termination (__tcf_hash_release / spinlock ) Date: Wed, 06 Jul 2016 23:14:22 +0300 Message-ID: <58100198b6d14f77ce95740e4e912fc3@nuclearcat.com> References: <9b35b4153ec398e17cfb7c2739eee1de@nuclearcat.com> <31e51065f1531b4fa819dac37dc0ba11@nuclearcat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII; format=flowed Content-Transfer-Encoding: 7bit Cc: Linux Kernel Network Developers , "David S. Miller" , Jamal Hadi Salim To: Cong Wang Return-path: Received: from nuclearcat.com ([144.76.183.226]:52304 "EHLO nuclearcat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751629AbcGFUOb (ORCPT ); Wed, 6 Jul 2016 16:14:31 -0400 In-Reply-To: Sender: netdev-owner@vger.kernel.org List-ID: On 2016-07-05 01:02, Cong Wang wrote: > On Mon, Jul 4, 2016 at 12:07 PM, wrote: >> A little bit more details: >> >> ~ # tc -s -d filter show dev ppp0 parent ffff: >> filter protocol ip pref 100 u32 >> filter protocol ip pref 100 u32 fh 800: ht divisor 1 >> filter protocol ip pref 100 u32 fh 800::1 order 1 key ht 800 bkt 0 >> flowid :1 >> (rule hit 156 success 156) >> match 00000000/00000000 at 12 (success 156 ) >> action order 1: police 0x1 rate 4024Kbit burst 503000b mtu >> 2Kb >> action drop overhead 0b linklayer unspec >> ref 1 bind 1 >> Action statistics: >> Sent 10564 bytes 156 pkt (dropped 0, overlimits 0 requeues 0) >> backlog 0b 0p requeues 0 >> >> >> Panic occurs when i delete ingress qdisc >> >> tc qdisc del dev ppp0 ingress >> >> After reversing commits (and some related commits) >> 1d4150c02c5709fdfd80f10368a31867de35e72e >> ddf97ccdd7cb7e00daba465a5c947b8d941dc2a4 >> >> Problem is not occurring. >> >> I'm not certainly sure this commits are causing issue, it is very hard >> to >> reverse patches on this particular system. >> > > Can you check if the following commit helps? > > commit a03e6fe569713fb3ff0714f8fd7c8785c0ca9e22 > Author: WANG Cong > Date: Mon Jun 6 09:54:30 2016 -0700 > > act_police: fix a crash during removal Yes, it fixes issue, thanks.