From mboxrd@z Thu Jan 1 00:00:00 1970 From: Johannes Berg Subject: Re: 2.6.35-rc2-git1 - net/mac80211/sta_info.c:125 invoked rcu_dereference_check() without protection! Date: Tue, 08 Jun 2010 09:26:09 +0200 Message-ID: <1275981969.3706.1.camel@jlt3.sipsolutions.net> References: <20100607235941.GD2387@linux.vnet.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Cc: Miles Lane , Vivek Goyal , Eric Paris , Lai Jiangshan , Ingo Molnar , Peter Zijlstra , LKML , nauman@google.com, eric.dumazet@gmail.com, netdev@vger.kernel.org, Jens Axboe , Gui Jianfeng , Li Zefan To: paulmck@linux.vnet.ibm.com Return-path: In-Reply-To: <20100607235941.GD2387@linux.vnet.ibm.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Mon, 2010-06-07 at 16:59 -0700, Paul E. McKenney wrote: > On Mon, Jun 07, 2010 at 02:25:44PM -0400, Miles Lane wrote: > > [ 43.478812] [ INFO: suspicious rcu_dereference_check() usage. ] > > [ 43.478815] --------------------------------------------------- > > [ 43.478820] net/mac80211/sta_info.c:125 invoked > > rcu_dereference_check() without protection! > > [ 43.478824] > > [ 43.478824] other info that might help us debug this: > > [ 43.478826] > > [ 43.478829] > > [ 43.478830] rcu_scheduler_active = 1, debug_locks = 1 > > [ 43.478834] no locks held by NetworkManager/4017. > > Hmmm... Johannes's update has been merged, and it requires that callers > either be in an RCU read-side critical section or hold either the > ->sta_lock or the ->sta_mtx, and this thread does none of this. > > Johannes, any thoughts? > > Thanx, Paul > > > [ 43.478837] stack backtrace: > > [ 43.478842] Pid: 4017, comm: NetworkManager Not tainted 2.6.35-rc2-git1 #8 > > [ 43.478846] Call Trace: > > [ 43.478849] [] lockdep_rcu_dereference+0x9d/0xa5 > > [ 43.478876] [] sta_info_get_bss+0x71/0x12d [mac80211] > > [ 43.478889] [] ieee80211_find_sta+0x15/0x2f [mac80211] > > [ 43.478902] [] iwlagn_tx_queue_reclaim+0xe7/0x1bb [iwlagn] iwlwifi wasn't using rcu protection here -- already sent a patch fixing it. My mistake, I think. Thanks for checking :) johannes