Messages in this thread | ![/](/images/icornerl.gif) | | Date | Wed, 28 Apr 2010 14:37:05 -0700 | From | "Paul E. McKenney" <> | Subject | Re: [PATCH] RCU: don't turn off lockdep when find suspicious rcu_dereference_check() usage |
| |
On Wed, Apr 28, 2010 at 04:58:54PM -0400, Miles Lane wrote: > On Sat, Apr 24, 2010 at 10:34 PM, Paul E. McKenney < > paulmck@linux.vnet.ibm.com> wrote: > On Fri, Apr 23, 2010 at 06:59:12PM -0400, Miles Lane wrote:
[ . . . ]
> > commit 0868dd631def762ba00c2f0f397a53c5cdf24ae2 > > Author: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > Date: Sat Apr 24 19:23:30 2010 -0700 > > > > block-cgroup: fix RCU-lockdep splat in blkiocg_add_blkio_group() > > > > It is necessary to be in an RCU read-side critical section when invoking > > css_id(), so this patch adds one to blkiocg_add_blkio_group(). This is > > actually a false positive, because this is called at initialization > > time, > > and hence always refers to the root cgroup, which cannot go away. > > > > Located-by: Miles Lane <miles.lane@gmail.com> > > Suggested-by: Vivek Goyal <vgoyal@redhat.com> > > Signed-off-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> > > > > diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c > > index 5fe03de..55c8c73 100644 > > --- a/block/blk-cgroup.c > > +++ b/block/blk-cgroup.c > > @@ -71,7 +71,9 @@ void blkiocg_add_blkio_group(struct blkio_cgroup *blkcg, > > > > spin_lock_irqsave(&blkcg->lock, flags); > > rcu_assign_pointer(blkg->key, key); > > + rcu_read_lock(); > > blkg->blkcg_id = css_id(&blkcg->css); > > + rcu_read_unlock(); > > hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list); > > spin_unlock_irqrestore(&blkcg->lock, flags); > > #ifdef CONFIG_DEBUG_BLK_CGROUP > > > > Hi Paul, > Did this patch make it into your patch set? Has your patch set gone into > the Linus tree? > I just tested 2.6.34-rc5-git8 and hit one of the issues again. I think this > patch is intended to correct this issue?
I replaced the above with an improved patch from Vivek Goyal, which has not yet reached mainline. I will resend my patch stack.
Thanx, Paul
> [ 2.289598] [ INFO: suspicious rcu_dereference_check() usage. ] > [ 2.289604] --------------------------------------------------- > [ 2.289610] kernel/cgroup.c:4438 invoked rcu_dereference_check() without > protection! > [ 2.289615] > [ 2.289617] other info that might help us debug this: > [ 2.289619] > [ 2.289624] > [ 2.289626] rcu_scheduler_active = 1, debug_locks = 1 > [ 2.289632] 2 locks held by async/1/575: > [ 2.289637] #0: (&shost->scan_mutex){+.+.+.}, at: [<c121063d>] > __scsi_add_device+0x5b/0xab > [ 2.289660] #1: (&(&blkcg->lock)->rlock){......}, at: [<c1143acb>] > blkiocg_add_blkio_group+0x1a/0x73 > [ 2.289678] > [ 2.289680] stack backtrace: > [ 2.289688] Pid: 575, comm: async/1 Not tainted 2.6.34-rc5-git8 #17 > [ 2.289693] Call Trace: > [ 2.289704] [<c12ee273>] ? printk+0xf/0x14 > [ 2.289715] [<c1050fbd>] lockdep_rcu_dereference+0x74/0x7d > [ 2.289725] [<c106227d>] css_id+0x37/0x46 > [ 2.289734] [<c1143adc>] blkiocg_add_blkio_group+0x2b/0x73 > [ 2.289744] [<c1146a19>] cfq_init_queue+0xd6/0x2a3 > [ 2.289755] [<c120d657>] ? scsi_request_fn+0x0/0x3ea > [ 2.289764] [<c113a316>] elevator_init+0xa1/0xd5 > [ 2.289774] [<c113bc3f>] blk_init_queue_node+0x103/0x109 > [ 2.289783] [<c113bc50>] blk_init_queue+0xb/0xd > [ 2.289792] [<c120da58>] __scsi_alloc_queue+0x17/0xef > [ 2.289802] [<c120db40>] scsi_alloc_queue+0x10/0x49 > [ 2.289811] [<c120f381>] scsi_alloc_sdev+0x14f/0x1ef > [ 2.289821] [<c120f617>] scsi_probe_and_add_lun+0xb5/0x7ed > [ 2.289831] [<c10517dc>] ? trace_hardirqs_on_caller+0x119/0x141 > [ 2.289843] [<c1210669>] __scsi_add_device+0x87/0xab > [ 2.289854] [<c1232dfe>] ata_scsi_scan_host+0x64/0x136 > [ 2.289865] [<c12312c3>] async_port_probe+0x9e/0xa4 > [ 2.289876] [<c10479c8>] async_thread+0xf0/0x1d4 > [ 2.289887] [<c102b474>] ? default_wake_function+0x0/0xd > [ 2.289896] [<c10478d8>] ? async_thread+0x0/0x1d4 > [ 2.289906] [<c1041a2a>] kthread+0x6a/0x6f > [ 2.289916] [<c10419c0>] ? kthread+0x0/0x6f > [ 2.289926] [<c1003742>] kernel_thread_helper+0x6/0x1a > [ 2.289934] > [ 2.289935] =================================================== > [ 2.289941] [ INFO: suspicious rcu_dereference_check() usage. ] > [ 2.289946] --------------------------------------------------- > [ 2.289951] kernel/cgroup.c:1651 invoked rcu_dereference_check() without > protection! > [ 2.289957] > [ 2.289958] other info that might help us debug this: > [ 2.289961] > [ 2.289966] > [ 2.289967] rcu_scheduler_active = 1, debug_locks = 1 > [ 2.289973] 1 lock held by async/1/575: > [ 2.289978] #0: (&shost->scan_mutex){+.+.+.}, at: [<c121063d>] > __scsi_add_device+0x5b/0xab > [ 2.289995] > [ 2.289996] stack backtrace: > [ 2.290003] Pid: 575, comm: async/1 Not tainted 2.6.34-rc5-git8 #17 > [ 2.290008] Call Trace: > [ 2.290016] [<c12ee273>] ? printk+0xf/0x14 > [ 2.290025] [<c1050fbd>] lockdep_rcu_dereference+0x74/0x7d > [ 2.290035] [<c106423a>] cgroup_path+0x4a/0x110 > [ 2.290045] [<c1143b12>] blkiocg_add_blkio_group+0x61/0x73 > [ 2.290055] [<c1146a19>] cfq_init_queue+0xd6/0x2a3 > [ 2.290065] [<c120d657>] ? scsi_request_fn+0x0/0x3ea > [ 2.290074] [<c113a316>] elevator_init+0xa1/0xd5 > [ 2.290083] [<c113bc3f>] blk_init_queue_node+0x103/0x109 > [ 2.290093] [<c113bc50>] blk_init_queue+0xb/0xd > [ 2.290102] [<c120da58>] __scsi_alloc_queue+0x17/0xef > [ 2.290111] [<c120db40>] scsi_alloc_queue+0x10/0x49 > [ 2.290120] [<c120f381>] scsi_alloc_sdev+0x14f/0x1ef > [ 2.290131] [<c120f617>] scsi_probe_and_add_lun+0xb5/0x7ed > [ 2.290140] [<c10517dc>] ? trace_hardirqs_on_caller+0x119/0x141 > [ 2.290152] [<c1210669>] __scsi_add_device+0x87/0xab > [ 2.290162] [<c1232dfe>] ata_scsi_scan_host+0x64/0x136 > [ 2.290172] [<c12312c3>] async_port_probe+0x9e/0xa4 > [ 2.290182] [<c10479c8>] async_thread+0xf0/0x1d4 > [ 2.290192] [<c102b474>] ? default_wake_function+0x0/0xd > [ 2.290202] [<c10478d8>] ? async_thread+0x0/0x1d4 > [ 2.290211] [<c1041a2a>] kthread+0x6a/0x6f > [ 2.290221] [<c10419c0>] ? kthread+0x0/0x6f > [ 2.290230] [<c1003742>] kernel_thread_helper+0x6/0x1a
| ![\](/images/icornerr.gif) |