Re: rcu: inet_timewait_sock.h:227 invoked rcu_dereference_check() without protection

["Paul E. McKenney" <paulmck@linux.vnet.ibm.com>]

On Wed, May 26, 2010 at 05:45:07PM +0200, Heiko Carstens wrote:
> Hi Paul,
> 
> we got the following RCU trace here while doing some tests with 2.6.34
> vanilla with some extra patches on top, but which do not touch RCU nor
> inet code.
> Would be nice if you could have a look at it.

Hello, Heiko,

This is fixed by commit 7ec75c582e639d956ce3afd499f67febe6f902a4 in
Linus's tree.

							Thanx, Paul

> May 21 09:05:16 h0515017 kernel: ===================================================
> May 21 09:05:16 h0515017 kernel: [ INFO: suspicious rcu_dereference_check() usage. ]
> May 21 09:05:16 h0515017 kernel: ---------------------------------------------------
> May 21 09:05:16 h0515017 kernel: /home/autobuild/BUILD/linux-2.6.34-20100520/include/net/inet_timewait_sock.h:227 invoked rcu_dereference_check() without protection!
> May 21 09:05:16 h0515017 kernel:
> May 21 09:05:16 h0515017 kernel: other info that might help us debug this:
> May 21 09:05:16 h0515017 kernel:
> May 21 09:05:16 h0515017 kernel:
> May 21 09:05:16 h0515017 kernel: rcu_scheduler_active = 1, debug_locks = 0
> May 21 09:05:16 h0515017 kernel: 1 lock held by swapper/0:
> May 21 09:05:16 h0515017 kernel: #0:  (/home/autobuild/BUILD/linux-2.6.34-20100520/net/ipv4/tcp_minisocks.c:41){+.-...}, at: [<0000000000158398>] run_timer_softirq+0x1b4/0x354
> May 21 09:05:16 h0515017 kernel:
> May 21 09:05:16 h0515017 kernel: stack backtrace:
> May 21 09:05:16 h0515017 kernel: CPU: 0 Not tainted 2.6.34-42.x.20100520-s390xdefault #1
> May 21 09:05:16 h0515017 kernel: Process swapper (pid: 0, task: 0000000000812a18, ksp: 0000000000807e00)
> May 21 09:05:16 h0515017 kernel: 0000000000000000 000000003ffa7bc8 0000000000000002 0000000000000000
> May 21 09:05:16 h0515017 kernel:       000000003ffa7c68 000000003ffa7be0 000000003ffa7be0 000000000056ba7c
> May 21 09:05:16 h0515017 kernel:       0000000000000000 0000000000000002 00000000000002d8 0000000000000000
> May 21 09:05:16 h0515017 kernel:       000000000000000d 0000000000000000 000000003ffa7c38 000000000000000e
> May 21 09:05:16 h0515017 kernel:       00000000005c48a8 0000000000105a20 000000003ffa7bc8 000000003ffa7c10
> May 21 09:05:16 h0515017 kernel: Call Trace:
> May 21 09:05:16 h0515017 kernel: ([<000000000010596e>] show_trace+0x142/0x148)
> May 21 09:05:16 h0515017 kernel: [<00000000004a5068>] __inet_twsk_kill+0x140/0x14c
> May 21 09:05:16 h0515017 kernel: [<00000000004a566c>] inet_twdr_do_twkill_work+0x90/0x130
> May 21 09:05:16 h0515017 kernel: [<00000000004a585a>] inet_twdr_hangman+0x5e/0xec
> May 21 09:05:16 h0515017 kernel: [<000000000015842a>] run_timer_softirq+0x246/0x354
> May 21 09:05:16 h0515017 kernel: [<00000000001514ae>] __do_softirq+0xfa/0x1b8
> May 21 09:05:16 h0515017 kernel: [<0000000000112608>] do_softirq+0xe4/0xe8
> May 21 09:05:16 h0515017 kernel: [<0000000000150f3c>] irq_exit+0xc0/0xe0
> May 21 09:05:16 h0515017 kernel: [<000000000010f9ce>] do_extint+0x126/0x17c
> May 21 09:05:16 h0515017 kernel: [<0000000000119e5c>] ext_no_vtime+0x28/0x2c
> May 21 09:05:16 h0515017 kernel: [<0000000000116c10>] vtime_stop_cpu+0xac/0x100
> May 21 09:05:16 h0515017 kernel: ([<0000000000116bb6>] vtime_stop_cpu+0x52/0x100)
> May 21 09:05:16 h0515017 kernel: [<000000000010c1a8>] cpu_idle+0xf8/0x190
> May 21 09:05:16 h0515017 kernel: [<0000000000859cf0>] start_kernel+0x3b4/0x418
> May 21 09:05:16 h0515017 kernel: [<0000000000100020>] _stext+0x20/0x80
> May 21 09:05:16 h0515017 kernel: INFO: lockdep is turned off.
> 
> Thanks,
> Heiko