Date:	Thu, 13 Sep 2012 18:17:56 +0530
From:	"Srivatsa S. Bhat" <srivatsa.bhat@...ux.vnet.ibm.com>
To:	Michael Wang <wangyun@...ux.vnet.ibm.com>,
	paulmck@...ux.vnet.ibm.com
CC:	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	Thomas Gleixner <tglx@...utronix.de>,
	"rusty@...tcorp.com.au" <rusty@...tcorp.com.au>,
	Peter Zijlstra <peterz@...radead.org>,
	Tejun Heo <tj@...nel.org>
Subject: Re: WARNING: at kernel/rcutree.c:1558 rcu_do_batch+0x386/0x3a0(),
 during CPU hotplug

On 09/13/2012 12:00 PM, Michael Wang wrote:
> On 09/12/2012 11:31 PM, Paul E. McKenney wrote:
>> On Wed, Sep 12, 2012 at 06:06:20PM +0530, Srivatsa S. Bhat wrote:
>>> On 07/19/2012 10:45 PM, Paul E. McKenney wrote:
>>>> On Thu, Jul 19, 2012 at 05:39:30PM +0530, Srivatsa S. Bhat wrote:
>>>>> Hi Paul,
>>>>>
>>>>> While running a CPU hotplug stress test on v3.5-rc7+
>>>>> (mainline commit 8a7298b7805ab) I hit this warning.
>>>>> I haven't tried to debug this yet...
>>>>>
>>>>> Line number 1550 maps to:
>>>>>
>>>>> WARN_ON_ONCE(cpu_is_offline(smp_processor_id()));
>>>>>
>>>>> inside rcu_do_batch().
>>>>
>>>> Hello, Srivatsa,
>>>>
>>>> I believe that you need commit a16b7a69 (Prevent __call_rcu() from
>>>> invoking RCU core on offline CPUs), which is currently in -tip, queued
>>>> for 3.6.  Please see below for the patch.
>>>>
>>>> Does this help?
>>>>
>>>
>>> Hi Paul,
>>>
>>> I am hitting the cpu_is_offline() warning in rcu_do_batch() (see 2 of the
>>> examples below) occasionally while testing CPU hotplug on Thomas' smp/hotplug
>>> branch in -tip. It does contain the commit that you had mentioned above.
>>>
>>> The stack trace suggests that we are not hitting this from the __call_rcu()
>>> path. So I guess this needs a different fix?
>>
>> So there was an interrupt from stop_machine_stop().  Because RCU complained
>> about offline, I presume that this was on exit from stop_machine_stop().
>> (Otherwise, on entry to stop_machine_stop(), the CPU has not yet marked
>> itself offline, right?)
>>

Yes, that's my understanding too.

>> So my question is:  Why didn't the CPU shut off all interrupts before
>> coming out of stop_machine_stop()?
>>
>> Or am I confused about what is really happening here?
> 
> I think Srivatsa may need patch a3716d2e too:
>

The changelog of this patch doesn't seem to suggest that it was written to
address the problem I am facing. Please see my thoughts below..

> commit a3716d2e5a50a9ed5268ae3d3c2f093968ff236a
> Author: Paul E. McKenney <paul.mckenney@...aro.org>
> Date:   Thu Jun 21 09:54:10 2012 -0700
> 
>     rcu: Prevent offline CPUs from executing RCU core code
>     
>     Earlier versions of RCU invoked the RCU core from the CPU_DYING notifier
>     in order to note a quiescent state for the outgoing CPU.  Because the
>     CPU is marked "offline" during the execution of the CPU_DYING notifiers,
>     the RCU core had to tolerate being invoked from an offline CPU.  However,
>     commit b1420f1c (Make rcu_barrier() less disruptive) left only tracing
>     code in the CPU_DYING notifier, so the RCU core need no longer execute
>     on offline CPUs.  This commit therefore enforces this restriction.
>     
>     Signed-off-by: Paul E. McKenney <paul.mckenney@...aro.org>
>     Signed-off-by: Paul E. McKenney <paulmck@...ux.vnet.ibm.com>
> 
> diff --git a/kernel/rcutree.c b/kernel/rcutree.c
> index 300aba6..84a6f55 100644
> --- a/kernel/rcutree.c
> +++ b/kernel/rcutree.c
> @@ -1892,6 +1892,8 @@ static void rcu_process_callbacks(struct softirq_action *unused)
>  {
>         struct rcu_state *rsp;
> 
> +       if (cpu_is_offline(smp_processor_id()))
> +               return;
>         trace_rcu_utilization("Start RCU core");
>         for_each_rcu_flavor(rsp)
>                 __rcu_process_callbacks(rsp);
> 
> With out it, the RCU_SOFTIRQ raised before offline won't
> have a chance to return.
> 

I'm sure this would help to get rid of the warning, but IMHO, like Paul pointed out,
its rather surprising that the CPU going offline got interrupts *after* coming out of
the stop_machine thing, iow, *after* it was marked offline.. So we need to examine
*this* issue right?

Regards,
Srivatsa S. Bhat

>>>
>>> [   53.882344] smpboot: CPU 7 is now offline
>>> [   53.891072] CPU 12 MCA banks CMCI:6 CMCI:8
>>> [   53.895621] CPU 15 MCA banks CMCI:2 CMCI:3 CMCI:5
>>> [   53.914738] Broke affinity for irq 81
>>> [   53.917769] do_IRQ: 8.211 No irq handler for vector (irq -1) 
>>> [   53.917769] ------------[ cut here ]------------
>>> [   53.917769] WARNING: at kernel/rcutree.c:1558 rcu_do_batch+0x386/0x3a0()
>>> [   53.917769] Hardware name: IBM System x -[7870C4Q]-
>>> [   53.917769] Modules linked in: ipv6 cpufreq_conservative cpufreq_userspace cpufreq_powersave acpi_cpufreq mperf fuse loop dm_mod iTCO_wdt iTCO_vendor_support coretemp kvm_intel kvm cdc_ether usbnet ioatdma lpc_ich mfd_core crc32c_intel microcode mii pcspkr i2c_i801 shpchp i2c_core serio_raw bnx2 tpm_tis dca tpm pci_hotplug i7core_edac tpm_bios edac_core sg rtc_cmos button uhci_hcd ehci_hcd usbcore usb_common sd_mod crc_t10dif edd ext3 mbcache jbd fan processor mptsas mptscsih mptbase scsi_transport_sas scsi_mod thermal thermal_sys hwmon
>>> [   53.917769] Pid: 47, comm: migration/8 Not tainted 3.6.0-rc1-tglx-hotplug-0.0.0.28.36b5ec9-default #1
>>> [   53.917769] Call Trace:
>>> [   53.917769]  <IRQ>  [<ffffffff810e7806>] ? rcu_do_batch+0x386/0x3a0
>>> [   53.917769]  [<ffffffff810e7806>] ? rcu_do_batch+0x386/0x3a0
>>> [   53.917769]  [<ffffffff8104338a>] warn_slowpath_common+0x7a/0xb0
>>> [   53.917769]  [<ffffffff810433d5>] warn_slowpath_null+0x15/0x20
>>> [   53.917769]  [<ffffffff810e7806>] rcu_do_batch+0x386/0x3a0
>>> [   53.917769]  [<ffffffff810ac8a0>] ? trace_hardirqs_on_caller+0x70/0x1b0
>>> [   53.917769]  [<ffffffff810ac9ed>] ? trace_hardirqs_on+0xd/0x10
>>> [   53.917769]  [<ffffffff810e9143>] __rcu_process_callbacks+0x1a3/0x200
>>> [   53.917769]  [<ffffffff810e9228>] rcu_process_callbacks+0x88/0x240
>>> [   53.917769]  [<ffffffff8104dc79>] __do_softirq+0x159/0x400
>>> [   53.917769]  [<ffffffff814c627c>] call_softirq+0x1c/0x30
>>> [   53.917769]  [<ffffffff810044f5>] do_softirq+0x95/0xd0
>>> [   53.917769]  [<ffffffff8104d745>] irq_exit+0xe5/0x100
>>> [   53.917769]  [<ffffffff81003c14>] do_IRQ+0x64/0xe0
>>> [   53.917769]  [<ffffffff814bc12f>] common_interrupt+0x6f/0x6f
>>> [   53.917769]  <EOI>  [<ffffffff810cfe7a>] ? stop_machine_cpu_stop+0xda/0x130
>>> [   53.917769]  [<ffffffff810cfda0>] ? stop_one_cpu_nowait+0x50/0x50
>>> [   53.917769]  [<ffffffff810cfab9>] cpu_stopper_thread+0xd9/0x1b0
>>> [   53.917769]  [<ffffffff814bbe4f>] ? _raw_spin_unlock_irqrestore+0x3f/0x80
>>> [   53.917769]  [<ffffffff810cf9e0>] ? res_counter_init+0x50/0x50
>>> [   53.917769]  [<ffffffff810ac95d>] ? trace_hardirqs_on_caller+0x12d/0x1b0
>>> [   53.917769]  [<ffffffff810ac9ed>] ? trace_hardirqs_on+0xd/0x10
>>> [   53.917769]  [<ffffffff810cf9e0>] ? res_counter_init+0x50/0x50
>>> [   53.917769]  [<ffffffff8106deae>] kthread+0xde/0xf0
>>> [   53.917769]  [<ffffffff814c6184>] kernel_thread_helper+0x4/0x10
>>> [   53.917769]  [<ffffffff814bc1f0>] ? retint_restore_args+0x13/0x13
>>> [   53.917769]  [<ffffffff8106ddd0>] ? __init_kthread_worker+0x70/0x70
>>> [   53.917769]  [<ffffffff814c6180>] ? gs_change+0x13/0x13
>>> [   53.917769] ---[ end trace f60a282810c4ce78 ]---
>>> [   54.170634] smpboot: CPU 8 is now offline
>>> [   54.192259] NOHZ: local_softirq_pending 200 
>>> [   54.197936] smpboot: CPU 9 is now offline
>>> [   54.219707] NOHZ: local_softirq_pending 200 
>>> [   54.225795] smpboot: CPU 10 is now offline
>>>
>>>
>>> ---
>>>
>>> [  372.482434] smpboot: CPU 11 is now offline
>>> [  372.534211] smpboot: CPU 12 is now offline
>>> [  372.539786] CPU 13 MCA banks CMCI:6 CMCI:8
>>> [  372.582474] smpboot: CPU 13 is now offline
>>> [  372.591006] CPU 14 MCA banks CMCI:6 CMCI:8
>>> [  372.629745] ------------[ cut here ]------------
>>> [  372.633736] WARNING: at kernel/rcutree.c:1558 rcu_do_batch+0x386/0x3a0()
>>> [  372.633736] Hardware name: IBM System x -[7870C4Q]-
>>> [  372.633736] Modules linked in: ipv6 cpufreq_conservative cpufreq_userspace cpufreq_powersave acpi_cpufreq mperf fuse loop dm_mod iTCO_wdt iTCO_vendor_support coretemp kvm_intel kvm crc32c_intel
>>> microcode serio_raw tpm_tis i7core_edac cdc_ether usbnet ioatdma mii lpc_ich pcspkr edac_core mfd_core bnx2 shpchp pci_hotplug i2c_i801 i2c_core dca tpm sg tpm_bios rtc_cmos button uhci_hcd ehci_hcd
>>> usbcore usb_common sd_mod crc_t10dif edd ext3 mbcache jbd fan processor mptsas mptscsih mptbase scsi_transport_sas scsi_mod thermal thermal_sys hwmon
>>> [  372.633736] Pid: 8625, comm: migration/14 Not tainted 3.6.0-rc1-tglx-hotplug-0.0.0.28.36b5ec9-default #1
>>> [  372.633736] Call Trace:
>>> [  372.633736]  <IRQ>  [<ffffffff810e7806>] ? rcu_do_batch+0x386/0x3a0
>>> [  372.633736]  [<ffffffff810e7806>] ? rcu_do_batch+0x386/0x3a0
>>> [  372.633736]  [<ffffffff8104338a>] warn_slowpath_common+0x7a/0xb0
>>> [  372.633736]  [<ffffffff810433d5>] warn_slowpath_null+0x15/0x20
>>> [  372.633736]  [<ffffffff810e7806>] rcu_do_batch+0x386/0x3a0
>>> [  372.633736]  [<ffffffff810ac8a0>] ? trace_hardirqs_on_caller+0x70/0x1b0
>>> [  372.633736]  [<ffffffff810ac9ed>] ? trace_hardirqs_on+0xd/0x10
>>> [  372.633736]  [<ffffffff810e9143>] __rcu_process_callbacks+0x1a3/0x200
>>> [  372.633736]  [<ffffffff810e9228>] rcu_process_callbacks+0x88/0x240
>>> [  372.633736]  [<ffffffff8104dc79>] __do_softirq+0x159/0x400
>>> [  372.633736]  [<ffffffff814c627c>] call_softirq+0x1c/0x30
>>> [  372.633736]  [<ffffffff810044f5>] do_softirq+0x95/0xd0
>>> [  372.633736]  [<ffffffff8104d745>] irq_exit+0xe5/0x100
>>> [  372.633736]  [<ffffffff81028df9>] smp_apic_timer_interrupt+0x69/0xa0
>>> [  372.633736]  [<ffffffff814c5aef>] apic_timer_interrupt+0x6f/0x80
>>> [  372.633736]  <EOI>  [<ffffffff810cfe7a>] ? stop_machine_cpu_stop+0xda/0x130
>>> [  372.633736]  [<ffffffff810cfda0>] ? stop_one_cpu_nowait+0x50/0x50
>>> [  372.633736]  [<ffffffff810cfab9>] cpu_stopper_thread+0xd9/0x1b0
>>> [  372.633736]  [<ffffffff814bbe4f>] ? _raw_spin_unlock_irqrestore+0x3f/0x80
>>> [  372.633736]  [<ffffffff810cf9e0>] ? res_counter_init+0x50/0x50
>>> [  372.633736]  [<ffffffff810ac95d>] ? trace_hardirqs_on_caller+0x12d/0x1b0
>>> [  372.633736]  [<ffffffff810ac9ed>] ? trace_hardirqs_on+0xd/0x10
>>> [  372.633736]  [<ffffffff810cf9e0>] ? res_counter_init+0x50/0x50
>>> [  372.633736]  [<ffffffff8106deae>] kthread+0xde/0xf0
>>> [  372.633736]  [<ffffffff814c6184>] kernel_thread_helper+0x4/0x10
>>> [  372.633736]  [<ffffffff814bc1f0>] ? retint_restore_args+0x13/0x13
>>> [  372.633736]  [<ffffffff8106ddd0>] ? __init_kthread_worker+0x70/0x70
>>> [  372.633736]  [<ffffffff814c6180>] ? gs_change+0x13/0x13
>>> [  372.633736] ---[ end trace a4296a31284c846d ]---
>>> [  372.883063] smpboot: CPU 14 is now offline
>>> [  372.892721] CPU 15 MCA banks CMCI:6 CMCI:8
>>> [  372.907250] smpboot: CPU 15 is now offline
>>> [  372.911545] SMP alternatives: lockdep: fixing up alternatives
>>> [  372.917292] SMP alternatives: switching to UP code
>>> [  372.941917] SMP alternatives: lockdep: fixing up alternatives

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Powered by blists - more mailing lists