[Skiboot] softlockup with irq/opal-elog

[Vasant Hegde]

On 06/23/2016 12:28 PM, Alistair Popple wrote:
> Stewart,
>
> It may be related to https://bugzilla.linux.ibm.com/show_bug.cgi?id=142332. I asked to Vasant to add a couple of prints to work out what was going on and there is also http://patchwork.ozlabs.org/patch/634230/ which looks relevant.


We have  an issue in ELOG OPAL code ... which is not disabling event flag in 
some scenario.

Will send fix soon.


-Vasant



>
> - Alistair
>
> On Thu, 23 Jun 2016 16:53:33 Stewart Smith wrote:
>> "Aneesh Kumar K.V" <aneesh.kumar at linux.vnet.ibm.com> writes:
>>> Hi,
>>>
>>> I am hitting this on a power7 big endian config. Once I clear the error,
>>> the host boots ok.
>>>
>>> NMI watchdog: BUG: soft lockup - CPU#27 stuck for 23s! [irq/34-opal-elo:509]
>>> Modules linked in:
>>> irq event stamp: 7326292
>>> hardirqs last  enabled at (7326291): [<c000000000dff258>] ._raw_spin_unlock_irq+0x38/0x70
>>> hardirqs last disabled at (7326292): [<c000000000002748>] decrementer_common+0x148/0x180
>>> softirqs last  enabled at (7323620): [<c0000000000ed170>] .__do_softirq+0x5b0/0x760
>>> softirqs last disabled at (7323615): [<c0000000000ed5c8>] .irq_exit+0xd8/0x120
>>> CPU: 27 PID: 509 Comm: irq/34-opal-elo Tainted: G        W       4.7.0-rc1-11926-g468e20a #1
>>> task: c000000ff0a08880 ti: c000000ff0a8c000 task.ti: c000000ff0a8c000
>>> NIP: c000000000010734 LR: c000000000010734 CTR: 00000000300551d0
>>> REGS: c000000ff0a8f760 TRAP: 0901   Tainted: G        W        (4.7.0-rc1-11926-g468e20a)
>>> MSR: 9000000000009032 <SF,HV,EE,ME,IR,DR,RI>  CR: 22000444  XER: 00000000
>>> CFAR: c00000000021dfc0 SOFTE: 1
>>> GPR00: c000000000dff264 c000000ff0a8f9e0 c0000000015b7300 0000000000000900
>>> GPR04: 0000000000000001 0000000000000001 9000000000009032 000000000000001b
>>> GPR08: 0000000000000000 c000000ff0a8c000 0000000000000000 0000000000000000
>>> GPR12: 0000000048000442 c00000000fb8f300
>>> NIP [c000000000010734] .arch_local_irq_restore.part.4+0x84/0xb0
>>> LR [c000000000010734] .arch_local_irq_restore.part.4+0x84/0xb0
>>> Call Trace:
>>> [c000000ff0a8fa50] [c000000000dff264] ._raw_spin_unlock_irq+0x44/0x70
>>> [c000000ff0a8fad0] [c00000000016e584] .irq_finalize_oneshot.part.2+0x84/0x1c0
>>> [c000000ff0a8fb60] [c00000000016e7f4] .irq_thread_fn+0x74/0xa0
>>> [c000000ff0a8fbf0] [c00000000016ec60] .irq_thread+0x1e0/0x280
>>> [c000000ff0a8fcd0] [c000000000116ad4] .kthread+0x114/0x140
>>> [c000000ff0a8fe30] [c0000000000095f0] .ret_from_kernel_thread+0x58/0x68
>>> Instruction dump:
>>> 409e002c e92d0020 61298000 7d210164 38210070 e8010010 7c0803a6 4e800020
>>> 60000000 60000000 60000000 4bff1a4d <60000000> 38210070 e8010010 7c0803a6
>>
>> Alistair/Vasant: something you've seen?
>>
>>
>