[Crash-utility] [PATCH 2/3] crash: ARM: get online CPUs from cpu_active_mask

[Liu hua]

于 2014/4/28 21:30, Dave Anderson 写道:
> 
> 
> ----- Original Message -----
>> 于 2014/4/27 7:43, Dave Anderson 写道:
>>>
>>>
>>> ----- Original Message -----
>>>> On 2014/4/26 3:20, Dave Anderson wrote:
>>>>>
>>>>>
>>>>> ----- Original Message -----
>>>>>> For ARM32 platfrom, The system will "offline" all CPUs except the
>>>>>> crashing one, by clear the cpu_online_mask. So we need to find
>>>>>> another way to get online-CPUs number for crash utility.
>>>>>>
>>>>>> This patch uses cpu_active_mask to get that value.
>>>>>>
>>>>>> Signed-off-by: Liu Hua <sdu.liu at huawei.com>
>>>>>> ---
>>>>>>  arm.c | 2 +-
>>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>>
>>>>>> diff --git a/arm.c b/arm.c
>>>>>> index 3c38cd5..c94d7bf 100644
>>>>>> --- a/arm.c
>>>>>> +++ b/arm.c
>>>>>> @@ -1518,7 +1518,7 @@ arm_display_machine_stats(void)
>>>>>>  static int
>>>>>>  arm_get_smp_cpus(void)
>>>>>>  {
>>>>>> -	return get_cpus_online();
>>>>>> +	return get_cpus_active();
>>>>>>  }
>>>>>>  
>>>>>>  /*
>>>>>
>>>>> When did this start happening?  I ask because no other ARM users have
>>>>> reported this until now.  I've only got one sample SMP vmcore, and
>>>>> clearly the non-crashing cpu was not offlined:
>>>>>
>>>>
>>>>
>>>> Hi Dave,
>>>>
>>>>  commit 36a2e5010fe87 "kexec: offline non panic CPUs on Kdump panic "
>>>>  introduces this change. <v3.10.27-1-g36a2e50>
>>>>
>>>>  The following messges come from the ARM boards on my hand. The crash
>>>>  utility has added my patchs.
>>>>
>>>> (1) ARM A15 16 CPUs
>>>>
>>>>   crash> sys
>>>>
>>>>       KERNEL: vmlinux
>>>>     DUMPFILE: vmcore_my  [PARTIAL DUMP]
>>>>         CPUS: 16
>>>>         DATE: Fri Feb 28 10:07:55 2014
>>>>       UPTIME: 00:03:57
>>>> LOAD AVERAGE: 0.03, 0.01, 0.01
>>>>        TASKS: 140
>>>>     NODENAME: arma15el
>>>>      RELEASE: 3.10.37+
>>>>      VERSION: #2 SMP Thu Apr 24 21:13:28 CST 2014
>>>>      MACHINE: armv7l  (unknown Mhz)
>>>>       MEMORY: 3.1 GB
>>>>        PANIC: "Internal error: Oops: a07 [#1] SMP ARM" (check log for
>>>>        details
>>>>
>>>>   crash> help -k
>>>>        ....
>>>>        cpu_possible_map: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
>>>>        cpu_present_map: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
>>>>        cpu_online_map: 0
>>>>        ...
>>>>
>>>>
>>>> (2) ARM A9 4 CPUs
>>>>
>>>> crash> sys
>>>>       KERNEL: vmlinux
>>>>     DUMPFILE: vmcore
>>>>         CPUS: 2
>>>>         DATE: Mon Apr 14 11:09:10 2014
>>>>       UPTIME: 00:06:17
>>>> LOAD AVERAGE: 0.16, 0.05, 0.02
>>>>        TASKS: 45
>>>>     NODENAME: arma9el
>>>>      RELEASE: 3.10.37+
>>>>      VERSION: #6 SMP Tue Apr 22 13:23:30 CST 2014
>>>>      MACHINE: armv7l  (unknown Mhz)
>>>>       MEMORY: 1.5 GB
>>>>        PANIC: "Internal error: Oops: 817 [#1] SMP ARM" (check log for
>>>>        details)
>>>>
>>>>   crash> help -k
>>>>        ....
>>>> 	cpu_possible_map: 0 1
>>>>         cpu_present_map: 0 1
>>>>         cpu_online_map: 1
>>>>
>>>>
>>>>    crash> p per_cpu__runqueues
>>>>
>>>>    PER-CPU DATA TYPE:
>>>>    struct rq runqueues;
>>>>    PER-CPU ADDRESSES:
>>>>    [0]: c069a680
>>>>    [1]: c06a2680
>>>>
>>>> With my patches, "sys" command is ok. And I can get per_cpu variables
>>>> correctly.
>>>> But something should be done for "help -k". May be I should add fully
>>>> support
>>>> for cpu_active_mask. If necessory, I will resend this patch series.
>>>
>>> It's not necessary, but it would be nice to have, and would be
>>> easy to add.
>>>
>>> Also, would it be possible that you can let me have a copy of your
>>> 16-cpu vmlinux/vmcore pair?  That would be a great sample for
>>> my testing.
>>>
>>> Thanks,
>>>   Dave
>>
>> Hi Dave,
>>
>> Sure, there seems no extra problems caused by the commit above. We can add it
>> when necessory.
> 
> Will you be re-posting the original patch?  (with or without the help -k support)


I will repost patches soon, with help -k support.
> 
>>
>> And, I am very sorry to tell you that I have no authority for sending binary files out of my
>> company, becasue of it's security rules. I have no ways to send that huge file out. Maybe
>> I can do some test for you on my platform.
> 
> OK I understand. 
> 
> Thanks,
>   Dave
> 
> .
>