ARM:kexec:offline panic_smp_self_stop CPU
diff mbox series

Message ID 1541071249-15660-1-git-send-email-wangyufen@huawei.com
State New
Headers show
Series
  • ARM:kexec:offline panic_smp_self_stop CPU
Related show

Commit Message

wangyufen Nov. 1, 2018, 11:20 a.m. UTC
From: Yufen Wang <wangyufen@huawei.com>

In case panic() and panic() called at the same time on different CPUS.
For example:
CPU 0:
  panic()
     __crash_kexec
       machine_crash_shutdown
         crash_smp_send_stop
       machine_kexec
         BUG_ON(num_online_cpus() > 1);

CPU 1:
  panic()
    local_irq_disable
    panic_smp_self_stop

If CPU 1 calls panic_smp_self_stop() before crash_smp_send_stop(), kdump
fails. CPU1 can't receive the ipi irq, CPU1 will be always online.
I changed BUG_ON to WARN in kexec crash as arm64 does, kdump also fails.
Because num_online_cpus() > 1, can't disable the L2 in _soft_restart.
To fix this problem, this patch split out the panic_smp_self_stop()
and add set_cpu_online(smp_processor_id(), false).

Signed-off-by: Yufen Wang <wangyufen@huawei.com>
---
 arch/arm/kernel/setup.c | 10 ++++++++++
 1 file changed, 10 insertions(+)

Comments

Russell King - ARM Linux admin Nov. 1, 2018, 11:34 a.m. UTC | #1
On Thu, Nov 01, 2018 at 07:20:49PM +0800, Wang Yufen wrote:
> From: Yufen Wang <wangyufen@huawei.com>
> 
> In case panic() and panic() called at the same time on different CPUS.
> For example:
> CPU 0:
>   panic()
>      __crash_kexec
>        machine_crash_shutdown
>          crash_smp_send_stop
>        machine_kexec
>          BUG_ON(num_online_cpus() > 1);
> 
> CPU 1:
>   panic()
>     local_irq_disable
>     panic_smp_self_stop
> 
> If CPU 1 calls panic_smp_self_stop() before crash_smp_send_stop(), kdump
> fails. CPU1 can't receive the ipi irq, CPU1 will be always online.
> I changed BUG_ON to WARN in kexec crash as arm64 does, kdump also fails.
> Because num_online_cpus() > 1, can't disable the L2 in _soft_restart.
> To fix this problem, this patch split out the panic_smp_self_stop()
> and add set_cpu_online(smp_processor_id(), false).

Thanks.

I think this may as well go into arch/arm/kernel/smp.c - it won't be
required for single-CPU systems, since there aren't "other" CPUs.

It's probably also worth a comment above the function as to why we
have this.

> 
> Signed-off-by: Yufen Wang <wangyufen@huawei.com>
> ---
>  arch/arm/kernel/setup.c | 10 ++++++++++
>  1 file changed, 10 insertions(+)
> 
> diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
> index 31940bd..151861f 100644
> --- a/arch/arm/kernel/setup.c
> +++ b/arch/arm/kernel/setup.c
> @@ -602,6 +602,16 @@ static void __init smp_build_mpidr_hash(void)
>  }
>  #endif
>  
> +void panic_smp_self_stop(void)
> +{
> +	printk(KERN_DEBUG "CPU %u will stop doing anything useful since another CPU has paniced\n",
> +			smp_processor_id());
> +	set_cpu_online(smp_processor_id(), false);
> +	while (1)
> +		cpu_relax();
> +
> +}
> +
>  static void __init setup_processor(void)
>  {
>  	struct proc_info_list *list;
> -- 
> 2.7.4
> 
>
wangyufen Nov. 2, 2018, 1:17 a.m. UTC | #2
On 2018/11/1 19:34, Russell King - ARM Linux wrote:
> On Thu, Nov 01, 2018 at 07:20:49PM +0800, Wang Yufen wrote:
>> From: Yufen Wang <wangyufen@huawei.com>
>>
>> In case panic() and panic() called at the same time on different CPUS.
>> For example:
>> CPU 0:
>>   panic()
>>      __crash_kexec
>>        machine_crash_shutdown
>>          crash_smp_send_stop
>>        machine_kexec
>>          BUG_ON(num_online_cpus() > 1);
>>
>> CPU 1:
>>   panic()
>>     local_irq_disable
>>     panic_smp_self_stop
>>
>> If CPU 1 calls panic_smp_self_stop() before crash_smp_send_stop(), kdump
>> fails. CPU1 can't receive the ipi irq, CPU1 will be always online.
>> I changed BUG_ON to WARN in kexec crash as arm64 does, kdump also fails.
>> Because num_online_cpus() > 1, can't disable the L2 in _soft_restart.
>> To fix this problem, this patch split out the panic_smp_self_stop()
>> and add set_cpu_online(smp_processor_id(), false).
> Thanks.
>
> I think this may as well go into arch/arm/kernel/smp.c - it won't be
> required for single-CPU systems, since there aren't "other" CPUs.
>
> It's probably also worth a comment above the function as to why we
> have this.

Thanks.

I will send v2.

>> Signed-off-by: Yufen Wang <wangyufen@huawei.com>
>> ---
>>  arch/arm/kernel/setup.c | 10 ++++++++++
>>  1 file changed, 10 insertions(+)
>>
>> diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
>> index 31940bd..151861f 100644
>> --- a/arch/arm/kernel/setup.c
>> +++ b/arch/arm/kernel/setup.c
>> @@ -602,6 +602,16 @@ static void __init smp_build_mpidr_hash(void)
>>  }
>>  #endif
>>  
>> +void panic_smp_self_stop(void)
>> +{
>> +	printk(KERN_DEBUG "CPU %u will stop doing anything useful since another CPU has paniced\n",
>> +			smp_processor_id());
>> +	set_cpu_online(smp_processor_id(), false);
>> +	while (1)
>> +		cpu_relax();
>> +
>> +}
>> +
>>  static void __init setup_processor(void)
>>  {
>>  	struct proc_info_list *list;
>> -- 
>> 2.7.4
>>
>>

Patch
diff mbox series

diff --git a/arch/arm/kernel/setup.c b/arch/arm/kernel/setup.c
index 31940bd..151861f 100644
--- a/arch/arm/kernel/setup.c
+++ b/arch/arm/kernel/setup.c
@@ -602,6 +602,16 @@  static void __init smp_build_mpidr_hash(void)
 }
 #endif
 
+void panic_smp_self_stop(void)
+{
+	printk(KERN_DEBUG "CPU %u will stop doing anything useful since another CPU has paniced\n",
+			smp_processor_id());
+	set_cpu_online(smp_processor_id(), false);
+	while (1)
+		cpu_relax();
+
+}
+
 static void __init setup_processor(void)
 {
 	struct proc_info_list *list;