[PATCH v2] x86: Fix an RT MCE crash

[PATCH v2] x86: Fix an RT MCE crash

[minyard]

From: Corey Minyard <cminyard@mvista.com>

On some x86 systems an MCE interrupt would come in before the kernel
was ready for it.  Looking at the latest RT code, it has similar
(but not quite the same) code, except it adds a bool that tells if
MCE handling is initialized.  That was required because they had
switched to use swork instead of a kernel thread.  Here, just
checking to see if the thread is NULL is good enough to see if
MCE handling is initialized.

Signed-off-by: Corey Minyard <cminyard@mvista.com>
---
 arch/x86/kernel/cpu/mcheck/mce.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index aaf4b9b..cc70d98 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -1391,7 +1391,8 @@ static int mce_notify_work_init(void)
 
 static void mce_notify_work(void)
 {
-	wake_up_process(mce_notify_helper);
+	if (mce_notify_helper)
+		wake_up_process(mce_notify_helper);
 }
 #else
 static void mce_notify_work(void)
-- 
2.7.4

Re: [PATCH v2] x86: Fix an RT MCE crash

[Corey Minyard]

I forgot to mention that this applies to v3.2-rt, v3.4-rt,
v3.10-rt, v3.12-rt, and v3.14-rt.  I didn't see a v3.16-rt.

-corey

On 07/06/2016 01:55 PM, minyard@acm.org wrote:
> From: Corey Minyard <cminyard@mvista.com>
>
> On some x86 systems an MCE interrupt would come in before the kernel
> was ready for it.  Looking at the latest RT code, it has similar
> (but not quite the same) code, except it adds a bool that tells if
> MCE handling is initialized.  That was required because they had
> switched to use swork instead of a kernel thread.  Here, just
> checking to see if the thread is NULL is good enough to see if
> MCE handling is initialized.
>
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> ---
>   arch/x86/kernel/cpu/mcheck/mce.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
> index aaf4b9b..cc70d98 100644
> --- a/arch/x86/kernel/cpu/mcheck/mce.c
> +++ b/arch/x86/kernel/cpu/mcheck/mce.c
> @@ -1391,7 +1391,8 @@ static int mce_notify_work_init(void)
>   
>   static void mce_notify_work(void)
>   {
> -	wake_up_process(mce_notify_helper);
> +	if (mce_notify_helper)
> +		wake_up_process(mce_notify_helper);
>   }
>   #else
>   static void mce_notify_work(void)

Re: [PATCH v2] x86: Fix an RT MCE crash

[Steven Rostedt]

On Wed,  6 Jul 2016 13:55:01 -0500
minyard@acm.org wrote:

> From: Corey Minyard <cminyard@mvista.com>
> 
> On some x86 systems an MCE interrupt would come in before the kernel
> was ready for it.  Looking at the latest RT code, it has similar
> (but not quite the same) code, except it adds a bool that tells if
> MCE handling is initialized.  That was required because they had
> switched to use swork instead of a kernel thread.  Here, just
> checking to see if the thread is NULL is good enough to see if
> MCE handling is initialized.
> 
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> ---
>  arch/x86/kernel/cpu/mcheck/mce.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
> index aaf4b9b..cc70d98 100644
> --- a/arch/x86/kernel/cpu/mcheck/mce.c
> +++ b/arch/x86/kernel/cpu/mcheck/mce.c
> @@ -1391,7 +1391,8 @@ static int mce_notify_work_init(void)
>  
>  static void mce_notify_work(void)
>  {
> -	wake_up_process(mce_notify_helper);
> +	if (mce_notify_helper)
> +		wake_up_process(mce_notify_helper);
>  }
>  #else
>  static void mce_notify_work(void)


Wait, is this all that's needed?

Also, Boris should be credited here somewhere.

-- Steve

Re: [PATCH v2] x86: Fix an RT MCE crash

[Corey Minyard]

On 07/11/2016 12:34 PM, Steven Rostedt wrote:
> On Wed,  6 Jul 2016 13:55:01 -0500
> minyard@acm.org wrote:
>
>> From: Corey Minyard <cminyard@mvista.com>
>>
>> On some x86 systems an MCE interrupt would come in before the kernel
>> was ready for it.  Looking at the latest RT code, it has similar
>> (but not quite the same) code, except it adds a bool that tells if
>> MCE handling is initialized.  That was required because they had
>> switched to use swork instead of a kernel thread.  Here, just
>> checking to see if the thread is NULL is good enough to see if
>> MCE handling is initialized.
>>
>> Signed-off-by: Corey Minyard <cminyard@mvista.com>
>> ---
>>   arch/x86/kernel/cpu/mcheck/mce.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
>> index aaf4b9b..cc70d98 100644
>> --- a/arch/x86/kernel/cpu/mcheck/mce.c
>> +++ b/arch/x86/kernel/cpu/mcheck/mce.c
>> @@ -1391,7 +1391,8 @@ static int mce_notify_work_init(void)
>>   
>>   static void mce_notify_work(void)
>>   {
>> -	wake_up_process(mce_notify_helper);
>> +	if (mce_notify_helper)
>> +		wake_up_process(mce_notify_helper);
>>   }
>>   #else
>>   static void mce_notify_work(void)
>
> Wait, is this all that's needed?

Yes, that's it.

> Also, Boris should be credited here somewhere.

Indeed, he should.  I'll issue V3.

-corey

> -- Steve