Re: [PATCH 1/4] rtmutex: avoid null derefence in WARN_ON

[Darren Hart <dvhltc@us.ibm.com>]

On 07/09/2010 05:29 PM, Steven Rostedt wrote:
> On Fri, 2010-07-09 at 15:32 -0700, Darren Hart wrote:
>> If the pi_blocked_on variable is NULL, the subsequent WARN_ON's
>> will cause an OOPS. Only perform the susequent checks if
>> pi_blocked_on is valid.
>>
>> Signed-off-by: Darren Hart<dvhltc@us.ibm.com>
>> Cc: Thomas Gleixner<tglx@linutronix.de>
>> Cc: Peter Zijlstra<peterz@infradead.org>
>> Cc: Ingo Molnar<mingo@elte.hu>
>> Cc: Eric Dumazet<eric.dumazet@gmail.com>
>> Cc: John Kacur<jkacur@redhat.com>
>> Cc: Steven Rostedt<rostedt@goodmis.org>
>> Cc: Mike Galbraith<efault@gmx.de>
>> ---
>>   kernel/rtmutex.c |    7 ++++---
>>   1 files changed, 4 insertions(+), 3 deletions(-)
>>
>> diff --git a/kernel/rtmutex.c b/kernel/rtmutex.c
>> index 23dd443..baac7d9 100644
>> --- a/kernel/rtmutex.c
>> +++ b/kernel/rtmutex.c
>> @@ -579,9 +579,10 @@ static void wakeup_next_waiter(struct rt_mutex *lock, int savestate)
>>
>>   	raw_spin_lock(&pendowner->pi_lock);
>>
>> -	WARN_ON(!pendowner->pi_blocked_on);
>> -	WARN_ON(pendowner->pi_blocked_on != waiter);
>> -	WARN_ON(pendowner->pi_blocked_on->lock != lock);
>> +	if (!WARN_ON(!pendowner->pi_blocked_on)) {
>> +		WARN_ON(pendowner->pi_blocked_on != waiter);
>
> The above actually has no issue if the pi_blocked_on is NULL.

It doesn't, but it's also redundant and makes the console noisier for no 
reason. Seemed worth while to drop it under the if in the same go.

--
Darren


> The below, well yeah.
>
> -- Steve
>
>> +		WARN_ON(pendowner->pi_blocked_on->lock != lock);
>> +	}
>>
>>   	pendowner->pi_blocked_on = NULL;
>>
>
>


-- 
Darren Hart
IBM Linux Technology Center
Real-Time Linux Team