Re: [PATCH bpf-next v5 2/3] perf: Add atomic operation in get_recursion_context

[Tao Chen <chen.dylane@linux.dev>]

在 2025/11/10 16:52, Peter Zijlstra 写道:
> On Mon, Nov 10, 2025 at 12:35:58AM +0800, Tao Chen wrote:
>>  From BPF side, preemption usually is enabled. Yonghong said, it is
>> possible that both tasks (at process level) may reach right before
>> "recursion[rctx]++;". In such cases, both tasks will be able to get
>> buffer and this is not right.
>>
>> Signed-off-by: Yonghong Song <yonghong.song@linux.dev>
>> Signed-off-by: Tao Chen <chen.dylane@linux.dev>
>> ---
> 
> Nope, this function really is meant to be used with preemption disabled.
> If BPF doesn't abide, fix that.
> 

Ok, let us use preempt_disable in bpf stackmap, thanks. I will change it 
in v6.

>>   kernel/events/internal.h | 5 +----
>>   1 file changed, 1 insertion(+), 4 deletions(-)
>>
>> diff --git a/kernel/events/internal.h b/kernel/events/internal.h
>> index d9cc5708309..684bde972ba 100644
>> --- a/kernel/events/internal.h
>> +++ b/kernel/events/internal.h
>> @@ -214,12 +214,9 @@ static inline int get_recursion_context(u8 *recursion)
>>   {
>>   	unsigned char rctx = interrupt_context_level();
>>   
>> -	if (recursion[rctx])
>> +	if (cmpxchg(&recursion[rctx], 0, 1) != 0)
>>   		return -1;
>>   
>> -	recursion[rctx]++;
>> -	barrier();
>> -
>>   	return rctx;
>>   }
>>   
>> -- 
>> 2.48.1
>>


-- 
Best Regards
Tao Chen