Re: [External] Re: [PATCH 2/2] sched: mark PRINTK_DEFERRED_CONTEXT_MASK in __schedule()

[Chengming Zhou <zhouchengming@bytedance.com>]

在 2020/9/28 下午3:32, Peter Zijlstra 写道:
> On Mon, Sep 28, 2020 at 12:11:30AM +0800, Chengming Zhou wrote:
>> The WARN_ON/WARN_ON_ONCE with rq lock held in __schedule() should be
>> deferred by marking the PRINTK_DEFERRED_CONTEXT_MASK, or will cause
>> deadlock on rq lock in the printk path.
> It also shouldn't happen in the first place, so who bloody cares.

Yes, but if our box deadlock just because a WARN_ON_ONCE, we have to
reboot : (

So these WARN_ON_ONCE have BUG_ON effect ?  Or we should change to use
BUG_ON ?

Thanks.

>> Signed-off-by: Chengming Zhou <zhouchengming@bytedance.com>
>> Signed-off-by: Muchun Song <songmuchun@bytedance.com>
>> ---
>>  kernel/sched/core.c | 2 ++
>>  1 file changed, 2 insertions(+)
>>
>> diff --git a/kernel/sched/core.c b/kernel/sched/core.c
>> index 2d95dc3f4644..81d8bf614225 100644
>> --- a/kernel/sched/core.c
>> +++ b/kernel/sched/core.c
>> @@ -4444,6 +4444,7 @@ static void __sched notrace __schedule(bool preempt)
>>  	 */
>>  	rq_lock(rq, &rf);
>>  	smp_mb__after_spinlock();
>> +	printk_deferred_enter();
>>  
>>  	/* Promote REQ to ACT */
>>  	rq->clock_update_flags <<= 1;
>> @@ -4530,6 +4531,7 @@ static void __sched notrace __schedule(bool preempt)
>>  		rq_unlock_irq(rq, &rf);
>>  	}
>>  
>> +	printk_deferred_exit();
>>  	balance_callback(rq);
>>  }
> NAK printk_deferred is an abomination, kill that.