Hello,

On 11/28/2010 09:30 PM, Oleg Nesterov wrote:
quoted text
> On 11/26, Tejun Heo wrote:
>>
>> Group stop notifications are unreliable if one or more tasks of the
>> task group are being ptraced.  If a ptraced task ends up finishing a
>> group stop, the notification is sent to the ptracer and the real
>> parent never gets notified.
> 
> Yes. I do not even know if this is bug or not, but certainly I agree,
> this doesn't look very nice.
> 
>>  	if (likely(!task_ptrace(current))) {
>> +		bool do_notify = false;
>> +
>> +		if (sig->flags & SIGNAL_NOTIFY_STOP) {
>> +			sig->flags &= ~SIGNAL_NOTIFY_STOP;
>> +			do_notify = true;
>> +		}
>> +
>>  		spin_unlock_irq(&current->sighand->siglock);
>>
>> -		if (notify) {
>> +		if (do_notify) {
>>  			read_lock(&tasklist_lock);
>> -			do_notify_parent_cldstop(current, notify);
>> +			do_notify_parent_cldstop(current, CLD_STOPPED);
> 
> This can race with ptrace_attach() in between.
> 
> IOW, this notification can go to the new tracer anyway.

Hmmm, okay, we should hold both locks when checking for notification
to remove that race, or we can just tell do_notify_parent_cldstop()
which parent to use.

-- 
tejun
--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/