Re: [PATCH v12] perf: Sharing PMU counters across compatible events

[Song Liu <songliubraving@fb.com>]

> On Apr 20, 2020, at 4:04 PM, Peter Zijlstra <peterz@infradead.org> wrote:
>> 
>> static inline u64 perf_event_count(struct perf_event *event)
>> {
>> -	return local64_read(&event->count) + atomic64_read(&event->child_count);
>> +	u64 count;
>> +
>> +	if (likely(event->dup_master != event))
>> +		count = local64_read(&event->count);
>> +	else
>> +		count = local64_read(&event->master_count);
>> +
>> +	return count + atomic64_read(&event->child_count);
>> }
> 
> So lsat time I said something about SMP ordering here. Where did that
> go?

I am not quite sure about this one. I guess the following is sufficient. 
Could you please help me double check? 

Thanks,
Song


static inline u64 perf_event_count(struct perf_event *event)
{
        struct perf_event *master;
        u64 count;

again:
        master = READ_ONCE(event->dup_master);
        if (master == EVENT_TOMBSTONE)
                goto again;

        if (likely(master != event))
                count = local64_read(&event->count);
        else
                count = local64_read(&event->master_count);

        return count + atomic64_read(&event->child_count);
}