> 
> However, perf_evlist__id_hdr_size() seems to be affected also.  Are you
> sure it's okay too?
> 
Yes.
id_hdr_size is calculated according to the first event. It impacts 
perf tool generated COMM, MMAP and FORK events.
But all of these events have same and position fixed IDENTIFIER, 0.
When parsing the events by perf_evlist__event2evsel, id 0 always
return first event.
So only first event will be involved to parse perf tool generated
COMM, MMAP and FORK events. TIME setting from other events
doesn’t impact.

Thanks,
Kan
���{.n�+�������+%�����ݶ��w��{.n�+����{��G�����{ay�ʇڙ�,j��f���h��������z_�(�階�ݢj"���m������G������?����&���~��iO��z��v�^�m����������?�I�