Re: (no subject)

Johnicholas Hines wrote:
> I have a trace where the reads and writes of the same message (as
> indicated by a unique message id) are timestamped first read then
> write.

I don't know the details of where the trace instrumentation
is for reads and writes, but it's possible that the log has
event start times and not event completion times.  If this
is the case, it's entirely possible for a read to have an
earlier timestamp than a write (even for the same data).
(That is, the read could have been started before the write,
then blocked, then completed after the write).

This is just a theory.
 -- Tim

=============================
Tim Bird
Architecture Group Chair, CE Linux Forum
Senior Staff Engineer, Sony Electronics
=============================