RE: [PATCH] optimize writer path in time_interpolator_get_counter()

["Luck, Tony" <tony.luck@intel.com>]

>Think about a threaded process that gets time on multiple processors 
>and then compares the times. This means that the time value obtained later 
>on one thread may indicate a time earlier than that obtained on another 
>thread. An essential requirement for time values is that they are 
>monotonically increasing. You are changing that basic nature.

But this comes down to how much time does it take for two threads
to perform some synchronization operation so that they know that one
thread made the call before the other[1]?  I think that it might be
possible to make the claim[2] that we have synchonized the ITC values
more closely than the fastest user synchronization method ... hence
a simplistic kernel implementation will be monotonic in practice.

-Tony

[1] It is pointless to expect one return value to be greater than
another unless we know that the calls were made in a particular sequence.

[2] Hands are waving so wildly here that it is a wonder that I can
even type.
-
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/