Re: [Linuxptp-devel] Peer delay regression in v1.6

[Richard C. <ric...@gm...>]

> The math in the tsproc code was supposed to be identical to the code
> it replaced. t2 and t3 in tsproc are local times. The mapping between
> the port and tsproc timestamps is (t1, t2, t3, t4) -> (t3, t4, t1, t2).

Oh, tricky!  Now I vaguely recall that from the original review.

> Maybe it wasn't the best idea to use the same variable names in all
> three (clock, port, tsproc) contexts?

The tsproc unifies the e2e and p2p methods in a good way.  But the
different meaning of t1-t4 could use a comment, in order keep people
like me from getting confused.
 
> Anyway, does this actually break something?

I guess not then.  I did try v1.5 an v1.6 on a test system, and I was
puzzled why I couldn't see the effect of this bug.

Thanks,
Richard