[x265] [PATCH] rc: fixes inconsistent output in linux because of RC Lock in CQP/CRF

[Steve Borho]

On 06/29, Deepthi Nandakumar wrote:
> Should m_startEndOrder be a member of Encoder, rather than ratecontrol,
> since it serializes RcStarts, slice context inits as well as   some CTU row
> encodes?

it might be an even worse layering violation for rate-control to
directly manipulate an object in the top-level encoder.

I'm thinking the best option would be to add a new RC method that is
called just before the CTU compress loop. which increments the counter
when RC is in the appropriate mode. Or perhaps just calling
updateStats() early for these modes.

-- 
Steve Borho