garrett_damore@tadpole.com wrote:

> Getting it in a *branch* ready for use is meaningful.  The loudest
> detractors (other than yourself), have been folks who've objected on the
> basis of performance.  So having the ability to actually test it makes
> sense.

Sorry but I didn't insist performance issue on regmap method.
I won't object regmap method if it's introduced to support aucom
even if it actually caused performance hit.

> By the way, this was discussed and I received other feedback than just
> yours.  Including discussion from Charles, Christos, Jason, etc.

Hmm. then it was much better to note it in your first post because
I couldn't hear it at all. If they accepted your COM_INIT_REGS
implementation, go ahead to commit it and no need to hear
my dumb comments ;-)
---
Izumi Tsutsui