garrett_damore@tadpole.com wrote:

> Anyway, what I have works, imposes no noticeable performance penalty,
> and is easily maintainable.  Is there some reason we are arguing about this?

Well, to be honest, I don't care about the performance
penalty so much, but I just care about API consistency.
I.e. introducing stride only for stride is stupid for me
(and maybe other people mentioned performance to object it),
but once some device which actually regmap method, it's
acceptable for device to use regmap instead of implementing
stride ops.

That's all, and it has already happened on wdc(4), which now
has multiple bus space handles for each register even though
many people had implemented stride ops for their attachments.
---
Izumi Tsutsui