[RFC] Improving udelay/ndelay on platforms where that is possible

[Marc Gonzalez]

On 01/11/2017 20:09, Linus Torvalds wrote:

> On Wed, Nov 1, 2017 at 12:03 PM, Marc Gonzalez wrote:
>
>> By default, ndelay is implemented in terms of udelay.
> 
> That's very much *NOT* the case.
> 
> Yes, there is a *fallback* for when somebody doesn't do ndelay() at
> all, but that doesn't make it the default.
> 
> It's just a "the architecture didn't implement ndelay at all, we'll
> work around it".

Yes, sorry, I wrote "default" when I meant "fallback".
(arm32 currently does not define a specific ndelay implementation.)

> So stop this idiocy already. About half of what I've seen in this
> thread has been pure garbage.

OK, I'll just send my patch, and then crawl back under my rock.

Regards.