[PATCH] ssb: cc: prepare clockmode support for cores rev 10+

Michael Büsch m at bues.ch
Wed Apr 27 17:37:49 EDT 2011

On Wed, 2011-04-27 at 22:52 +0200, Rafał Miłecki wrote: 
> 2011/4/27 Michael Büsch <m at bues.ch>:
> >> +            /* udelay(150); TODO: not available in early init */
> >
> > udelay() should be available at this point. Are you confusing it with
> > msleep(), which is not available that early?
> I had no idea about this difference, thanks.
> However I still think about initializing ChipCommon later, to let it
> *sleep* and no to delay boot process. Should this be fine? Or are
> there some important reasons to init ChipCommon early?

I don't think it matters it all, if the boot process takes
150 microseconds longer or not. So I'd rather not change it for
no good reason.
This new 0.00015 seconds (150 us) delay in chipcommon seems perfectly

Even a few MMIO bus accesses will take longer than this tiny delay.
And there are thousands of MMIO accesses on bootup.
So it will hardly be measureable.

Greetings Michael.

More information about the b43-dev mailing list