Hi!

> As can be seen this divides llu by llu in few warnings and error.
> At the time of sending i didn't realize it but this fails on 32 bit
> architectures.
> 
> So i would like to ask how would you like this fixed ?
> Using macro or some other way ?

+#include <linux/math64.h>
-               gain_q23 = (gain_q23 * dmic->boost_gain) / 100;
+               gain_q23 = div_u64(gain_q23 * dmic->boost_gain, 100);

Best regards,
									Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html