Re: [PATCH] __div64_32: implement division by multiplication for 32-bit arches

From: Alexey Brodkin
Date: Fri Oct 30 2015 - 11:54:17 EST

Hi Nicolas,

On Fri, 2015-10-30 at 11:17 -0400, Nicolas Pitre wrote:
> On Fri, 30 Oct 2015, MÃns RullgÃrd wrote:
> > Nicolas Pitre <nicolas.pitre@xxxxxxxxxx> writes:
> >
> > > OK... I was intrigued, so I adapted my ARM code to the generic case,
> > > including the overflow avoidance optimizations. Please have look and
> > > tell me how this works for you.
> > >
> > > If this patch is accepted upstream, then it could be possible to
> > > abstract only the actual multiplication part with some architecture
> > > specific assembly.
> >
> > Good idea.
> Could you please provide a reviewed-by or acked-by tag?


Acked-by: Alexey Brodkin <abrodkin@xxxxxxxxxxxx>

BTW I thought about that optimization a bit more and now I think
we may even skip addition of arch-specific assembly insertions.

That's because that kind of division as discussed many times
should be used as limited as possible, in other words there should be
just a very few usages of it especially in very frequently used code paths.
And in that case there might be not much of benefit having do_div()
even faster and smaller than the one we're about to get with your change.