Re: [PATCH net-next] net: replace macros net_random and net_srandom with direct calls to prandom

[Joe Perches <joe@perches.com>]

On Sat, 2014-01-11 at 19:52 +0100, Daniel Borkmann wrote:
> On 01/11/2014 07:00 PM, Hannes Frederic Sowa wrote:
> > On Sat, Jan 11, 2014 at 09:52:37AM -0800, Joe Perches wrote:
> >> On Sat, 2014-01-11 at 07:15 -0500, Aruna-Hewapathirane wrote:
> >>> This patch removes the net_random and net_srandom macros and replaces
> >>> them with direct calls to the prandom ones. As new commits only seem to
> >>> use prandom_u32 there is no use to keep them around.
> >>> This change makes it easier to grep for users of prandom_u32.
> >>
> >> Seems sensible.
> >>
> >> Also, there may be some value in a future patch
> >> to use reciprocal_divide in a few places
> >>
> >>> diff --git a/net/802/garp.c b/net/802/garp.c
> >>> @@ -397,7 +397,7 @@ static void garp_join_timer_arm(struct garp_applicant *app)
> >>>   {
> >>>   	unsigned long delay;
> >>>
> >>> -	delay = (u64)msecs_to_jiffies(garp_join_time) * net_random() >> 32;
> >>> +	delay = (u64)msecs_to_jiffies(garp_join_time) * prandom_u32() >> 32;
> >>
> >> reciprocal_divide()
> >
> > Does reciprocal_divide() make sense without reciprocal_value() from a
> > stylish point of view?
> 
> No. ;-)
> 
> There was already some work, but I didn't have time to finish it:
> 
> http://www.spinics.net/lists/netdev/msg249395.html

I had forgotten about this.

> Maybe it's time that we should revisit that.

It'd be nice if you found some time to finish it.

	delay = (u64)msecs_to_jiffies(garp_join_time) * prandom_u32() >> 32;
vs
	delay = prandom_u32_max(msecs_to_jiffies(garp_join_time));
or
	delay = msecs_to_jiffies(prandom_u32_max(garp_join_time));

I think the first is harder to read.