Messages in this thread | | | Date | Thu, 16 Jan 2014 10:30:37 +0100 | From | Daniel Borkmann <> | Subject | Re: [PATCH net-next 1/2] random32: add prandom_u32_lt_N and convert "misuses" of reciprocal_divide |
| |
On 01/16/2014 04:14 AM, Eric Dumazet wrote: > On Thu, 2014-01-16 at 00:23 +0100, Daniel Borkmann wrote: > >> @@ -1220,7 +1219,7 @@ static unsigned int fanout_demux_hash(struct packet_fanout *f, >> struct sk_buff *skb, >> unsigned int num) >> { >> - return reciprocal_divide(skb->rxhash, num); >> + return (u32)(((u64) skb->rxhash * num) >> 32); >> } >> > > This is unfortunate.
Hm, I agree, it would be better to have a function here. It was mostly for the purpose of the 2nd patch to migrate users since the structure was introduced, and thus they become incompatible.
> (This reverts one of your patch : f55d112e529386 ) > > Please add a helper to explain what's going on here, and on many other > spots we do this computation (as in get_rps_cpu()). > Few people really understand this.
Indeed, that's why we also thought that we should fix some spots in the second patch, also with the thought in mind, when upcoming users have to use a structure to enforce this call combination.
Anyway, we'll wait for some more feedback anyway regarding the 2nd patch and then respin.
> Or keep reciprocal_divide() as is, and introduce a new set of functions > for people really wanting the precise divides.
That might also be a possible option, we could consider. Thanks!
> > >
| |