Re: [B.A.T.M.A.N.] [PATCH v5 3/4] batman-adv: B.A.T.M.A.N. V - implement GW selection logic

[Antonio Quartulli <a@unstable.cc>]

[-- Attachment #1: Type: text/plain, Size: 1440 bytes --]

On Mon, Jun 27, 2016 at 08:19:22AM +0200, Sven Eckelmann wrote:
> On Monday 27 June 2016 10:36:13 Antonio Quartulli wrote:
> > On Mon, Jun 13, 2016 at 01:26:31PM +0200, Sven Eckelmann wrote:
> > > On Sunday 12 June 2016 12:14:25 Antonio Quartulli wrote:
> > > > +       if (orig_throughput < (gw_throughput + threshold))
> > > > +               goto out;
> > > 
> > > Possible overflow problem in batadv_v_gw_is_eligible. We don't
> > > know what the user will add here and what the gw_throughput is.
> > > 
> > > We already had a similar problem in the BATMAN_IV gw selection
> > > code which resulted in weird gw selections. See f63c54bba31d
> > > ("batman-adv: Avoid u32 overflow during gateway select").
> > 
> > Sven, correct me if I am wrong, but in f63c54bba31d we are changing the type of
> > the variables aimed to contain the value computed by the "overflowing"
> > operation, while in this case we have everything in a if condition: should the
> > temporary result be implicitly casted to a larger variable as required ?
> 
> No, it isn't casted to a larger type. Yes, C has some wild casting rules but I
> would not be aware of an "help the programmer to avoid overflows" casting
> rule.
> 
> We had no other idea how to fix it in the mentioned commit. But here we can do
> it without casting to u64.

ok, thanks !

I will modify the patch in the next series.

Cheers,


-- 
Antonio Quartulli

[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 819 bytes --]