Re: [PATCH] cfg80211: fix cmp_hidden_bss

[Johannes Berg <johannes@sipsolutions.net>]

On Thu, 2012-11-29 at 02:07 +0200, Eliad Peller wrote:
> On Thu, Nov 29, 2012 at 2:00 AM, Johannes Berg
> <johannes@sipsolutions.net> wrote:
> > On Thu, 2012-11-29 at 01:53 +0200, Eliad Peller wrote:
> >> On Wed, Nov 28, 2012 at 11:44 PM, Johannes Berg
> >> <johannes@sipsolutions.net> wrote:
> >> > From: Johannes Berg <johannes.berg@intel.com>
> >> >
> >> > The cmp_bss() comparator function uses memcmp() to
> >> > compare the SSID. This means that cmp_hidden_bss()
> >> > needs to similarly return a number bigger than zero
> >> > (use 1) instead of -1 when ie1 is bigger than ie2,
> >> > which is the case if an ie2 byte is non-zero.
> >
> >> > +++ b/net/wireless/scan.c
> >> > @@ -472,7 +472,7 @@ static int cmp_hidden_bss(struct cfg80211_bss *a, struct cfg80211_bss *b)
> >> >         /* zeroed SSID ie is another indication of a hidden bss */
> >> >         for (i = 0; i < ie2[1]; i++)
> >> >                 if (ie2[i + 2])
> >> > -                       return -1;
> >> > +                       return 1;
> >> >
> >> >         return 0;
> >> >  }
> >>
> >> maybe i'm misreading it, but there is still no comparison here at all,
> >> so it seems wrong as well :)
> >> (practically, i guess it doesn't really matter, because if there is
> >> some non-null char, this is not a hidden ssid anyway)
> >
> > Hmm yeah you're right, it should be
> >         return ie2[i + 1] - ie1[i + 1];
> >
> > (or the other way around? need to sleep on that first...)
> >
> that's also not enough, because you ignore the previous chars.
> ("A" < "B" vs. "AA" > "\x00B")
> 
> i think returning memcmp() is the easiest solution, but sleeping on
> that is definitely a good idea :)

Hmm, yeah, indeed. If we find a non-zero byte we should return memcmp().
I'll fix it in the morning :)

johannes