On Fri, 2005-12-16 at 16:30 +0100, francesco rana wrote: > --- ieee80211softmac_wx.c 2005-12-11 18:01:54.000000000 +0100 > +++ ieee80211softmac_wx.c_orig 2005-12-14 17: 59:48.000000000 +0100
Agh. And I was wondering what was wrong with the patch. Hint: use diff
-u orig new and not the other way around :)
> @@ -79,22 +79,19 @@:
> /* If all fails, return ANY (empty) */
> data->essid.length = 0;
> data->essid.flags = 0; /* active */
> -:
> - if(extra) {:
IMHO that ought to be
| if (!extra) {
| // unlock if necessary, not sure right now.
| return -EINVAL;
| }
instead.
> - /* If we have a statically configured ESSID then return :
> it */:
> - if (sm->associnfo.static_essid) {:
> - data->essid.length = sm->associnfo.req_essid_len;:
> - data->essid.flags = 1; /* active */:
> - memcpy(extra, sm->associnfo.req_essid, :
> sm->associnfo.req_essid_len);:
> - }:
> -:
> - /* If we're associating/associated, return that */:
> - if (sm->associated || sm->associnfo.associating) {:
> - :
> sm->associnfo.associate_essid_len=strlen(sm->associnfo.associate_essid);:
Your patch is word-wrapped here, please make sure that doesn't happen.
> - data->essid.length = :
> sm->associnfo.associate_essid_len;:
> - data->essid.flags = 1; /* active */:
> - memcpy(extra, sm->associnfo.associate_essid, :
> sm->associnfo.associate_essid_len);:
> - }:
> +:
> + /* If we have a statically configured ESSID then return it */:
> + if (sm->associnfo.static_essid) {:
> + data->essid.length = sm->associnfo.req_essid_len;:
> + data->essid.flags = 1; /* active */:
> + memcpy(extra, sm->associnfo.req_essid, :
> sm->associnfo.req_essid_len);:
> + }:
> +:
> + /* If we're associating/associated, return that */:
> + if (sm->associated || sm->associnfo.associating) {:
> + data->essid.length = sm->associnfo.associate_essid_len;:
> + data->essid.flags = 1; /* active */:
> + memcpy(extra, sm->associnfo.associate_essid, :
> sm->associnfo.associate_essid_len);:
> }
> spin_unlock_irqrestore(&sm->lock, flags);
> return 0;
Hmm, ok, I don't think I can make out what you actually changed. Did you
only add the extra check?
> I hope this would helps:
Ah.
> the main problem is that if I set an "any" ESSID the req_essid_len and
> the associate_essid_len are set both to zero. So we copy nothing.
Hm? I really don't see where the bug was. I don't even see an actual
code change. Maybe it helps if you can post a new patch that doesn't
change the indentation and thus only shows the few things you changed as
change.
> The
> control on "extra" value is necessary because it happens to be <NULL> or
> 0x000000 value. So it would be a non-sense to operate in such case.
Yeah, ok, but please do as I wrote above, it makes the indentation
smaller ;)
johannes
signature.asc
Description: This is a digitally signed message part
