[PATCH] : Scan in master mode
jt at bougret.hpl.hp.com
Mon Dec 9 13:37:35 EST 2002
On Sat, Dec 07, 2002 at 08:20:12PM -0800, Jouni Malinen wrote:
> On Thu, Dec 05, 2002 at 06:40:56PM -0800, Jean Tourrilhes wrote:
> > As part of WE-16, I'm improving the iwspy support. HostAP is
> > currently problematic as I want to "reclaim" the SIOCGIWSPY call, and
> > HostAP currently uses it.
> Hmm.. If I understood your design correctly, iwspy ioctl's were from
> driver code to main kernel. That is fine for me. It will make the driver
> code a bit more complex with all those conditional #if blocks for
> different wireless extension versions since I'm going to keep supporting
> the older versions, but I think I can manage with it.
The patch listed on my web page does it with minimal
fuss. Actually, I was impressed by how small it was. Notice that for
WE < 16, it uses the old iwspy support.
If you look carefully, I don't really "reclaim" the ioctl, and
it's as usual up to the driver if it wants to use the new scheme or
> > So, I've decided 2 things :
> > 1) I've un-deprecated SIOCGIWAPLIST and broaden its
> > definition. You can now return any list of address you want, not only
> > APs, and will remain supported.
> 'iwlist wlan0 ap' shows "Access Point in range".. Will this also change
> to show something else? And will there be some kind of flag that will
> differentiate APs from associated stations?
New wtools (24-pre4) shows : "Peers/Access Points in range".
There won't be any flags. Well, there can't be
miracles. Anything I would add would be WE-16 and later, so would not
help those older versions of WE, and starting with WE-14 we have
scanning support that has all the flags we need (cf. patch we are
In fact, if you look at the current output of iwspy with
HostAP, you have all stations and APs together without any flag. My
idea was only to move this info from iwspy to aplist.
But as you have seen, I didn't implement it in my HostAP
patch. It's because I don't know which is most important, getting only
APs or getting all stations and APs.
> > 2) I've made the little patch below (for 2002-10-12) and I'm
> > suggesting its inclusion in HostAP. The patch was *very* quickly
> > tested.
> Merged into Host AP CVS with couple of minor changes due to differences
> between 2002-10-12 release and CVS head.
> > Jouni, any comments ?
> Above comments were about this patch and I already replied to your
> private message about WE-16 in a separate email.
I'll update with your comments and re-spin.
> Jouni Malinen PGP id EFC895FA
More information about the HostAP