Query: can wpa_supplicant set up an ad-hoc network
dcbw at redhat.com
Fri Oct 16 12:08:43 EDT 2009
On Fri, 2009-10-16 at 16:58 +0800, Soh Kam Yung wrote:
> On Fri, Oct 16, 2009 at 2:28 PM, Holger Schurig
> <hs4233 at mail.mn-solutions.de> wrote:
> >> with a Marvell wifi driver.
> > Which one, a Marvell specific one or the one that is included in
> > the kernel?
> It's a Marvell specific one.
> >> Could this be a problem with the kernel, driver
> >> or wpa_supplicant?
> > Yes.
> > (Couldn't resist, as the question here is way too broad)
> Problem resolved on my end: I forgot to disable a background network
> manager related process during the wpa_supplicant test that was
> killing the connection.
> A general question about ad-hoc networks: wpa_supplicant does not
> distinguish between joining an existing ad-hoc network and creating a
> new ad-hoc network. Is this how ad-hoc networks are handled in
Yes. There is no real difference between starting and joining an adhoc
network. Except a scan is done first (usually in the driver) to
determine whether there is an existing IBSS with that SSID and security
settings, and if so, that IBSS's BSSID is used.
Remember, ad-hoc is *ad-hoc*. There's no central controller to say who
can join the network and who cannot. The only access restriction is by
SSID and encryption key.
Is there some specific reason you are thinking you need to distinguish
between the operations?
More information about the HostAP