[PATCH] hostapd: fixes for support the bcm43xx (and, probably, other dscape drivers) in master mode
jkmaline at cc.hut.fi
Tue May 9 00:01:35 EDT 2006
On Sun, May 07, 2006 at 10:35:06AM +0300, Alexander V. Tsvyashchenko wrote:
> Recently I've spent some time figuring out the way to use bcm43xx in master
> mode (on my router Asus Wl-500g Deluxe). I used bcm43xx-d80211 + hostapd. It
> took much longer than expected, but finally I have working master mode
> (though, there are quite some restrictions).
> 1) Fix for the name changes introduced in the latest version of d80211 stack
> at wireless-dev.git tree
> (i.e., wlan0 -> wmaster0, wlan0ap -> wmaster0ap, wlan0.11 -> wlan0).
> I had to hard-code wmaster0ap name in the code, but it seems to be safe for
> now, as it is curently the only interface
> d80211 creates for that purpose (it seems, though, it can change in the
I had already done this a week or so ago, so I skipped this part.
> 2) Support for disabling probe response sending - bcm43xx driver handles
> probe responses on his own,
> but passes probe response packets to hostapd, so hostapd also sends probe
> replies - this confuses clients.
> I added config parameter "send_probe_response" (1 by default) which allows
> to disable probe responses handing
> by hostapd.
Hmm.. I have not noticed extra probe responses, but then again, I
haven't looked very closely on sniffer traces. Do you know whether the
driver/microcode can be modified not to reply to probes? Anyway, applied
with the if statement moved into beacon.c.
> 3) I moved set_beacon call later in the code, because at the point it was
> called, there's no yet info about supported
> rates - thus, this info is not included in the beacon, which seems to be
> wrong ...
> 4) I added set_beacon_int call, as I did not find it anywhere in the code.
I had missed this when merging in beacon configuration from Devicescape
> 5) Change PRISM2_PARAM_ADM_STATUS to PRISM2_PARAM_RADIO_ENABLED (this
> corresponds to the change in d80211 stack).
I had already done this.
> The patch is against one of the latest snapshots before 0.5.3 release - I
> beleive, it should apply correctly to 0.5.3 also.
Thanks! I applied the changes apart from 1 and 5 which already done.
Jouni Malinen PGP id EFC895FA
More information about the HostAP