Re: [rfc] Implement ath9k mesh point operation

[Pat Erley <pat-lkml@erley.org>]

Andrey Yurovsky wrote:
> Hi Pat,
> 
> On Mon, Mar 16, 2009 at 3:36 PM, Pat Erley <pat-lkml@erley.org> wrote:
>> Currently, you need to trigger a scan on the mesh point interface to
>> enable beaconing.  This is the same behavior the ath5k, rt2500usb and
>> rt61pci drivers exhibit.  Is this a mac80211 design flaw, or something
>> in all of these drivers?
> 
> Are you really seeing this behavior on ath5k?  I have never seen that
> happen, if it's new then it's a regression.  rt2x00 does have this
> problem at this time.
> 
>   -Andrey

Yep, here's my lspci output and relevant dmesg messages when I load ath5k:

Atheros Communications Inc. AR5212/AR5213 Multiprotocol MAC/baseband processor (rev 01)

ath5k 0000:02:0a.0: PCI INT A -> Link[APC3] -> GSI 18 (level, low) -> IRQ 18
ath5k 0000:02:0a.0: registered as 'phy5'
phy5: Selected rate control algorithm 'minstrel'
ath5k phy5: Atheros AR5213A chip found (MAC: 0x59, PHY: 0x43)
ath5k phy5: RF5112B multiband radio found (0x36)
udev: renamed network interface wlan1 to wlan3

(yes, phy5, I have a ton of wireless stuff plugged into this dev system)

and here's the procedure I use:

iw dev wlan3 del
iw phy phy5 interface add mp0 type mp mesh_id FooBarMesh
ifconfig mp0 10.0.0.5 netmask 255.255.255.0 up
(tshark on rtl8187 shows no beaconing)
iw dev mp0 scan trigger
(tshark on rtl8187 now shows beaconing)

Out of all the devices I have, only b43 seems to beacon before a scan.  Is
there a chance I need to wait like 45 seconds for beaconing to start on ath5k
like how ibss mode tries to find a peer already broadcasting before it enables
beacons?

Pat Erley