> On 2/24/14, 9:46 AM, Matti Laakso wrote: > >>/ On 2/17/14, 09:05 PM, Matti Laakso wrote: > />>/ >/ The latest hostapd and some of the wireless drivers support DFS, which > />>/ />/ is required on the 5 GHz band in many countries. This patch adds the > />>/ />/ correct option to the hostapd configuration if dfs=1 is set (along > with > />>/ />/ the country code) in the wireless configuration. > />>/ />/ > />>/ / > />>/ It is good to see that there are more people working on DFS support. One > />>/ month ago we also added a ticket to solve the same "ieee80211d=1" > />>/ problem (https://dev.openwrt.org/ticket/14867). > />>/ Our patch reuses the "doth" variable which already exists (see below). > />>/ > />>/ Index: network/services/hostapd/files/netifd.sh > />>/ =================================================================== > />>/ --- network/services/hostapd/files/netifd.sh (revision 39386) > />>/ +++ network/services/hostapd/files/netifd.sh (working copy) > />>/ @@ -140,7 +140,7 @@ > />>/ maxassoc max_inactivity disassoc_low_ack isolate > auth_cache \ > />>/ wps_pushbutton wps_label ext_registrar \ > />>/ wps_device_type wps_device_name wps_manufacturer > wps_pin \ > />>/ - macfilter ssid wmm hidden short_preamble > />>/ + macfilter ssid wmm hidden short_preamble doth > />>/ > />>/ set_default isolate 0 > />>/ set_default maxassoc 0 > />>/ @@ -149,6 +149,7 @@ > />>/ set_default disassoc_low_ack 1 > />>/ set_default hidden 0 > />>/ set_default wmm 1 > />>/ + set_default doth 1 > />>/ > />>/ append bss_conf "ctrl_interface=/var/run/hostapd" > />>/ if [ "$isolate" -gt 0 ]; then > />>/ @@ -160,6 +161,9 @@ > />>/ if [ "$max_inactivity" -gt 0 ]; then > />>/ append bss_conf "ap_max_inactivity=$max_inactivity" "$N" > />>/ fi > />>/ + if [ "$doth" -gt 0 ]; then > />>/ + append bss_conf "ieee80211h=1" "$N" > />>/ + fi > />>/ > />>/ append bss_conf "disassoc_low_ack=$disassoc_low_ack" "$N" > />>/ append bss_conf "preamble=$short_preamble" "$N" > />/ > />/ I wasn't actually even aware of doth, but now looking at wiki, it indeed > />/ is there. As an interface level attribute it is a bit strange, though, > />/ as the channel is set at the device level. DFS support depends also on > />/ the country code, which is a device level attribute as well. I also > />/ think that 'dfs' is a more obvious name for DFS support for the casual > />/ user :P > / > I agree that 'doth' is not the best name. > > >/ > />/ From your trac ticket I got the impression that you were able to enable > />/ DFS even without enabling the ATH9K_DFS_CERTIFIED -configuration option. > />/ Is that really the case? Have you witnessed whether it actually changes > />/ channels? > / > Yes it is working. The following worked for me today with a Ubiquiti > Nanostation M5: > * checkout trunk OpenWRT > * apply our patch from ticket (https://dev.openwrt.org/ticket/14867) > * compile and flash > * edit /etc/config/wireless to the following > ---- > config wifi-device radio0 > option type mac80211 > option channel 60 > option hwmode 11na > option country 'DE' > option path 'pci0000:00/0000:00:00.0' > list ht_capab SHORT-GI-40 > list ht_capab TX-STBC > list ht_capab RX-STBC1 > list ht_capab DSSS_CCK-40 > option htmode HT20 > # REMOVE THIS LINE TO ENABLE WIFI: > option disabled 0 > ---- > * wifi up > > And now you should see an OpenWRT network on channel 60.
It seems that ath9k advertises radar_detect_widths in /drivers/net/wireless/ath/ath9k/init.c even when ATH9K_DFS_CERTIFIED is not set. This allows hostapd to use DFS-channels even when the DFS radar detection is not actually working. Looks like a bug. Matti _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel
