Hi Arend, Thanks for your suggestion, but it looks the ath9k_htc didn't get probe request frame. I added debug message in "ath9k_rx_tasklet()"(htc_drv_txrx.c), but no probe request frame arrived there. After more test, it looks this issue occurred after "compat-wireless-3.6.2-1", and this issue couldn't be observed in "compat-wireless-3.5.4-1". Do you have other suggestion for this issue? Thanks Flavian Deng -----Original Message----- From: Arend van Spriel [mailto:arend@xxxxxxxxxxxx] Sent: 2013年5月23日 16:13 To: Zhou, Robie Cc: linux-wireless@xxxxxxxxxxxxxxx; Deng, Flavian Subject: Re: ath9k_htc p2p finding issue On 05/23/2013 07:58 AM, Zhou, Robie wrote: > Hi all, > > We meet some problem when try P2P by using ath9k_htc compat-wireless-v3.6.8-1. > > When using the compat-wireless-v3.6.8-1, AR9271 could find other P2P devices, but other P2P devices couldn't find AR9271 (runs ath9k_htc) because driver couldn't report Probe Request to wpa_supplicant. > > And this issue could not be observed when using compat-wireless-v3.0.9-1, by compare with compat-wireless-v3.0.9-1, ATH9K RX filter is set correctly after got "Report Probe Request" command from wpa_supplicant, but ATH9K driver couldn't get "Probe Request" frame(I add debug msg in ieee80211_rx_handlers() function). > > <7>[ 205.784118] [DF_DBG] [ieee80211_mgmt_frame_register] : local->probe_req_reg = 1; > <7>[ 205.830232] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x17; -> ath9k_hw_setrxfilter > <7>[ 205.832858] [DF_DBG] [ieee80211_configure_filter] : new_flags |= FIF_PROBE_REQ; > <7>[ 205.833101] [DF_DBG] [ath9k_htc_configure_filter] rfilt = 0x287; -> ath9k_hw_setrxfilter > <7>[ 205.878230] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x287; -> ath9k_hw_setrxfilter > <7>[ 205.882556] [DF_DBG] [ieee80211_configure_filter] : new_flags |= FIF_PROBE_REQ; > <7>[ 205.883606] [DF_DBG] [ath9k_htc_configure_filter] rfilt = 0x287; -> ath9k_hw_setrxfilter > > <7>[ 206.010886] [DF_DBG] [ieee80211_mgmt_frame_register] : local->probe_req_reg = 0; > <7>[ 206.036237] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x287; -> ath9k_hw_setrxfilter > <7>[ 206.043357] [DF_DBG] [ath9k_htc_configure_filter] rfilt = 0x17; -> ath9k_hw_setrxfilter > <7>[ 206.044107] [DF_DBG] [ath9k_htc_configure_filter] rfilt = 0x17; -> ath9k_hw_setrxfilter > <7>[ 206.090106] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x17; -> ath9k_hw_setrxfilter > <7>[ 206.160734] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x17; -> ath9k_hw_setrxfilter > <7>[ 206.230740] [DF_DBG] [ath9k_htc_opmode_init] rfilt = 0x17; -> ath9k_hw_setrxfilter > > > Did anyone ever meet this issue? Any suggestion for further debugging this issue? Did you confirm ath9k_htc is handing over probe requests to mac80211? Might be useful to add debugging in prepare_for_handlers() instead. Regards, Arend ?韬{.n?????%??檩??w?{.n???{??W????塄}?财??j:+v?? ? ???2??璀??摺?囤??z夸z罐?+?????w棹f
