Return-path: Received: from mail-yw0-f195.google.com ([209.85.161.195]:36348 "EHLO mail-yw0-f195.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751510AbcGTKwv (ORCPT ); Wed, 20 Jul 2016 06:52:51 -0400 Received: by mail-yw0-f195.google.com with SMTP id u134so3004891ywg.3 for ; Wed, 20 Jul 2016 03:52:50 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1467972560885.99812@qti.qualcomm.com> <1467982766213.21223@qti.qualcomm.com> From: Krishna Chaitanya Date: Wed, 20 Jul 2016 16:22:30 +0530 Message-ID: (sfid-20160720_125254_166293_14DE49FA) Subject: Re: ath10k + iw set bitrates is causing FW crash To: "Manoharan, Rajkumar" Cc: linux-wireless , ath10k Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Fri, Jul 8, 2016 at 9:07 PM, Krishna Chaitanya wrote: > > On Fri, Jul 8, 2016 at 8:57 PM, Krishna Chaitanya > wrote: > > On Fri, Jul 8, 2016 at 6:29 PM, Manoharan, Rajkumar > > wrote: > >> I think security failures are due to peer unmap & map upon reassoc. Can you please try below change? > >> > >> diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c > >> index 5e1cc8f4c43c..f7f04bb46fc8 100644 > >> --- a/drivers/net/wireless/ath/ath10k/mac.c > >> +++ b/drivers/net/wireless/ath/ath10k/mac.c > >> @@ -5745,8 +5745,9 @@ static void ath10k_sta_rc_update_wk(struct work_struct *wk) > >> sta->addr, smps, err); > >> } > >> > >> - if (changed & IEEE80211_RC_SUPP_RATES_CHANGED || > >> - changed & IEEE80211_RC_NSS_CHANGED) { > >> + if ((changed & IEEE80211_RC_SUPP_RATES_CHANGED || > >> + changed & IEEE80211_RC_NSS_CHANGED) && > >> + (arvif->vif->type == NL80211_IFTYPE_ADHOC)) { > >> ath10k_dbg(ar, ATH10K_DBG_MAC, "mac update sta %pM supp rates/nss\n", > >> sta->addr); > > > > Bottom posting please. I have tried your change i still see the same issue. > > Right after connecting i see that packets are encrypted but once it > > unset/set "bitrates" plain packets are seen in sniffer. > > > > BTW, i am not rebooting for each driver update, just re-inserting the module, > > do i need to do a reboot for each driver change? > > After reboot i dont see the "unencrypted" packet issue, i will do > some more testing on limiting the rates, currently in my setup its > not reaching MCS9 so cannot verify now. > > Thanks for the fixes and prompt reply. > > ________________________________________ > >> From: Krishna Chaitanya > >> Sent: Friday, July 8, 2016 5:26 PM > >> To: Manoharan, Rajkumar > >> Cc: linux-wireless; ath10k > >> Subject: Re: ath10k + iw set bitrates is causing FW crash > >> > >> On Fri, Jul 8, 2016 at 3:39 PM, Manoharan, Rajkumar > >> wrote: > >>>>> I am using ath10k driver with qca988x hw2.0 and trying to limit it to use > >>>>> VHT MCS0-7 (iw set bitrates vht-mcs-5 2:0-7). > >>>>> > >>>>> But the command it causing a FW crash, if it disable HW_HAS_RATE_CONTROL > >>>>> no crash is observed but it still uses MCS9. > >>>>> > >>>>> tree: wireless-drivers-next: commit#535633a5ba4ea2504fa6c33176633becf0e59339 > >>>>> > >>>>> 1) If i disable HW_RATE_CONTROL, will ath10k honor > >>>>> the mac80211 rates? > >>>>> > >>>> > >>> Thanks for reporting the issue. Could you please try with below change? > >>> > >>> diff --git a/drivers/net/wireless/ath/ath10k/mac.c b/drivers/net/wireless/ath/ath10k/mac.c > >>> index 5e1cc8f4c43c..cfa7e01a6103 100644 > >>> --- a/drivers/net/wireless/ath/ath10k/mac.c > >>> +++ b/drivers/net/wireless/ath/ath10k/mac.c > >>> @@ -2812,6 +2812,9 @@ static int ath10k_station_assoc(struct ath10k *ar, > >>> return ret; > >>> } > >>> > >>> + if (vif->type != NL80211_IFTYPE_ADHOC) > >>> + peer_arg.peer_reassoc = reassoc; > >>> + > >>> ret = ath10k_wmi_peer_assoc(ar, &peer_arg); > >>> if (ret) { > >>> ath10k_warn(ar, "failed to run peer assoc for STA %pM vdev %i: %d\n", > >> > >> Thanks Raj, with this fix the rates are 0-7, if i disable then i am > >> seeing 0-9, so its > >> working. > >> > >> But i am seeing a weird issues, the moment i give bitrates command, > >> ath10k no longer does encryption, link is a WPA2-PSK: AES. Even after > >> interface up/down > >> it doesn't work. Rajkumar, I am still not able to make mcs9 work, so i have tried to change the Nss to 1 using iw, but that is not taking affect. This command is not taking effect.