Return-path: Received: from mail-pz0-f200.google.com ([209.85.222.200]:58980 "EHLO mail-pz0-f200.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S965811Ab0CPAJZ convert rfc822-to-8bit (ORCPT ); Mon, 15 Mar 2010 20:09:25 -0400 Received: by pzk38 with SMTP id 38so2601420pzk.33 for ; Mon, 15 Mar 2010 17:09:24 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <4B9ECB3E.3010802@openwrt.org> References: <1268459216-22372-1-git-send-email-lrodriguez@atheros.com> <20100315214338.GB4000@tux> <43e72e891003151625t795b8ecdr2ce2e6927e8632a4@mail.gmail.com> <20100315234909.GA2662@tux> <4B9EC85B.8070800@openwrt.org> <43e72e891003151655i7dc051d0lc1da26267e9c706@mail.gmail.com> <4B9ECB3E.3010802@openwrt.org> From: "Luis R. Rodriguez" Date: Mon, 15 Mar 2010 17:09:04 -0700 Message-ID: <43e72e891003151709r42745e27q6e42c131d202a4cb@mail.gmail.com> Subject: Re: [PATCH v3 00/12] ath9k_hw: complete support for AR9271 To: Felix Fietkau Cc: "linville@tuxdriver.com" , Vivek Natarajan , Sujith Manoharan , "linux-wireless@vger.kernel.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-wireless-owner@vger.kernel.org List-ID: On Mon, Mar 15, 2010 at 5:05 PM, Felix Fietkau wrote: > On 2010-03-16 12:55 AM, Luis R. Rodriguez wrote: >> On Mon, Mar 15, 2010 at 4:52 PM, Felix Fietkau wrote: >>> On 2010-03-16 12:49 AM, Luis R. Rodriguez wrote: >>>> diff --git a/drivers/net/wireless/ath/ath9k/hw.c b/drivers/net/wireless/ath/ath9k/hw.c >>>> index 00570fb..867cfaf 100644 >>>> --- a/drivers/net/wireless/ath/ath9k/hw.c >>>> +++ b/drivers/net/wireless/ath/ath9k/hw.c >>>> @@ -807,6 +804,10 @@ void ath9k_hw_init_global_settings(struct ath_hw *ah) >>>>               acktimeout += 64 - sifstime - ah->slottime; >>>> >>>>       ath9k_hw_setslottime(ah, slottime); >>>> + >>>> +     if (AR_SREV_9271(ah)) >>>> +             return; >>>> + >>>>       ath9k_hw_set_ack_timeout(ah, acktimeout); >>>>       ath9k_hw_set_cts_timeout(ah, acktimeout); >>>>       if (ah->globaltxtimeout != (u32) -1) >>> >>> Why leave out the ACK/CTS timeout setting? How does the hw handle ACK >>> and CTS timeouts? >> >> Right, so that's why I originally removed this hunk completely, I was >> under the impression you latest fix would have addressed any issues >> but it does not, we get association time outs. I think we'll have to >> review this further in order to support coverage class. > Are you sure the slot time was set properly by the callback? Well ath9k_htc did not have it implemented, I just added the code for it, you know what.. I did not test modifying the ACK/CTS timeout *and* with the new slot time code I added to ath9k_htc, let me try that. > Can you > print out the values that it's setting? Yeah, 20 upon initial association and then I see another setting for 9. > The initvals for 9271 don't > really look all that different. Indeed, odd... this should work with the fudge factor you added. Luis