From mboxrd@z Thu Jan 1 00:00:00 1970 Return-path: Received: from wolverine01.qualcomm.com ([199.106.114.254]:55721 "EHLO wolverine01.qualcomm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751268AbaILGH7 (ORCPT ); Fri, 12 Sep 2014 02:07:59 -0400 Date: Fri, 12 Sep 2014 11:39:38 +0530 From: Rajkumar Manoharan To: Sujith Manoharan CC: John Linville , , , Subject: Re: [PATCH] Revert "ath9k: Calculate sleep duration" Message-ID: <20140912060936.GA20575@qca.qualcomm.com> (sfid-20140912_080803_728370_BC9C72C6) References: <1410442548-15782-1-git-send-email-sujith@msujith.org> <20140911141338.GA21543@qca.qualcomm.com> <21521.45335.848659.125183@gargle.gargle.HOWL> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" In-Reply-To: <21521.45335.848659.125183@gargle.gargle.HOWL> Sender: linux-wireless-owner@vger.kernel.org List-ID: On Thu, Sep 11, 2014 at 07:56:31PM +0530, Sujith Manoharan wrote: > Rajkumar Manoharan wrote: > > Some of the IOE customer using ath9k with power save enabled. IIRC the > > expectation is that when the station is idle, it should wakeup on DTIM > > beacon not for every beacons. Are you seeing frequest bmiss events on > > higher beacon interval? > > I don't think ath9k has ever been tested with longer sleep durations > based on dtim period. We have always programmed the hardware to wake up > for every beacon interval, using the TIM_TIMER interrupt. And since the > TIM bit needs to be checked for every beacon, the sleep duration has > to be the beacon interval. > > Also, the bmiss threshold needs to be adjusted based on the sleep > duration. We are hardcoding it currently in the driver using > ATH_DEFAULT_BMISS_LIMIT. This will break, too. > > Using dtim for PS needs more testing before it can be used, > so we can have the original behavior for now. > Sounds clear. Acked-by: Rajkumar Manoharan