Return-path: Received: from mx1.redhat.com ([209.132.183.28]:34258 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754435AbcKVPBr (ORCPT ); Tue, 22 Nov 2016 10:01:47 -0500 From: Jes Sorensen To: Barry Day Cc: Kalle Valo , linux-wireless@vger.kernel.org Subject: Re: [PATCH] rtl8xxxu: Fix failure to reconnect to AP References: <20161113111705.GA2393@box64.home.org> <20161122125300.GB23278@testbox> Date: Tue, 22 Nov 2016 10:01:45 -0500 In-Reply-To: <20161122125300.GB23278@testbox> (Barry Day's message of "Tue, 22 Nov 2016 22:53:01 +1000") Message-ID: (sfid-20161122_160150_571444_E33F29B1) MIME-Version: 1.0 Content-Type: text/plain Sender: linux-wireless-owner@vger.kernel.org List-ID: Barry Day writes: > On Mon, Nov 21, 2016 at 11:57:22AM -0500, Jes Sorensen wrote: >> Jes Sorensen writes: >> >> void rtl8xxxu_gen1_init_aggregation(struct rtl8xxxu_priv *priv) >> >> @@ -4515,6 +4523,8 @@ rtl8xxxu_bss_info_changed(struct ieee80211_hw *hw, struct ieee80211_vif *vif, >> >> sgi = 1; >> >> rcu_read_unlock(); >> >> >> >> + rtl8xxxu_write16(priv, REG_RXFLTMAP2, 0xffff); >> >> + >> >> priv->fops->update_rate_mask(priv, ramask, sgi); >> >> >> >> rtl8xxxu_write8(priv, REG_BCN_MAX_ERR, 0xff); >> >> I believe this change only matters because you disable RXFLTMAP2 >> above. If we really have to write to RXFLTMAP2 to make this work, I >> suspect we need to keep some sort of state information. >> >> I would also be curious if RXFLTMAP2 gets reset somehow by the firmware, >> and we do not account for that. >> >> Cheers, >> Jes > > I'll redo the patch without touching REG_RXFLTMAP2. I don't think it's needed > to fix the fail to reconnect issue. > > I haven't had a proper look at the 8723 chips yet but the vendor drivers for > the others don't do a h2c cmd for disconnect but I'll test leaving it in to see > if it makes any difference. A 8723bu arrived in the mail today so now I can > test it too and I discovered yesterday I have a 8723au but it's in a cheap > Android tablet. Let me know what you find out - if the h2c command causes the failure that would be very bizarre but certainly interesting to learn. Cheers, Jes