From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757578AbdKOKxR convert rfc822-to-8bit (ORCPT ); Wed, 15 Nov 2017 05:53:17 -0500 Received: from shards.monkeyblade.net ([184.105.139.130]:55976 "EHLO shards.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752289AbdKOKxL (ORCPT ); Wed, 15 Nov 2017 05:53:11 -0500 Date: Wed, 15 Nov 2017 19:53:06 +0900 (KST) Message-Id: <20171115.195306.1005850318419905294.davem@davemloft.net> To: kai.heng.feng@canonical.com Cc: nic_swsd@realtek.com, netdev@vger.kernel.org, linux-kernel@vger.kernel.org, romieu@fr.zoreil.com, hayeswang@realtek.com, jrg.otte@gmail.com Subject: Re: [PATCH 1/2] r8169: reinstate ALDPS for power saving From: David Miller In-Reply-To: <20171115090019.23972-1-kai.heng.feng@canonical.com> References: <20171115090019.23972-1-kai.heng.feng@canonical.com> X-Mailer: Mew version 6.7 on Emacs 25.3 / Mule 6.0 (HANACHIRUSATO) Mime-Version: 1.0 Content-Type: Text/Plain; charset=iso-8859-1 Content-Transfer-Encoding: 8BIT X-Greylist: Sender succeeded SMTP AUTH, not delayed by milter-greylist-4.5.12 (shards.monkeyblade.net [149.20.54.216]); Wed, 15 Nov 2017 02:53:10 -0800 (PST) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org From: Kai-Heng Feng Date: Wed, 15 Nov 2017 04:00:18 -0500 > Commit ("r8169: enable ALDPS for power saving") caused a regression on > RTL8168evl/8111evl [1], so it got reverted. > > Instead of reverting the whole commit, let's reinstate ALDPS for all > models other than RTL8168evl/8111evl. > > [1] https://lkml.org/lkml/2013/1/5/36 > > Cc: Francois Romieu > Cc: Hayes Wang > Cc: Jörg Otte > Signed-off-by: Kai-Heng Feng Sorry, this isn't going to work. The problem is that we have no idea what chips this really works for. All we know is that it definitely does not work for one particular chip variant. The amount of coverage this change is going to get is very small as well, meaning an even greater change of regressions. Therefore the only acceptable way to handle this is to have a white-list, specific chips that have been explicitly tested and are known to work with this feature, rather than the other way around. Furthermore, you're not even checking the chip version, you're checking instead whether the firmware is loaded or not. That doesn't seem like a safe way to guard this at all.