From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Rosin Subject: Re: [PATCH v1 1/2] Revert "i2c: mux: pca954x: Add ACPI support for pca954x" Date: Thu, 23 Mar 2017 12:21:59 +0100 Message-ID: References: <20170321191310.32957-1-andriy.shevchenko@linux.intel.com> <01b7c79e-c52f-8e87-59a8-2eb17a72d733@axentia.se> <1490187942.19767.161.camel@linux.intel.com> <7d5fba14-d3f1-dfaa-ac7d-5001ccf370a1@axentia.se> <20170323100403.GA9307@amd> Mime-Version: 1.0 Content-Type: text/plain; charset="windows-1252" Content-Transfer-Encoding: 7bit Return-path: Received: from mail-he1eur01on0105.outbound.protection.outlook.com ([104.47.0.105]:11124 "EHLO EUR01-HE1-obe.outbound.protection.outlook.com" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S932780AbdCWLWI (ORCPT ); Thu, 23 Mar 2017 07:22:08 -0400 In-Reply-To: <20170323100403.GA9307@amd> Sender: linux-leds-owner@vger.kernel.org List-Id: linux-leds@vger.kernel.org To: Pavel Machek Cc: Andy Shevchenko , linux-i2c@vger.kernel.org, Wolfram Sang , Tin Huynh , Richard Purdie , Jacek Anaszewski , linux-leds@vger.kernel.org On 2017-03-23 11:04, Pavel Machek wrote: > On Thu 2017-03-23 08:45:58, Peter Rosin wrote: >> On 2017-03-22 14:05, Andy Shevchenko wrote: >>> On Wed, 2017-03-22 at 11:23 +0100, Peter Rosin wrote: >>>> On 2017-03-21 20:13, Andy Shevchenko wrote: >>>>> In ACPI world any ID should be carefully chosen and registered >>>>> officially. The commit bbf9d262a147 seems did a wrong assumption >>>>> because >>>>> PCA is the registered PNP ID for "PHILIPS BU ADD ON CARD". I'm >>>>> pretty >>>>> sure this prefix has nothing to do with the driver in question. >>>> >>>> [Cc: leds people, in case they know any details] >>>> >>>> Hmmm, a couple of questions about that "pretty sure"... >>> >>> I didn't neither see the *real* excerpt from DSDT nor hear anything >>> about official IDs from Phillips. >>> >>>> Philips and NXP are pretty much just different faces of the same coin, >>>> IIUC. >>> >>> Good to know. >>> >>> While I might be mistaken, I would like to remove a confusion until we >>> get an official confirmation either in *real* existing product on the >>> market or letter from Phillips representatives (see above). >> >> Right, I don't disagree with the revert at all. The IDs were >> apparently just grabbed and, as you point out, that is not the ACPI >> way. >> >> One more question though, the revert (patch 1/2) should probably be >> queued up for current (4.11) and sent to stable as well (4.10 is the >> only version affected), but what about patch 2/2? Is that 4.12 > > Sent to stable? What serious bug it fixes? Hi Pavel, It obviously fixes the bug that someone might start depending on these ACPI IDs when they shouldn't, which potentially stops that someone from later complaining when it no longer works. I thought that this was probably serious and qualified under "oh, that's not good". Because regressions are nasty. Now that I look closer on the stable rules I suppose it also disqualifies as it is on the theoretical side. Also, the general tone from Andy indicates a certain amount of frustration with the whole issue, which perhaps has no bearing on the seriousness, but what do I know? So, I'm still seeking guidance on how to handle these two patches. Cheers, peda