From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jean Delvare Subject: Re: [Bug 203297] Synaptics touchpad TM-3127 functionality broken by PCI runtime power management patch on 4.20.2 Date: Mon, 29 Apr 2019 11:45:38 +0200 Message-ID: <1556531138.5647.1.camel@suse.de> References: <20190422130814.GJ173520@google.com> <3a1139ef-10ed-6923-73c5-30fbf0c065c3@linux.intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Benjamin Tissoires , Jarkko Nikula Cc: Bjorn Helgaas , Keijo Vaara , "Rafael J. Wysocki" , Dmitry Torokhov , linux-pm@vger.kernel.org, linux-pci@vger.kernel.org, lkml , linux-input@vger.kernel.org, Wolfram Sang List-Id: linux-input@vger.kernel.org On Mon, 2019-04-29 at 10:53 +0200, Benjamin Tissoires wrote: > On Mon, Apr 29, 2019 at 10:38 AM Jarkko Nikula > wrote: > > I got another thought about this. I noticed these input drivers need > > SMBus Host Notify, maybe that explain the PM dependency? If that's the > > only dependency then we could prevent the controller suspend if there is > > a client needing host notify mechanism. IMHO that's less hack than the > > patch to rmi_smbus.c. > > So currently, AFAIK, only Synaptics (rmi4) and Elantech are using > SMBus Host Notify. > So this patch would prevent the same bugs for those 2 vendors, which is good. > > It took me some time to understand why this would be less than a hack. > And indeed, given that Host Notify relies on the I2C connection to be > ready for the IRQ, we can not put the controller in suspend like we do > for others where the IRQ controller is still ready. > > So yes, that could work from me. Not sure what Wolfram and Jean would > say though. I would say OK with me, this looks like the cleanest solution to me, so if testing is positive, let's go with it. -- Jean Delvare SUSE L3 Support