From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Return-Path: Date: Wed, 1 Jun 2016 13:42:35 +0200 From: Lukas Wunner To: Mika Westerberg Cc: Andreas Noever , "Rafael J. Wysocki" , Bjorn Helgaas , Peter Wu , "linux-pci@vger.kernel.org" , Linux PM list , Valdis Kletnieks , Dave Airlie Subject: Re: [PATCH] PCI: Wait for 50ms after bridge is powered up Message-ID: <20160601114235.GA15191@wunner.de> References: <1576190.gfFb6HhZV6@vostro.rjw.lan> <20160530093326.GH1789@lahna.fi.intel.com> <20160530144445.GC1743@lahna.fi.intel.com> <20160531083349.GG1743@lahna.fi.intel.com> <20160531085805.GI1743@lahna.fi.intel.com> <20160531104051.GA13958@wunner.de> <20160531104714.GL1743@lahna.fi.intel.com> <20160531110757.GB14008@wunner.de> <20160601091145.GP1743@lahna.fi.intel.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <20160601091145.GP1743@lahna.fi.intel.com> List-ID: On Wed, Jun 01, 2016 at 12:11:45PM +0300, Mika Westerberg wrote: > I'm going to change the code to look like following (pm_runtime_mark_last_busy() > gets called before pm_runtime_put_autosuspend() even if not strictly needed): > > pm_runtime_set_autosuspend_delay(&dev->dev, 100); > pm_runtime_use_autosuspend(&dev->dev); > pm_runtime_mark_last_busy(&dev->dev); > pm_runtime_put_autosuspend(&dev->dev); > pm_runtime_allow(&dev->dev); > > Note I'm still increasing default autosuspend delay from 10ms to 100ms. > > Does the above work for you? Yes, tested it and couldn't spot any issues. Thanks, Lukas