From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Return-Path: Date: Wed, 4 Jan 2017 14:29:39 +0200 From: Mika Westerberg To: Kilian Singer Cc: Lukas Wunner , Bjorn Helgaas , "Rafael J. Wysocki" , Peter Wu , linux-pci , Alex Deucher , Dave Airlie Subject: Re: PCI: Revert "PCI: Add runtime PM support for PCIe ports" Message-ID: <20170104122939.GQ3353@lahna.fi.intel.com> References: <20161228161816.GA19653@bhelgaas-glaptop.roam.corp.google.com> <3913750.OA6IItgEjC@aspire.rjw.lan> <20170103222509.GA4499@bhelgaas-glaptop.roam.corp.google.com> <16155135.jUckSz7QPL@aspire.rjw.lan> <20170104000557.GA13950@bhelgaas-glaptop.roam.corp.google.com> <20170104081639.GA21076@wunner.de> <83901bcb-1315-fbca-e1e3-960bfefbd9c0@quantumtechnology.info> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: <83901bcb-1315-fbca-e1e3-960bfefbd9c0@quantumtechnology.info> List-ID: On Wed, Jan 04, 2017 at 11:33:16AM +0100, Kilian Singer wrote: > Dear all, > > the weird thing is also that when calling: > > echo 0 > /sys/bus/pci/devices/0000:01:00.0/d3cold_allowed > or > echo on > /sys/bus/pci/devices/0000:01:00.0/power/control > > on a command line then the command line crashes. The system stays responsive but still > becomes unresponsive when I lock the screen. Most probably because the device has already been runtime suspended. Setting them from command line may be too late. > Maybe a similar thing are happening in the kernel. Maybe one could use the above fact > and a timeout to test for the problem and then take measures to deactivate the pm. > > I hope this gives you some more clues. You could try to run 'lspci -vv' once the machine is unresponsive (if you can do anything anymore). That should show whether the device and the root port are still in D3.