On Tue, Apr 04, 2023 at 01:09:19PM +0200, Oliver Neukum wrote: > On 04.04.23 13:05, Johan Hovold wrote: > > > This is backwards. If you need the device to be active to access these > > registers then you should resume it unconditionally instead of failing. > usually you would be right. But this is debugfs. It is intended to observe > the system in the state it is actually in. If by the act of observation you > wake up the device, you change the experiment. I admittedly didn't look to closely at what this particular debugfs interface is used for, but I generally do not agree with that reasoning. The device is being used; by the driver and ultimately by a user telling the driver to do something on their behalf. The fact that the user is initiating an action through an interface which intended for debugging should not matter (and the user always has the option to check the runtime pm state before initiating the action if that matters at all). Johan
