Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932266AbaFSOeF (ORCPT ); Thu, 19 Jun 2014 10:34:05 -0400 Received: from iolanthe.rowland.org ([192.131.102.54]:59860 "HELO iolanthe.rowland.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1757420AbaFSOeD (ORCPT ); Thu, 19 Jun 2014 10:34:03 -0400 Date: Thu, 19 Jun 2014 10:34:01 -0400 (EDT) From: Alan Stern X-X-Sender: stern@iolanthe.rowland.org To: "Rafael J. Wysocki" cc: Allen Yu , Pavel Machek , Len Brown , Greg Kroah-Hartman , , Subject: Re: [PATCH 1/1] PM / Runtime: let rpm_resume fail if rpm disabled and device suspended. In-Reply-To: <3528855.Z370J7O75B@vostro.rjw.lan> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 19 Jun 2014, Rafael J. Wysocki wrote: > Well, we used to have the notion that runtime_status is not meaningful for > devices with dev->power.disable_depth greater than 0 (except for the special > case in the suspend code path where we know why it is greater than 0). I think > it was useful. :-) Did we really have that notion? My memory is a little cloudy, but I thought we decided that runtime_status would not be meaningful when dev->power.runtime_error was set -- not when dev->power.disable_depth was greater than 0. Am I mixed up? In any case, I think it is reasonable to regard runtime_status as meaningful when disable_depth > 0. The PM core isn't allowed to invoke the runtime callbacks at such times, that's all. This makes perfect sense for a device that doesn't support power management and hence must always be at full power. Or when a driver knows that runtime_status is out of agreement with the device's actual power state and wants to update runtime_status directly. > > So pm_runtime_resume() and pm_request_resume() would still fail, but > > pm_runtime_get() and pm_runtime_get_sync() would work? I'm not sure > > about the reason for this distinction. > > The meaning of pm_runtime_get()/pm_runtime_get_sync() is "prevent the > device from being suspended from now on and resume it if necessary" while > "runtime PM disabled and runtime_status == RPM_ACTIVE" may be interpreted > as "not necessary to resume", so it is reasonable to special case this > particular situation for these particular routines IMHO. By the same reasoning, the meaning of pm_runtime_resume() is "resume the device now if necsesary". Since "runtime PM disabled and runtime_status == RPM_ACTIVE" means "not necessary to resume", isn't it logical for pm_runtime_resume() also to succeed under that condition? Alan Stern -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/