From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Rafael J. Wysocki" Subject: Re: [RFC][PATCH -mm 3/3] PM: Disable _request_firmware before hibernation/suspend Date: Mon, 28 May 2007 19:43:24 +0200 Message-ID: <200705281943.25129.rjw@sisk.pl> References: Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Content-Disposition: inline List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-pm-bounces@lists.linux-foundation.org Errors-To: linux-pm-bounces@lists.linux-foundation.org To: Alan Stern Cc: Matthew Garrett , LKML , Pavel Machek , pm list List-Id: linux-pm@vger.kernel.org On Monday, 28 May 2007 17:55, Alan Stern wrote: > On Sun, 27 May 2007, Rafael J. Wysocki wrote: > > > From: Rafael J. Wysocki > > > > Use a hibernation and suspend notifier to disable the firmware requesting > > mechanism before a hibernation/suspend and enable it after the operation. > > You're using the PM_PRE_FREEZE and PM_POST_THAW notifiers for both this > and the userspace helper change. Is it your intention that drivers > should continue to request these services but encounter an error if the > request occurs at the wrong time? Or do you expect drivers to use the > notifier chains to know when they shouldn't make any requests? In fact, I'd like drivers to use notifiers to actually load the firmware into memory before hibernation/suspend. Namely, if there's PM_PRE_FREEZE, the driver calls request_firmware() from within the notifier and saves the firmware in memory for future use, if need be. Later, when PM_POST_THAW comes, the memory holding the firmware is released. Unfortunately there are drivers that call request_firmware() directly from .resume() which blocks until timeout expires and fails anyway. I just wanted this to fail immediately, without waiting. > In the second case you may have a problem, because there's no > specification about the order in which the notifiers are sent. The > service may get disabled before the driver learns it isn't available, > or the driver may think the service is once again available before it > gets enabled. Yes. Greetings, Rafael