From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Mackerras Subject: Re: [PATCH v4] pm_ops: add system quiesce/activate hooks Date: Sun, 15 Apr 2007 02:55:13 +1000 Message-ID: <17953.1905.510261.169936@cargo.ozlabs.ibm.com> References: <1175810054.3489.34.camel@johannes.berg> <1176499086.7052.124.camel@johannes.berg> <20070413213349.GL28264@elf.ucw.cz> <200704140009.33298.rjw@sisk.pl> <1176502406.7052.134.camel@johannes.berg> <20070413221652.GR28264@elf.ucw.cz> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <20070413221652.GR28264@elf.ucw.cz> 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: Pavel Machek Cc: Johannes Berg , linux-pm@lists.linux-foundation.org List-Id: linux-pm@vger.kernel.org Pavel Machek writes: > I see ppc as being really special here. It is, I guess. The salient point is that once the decrementer has an interrupt pending, the only way to clear it is to take the interrupt. There is no explicit access to the interrupt pending bit and no separate enable bit for its interrupt, only the global interrupt enable for the whole CPU. We have to turn on the global interrupt enable when finally putting the system into the suspend state (for suspend to RAM), otherwise on some machines we'll never wake up again. Effectively, when dealing with the decrementer during preparation for suspend-to-RAM, we may need to have interrupts enabled briefly in order to clear a pending decrementer interrupt. Paul.