From mboxrd@z Thu Jan 1 00:00:00 1970 From: Brian King Subject: Re: [PATCH 1/1] PM: Add arch_suspend_disable_nonboot_cpus Date: Tue, 23 Feb 2010 10:41:33 -0600 Message-ID: <4B84053D.60703@linux.vnet.ibm.com> References: <201002211631.o1LGVsw8022630@d01av02.pok.ibm.com> <201002212337.10462.rjw@sisk.pl> <4B81B7C5.5060301@linux.vnet.ibm.com> <201002222014.07280.rjw@sisk.pl> <4B8313D4.4050802@linux.vnet.ibm.com> <20100223154359.GA6220@elf.ucw.cz> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20100223154359.GA6220@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: linux-pm@lists.linux-foundation.org List-Id: linux-pm@vger.kernel.org On 02/23/2010 09:43 AM, Pavel Machek wrote: > Hi! > >>> OK >>> >>>> Here is an alternative implementation of the patch. My test machine is >>>> currently unavailable, so it is not yet been tested. How does this one look? >>> >>> Well, I'd like to do that cleanly from the start. >>> >>> Now, the problem is that PM_SLEEP_SMP selects HOTPLUG_CPU, because >>> that's necessary for the other architectures to make SMP suspend work, but it's >>> not necessary on your architecture. Moreover, you don't need to compile >>> enable_nonboot_cpus() at all. > >> At least for the architecture I am enabling this support for >> (PPC_PSERIES), upon looking closer, it looks like PM_SLEEP_SMP was >> never defined, so enable_nonboot_cpus and disable_nonboot_cpus were >> always nooped before, which I didn't previously realize. We probably >> want to retain this behavior. > > > (Please wrap at column 80) > > This patch is already way better than the original one, but... Why do > you want enable/disable_nonboot_cpus to be noped out? > Pavel Today for PPC_PSERIES, PM_SLEEP_SMP is never defined, so for all the current code paths that call enable/disable_nonboot_cpus (power off, kexec), these functions are noops. I don't want to change that behavior. I figured I can just use the prepare_late and wake pm functions to do the work I need to do. Let me know if you think this is a reasonable approach and I'll be happy to resend the patch with an appropriate subject line and description. Thanks, Brian -- Brian King Linux on Power Virtualization IBM Linux Technology Center