From mboxrd@z Thu Jan 1 00:00:00 1970 From: ethan zhao Subject: Re: [PATCH] cpufreq: Set cpufreq_cpu_data to NULL before putting kobject Date: Mon, 02 Feb 2015 09:54:23 +0800 Message-ID: <54CED8CF.9060407@oracle.com> References: <54CADEAE.2090305@oracle.com> <54CAE80C.4060406@oracle.com> <54CAEABB.4060508@oracle.com> <54CAFE88.2080508@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from aserp1040.oracle.com ([141.146.126.69]:24889 "EHLO aserp1040.oracle.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754601AbbBBByk (ORCPT ); Sun, 1 Feb 2015 20:54:40 -0500 In-Reply-To: Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: Viresh Kumar Cc: Rafael Wysocki , santosh shilimkar , Linaro Kernel Mailman List , "linux-pm@vger.kernel.org" , "# 3.13.x" Viresh, On 2015/1/30 12:14, Viresh Kumar wrote: > On 30 January 2015 at 09:16, ethan zhao wrote: >> You mean the policy is shared by all CPUs, so PPC notification about one > A policy may or maynot be shared by multiple CPUs. It all depends on your > systems configuration. All CPUs which share clock line, share a policy. > > You can verify that from /sys/devices/system/cpu/cpu*/related_cpus. This > gives the CPUs that share policy. > >> CPU should update all CPU's policy, right ? even the requested CPU is >> shutting down. > CPUs sharing policy have a single policy structure. And so policy would > be updated for all CPUs that share the poilcy. Even if some cpu goes down, > the policy might still be up and running. Let' me check ACPI spec and BIOS to match your implementation. About that patch you suggested, there is another question left pending: Policy will be freed only when that's last CPU shutting down, how does it happen when system booting ? The description of the patch would be wrong (the Xen_bus call stack) -- Did the xen_bus shut down every CPU till the last during booting ? Thanks, Ethan > > -- > viresh