From mboxrd@z Thu Jan 1 00:00:00 1970 From: Chen Gang Subject: Re: [PATCH] arch: s390: kernel: reset 'c->hotpluggable' when failure occurs Date: Wed, 26 Jun 2013 09:44:11 +0800 Message-ID: <51CA476B.3080208@asianux.com> References: <51C8F685.6000209@asianux.com> <20130625064832.GA4999@osiris> <51C94599.1020707@asianux.com> <20130625090928.GA8981@osiris> <51C9636F.1050709@asianux.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Return-path: Received: from intranet.asianux.com ([58.214.24.6]:7290 "EHLO intranet.asianux.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750766Ab3FZFQB (ORCPT ); Wed, 26 Jun 2013 01:16:01 -0400 In-Reply-To: <51C9636F.1050709@asianux.com> Sender: linux-arch-owner@vger.kernel.org List-ID: To: Heiko Carstens Cc: Martin Schwidefsky , linux390@de.ibm.com, cornelia.huck@de.ibm.com, mtosatti@redhat.com, Thomas Gleixner , linux-s390@vger.kernel.org, "linux-kernel@vger.kernel.org" , Linux-Arch On 06/25/2013 05:31 PM, Chen Gang wrote: > On 06/25/2013 05:09 PM, Heiko Carstens wrote: >> On Tue, Jun 25, 2013 at 03:24:09PM +0800, Chen Gang wrote: >>> On 06/25/2013 02:48 PM, Heiko Carstens wrote: >>>> On Tue, Jun 25, 2013 at 09:46:45AM +0800, Chen Gang wrote: >>>>>> When smp_add_present_cpu() fails, it has reset all things excluding >>>>>> 'c->hotpluggable', so need reset it as original state completely. >>>>>> >>>>>> + c->hotpluggable = 0; >>>>>> return rc; >>>> No, that doesn't make sense. All cpus on s390 are always hotplugable. >>>> It really doesn't matter if the value of this field is 0 or 1 after >>>> an error. >>>> >>> >>> If so, is it better to set 'c->hotpluggable' for all cpus on s390 during >>> initializing ? >> > No, just leave the code as it is. >> > Please help check the diff below whether is valuable, thanks. For architectures which may support 'hotpluggable', can loop all cpus during subsys_initcall(). The caller will skip the return value of subsys_initcall(), so can only return 0. --------------------------------diff begin------------------------------ diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c index 15a016c..147cc18 100644 --- a/arch/s390/kernel/smp.c +++ b/arch/s390/kernel/smp.c @@ -1065,8 +1065,9 @@ static int __init s390_smp_init(void) #endif for_each_present_cpu(cpu) { rc = smp_add_present_cpu(cpu); - if (rc) - return rc; + if (unlikely(rc)) + printk(KERN_WARNING "%s: register_cpu %d failed (%d)\n", + __func__, cpu, rc); } return 0; } --------------------------------diff end-------------------------------- > Currently we have 3 possible states for 'struct cpu': > initialized, and set hotplugable. > unitialized, and set hotplugable. > unitialized, and unset hotplugable. > > Either 2nd or 3rd is inconsistent. > > The cpu_is_hotplugable() will be used in kernel/rcutorture.c. I do not > know whether it will cause issue or not, but it seems dangerous when > __smp_rescan_cpus() has already called, and has some unsucessfull call > of smp_add_present_cpu(). > > Please help check. > > Thanks. > -- Chen Gang Asianux Corporation