From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752696AbaBNH6M (ORCPT ); Fri, 14 Feb 2014 02:58:12 -0500 Received: from e23smtp08.au.ibm.com ([202.81.31.141]:39910 "EHLO e23smtp08.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752010AbaBNH6I (ORCPT ); Fri, 14 Feb 2014 02:58:08 -0500 From: "Srivatsa S. Bhat" Subject: [PATCH v2 14/52] powerpc, sysfs: Fix CPU hotplug callback registration To: paulus@samba.org, oleg@redhat.com, mingo@kernel.org, rusty@rustcorp.com.au, peterz@infradead.org, tglx@linutronix.de, akpm@linux-foundation.org Cc: paulmck@linux.vnet.ibm.com, tj@kernel.org, walken@google.com, ego@linux.vnet.ibm.com, linux@arm.linux.org.uk, rjw@rjwysocki.net, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, srivatsa.bhat@linux.vnet.ibm.com, Benjamin Herrenschmidt , Paul Mackerras , Olof Johansson , Wang Dongsheng , Ingo Molnar , linuxppc-dev@lists.ozlabs.org, Madhavan Srinivasan , "Srivatsa S. Bhat" Date: Fri, 14 Feb 2014 13:22:15 +0530 Message-ID: <20140214075215.22701.23653.stgit@srivatsabhat.in.ibm.com> In-Reply-To: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> References: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> User-Agent: StGIT/0.14.3 MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit X-TM-AS-MML: disable X-Content-Scanned: Fidelis XPS MAILER x-cbid: 14021407-5140-0000-0000-0000048EE356 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Subsystems that want to register CPU hotplug callbacks, as well as perform initialization for the CPUs that are already online, often do it as shown below: get_online_cpus(); for_each_online_cpu(cpu) init_cpu(cpu); register_cpu_notifier(&foobar_cpu_notifier); put_online_cpus(); This is wrong, since it is prone to ABBA deadlocks involving the cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently with CPU hotplug operations). Instead, the correct and race-free way of performing the callback registration is: cpu_notifier_register_begin(); for_each_online_cpu(cpu) init_cpu(cpu); /* Note the use of the double underscored version of the API */ __register_cpu_notifier(&foobar_cpu_notifier); cpu_notifier_register_done(); Fix the sysfs code in powerpc by using this latter form of callback registration. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Olof Johansson Cc: Wang Dongsheng Cc: Ingo Molnar Cc: linuxppc-dev@lists.ozlabs.org Acked-by: Madhavan Srinivasan Signed-off-by: Srivatsa S. Bhat --- arch/powerpc/kernel/sysfs.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c index 97e1dc9..d90d4b7 100644 --- a/arch/powerpc/kernel/sysfs.c +++ b/arch/powerpc/kernel/sysfs.c @@ -975,7 +975,8 @@ static int __init topology_init(void) int cpu; register_nodes(); - register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_begin(); for_each_possible_cpu(cpu) { struct cpu *c = &per_cpu(cpu_devices, cpu); @@ -999,6 +1000,11 @@ static int __init topology_init(void) if (cpu_online(cpu)) register_cpu_online(cpu); } + + __register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_done(); + #ifdef CONFIG_PPC64 sysfs_create_dscr_default(); #endif /* CONFIG_PPC64 */ From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Srivatsa S. Bhat" Subject: [PATCH v2 14/52] powerpc, sysfs: Fix CPU hotplug callback registration Date: Fri, 14 Feb 2014 13:22:15 +0530 Message-ID: <20140214075215.22701.23653.stgit@srivatsabhat.in.ibm.com> References: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> Sender: linux-kernel-owner@vger.kernel.org To: oleg@redhat.com, rusty@rustcorp.com.au, peterz@infradead.org, tglx@linutronix.de, akpm@linux-foundation.org Cc: paulmck@linux.vnet.ibm.com, tj@kernel.org, walken@google.com, ego@linux.vnet.ibm.com, linux@arm.linux.org.uk, rjw@rjwysocki.net, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, srivatsa.bhat@linux.vnet.ibm.com, Benjamin Herrenschmidt , Paul Mackerras , Olof Johansson , Wang Dongsheng , Ingo Molnar , linuxppc-dev@lists.ozlabs.org, Madhavan Srinivasan "Srivatsa S. Bhat" List-Id: linux-arch.vger.kernel.org Subsystems that want to register CPU hotplug callbacks, as well as perform initialization for the CPUs that are already online, often do it as shown below: get_online_cpus(); for_each_online_cpu(cpu) init_cpu(cpu); register_cpu_notifier(&foobar_cpu_notifier); put_online_cpus(); This is wrong, since it is prone to ABBA deadlocks involving the cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently with CPU hotplug operations). Instead, the correct and race-free way of performing the callback registration is: cpu_notifier_register_begin(); for_each_online_cpu(cpu) init_cpu(cpu); /* Note the use of the double underscored version of the API */ __register_cpu_notifier(&foobar_cpu_notifier); cpu_notifier_register_done(); Fix the sysfs code in powerpc by using this latter form of callback registration. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Olof Johansson Cc: Wang Dongsheng Cc: Ingo Molnar Cc: linuxppc-dev@lists.ozlabs.org Acked-by: Madhavan Srinivasan Signed-off-by: Srivatsa S. Bhat --- arch/powerpc/kernel/sysfs.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c index 97e1dc9..d90d4b7 100644 --- a/arch/powerpc/kernel/sysfs.c +++ b/arch/powerpc/kernel/sysfs.c @@ -975,7 +975,8 @@ static int __init topology_init(void) int cpu; register_nodes(); - register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_begin(); for_each_possible_cpu(cpu) { struct cpu *c = &per_cpu(cpu_devices, cpu); @@ -999,6 +1000,11 @@ static int __init topology_init(void) if (cpu_online(cpu)) register_cpu_online(cpu); } + + __register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_done(); + #ifdef CONFIG_PPC64 sysfs_create_dscr_default(); #endif /* CONFIG_PPC64 */ From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e23smtp08.au.ibm.com ([202.81.31.141]:39908 "EHLO e23smtp08.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752375AbaBNH6I (ORCPT ); Fri, 14 Feb 2014 02:58:08 -0500 Received: from /spool/local by e23smtp08.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 14 Feb 2014 17:57:49 +1000 From: "Srivatsa S. Bhat" Subject: [PATCH v2 14/52] powerpc, sysfs: Fix CPU hotplug callback registration Date: Fri, 14 Feb 2014 13:22:15 +0530 Message-ID: <20140214075215.22701.23653.stgit@srivatsabhat.in.ibm.com> In-Reply-To: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> References: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 7bit Sender: linux-arch-owner@vger.kernel.org List-ID: To: paulus@samba.org, oleg@redhat.com, mingo@kernel.org, rusty@rustcorp.com.au, peterz@infradead.org, tglx@linutronix.de, akpm@linux-foundation.org Cc: paulmck@linux.vnet.ibm.com, tj@kernel.org, walken@google.com, ego@linux.vnet.ibm.com, linux@arm.linux.org.uk, rjw@rjwysocki.net, linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org, srivatsa.bhat@linux.vnet.ibm.com, Benjamin Herrenschmidt , Olof Johansson , Wang Dongsheng , linuxppc-dev@lists.ozlabs.org, Madhavan Srinivasan "Srivatsa S. Bhat" Message-ID: <20140214075215.7d-SFeTNgbZAuG9YezTl-Z1RmNVNGjbDcYTUOpD1ZIU@z> Subsystems that want to register CPU hotplug callbacks, as well as perform initialization for the CPUs that are already online, often do it as shown below: get_online_cpus(); for_each_online_cpu(cpu) init_cpu(cpu); register_cpu_notifier(&foobar_cpu_notifier); put_online_cpus(); This is wrong, since it is prone to ABBA deadlocks involving the cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently with CPU hotplug operations). Instead, the correct and race-free way of performing the callback registration is: cpu_notifier_register_begin(); for_each_online_cpu(cpu) init_cpu(cpu); /* Note the use of the double underscored version of the API */ __register_cpu_notifier(&foobar_cpu_notifier); cpu_notifier_register_done(); Fix the sysfs code in powerpc by using this latter form of callback registration. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Olof Johansson Cc: Wang Dongsheng Cc: Ingo Molnar Cc: linuxppc-dev@lists.ozlabs.org Acked-by: Madhavan Srinivasan Signed-off-by: Srivatsa S. Bhat --- arch/powerpc/kernel/sysfs.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c index 97e1dc9..d90d4b7 100644 --- a/arch/powerpc/kernel/sysfs.c +++ b/arch/powerpc/kernel/sysfs.c @@ -975,7 +975,8 @@ static int __init topology_init(void) int cpu; register_nodes(); - register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_begin(); for_each_possible_cpu(cpu) { struct cpu *c = &per_cpu(cpu_devices, cpu); @@ -999,6 +1000,11 @@ static int __init topology_init(void) if (cpu_online(cpu)) register_cpu_online(cpu); } + + __register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_done(); + #ifdef CONFIG_PPC64 sysfs_create_dscr_default(); #endif /* CONFIG_PPC64 */ From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e23smtp09.au.ibm.com (e23smtp09.au.ibm.com [202.81.31.142]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by ozlabs.org (Postfix) with ESMTPS id 6807E2C00AC for ; Fri, 14 Feb 2014 18:58:08 +1100 (EST) Received: from /spool/local by e23smtp09.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 14 Feb 2014 17:57:48 +1000 Received: from d23relay03.au.ibm.com (d23relay03.au.ibm.com [9.190.235.21]) by d23dlp02.au.ibm.com (Postfix) with ESMTP id 2DF7B2BB0052 for ; Fri, 14 Feb 2014 18:57:45 +1100 (EST) Received: from d23av03.au.ibm.com (d23av03.au.ibm.com [9.190.234.97]) by d23relay03.au.ibm.com (8.13.8/8.13.8/NCO v10.0) with ESMTP id s1E7vVQR2163180 for ; Fri, 14 Feb 2014 18:57:32 +1100 Received: from d23av03.au.ibm.com (localhost [127.0.0.1]) by d23av03.au.ibm.com (8.14.4/8.14.4/NCO v10.0 AVout) with ESMTP id s1E7vhg7023017 for ; Fri, 14 Feb 2014 18:57:44 +1100 From: "Srivatsa S. Bhat" Subject: [PATCH v2 14/52] powerpc, sysfs: Fix CPU hotplug callback registration To: paulus@samba.org, oleg@redhat.com, mingo@kernel.org, rusty@rustcorp.com.au, peterz@infradead.org, tglx@linutronix.de, akpm@linux-foundation.org Date: Fri, 14 Feb 2014 13:22:15 +0530 Message-ID: <20140214075215.22701.23653.stgit@srivatsabhat.in.ibm.com> In-Reply-To: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> References: <20140214074750.22701.47330.stgit@srivatsabhat.in.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Cc: linux-arch@vger.kernel.org, ego@linux.vnet.ibm.com, walken@google.com, linux@arm.linux.org.uk, rjw@rjwysocki.net, linux-kernel@vger.kernel.org, Wang Dongsheng , Olof Johansson , Madhavan Srinivasan , Paul Mackerras , "Srivatsa S. Bhat" , tj@kernel.org, paulmck@linux.vnet.ibm.com, linuxppc-dev@lists.ozlabs.org, Ingo Molnar List-Id: Linux on PowerPC Developers Mail List List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Subsystems that want to register CPU hotplug callbacks, as well as perform initialization for the CPUs that are already online, often do it as shown below: get_online_cpus(); for_each_online_cpu(cpu) init_cpu(cpu); register_cpu_notifier(&foobar_cpu_notifier); put_online_cpus(); This is wrong, since it is prone to ABBA deadlocks involving the cpu_add_remove_lock and the cpu_hotplug.lock (when running concurrently with CPU hotplug operations). Instead, the correct and race-free way of performing the callback registration is: cpu_notifier_register_begin(); for_each_online_cpu(cpu) init_cpu(cpu); /* Note the use of the double underscored version of the API */ __register_cpu_notifier(&foobar_cpu_notifier); cpu_notifier_register_done(); Fix the sysfs code in powerpc by using this latter form of callback registration. Cc: Benjamin Herrenschmidt Cc: Paul Mackerras Cc: Olof Johansson Cc: Wang Dongsheng Cc: Ingo Molnar Cc: linuxppc-dev@lists.ozlabs.org Acked-by: Madhavan Srinivasan Signed-off-by: Srivatsa S. Bhat --- arch/powerpc/kernel/sysfs.c | 8 +++++++- 1 file changed, 7 insertions(+), 1 deletion(-) diff --git a/arch/powerpc/kernel/sysfs.c b/arch/powerpc/kernel/sysfs.c index 97e1dc9..d90d4b7 100644 --- a/arch/powerpc/kernel/sysfs.c +++ b/arch/powerpc/kernel/sysfs.c @@ -975,7 +975,8 @@ static int __init topology_init(void) int cpu; register_nodes(); - register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_begin(); for_each_possible_cpu(cpu) { struct cpu *c = &per_cpu(cpu_devices, cpu); @@ -999,6 +1000,11 @@ static int __init topology_init(void) if (cpu_online(cpu)) register_cpu_online(cpu); } + + __register_cpu_notifier(&sysfs_cpu_nb); + + cpu_notifier_register_done(); + #ifdef CONFIG_PPC64 sysfs_create_dscr_default(); #endif /* CONFIG_PPC64 */