From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752973AbaBNIBG (ORCPT ); Fri, 14 Feb 2014 03:01:06 -0500 Received: from e23smtp05.au.ibm.com ([202.81.31.147]:56056 "EHLO e23smtp05.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752948AbaBNIBD (ORCPT ); Fri, 14 Feb 2014 03:01:03 -0500 From: "Srivatsa S. Bhat" Subject: [PATCH v2 28/52] x86, oprofile, nmi: 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, Robert Richter , Thomas Gleixner , Ingo Molnar , "H. Peter Anvin" , x86@kernel.org, "Srivatsa S. Bhat" Date: Fri, 14 Feb 2014 13:25:29 +0530 Message-ID: <20140214075528.22701.64691.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: 14021408-1396-0000-0000-0000044B36C6 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 oprofile code in x86 by using this latter form of callback registration. But retain the calls to get/put_online_cpus(), since they are used in other places as well, to protect the variables 'nmi_enabled' and 'ctr_running'. Strictly speaking, this is not necessary since cpu_notifier_register_begin/done() provide a stronger synchronization with CPU hotplug than get/put_online_cpus(). However, let's retain the calls to get/put_online_cpus() to be consistent with the other call-sites. By nesting get/put_online_cpus() *inside* cpu_notifier_register_begin/done(), we avoid the ABBA deadlock possibility mentioned above. Cc: Robert Richter Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: x86@kernel.org Signed-off-by: Srivatsa S. Bhat --- arch/x86/oprofile/nmi_int.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 6890d84..379e8bd 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -494,14 +494,19 @@ static int nmi_setup(void) if (err) goto fail; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' */ get_online_cpus(); - register_cpu_notifier(&oprofile_cpu_nb); nmi_enabled = 1; /* make nmi_enabled visible to the nmi handler: */ smp_mb(); on_each_cpu(nmi_cpu_setup, NULL, 1); + __register_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + cpu_notifier_register_done(); + return 0; fail: free_msrs(); @@ -512,12 +517,18 @@ static void nmi_shutdown(void) { struct op_msrs *msrs; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' & 'ctr_running' */ get_online_cpus(); - unregister_cpu_notifier(&oprofile_cpu_nb); on_each_cpu(nmi_cpu_shutdown, NULL, 1); nmi_enabled = 0; ctr_running = 0; + __unregister_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + + cpu_notifier_register_done(); + /* make variables visible to the nmi handler: */ smp_mb(); unregister_nmi_handler(NMI_LOCAL, "oprofile"); From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Srivatsa S. Bhat" Subject: [PATCH v2 28/52] x86, oprofile, nmi: Fix CPU hotplug callback registration Date: Fri, 14 Feb 2014 13:25:29 +0530 Message-ID: <20140214075528.22701.64691.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: paulus@samba.org, oleg@redhat.com, rusty@rustcorp.com.au, peterz@infradead.org, 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, Robert Richter , Thomas Gleixner , Ingo Molnar , "H. Peter Anvin" , x86@kernel.org"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 oprofile code in x86 by using this latter form of callback registration. But retain the calls to get/put_online_cpus(), since they are used in other places as well, to protect the variables 'nmi_enabled' and 'ctr_running'. Strictly speaking, this is not necessary since cpu_notifier_register_begin/done() provide a stronger synchronization with CPU hotplug than get/put_online_cpus(). However, let's retain the calls to get/put_online_cpus() to be consistent with the other call-sites. By nesting get/put_online_cpus() *inside* cpu_notifier_register_begin/done(), we avoid the ABBA deadlock possibility mentioned above. Cc: Robert Richter Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: x86@kernel.org Signed-off-by: Srivatsa S. Bhat --- arch/x86/oprofile/nmi_int.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 6890d84..379e8bd 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -494,14 +494,19 @@ static int nmi_setup(void) if (err) goto fail; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' */ get_online_cpus(); - register_cpu_notifier(&oprofile_cpu_nb); nmi_enabled = 1; /* make nmi_enabled visible to the nmi handler: */ smp_mb(); on_each_cpu(nmi_cpu_setup, NULL, 1); + __register_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + cpu_notifier_register_done(); + return 0; fail: free_msrs(); @@ -512,12 +517,18 @@ static void nmi_shutdown(void) { struct op_msrs *msrs; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' & 'ctr_running' */ get_online_cpus(); - unregister_cpu_notifier(&oprofile_cpu_nb); on_each_cpu(nmi_cpu_shutdown, NULL, 1); nmi_enabled = 0; ctr_running = 0; + __unregister_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + + cpu_notifier_register_done(); + /* make variables visible to the nmi handler: */ smp_mb(); unregister_nmi_handler(NMI_LOCAL, "oprofile"); From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from e23smtp05.au.ibm.com ([202.81.31.147]:56054 "EHLO e23smtp05.au.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752947AbaBNIBD (ORCPT ); Fri, 14 Feb 2014 03:01:03 -0500 Received: from /spool/local by e23smtp05.au.ibm.com with IBM ESMTP SMTP Gateway: Authorized Use Only! Violators will be prosecuted for from ; Fri, 14 Feb 2014 18:01:00 +1000 From: "Srivatsa S. Bhat" Subject: [PATCH v2 28/52] x86, oprofile, nmi: Fix CPU hotplug callback registration Date: Fri, 14 Feb 2014 13:25:29 +0530 Message-ID: <20140214075528.22701.64691.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, Robert Richter , "H. Peter Anvin" , x86@kernel.org"Srivatsa S. Bhat" Message-ID: <20140214075529.2-ICuzumxg765Wipa-hL_V8OujKVtIaqi_kmGzWIh38@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 oprofile code in x86 by using this latter form of callback registration. But retain the calls to get/put_online_cpus(), since they are used in other places as well, to protect the variables 'nmi_enabled' and 'ctr_running'. Strictly speaking, this is not necessary since cpu_notifier_register_begin/done() provide a stronger synchronization with CPU hotplug than get/put_online_cpus(). However, let's retain the calls to get/put_online_cpus() to be consistent with the other call-sites. By nesting get/put_online_cpus() *inside* cpu_notifier_register_begin/done(), we avoid the ABBA deadlock possibility mentioned above. Cc: Robert Richter Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: x86@kernel.org Signed-off-by: Srivatsa S. Bhat --- arch/x86/oprofile/nmi_int.c | 15 +++++++++++++-- 1 file changed, 13 insertions(+), 2 deletions(-) diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c index 6890d84..379e8bd 100644 --- a/arch/x86/oprofile/nmi_int.c +++ b/arch/x86/oprofile/nmi_int.c @@ -494,14 +494,19 @@ static int nmi_setup(void) if (err) goto fail; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' */ get_online_cpus(); - register_cpu_notifier(&oprofile_cpu_nb); nmi_enabled = 1; /* make nmi_enabled visible to the nmi handler: */ smp_mb(); on_each_cpu(nmi_cpu_setup, NULL, 1); + __register_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + cpu_notifier_register_done(); + return 0; fail: free_msrs(); @@ -512,12 +517,18 @@ static void nmi_shutdown(void) { struct op_msrs *msrs; + cpu_notifier_register_begin(); + + /* Use get/put_online_cpus() to protect 'nmi_enabled' & 'ctr_running' */ get_online_cpus(); - unregister_cpu_notifier(&oprofile_cpu_nb); on_each_cpu(nmi_cpu_shutdown, NULL, 1); nmi_enabled = 0; ctr_running = 0; + __unregister_cpu_notifier(&oprofile_cpu_nb); put_online_cpus(); + + cpu_notifier_register_done(); + /* make variables visible to the nmi handler: */ smp_mb(); unregister_nmi_handler(NMI_LOCAL, "oprofile");