From: Matthew Garrett <mjg@redhat.com>
Date: Tue, 10 Nov 2009 19:33:03 -0500
Subject: [acpi] run events on cpu 0
Message-id: <1257881583-1907-1-git-send-email-mjg@redhat.com>
Patchwork-id: 21353
O-Subject: [PATCH] [RHEL 5.5] acpi: Run events on CPU 0
Bugzilla: 485016
RH-Acked-by: Danny Feng <dfeng@redhat.com>
RH-Acked-by: Prarit Bhargava <prarit@redhat.com>

Some ACPI events will trigger SMIs. Some machines (such as HP laptops) are
unable to cope with SMIs being triggered on CPUs other than CPU 0, and fail
to restore state correctly resulting in a crash. This is a backport of
the upstream code to bind the event runqueues to CPU 0.

diff --git a/drivers/acpi/osl.c b/drivers/acpi/osl.c
index 261621d..b6971a1 100644
--- a/drivers/acpi/osl.c
+++ b/drivers/acpi/osl.c
@@ -74,6 +74,21 @@ static acpi_osd_handler acpi_irq_handler;
 static void *acpi_irq_context;
 static struct workqueue_struct *kacpid_wq;
 
+static void bind_to_cpu0(void *context)
+{
+	set_cpus_allowed(current, cpumask_of_cpu(0));
+	kfree(context);
+}
+
+static void bind_workqueue(struct workqueue_struct *wq)
+{
+	struct work_struct *work;
+
+	work = kzalloc(sizeof(struct work_struct), GFP_KERNEL);
+	INIT_WORK(work, bind_to_cpu0, work);
+	queue_work(wq, work);
+}
+
 acpi_status acpi_os_initialize(void)
 {
 	return AE_OK;
@@ -92,6 +107,7 @@ acpi_status acpi_os_initialize1(void)
 	}
 	kacpid_wq = create_singlethread_workqueue("kacpid");
 	BUG_ON(!kacpid_wq);
+	bind_workqueue(kacpid_wq);
 
 	return AE_OK;
 }