Revert "Move xen patchset to new version's subdir."

[people/pmueller/ipfire-2.x.git] / src / patches / suse-2.6.27.25 / patches.arch / s390-07-03-topology-fix.diff

From: Gerald Schaefer <geraldsc@de.ibm.com>
Subject: kernel: fix cpu topology support
References: bnc#464466

Symptom:     CPU topology changes aren't recognized by the scheduler.
Problem:     The common code scheduler used to have a hook which could be
             called from architecture code to trigger a rebuild of all
             scheduling domains when cpu topology changed. This hook got
             removed errorneously. So cpu topology change notifications
             got lost.
Solution:    Readd the hook. This patch also removes some unused code
             from the s390 specific cpu topology code.

Acked-by: John Jolly <jjolly@suse.de>
---
 arch/s390/kernel/topology.c |   35 ++++++++++-------------------------
 include/linux/topology.h    |    2 +-
 kernel/sched.c              |   16 +++++++++++++---
 3 files changed, 24 insertions(+), 29 deletions(-)

--- a/arch/s390/kernel/topology.c
+++ b/arch/s390/kernel/topology.c
@@ -14,6 +14,7 @@
 #include <linux/workqueue.h>
 #include <linux/cpu.h>
 #include <linux/smp.h>
+#include <linux/cpuset.h>
 #include <asm/delay.h>
 #include <asm/s390_ext.h>
 #include <asm/sysinfo.h>
@@ -64,7 +65,6 @@ static void topology_work_fn(struct work
 static struct tl_info *tl_info;
 static struct core_info core_info;
 static int machine_has_topology;
-static int machine_has_topology_irq;
 static struct timer_list topology_timer;
 static void set_topology_timer(void);
 static DECLARE_WORK(topology_work, topology_work_fn);
@@ -81,7 +81,7 @@ cpumask_t cpu_coregroup_map(unsigned int
 
 	cpus_clear(mask);
 	if (!topology_enabled || !machine_has_topology)
-		return cpu_present_map;
+		return cpu_possible_map;
 	spin_lock_irqsave(&topology_lock, flags);
 	while (core) {
 		if (cpu_isset(cpu, core->mask)) {
@@ -171,7 +171,7 @@ static void topology_update_polarization
 	int cpu;
 
 	mutex_lock(&smp_cpu_state_mutex);
-	for_each_present_cpu(cpu)
+	for_each_possible_cpu(cpu)
 		smp_cpu_polarization[cpu] = POLARIZATION_HRZ;
 	mutex_unlock(&smp_cpu_state_mutex);
 }
@@ -202,7 +202,7 @@ int topology_set_cpu_management(int fc)
 		rc = ptf(PTF_HORIZONTAL);
 	if (rc)
 		return -EBUSY;
-	for_each_present_cpu(cpu)
+	for_each_possible_cpu(cpu)
 		smp_cpu_polarization[cpu] = POLARIZATION_UNKNWN;
 	return rc;
 }
@@ -211,11 +211,11 @@ static void update_cpu_core_map(void)
 {
 	int cpu;
 
-	for_each_present_cpu(cpu)
+	for_each_possible_cpu(cpu)
 		cpu_core_map[cpu] = cpu_coregroup_map(cpu);
 }
 
-void arch_update_cpu_topology(void)
+int arch_update_cpu_topology(void)
 {
 	struct tl_info *info = tl_info;
 	struct sys_device *sysdev;
@@ -224,7 +224,7 @@ void arch_update_cpu_topology(void)
 	if (!machine_has_topology) {
 		update_cpu_core_map();
 		topology_update_polarization_simple();
-		return;
+		return 0;
 	}
 	stsi(info, 15, 1, 2);
 	tl_to_cores(info);
@@ -233,11 +233,12 @@ void arch_update_cpu_topology(void)
 		sysdev = get_cpu_sysdev(cpu);
 		kobject_uevent(&sysdev->kobj, KOBJ_CHANGE);
 	}
+	return 1;
 }
 
 static void topology_work_fn(struct work_struct *work)
 {
-	arch_reinit_sched_domains();
+	rebuild_sched_domains();
 }
 
 void topology_schedule_update(void)
@@ -260,11 +261,6 @@ static void set_topology_timer(void)
 	add_timer(&topology_timer);
 }
 
-static void topology_interrupt(__u16 code)
-{
-	schedule_work(&topology_work);
-}
-
 static int __init early_parse_topology(char *p)
 {
 	if (strncmp(p, "on", 2))
@@ -284,14 +280,7 @@ static int __init init_topology_update(v
 		goto out;
 	}
 	init_timer_deferrable(&topology_timer);
-	if (machine_has_topology_irq) {
-		rc = register_external_interrupt(0x2005, topology_interrupt);
-		if (rc)
-			goto out;
-		ctl_set_bit(0, 8);
-	}
-	else
-		set_topology_timer();
+	set_topology_timer();
 out:
 	update_cpu_core_map();
 	return rc;
@@ -312,9 +301,6 @@ void __init s390_init_cpu_topology(void)
 		return;
 	machine_has_topology = 1;
 
-	if (facility_bits & (1ULL << 51))
-		machine_has_topology_irq = 1;
-
 	tl_info = alloc_bootmem_pages(PAGE_SIZE);
 	info = tl_info;
 	stsi(info, 15, 1, 2);
@@ -338,5 +324,4 @@ void __init s390_init_cpu_topology(void)
 	return;
 error:
 	machine_has_topology = 0;
-	machine_has_topology_irq = 0;
 }
--- a/include/linux/topology.h
+++ b/include/linux/topology.h
@@ -49,7 +49,7 @@
 	for_each_online_node(node)			\
 		if (nr_cpus_node(node))
 
-void arch_update_cpu_topology(void);
+int arch_update_cpu_topology(void);
 
 /* Conform to ACPI 2.0 SLIT distance definitions */
 #define LOCAL_DISTANCE		10
--- a/kernel/sched.c
+++ b/kernel/sched.c
@@ -7640,8 +7640,14 @@ static struct sched_domain_attr *dattr_c
  */
 static cpumask_t fallback_doms;
 
-void __attribute__((weak)) arch_update_cpu_topology(void)
+/*
+ * arch_update_cpu_topology lets virtualized architectures update the
+ * cpu core maps. It is supposed to return 1 if the topology changed
+ * or 0 if it stayed the same.
+ */
+int __attribute__((weak)) arch_update_cpu_topology(void)
 {
+	return 0;
 }
 
 /*
@@ -7735,17 +7741,21 @@ void partition_sched_domains(int ndoms_n
 			     struct sched_domain_attr *dattr_new)
 {
 	int i, j, n;
+	int top_changed;
 
 	mutex_lock(&sched_domains_mutex);
 
 	/* always unregister in case we don't destroy any domains */
 	unregister_sched_domain_sysctl();
 
+	/* Let architecture update cpu core mappings. */
+	top_changed = arch_update_cpu_topology();
+
 	n = doms_new ? ndoms_new : 0;
 
 	/* Destroy deleted domains */
 	for (i = 0; i < ndoms_cur; i++) {
-		for (j = 0; j < n; j++) {
+		for (j = 0; j < n && !top_changed; j++) {
 			if (cpus_equal(doms_cur[i], doms_new[j])
 			    && dattrs_equal(dattr_cur, i, dattr_new, j))
 				goto match1;
@@ -7765,7 +7775,7 @@ match1:
 
 	/* Build new domains */
 	for (i = 0; i < ndoms_new; i++) {
-		for (j = 0; j < ndoms_cur; j++) {
+		for (j = 0; j < ndoms_cur && !top_changed; j++) {
 			if (cpus_equal(doms_new[i], doms_cur[j])
 			    && dattrs_equal(dattr_new, i, dattr_cur, j))
 				goto match2;