linux/arch/s390/kernel/topology.c
Heiko Carstens c10fde0d9e [S390] Vertical cpu management.
If vertical cpu polarization is active then the hypervisor will
dispatch certain cpus for a longer time than other cpus for maximum
performance. For example if a guest would have three virtual cpus,
each of them with a share of 33 percent, then in case of vertical
cpu polarization all of the processing time would be combined to a
single cpu which would run all the time, while the other two cpus
would get nearly no cpu time.

There are three different types of vertical cpus: high, medium and
low. Low cpus hardly get any real cpu time, while high cpus get a
full real cpu. Medium cpus get something in between.

In order to switch between the two possible modes (default is
horizontal) a 0 for horizontal polarization or a 1 for vertical
polarization must be written to the dispatching sysfs attribute:

/sys/devices/system/cpu/dispatching

The polarization of each single cpu can be figured out by the
polarization sysfs attribute of each cpu:

/sys/devices/system/cpu/cpuX/polarization

horizontal, vertical:high, vertical:medium, vertical:low or unknown.

When switching polarization the polarization attribute may contain
the value unknown until the configuration change is done and the
kernel has figured out the new polarization of each cpu.

Note that running a system with different types of vertical cpus may
result in significant performance regressions. If possible only one
type of vertical cpus should be used. All other cpus should be
offlined.

Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
2008-04-17 07:47:01 +02:00

315 lines
6.1 KiB
C

/*
* Copyright IBM Corp. 2007
* Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
*/
#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/init.h>
#include <linux/device.h>
#include <linux/bootmem.h>
#include <linux/sched.h>
#include <linux/workqueue.h>
#include <linux/cpu.h>
#include <linux/smp.h>
#include <asm/delay.h>
#include <asm/s390_ext.h>
#include <asm/sysinfo.h>
#define CPU_BITS 64
#define NR_MAG 6
#define PTF_HORIZONTAL (0UL)
#define PTF_VERTICAL (1UL)
#define PTF_CHECK (2UL)
struct tl_cpu {
unsigned char reserved0[4];
unsigned char :6;
unsigned char pp:2;
unsigned char reserved1;
unsigned short origin;
unsigned long mask[CPU_BITS / BITS_PER_LONG];
};
struct tl_container {
unsigned char reserved[8];
};
union tl_entry {
unsigned char nl;
struct tl_cpu cpu;
struct tl_container container;
};
struct tl_info {
unsigned char reserved0[2];
unsigned short length;
unsigned char mag[NR_MAG];
unsigned char reserved1;
unsigned char mnest;
unsigned char reserved2[4];
union tl_entry tle[0];
};
struct core_info {
struct core_info *next;
cpumask_t mask;
};
static void topology_work_fn(struct work_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);
cpumask_t cpu_coregroup_map(unsigned int cpu)
{
struct core_info *core = &core_info;
cpumask_t mask;
cpus_clear(mask);
if (!machine_has_topology)
return cpu_present_map;
mutex_lock(&smp_cpu_state_mutex);
while (core) {
if (cpu_isset(cpu, core->mask)) {
mask = core->mask;
break;
}
core = core->next;
}
mutex_unlock(&smp_cpu_state_mutex);
if (cpus_empty(mask))
mask = cpumask_of_cpu(cpu);
return mask;
}
static void add_cpus_to_core(struct tl_cpu *tl_cpu, struct core_info *core)
{
unsigned int cpu;
for (cpu = find_first_bit(&tl_cpu->mask[0], CPU_BITS);
cpu < CPU_BITS;
cpu = find_next_bit(&tl_cpu->mask[0], CPU_BITS, cpu + 1))
{
unsigned int rcpu, lcpu;
rcpu = CPU_BITS - 1 - cpu + tl_cpu->origin;
for_each_present_cpu(lcpu) {
if (__cpu_logical_map[lcpu] == rcpu) {
cpu_set(lcpu, core->mask);
smp_cpu_polarization[lcpu] = tl_cpu->pp;
}
}
}
}
static void clear_cores(void)
{
struct core_info *core = &core_info;
while (core) {
cpus_clear(core->mask);
core = core->next;
}
}
static union tl_entry *next_tle(union tl_entry *tle)
{
if (tle->nl)
return (union tl_entry *)((struct tl_container *)tle + 1);
else
return (union tl_entry *)((struct tl_cpu *)tle + 1);
}
static void tl_to_cores(struct tl_info *info)
{
union tl_entry *tle, *end;
struct core_info *core = &core_info;
mutex_lock(&smp_cpu_state_mutex);
clear_cores();
tle = info->tle;
end = (union tl_entry *)((unsigned long)info + info->length);
while (tle < end) {
switch (tle->nl) {
case 5:
case 4:
case 3:
case 2:
break;
case 1:
core = core->next;
break;
case 0:
add_cpus_to_core(&tle->cpu, core);
break;
default:
clear_cores();
machine_has_topology = 0;
return;
}
tle = next_tle(tle);
}
mutex_unlock(&smp_cpu_state_mutex);
}
static void topology_update_polarization_simple(void)
{
int cpu;
mutex_lock(&smp_cpu_state_mutex);
for_each_present_cpu(cpu)
smp_cpu_polarization[cpu] = POLARIZATION_HRZ;
mutex_unlock(&smp_cpu_state_mutex);
}
static int ptf(unsigned long fc)
{
int rc;
asm volatile(
" .insn rre,0xb9a20000,%1,%1\n"
" ipm %0\n"
" srl %0,28\n"
: "=d" (rc)
: "d" (fc) : "cc");
return rc;
}
int topology_set_cpu_management(int fc)
{
int cpu;
int rc;
if (!machine_has_topology)
return -EOPNOTSUPP;
if (fc)
rc = ptf(PTF_VERTICAL);
else
rc = ptf(PTF_HORIZONTAL);
if (rc)
return -EBUSY;
for_each_present_cpu(cpu)
smp_cpu_polarization[cpu] = POLARIZATION_UNKNWN;
return rc;
}
void arch_update_cpu_topology(void)
{
struct tl_info *info = tl_info;
struct sys_device *sysdev;
int cpu;
if (!machine_has_topology) {
topology_update_polarization_simple();
return;
}
stsi(info, 15, 1, 2);
tl_to_cores(info);
for_each_online_cpu(cpu) {
sysdev = get_cpu_sysdev(cpu);
kobject_uevent(&sysdev->kobj, KOBJ_CHANGE);
}
}
static void topology_work_fn(struct work_struct *work)
{
arch_reinit_sched_domains();
}
void topology_schedule_update(void)
{
schedule_work(&topology_work);
}
static void topology_timer_fn(unsigned long ignored)
{
if (ptf(PTF_CHECK))
topology_schedule_update();
set_topology_timer();
}
static void set_topology_timer(void)
{
topology_timer.function = topology_timer_fn;
topology_timer.data = 0;
topology_timer.expires = jiffies + 60 * HZ;
add_timer(&topology_timer);
}
static void topology_interrupt(__u16 code)
{
schedule_work(&topology_work);
}
static int __init init_topology_update(void)
{
int rc;
if (!machine_has_topology) {
topology_update_polarization_simple();
return 0;
}
init_timer_deferrable(&topology_timer);
if (machine_has_topology_irq) {
rc = register_external_interrupt(0x2005, topology_interrupt);
if (rc)
return rc;
ctl_set_bit(0, 8);
}
else
set_topology_timer();
return 0;
}
__initcall(init_topology_update);
void __init s390_init_cpu_topology(void)
{
unsigned long long facility_bits;
struct tl_info *info;
struct core_info *core;
int nr_cores;
int i;
if (stfle(&facility_bits, 1) <= 0)
return;
if (!(facility_bits & (1ULL << 52)) || !(facility_bits & (1ULL << 61)))
return;
machine_has_topology = 1;
if (facility_bits & (1ULL << 51))
machine_has_topology_irq = 1;
tl_info = alloc_bootmem_pages(PAGE_SIZE);
if (!tl_info)
goto error;
info = tl_info;
stsi(info, 15, 1, 2);
nr_cores = info->mag[NR_MAG - 2];
for (i = 0; i < info->mnest - 2; i++)
nr_cores *= info->mag[NR_MAG - 3 - i];
printk(KERN_INFO "CPU topology:");
for (i = 0; i < NR_MAG; i++)
printk(" %d", info->mag[i]);
printk(" / %d\n", info->mnest);
core = &core_info;
for (i = 0; i < nr_cores; i++) {
core->next = alloc_bootmem(sizeof(struct core_info));
core = core->next;
if (!core)
goto error;
}
return;
error:
machine_has_topology = 0;
machine_has_topology_irq = 0;
}