aboutsummaryrefslogtreecommitdiff
path: root/drivers/acpi/acpi_pad.c
diff options
context:
space:
mode:
Diffstat (limited to 'drivers/acpi/acpi_pad.c')
-rw-r--r--drivers/acpi/acpi_pad.c214
1 files changed, 111 insertions, 103 deletions
diff --git a/drivers/acpi/acpi_pad.c b/drivers/acpi/acpi_pad.c
index 97991ac6f5f..f148a0580e0 100644
--- a/drivers/acpi/acpi_pad.c
+++ b/drivers/acpi/acpi_pad.c
@@ -27,21 +27,23 @@
#include <linux/freezer.h>
#include <linux/cpu.h>
#include <linux/clockchips.h>
-#include <acpi/acpi_bus.h>
-#include <acpi/acpi_drivers.h>
+#include <linux/slab.h>
+#include <linux/acpi.h>
+#include <asm/mwait.h>
-#define ACPI_PROCESSOR_AGGREGATOR_CLASS "processor_aggregator"
+#define ACPI_PROCESSOR_AGGREGATOR_CLASS "acpi_pad"
#define ACPI_PROCESSOR_AGGREGATOR_DEVICE_NAME "Processor Aggregator"
#define ACPI_PROCESSOR_AGGREGATOR_NOTIFY 0x80
static DEFINE_MUTEX(isolated_cpus_lock);
+static DEFINE_MUTEX(round_robin_lock);
-#define MWAIT_SUBSTATE_MASK (0xf)
-#define MWAIT_CSTATE_MASK (0xf)
-#define MWAIT_SUBSTATE_SIZE (4)
-#define CPUID_MWAIT_LEAF (5)
-#define CPUID5_ECX_EXTENSIONS_SUPPORTED (0x1)
-#define CPUID5_ECX_INTERRUPT_BREAK (0x2)
static unsigned long power_saving_mwait_eax;
+
+static unsigned char tsc_detected_unstable;
+static unsigned char tsc_marked_unstable;
+static unsigned char lapic_detected_unstable;
+static unsigned char lapic_marked_unstable;
+
static void power_saving_mwait_init(void)
{
unsigned int eax, ebx, ecx, edx;
@@ -70,10 +72,7 @@ static void power_saving_mwait_init(void)
power_saving_mwait_eax = (highest_cstate << MWAIT_SUBSTATE_SIZE) |
(highest_subcstate - 1);
- for_each_online_cpu(i)
- clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ON, &i);
-
-#if defined(CONFIG_GENERIC_TIME) && defined(CONFIG_X86)
+#if defined(CONFIG_X86)
switch (boot_cpu_data.x86_vendor) {
case X86_VENDOR_AMD:
case X86_VENDOR_INTEL:
@@ -81,13 +80,15 @@ static void power_saving_mwait_init(void)
* AMD Fam10h TSC will tick in all
* C/P/S0/S1 states when this bit is set.
*/
- if (boot_cpu_has(X86_FEATURE_NONSTOP_TSC))
- return;
-
- /*FALL THROUGH*/
+ if (!boot_cpu_has(X86_FEATURE_NONSTOP_TSC))
+ tsc_detected_unstable = 1;
+ if (!boot_cpu_has(X86_FEATURE_ARAT))
+ lapic_detected_unstable = 1;
+ break;
default:
- /* TSC could halt in idle, so notify users */
- mark_tsc_unstable("TSC halts in idle");
+ /* TSC & LAPIC could halt in idle */
+ tsc_detected_unstable = 1;
+ lapic_detected_unstable = 1;
}
#endif
}
@@ -106,7 +107,7 @@ static void round_robin_cpu(unsigned int tsk_index)
if (!alloc_cpumask_var(&tmp, GFP_KERNEL))
return;
- mutex_lock(&isolated_cpus_lock);
+ mutex_lock(&round_robin_lock);
cpumask_clear(tmp);
for_each_cpu(cpu, pad_busy_cpus)
cpumask_or(tmp, tmp, topology_thread_cpumask(cpu));
@@ -115,7 +116,7 @@ static void round_robin_cpu(unsigned int tsk_index)
if (cpumask_empty(tmp))
cpumask_andnot(tmp, cpu_online_mask, pad_busy_cpus);
if (cpumask_empty(tmp)) {
- mutex_unlock(&isolated_cpus_lock);
+ mutex_unlock(&round_robin_lock);
return;
}
for_each_cpu(cpu, tmp) {
@@ -130,7 +131,7 @@ static void round_robin_cpu(unsigned int tsk_index)
tsk_in_cpu[tsk_index] = preferred_cpu;
cpumask_set_cpu(preferred_cpu, pad_busy_cpus);
cpu_weight[preferred_cpu]++;
- mutex_unlock(&isolated_cpus_lock);
+ mutex_unlock(&round_robin_lock);
set_cpus_allowed_ptr(current, cpumask_of(preferred_cpu));
}
@@ -143,7 +144,7 @@ static void exit_round_robin(unsigned int tsk_index)
}
static unsigned int idle_pct = 5; /* percentage */
-static unsigned int round_robin_time = 10; /* second */
+static unsigned int round_robin_time = 1; /* second */
static int power_saving_thread(void *data)
{
struct sched_param param = {.sched_priority = 1};
@@ -155,52 +156,57 @@ static int power_saving_thread(void *data)
while (!kthread_should_stop()) {
int cpu;
- u64 expire_time;
+ unsigned long expire_time;
try_to_freeze();
/* round robin to cpus */
- if (last_jiffies + round_robin_time * HZ < jiffies) {
+ expire_time = last_jiffies + round_robin_time * HZ;
+ if (time_before(expire_time, jiffies)) {
last_jiffies = jiffies;
round_robin_cpu(tsk_index);
}
do_sleep = 0;
- current_thread_info()->status &= ~TS_POLLING;
- /*
- * TS_POLLING-cleared state must be visible before we test
- * NEED_RESCHED:
- */
- smp_mb();
-
expire_time = jiffies + HZ * (100 - idle_pct) / 100;
while (!need_resched()) {
+ if (tsc_detected_unstable && !tsc_marked_unstable) {
+ /* TSC could halt in idle, so notify users */
+ mark_tsc_unstable("TSC halts in idle");
+ tsc_marked_unstable = 1;
+ }
+ if (lapic_detected_unstable && !lapic_marked_unstable) {
+ int i;
+ /* LAPIC could halt in idle, so notify users */
+ for_each_online_cpu(i)
+ clockevents_notify(
+ CLOCK_EVT_NOTIFY_BROADCAST_ON,
+ &i);
+ lapic_marked_unstable = 1;
+ }
local_irq_disable();
cpu = smp_processor_id();
- clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER,
- &cpu);
+ if (lapic_marked_unstable)
+ clockevents_notify(
+ CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &cpu);
stop_critical_timings();
- __monitor((void *)&current_thread_info()->flags, 0, 0);
- smp_mb();
- if (!need_resched())
- __mwait(power_saving_mwait_eax, 1);
+ mwait_idle_with_hints(power_saving_mwait_eax, 1);
start_critical_timings();
- clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT,
- &cpu);
+ if (lapic_marked_unstable)
+ clockevents_notify(
+ CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &cpu);
local_irq_enable();
- if (jiffies > expire_time) {
+ if (time_before(expire_time, jiffies)) {
do_sleep = 1;
break;
}
}
- current_thread_info()->status |= TS_POLLING;
-
/*
* current sched_rt has threshold for rt task running time.
* When a rt task uses 95% CPU time, the rt thread will be
@@ -208,10 +214,17 @@ static int power_saving_thread(void *data)
* the mechanism only works when all CPUs have RT task running,
* as if one CPU hasn't RT task, RT task from other CPUs will
* borrow CPU time from this CPU and cause RT task use > 95%
- * CPU time. To make 'avoid staration' work, takes a nap here.
+ * CPU time. To make 'avoid starvation' work, takes a nap here.
*/
- if (do_sleep)
+ if (unlikely(do_sleep))
schedule_timeout_killable(HZ * idle_pct / 100);
+
+ /* If an external event has set the need_resched flag, then
+ * we need to deal with it, or this loop will continue to
+ * spin without calling __mwait().
+ */
+ if (unlikely(need_resched()))
+ schedule();
}
exit_round_robin(tsk_index);
@@ -222,14 +235,21 @@ static struct task_struct *ps_tsks[NR_CPUS];
static unsigned int ps_tsk_num;
static int create_power_saving_task(void)
{
+ int rc;
+
ps_tsks[ps_tsk_num] = kthread_run(power_saving_thread,
(void *)(unsigned long)ps_tsk_num,
- "power_saving/%d", ps_tsk_num);
- if (ps_tsks[ps_tsk_num]) {
+ "acpi_pad/%d", ps_tsk_num);
+
+ if (IS_ERR(ps_tsks[ps_tsk_num])) {
+ rc = PTR_ERR(ps_tsks[ps_tsk_num]);
+ ps_tsks[ps_tsk_num] = NULL;
+ } else {
+ rc = 0;
ps_tsk_num++;
- return 0;
}
- return -EINVAL;
+
+ return rc;
}
static void destroy_power_saving_task(void)
@@ -237,6 +257,7 @@ static void destroy_power_saving_task(void)
if (ps_tsk_num > 0) {
ps_tsk_num--;
kthread_stop(ps_tsks[ps_tsk_num]);
+ ps_tsks[ps_tsk_num] = NULL;
}
}
@@ -253,7 +274,7 @@ static void set_power_saving_task_num(unsigned int num)
}
}
-static int acpi_pad_idle_cpus(unsigned int num_cpus)
+static void acpi_pad_idle_cpus(unsigned int num_cpus)
{
get_online_cpus();
@@ -261,7 +282,6 @@ static int acpi_pad_idle_cpus(unsigned int num_cpus)
set_power_saving_task_num(num_cpus);
put_online_cpus();
- return 0;
}
static uint32_t acpi_pad_idle_cpus_num(void)
@@ -273,7 +293,7 @@ static ssize_t acpi_pad_rrtime_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
unsigned long num;
- if (strict_strtoul(buf, 0, &num))
+ if (kstrtoul(buf, 0, &num))
return -EINVAL;
if (num < 1 || num >= 100)
return -EINVAL;
@@ -286,7 +306,7 @@ static ssize_t acpi_pad_rrtime_store(struct device *dev,
static ssize_t acpi_pad_rrtime_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
- return scnprintf(buf, PAGE_SIZE, "%d", round_robin_time);
+ return scnprintf(buf, PAGE_SIZE, "%d\n", round_robin_time);
}
static DEVICE_ATTR(rrtime, S_IRUGO|S_IWUSR,
acpi_pad_rrtime_show,
@@ -296,7 +316,7 @@ static ssize_t acpi_pad_idlepct_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
unsigned long num;
- if (strict_strtoul(buf, 0, &num))
+ if (kstrtoul(buf, 0, &num))
return -EINVAL;
if (num < 1 || num >= 100)
return -EINVAL;
@@ -309,7 +329,7 @@ static ssize_t acpi_pad_idlepct_store(struct device *dev,
static ssize_t acpi_pad_idlepct_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
- return scnprintf(buf, PAGE_SIZE, "%d", idle_pct);
+ return scnprintf(buf, PAGE_SIZE, "%d\n", idle_pct);
}
static DEVICE_ATTR(idlepct, S_IRUGO|S_IWUSR,
acpi_pad_idlepct_show,
@@ -319,7 +339,7 @@ static ssize_t acpi_pad_idlecpus_store(struct device *dev,
struct device_attribute *attr, const char *buf, size_t count)
{
unsigned long num;
- if (strict_strtoul(buf, 0, &num))
+ if (kstrtoul(buf, 0, &num))
return -EINVAL;
mutex_lock(&isolated_cpus_lock);
acpi_pad_idle_cpus(num);
@@ -330,8 +350,11 @@ static ssize_t acpi_pad_idlecpus_store(struct device *dev,
static ssize_t acpi_pad_idlecpus_show(struct device *dev,
struct device_attribute *attr, char *buf)
{
- return cpumask_scnprintf(buf, PAGE_SIZE,
- to_cpumask(pad_busy_cpus_bits));
+ int n = 0;
+ n = cpumask_scnprintf(buf, PAGE_SIZE-2, to_cpumask(pad_busy_cpus_bits));
+ buf[n++] = '\n';
+ buf[n] = '\0';
+ return n;
}
static DEVICE_ATTR(idlecpus, S_IRUGO|S_IWUSR,
acpi_pad_idlecpus_show,
@@ -365,65 +388,52 @@ static void acpi_pad_remove_sysfs(struct acpi_device *device)
device_remove_file(&device->dev, &dev_attr_rrtime);
}
-/* Query firmware how many CPUs should be idle */
-static int acpi_pad_pur(acpi_handle handle, int *num_cpus)
+/*
+ * Query firmware how many CPUs should be idle
+ * return -1 on failure
+ */
+static int acpi_pad_pur(acpi_handle handle)
{
struct acpi_buffer buffer = {ACPI_ALLOCATE_BUFFER, NULL};
- acpi_status status;
union acpi_object *package;
- int rev, num, ret = -EINVAL;
+ int num = -1;
+
+ if (ACPI_FAILURE(acpi_evaluate_object(handle, "_PUR", NULL, &buffer)))
+ return num;
+
+ if (!buffer.length || !buffer.pointer)
+ return num;
- status = acpi_evaluate_object(handle, "_PUR", NULL, &buffer);
- if (ACPI_FAILURE(status))
- return -EINVAL;
package = buffer.pointer;
- if (package->type != ACPI_TYPE_PACKAGE || package->package.count != 2)
- goto out;
- rev = package->package.elements[0].integer.value;
- num = package->package.elements[1].integer.value;
- if (rev != 1)
- goto out;
- *num_cpus = num;
- ret = 0;
-out:
- kfree(buffer.pointer);
- return ret;
-}
-/* Notify firmware how many CPUs are idle */
-static void acpi_pad_ost(acpi_handle handle, int stat,
- uint32_t idle_cpus)
-{
- union acpi_object params[3] = {
- {.type = ACPI_TYPE_INTEGER,},
- {.type = ACPI_TYPE_INTEGER,},
- {.type = ACPI_TYPE_BUFFER,},
- };
- struct acpi_object_list arg_list = {3, params};
+ if (package->type == ACPI_TYPE_PACKAGE &&
+ package->package.count == 2 &&
+ package->package.elements[0].integer.value == 1) /* rev 1 */
+
+ num = package->package.elements[1].integer.value;
- params[0].integer.value = ACPI_PROCESSOR_AGGREGATOR_NOTIFY;
- params[1].integer.value = stat;
- params[2].buffer.length = 4;
- params[2].buffer.pointer = (void *)&idle_cpus;
- acpi_evaluate_object(handle, "_OST", &arg_list, NULL);
+ kfree(buffer.pointer);
+ return num;
}
static void acpi_pad_handle_notify(acpi_handle handle)
{
- int num_cpus, ret;
+ int num_cpus;
uint32_t idle_cpus;
+ struct acpi_buffer param = {
+ .length = 4,
+ .pointer = (void *)&idle_cpus,
+ };
mutex_lock(&isolated_cpus_lock);
- if (acpi_pad_pur(handle, &num_cpus)) {
+ num_cpus = acpi_pad_pur(handle);
+ if (num_cpus < 0) {
mutex_unlock(&isolated_cpus_lock);
return;
}
- ret = acpi_pad_idle_cpus(num_cpus);
+ acpi_pad_idle_cpus(num_cpus);
idle_cpus = acpi_pad_idle_cpus_num();
- if (!ret)
- acpi_pad_ost(handle, 0, idle_cpus);
- else
- acpi_pad_ost(handle, 1, 0);
+ acpi_evaluate_ost(handle, ACPI_PROCESSOR_AGGREGATOR_NOTIFY, 0, &param);
mutex_unlock(&isolated_cpus_lock);
}
@@ -435,12 +445,11 @@ static void acpi_pad_notify(acpi_handle handle, u32 event,
switch (event) {
case ACPI_PROCESSOR_AGGREGATOR_NOTIFY:
acpi_pad_handle_notify(handle);
- acpi_bus_generate_proc_event(device, event, 0);
acpi_bus_generate_netlink_event(device->pnp.device_class,
dev_name(&device->dev), event, 0);
break;
default:
- printk(KERN_WARNING"Unsupported event [0x%x]\n", event);
+ pr_warn("Unsupported event [0x%x]\n", event);
break;
}
}
@@ -465,8 +474,7 @@ static int acpi_pad_add(struct acpi_device *device)
return 0;
}
-static int acpi_pad_remove(struct acpi_device *device,
- int type)
+static int acpi_pad_remove(struct acpi_device *device)
{
mutex_lock(&isolated_cpus_lock);
acpi_pad_idle_cpus(0);
generated by cgit v1.2.3-70-g09d2 (git 2.46.0) at 2025-11-09 06:55:41 +0000