1
linux/drivers/acpi/processor_idle.c

1318 lines
34 KiB
C
Raw Normal View History

/*
* processor_idle - idle state submodule to the ACPI processor driver
*
* Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com>
* Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
* Copyright (C) 2004, 2005 Dominik Brodowski <linux@brodo.de>
* Copyright (C) 2004 Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
* - Added processor hotplug support
* Copyright (C) 2005 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
* - Added support for C3 on SMP
*
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or (at
* your option) any later version.
*
* This program is distributed in the hope that it will be useful, but
* WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* General Public License for more details.
*
* You should have received a copy of the GNU General Public License along
* with this program; if not, write to the Free Software Foundation, Inc.,
* 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA.
*
* ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*/
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/cpufreq.h>
#include <linux/proc_fs.h>
#include <linux/seq_file.h>
#include <linux/acpi.h>
#include <linux/dmi.h>
#include <linux/moduleparam.h>
#include <linux/sched.h> /* need_resched() */
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
#include <linux/latency.h>
#include <linux/clockchips.h>
/*
* Include the apic definitions for x86 to have the APIC timer related defines
* available also for UP (on SMP it gets magically included via linux/smp.h).
* asm/acpi.h is not an option, as it would require more include magic. Also
* creating an empty asm-ia64/apic.h would just trade pest vs. cholera.
*/
#ifdef CONFIG_X86
#include <asm/apic.h>
#endif
/*
* Include the apic definitions for x86 to have the APIC timer related defines
* available also for UP (on SMP it gets magically included via linux/smp.h).
*/
#ifdef CONFIG_X86
#include <asm/apic.h>
#endif
#include <asm/io.h>
#include <asm/uaccess.h>
#include <acpi/acpi_bus.h>
#include <acpi/processor.h>
#define ACPI_PROCESSOR_COMPONENT 0x01000000
#define ACPI_PROCESSOR_CLASS "processor"
#define _COMPONENT ACPI_PROCESSOR_COMPONENT
ACPI_MODULE_NAME("processor_idle");
#define ACPI_PROCESSOR_FILE_POWER "power"
#define US_TO_PM_TIMER_TICKS(t) ((t * (PM_TIMER_FREQUENCY/1000)) / 1000)
#define C2_OVERHEAD 4 /* 1us (3.579 ticks per us) */
#define C3_OVERHEAD 4 /* 1us (3.579 ticks per us) */
static void (*pm_idle_save) (void) __read_mostly;
module_param(max_cstate, uint, 0644);
static unsigned int nocst __read_mostly;
module_param(nocst, uint, 0000);
/*
* bm_history -- bit-mask with a bit per jiffy of bus-master activity
* 1000 HZ: 0xFFFFFFFF: 32 jiffies = 32ms
* 800 HZ: 0xFFFFFFFF: 32 jiffies = 40ms
* 100 HZ: 0x0000000F: 4 jiffies = 40ms
* reduce history for more aggressive entry into C3
*/
static unsigned int bm_history __read_mostly =
(HZ >= 800 ? 0xFFFFFFFF : ((1U << (HZ / 25)) - 1));
module_param(bm_history, uint, 0644);
/* --------------------------------------------------------------------------
Power Management
-------------------------------------------------------------------------- */
/*
* IBM ThinkPad R40e crashes mysteriously when going into C2 or C3.
* For now disable this. Probably a bug somewhere else.
*
* To skip this limit, boot/load with a large max_cstate limit.
*/
static int set_max_cstate(struct dmi_system_id *id)
{
if (max_cstate > ACPI_PROCESSOR_MAX_POWER)
return 0;
printk(KERN_NOTICE PREFIX "%s detected - limiting to C%ld max_cstate."
" Override with \"processor.max_cstate=%d\"\n", id->ident,
(long)id->driver_data, ACPI_PROCESSOR_MAX_POWER + 1);
max_cstate = (long)id->driver_data;
return 0;
}
/* Actually this shouldn't be __cpuinitdata, would be better to fix the
callers to only run once -AK */
static struct dmi_system_id __cpuinitdata processor_power_dmi_table[] = {
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET70WW")}, (void *)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET60WW")}, (void *)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET43WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET45WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET47WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET50WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET52WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET55WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET56WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET59WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET60WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET61WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET62WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET64WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET65WW") }, (void*)1},
{ set_max_cstate, "IBM ThinkPad R40e", {
DMI_MATCH(DMI_BIOS_VENDOR,"IBM"),
DMI_MATCH(DMI_BIOS_VERSION,"1SET68WW") }, (void*)1},
{ set_max_cstate, "Medion 41700", {
DMI_MATCH(DMI_BIOS_VENDOR,"Phoenix Technologies LTD"),
DMI_MATCH(DMI_BIOS_VERSION,"R01-A1J")}, (void *)1},
{ set_max_cstate, "Clevo 5600D", {
DMI_MATCH(DMI_BIOS_VENDOR,"Phoenix Technologies LTD"),
DMI_MATCH(DMI_BIOS_VERSION,"SHE845M0.86C.0013.D.0302131307")},
(void *)2},
{},
};
static inline u32 ticks_elapsed(u32 t1, u32 t2)
{
if (t2 >= t1)
return (t2 - t1);
else if (!(acpi_gbl_FADT.flags & ACPI_FADT_32BIT_TIMER))
return (((0x00FFFFFF - t1) + t2) & 0x00FFFFFF);
else
return ((0xFFFFFFFF - t1) + t2);
}
static void
acpi_processor_power_activate(struct acpi_processor *pr,
struct acpi_processor_cx *new)
{
struct acpi_processor_cx *old;
if (!pr || !new)
return;
old = pr->power.state;
if (old)
old->promotion.count = 0;
new->demotion.count = 0;
/* Cleanup from old state. */
if (old) {
switch (old->type) {
case ACPI_STATE_C3:
/* Disable bus master reload */
if (new->type != ACPI_STATE_C3 && pr->flags.bm_check)
acpi_set_register(ACPI_BITREG_BUS_MASTER_RLD, 0);
break;
}
}
/* Prepare to use new state. */
switch (new->type) {
case ACPI_STATE_C3:
/* Enable bus master reload */
if (old->type != ACPI_STATE_C3 && pr->flags.bm_check)
acpi_set_register(ACPI_BITREG_BUS_MASTER_RLD, 1);
break;
}
pr->power.state = new;
return;
}
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
static void acpi_safe_halt(void)
{
current_thread_info()->status &= ~TS_POLLING;
[PATCH] sched: fix bad missed wakeups in the i386, x86_64, ia64, ACPI and APM idle code Fernando Lopez-Lezcano reported frequent scheduling latencies and audio xruns starting at the 2.6.18-rt kernel, and those problems persisted all until current -rt kernels. The latencies were serious and unjustified by system load, often in the milliseconds range. After a patient and heroic multi-month effort of Fernando, where he tested dozens of kernels, tried various configs, boot options, test-patches of mine and provided latency traces of those incidents, the following 'smoking gun' trace was captured by him: _------=> CPU# / _-----=> irqs-off | / _----=> need-resched || / _---=> hardirq/softirq ||| / _--=> preempt-depth |||| / ||||| delay cmd pid ||||| time | caller \ / ||||| \ | / IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup (try_to_wake_up) IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup <<...>-5856> (37 0) IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup (c01262ba 0 0) IRQ_19-1479 1D..1 0us : resched_task (try_to_wake_up) IRQ_19-1479 1D..1 0us : __spin_unlock_irqrestore (try_to_wake_up) ... <idle>-0 1...1 11us!: default_idle (cpu_idle) ... <idle>-0 0Dn.1 602us : smp_apic_timer_interrupt (c0103baf 1 0) ... <...>-5856 0D..2 618us : __switch_to (__schedule) <...>-5856 0D..2 618us : __schedule <<idle>-0> (20 162) <...>-5856 0D..2 619us : __spin_unlock_irq (__schedule) <...>-5856 0...1 619us : trace_stop_sched_switched (__schedule) <...>-5856 0D..1 619us : trace_stop_sched_switched <<...>-5856> (37 0) what is visible in this trace is that CPU#1 ran try_to_wake_up() for PID:5856, it placed PID:5856 on CPU#0's runqueue and ran resched_task() for CPU#0. But it decided to not send an IPI that no CPU - due to TS_POLLING. But CPU#0 never woke up after its NEED_RESCHED bit was set, and only rescheduled to PID:5856 upon the next lapic timer IRQ. The result was a 600+ usecs latency and a missed wakeup! the bug turned out to be an idle-wakeup bug introduced into the mainline kernel this summer via an optimization in the x86_64 tree: commit 495ab9c045e1b0e5c82951b762257fe1c9d81564 Author: Andi Kleen <ak@suse.de> Date: Mon Jun 26 13:59:11 2006 +0200 [PATCH] i386/x86-64/ia64: Move polling flag into thread_info_status During some profiling I noticed that default_idle causes a lot of memory traffic. I think that is caused by the atomic operations to clear/set the polling flag in thread_info. There is actually no reason to make this atomic - only the idle thread does it to itself, other CPUs only read it. So I moved it into ti->status. the problem is this type of change: if (!hlt_counter && boot_cpu_data.hlt_works_ok) { - clear_thread_flag(TIF_POLLING_NRFLAG); + current_thread_info()->status &= ~TS_POLLING; smp_mb__after_clear_bit(); while (!need_resched()) { local_irq_disable(); this changes clear_thread_flag() to an explicit clearing of TS_POLLING. clear_thread_flag() is defined as: clear_bit(flag, &ti->flags); and clear_bit() is a LOCK-ed atomic instruction on all x86 platforms: static inline void clear_bit(int nr, volatile unsigned long * addr) { __asm__ __volatile__( LOCK_PREFIX "btrl %1,%0" hence smp_mb__after_clear_bit() is defined as a simple compile barrier: #define smp_mb__after_clear_bit() barrier() but the explicit TS_POLLING clearing introduced by the patch: + current_thread_info()->status &= ~TS_POLLING; is not an atomic op! So the clearing of the TS_POLLING bit is freely reorderable with the reading of the NEED_RESCHED bit - and both now reside in different memory addresses. CPU idle wakeup very much depends on ordered memory ops, the clearing of the TS_POLLING flag must always be done before we test need_resched() and hit the idle instruction(s). [Symmetrically, the wakeup code needs to set NEED_RESCHED before it tests the TS_POLLING flag, so memory ordering is paramount.] Fernando's dual-core Athlon64 system has a sufficiently advanced memory ordering model so that it triggered this scenario very often. ( And it also turned out that the reason why these latencies never triggered on my testsystems is that i routinely use idle=poll, which was the only idle variant not affected by this bug. ) The fix is to change the smp_mb__after_clear_bit() to an smp_mb(), to act as an absolute barrier between the TS_POLLING write and the NEED_RESCHED read. This affects almost all idling methods (default, ACPI, APM), on all 3 x86 architectures: i386, x86_64, ia64. Signed-off-by: Ingo Molnar <mingo@elte.hu> Tested-by: Fernando Lopez-Lezcano <nando@ccrma.Stanford.EDU> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-22 02:11:56 -07:00
/*
* TS_POLLING-cleared state must be visible before we
* test NEED_RESCHED:
*/
smp_mb();
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
if (!need_resched())
safe_halt();
current_thread_info()->status |= TS_POLLING;
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
}
static atomic_t c3_cpu_count;
/* Common C-state entry for C2, C3, .. */
static void acpi_cstate_enter(struct acpi_processor_cx *cstate)
{
if (cstate->space_id == ACPI_CSTATE_FFH) {
/* Call into architectural FFH based C-state */
acpi_processor_ffh_cstate_enter(cstate);
} else {
int unused;
/* IO port based C-state */
inb(cstate->address);
/* Dummy wait op - must do something useless after P_LVL2 read
because chipsets cannot guarantee that STPCLK# signal
gets asserted in time to freeze execution properly. */
unused = inl(acpi_gbl_FADT.xpm_timer_block.address);
}
}
#ifdef ARCH_APICTIMER_STOPS_ON_C3
/*
* Some BIOS implementations switch to C3 in the published C2 state.
* This seems to be a common problem on AMD boxen, but other vendors
* are affected too. We pick the most conservative approach: we assume
* that the local APIC stops in both C2 and C3.
*/
static void acpi_timer_check_state(int state, struct acpi_processor *pr,
struct acpi_processor_cx *cx)
{
struct acpi_processor_power *pwr = &pr->power;
u8 type = local_apic_timer_c2_ok ? ACPI_STATE_C3 : ACPI_STATE_C2;
/*
* Check, if one of the previous states already marked the lapic
* unstable
*/
if (pwr->timer_broadcast_on_state < state)
return;
if (cx->type >= type)
pr->power.timer_broadcast_on_state = state;
}
static void acpi_propagate_timer_broadcast(struct acpi_processor *pr)
{
#ifdef CONFIG_GENERIC_CLOCKEVENTS
unsigned long reason;
reason = pr->power.timer_broadcast_on_state < INT_MAX ?
CLOCK_EVT_NOTIFY_BROADCAST_ON : CLOCK_EVT_NOTIFY_BROADCAST_OFF;
clockevents_notify(reason, &pr->id);
#else
cpumask_t mask = cpumask_of_cpu(pr->id);
if (pr->power.timer_broadcast_on_state < INT_MAX)
on_each_cpu(switch_APIC_timer_to_ipi, &mask, 1, 1);
else
on_each_cpu(switch_ipi_to_APIC_timer, &mask, 1, 1);
#endif
}
/* Power(C) State timer broadcast control */
static void acpi_state_timer_broadcast(struct acpi_processor *pr,
struct acpi_processor_cx *cx,
int broadcast)
{
#ifdef CONFIG_GENERIC_CLOCKEVENTS
int state = cx - pr->power.states;
if (state >= pr->power.timer_broadcast_on_state) {
unsigned long reason;
reason = broadcast ? CLOCK_EVT_NOTIFY_BROADCAST_ENTER :
CLOCK_EVT_NOTIFY_BROADCAST_EXIT;
clockevents_notify(reason, &pr->id);
}
#endif
}
#else
static void acpi_timer_check_state(int state, struct acpi_processor *pr,
struct acpi_processor_cx *cstate) { }
static void acpi_propagate_timer_broadcast(struct acpi_processor *pr) { }
static void acpi_state_timer_broadcast(struct acpi_processor *pr,
struct acpi_processor_cx *cx,
int broadcast)
{
}
#endif
static void acpi_processor_idle(void)
{
struct acpi_processor *pr = NULL;
struct acpi_processor_cx *cx = NULL;
struct acpi_processor_cx *next_state = NULL;
int sleep_ticks = 0;
u32 t1, t2 = 0;
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
pr = processors[smp_processor_id()];
if (!pr)
return;
/*
* Interrupts must be disabled during bus mastering calculations and
* for C2/C3 transitions.
*/
local_irq_disable();
/*
* Check whether we truly need to go idle, or should
* reschedule:
*/
if (unlikely(need_resched())) {
local_irq_enable();
return;
}
cx = pr->power.state;
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
if (!cx) {
if (pm_idle_save)
pm_idle_save();
else
acpi_safe_halt();
return;
}
/*
* Check BM Activity
* -----------------
* Check for bus mastering activity (if required), record, and check
* for demotion.
*/
if (pr->flags.bm_check) {
u32 bm_status = 0;
unsigned long diff = jiffies - pr->power.bm_check_timestamp;
if (diff > 31)
diff = 31;
pr->power.bm_activity <<= diff;
acpi_get_register(ACPI_BITREG_BUS_MASTER_STATUS, &bm_status);
if (bm_status) {
pr->power.bm_activity |= 0x1;
acpi_set_register(ACPI_BITREG_BUS_MASTER_STATUS, 1);
}
/*
* PIIX4 Erratum #18: Note that BM_STS doesn't always reflect
* the true state of bus mastering activity; forcing us to
* manually check the BMIDEA bit of each IDE channel.
*/
else if (errata.piix4.bmisx) {
if ((inb_p(errata.piix4.bmisx + 0x02) & 0x01)
|| (inb_p(errata.piix4.bmisx + 0x0A) & 0x01))
pr->power.bm_activity |= 0x1;
}
pr->power.bm_check_timestamp = jiffies;
/*
* If bus mastering is or was active this jiffy, demote
* to avoid a faulty transition. Note that the processor
* won't enter a low-power state during this call (to this
* function) but should upon the next.
*
* TBD: A better policy might be to fallback to the demotion
* state (use it for this quantum only) istead of
* demoting -- and rely on duration as our sole demotion
* qualification. This may, however, introduce DMA
* issues (e.g. floppy DMA transfer overrun/underrun).
*/
if ((pr->power.bm_activity & 0x1) &&
cx->demotion.threshold.bm) {
local_irq_enable();
next_state = cx->demotion.state;
goto end;
}
}
#ifdef CONFIG_HOTPLUG_CPU
/*
* Check for P_LVL2_UP flag before entering C2 and above on
* an SMP system. We do it here instead of doing it at _CST/P_LVL
* detection phase, to work cleanly with logical CPU hotplug.
*/
if ((cx->type != ACPI_STATE_C1) && (num_online_cpus() > 1) &&
!pr->flags.has_cst && !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED))
cx = &pr->power.states[ACPI_STATE_C1];
#endif
/*
* Sleep:
* ------
* Invoke the current Cx state to put the processor to sleep.
*/
if (cx->type == ACPI_STATE_C2 || cx->type == ACPI_STATE_C3) {
current_thread_info()->status &= ~TS_POLLING;
[PATCH] sched: fix bad missed wakeups in the i386, x86_64, ia64, ACPI and APM idle code Fernando Lopez-Lezcano reported frequent scheduling latencies and audio xruns starting at the 2.6.18-rt kernel, and those problems persisted all until current -rt kernels. The latencies were serious and unjustified by system load, often in the milliseconds range. After a patient and heroic multi-month effort of Fernando, where he tested dozens of kernels, tried various configs, boot options, test-patches of mine and provided latency traces of those incidents, the following 'smoking gun' trace was captured by him: _------=> CPU# / _-----=> irqs-off | / _----=> need-resched || / _---=> hardirq/softirq ||| / _--=> preempt-depth |||| / ||||| delay cmd pid ||||| time | caller \ / ||||| \ | / IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup (try_to_wake_up) IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup <<...>-5856> (37 0) IRQ_19-1479 1D..1 0us : __trace_start_sched_wakeup (c01262ba 0 0) IRQ_19-1479 1D..1 0us : resched_task (try_to_wake_up) IRQ_19-1479 1D..1 0us : __spin_unlock_irqrestore (try_to_wake_up) ... <idle>-0 1...1 11us!: default_idle (cpu_idle) ... <idle>-0 0Dn.1 602us : smp_apic_timer_interrupt (c0103baf 1 0) ... <...>-5856 0D..2 618us : __switch_to (__schedule) <...>-5856 0D..2 618us : __schedule <<idle>-0> (20 162) <...>-5856 0D..2 619us : __spin_unlock_irq (__schedule) <...>-5856 0...1 619us : trace_stop_sched_switched (__schedule) <...>-5856 0D..1 619us : trace_stop_sched_switched <<...>-5856> (37 0) what is visible in this trace is that CPU#1 ran try_to_wake_up() for PID:5856, it placed PID:5856 on CPU#0's runqueue and ran resched_task() for CPU#0. But it decided to not send an IPI that no CPU - due to TS_POLLING. But CPU#0 never woke up after its NEED_RESCHED bit was set, and only rescheduled to PID:5856 upon the next lapic timer IRQ. The result was a 600+ usecs latency and a missed wakeup! the bug turned out to be an idle-wakeup bug introduced into the mainline kernel this summer via an optimization in the x86_64 tree: commit 495ab9c045e1b0e5c82951b762257fe1c9d81564 Author: Andi Kleen <ak@suse.de> Date: Mon Jun 26 13:59:11 2006 +0200 [PATCH] i386/x86-64/ia64: Move polling flag into thread_info_status During some profiling I noticed that default_idle causes a lot of memory traffic. I think that is caused by the atomic operations to clear/set the polling flag in thread_info. There is actually no reason to make this atomic - only the idle thread does it to itself, other CPUs only read it. So I moved it into ti->status. the problem is this type of change: if (!hlt_counter && boot_cpu_data.hlt_works_ok) { - clear_thread_flag(TIF_POLLING_NRFLAG); + current_thread_info()->status &= ~TS_POLLING; smp_mb__after_clear_bit(); while (!need_resched()) { local_irq_disable(); this changes clear_thread_flag() to an explicit clearing of TS_POLLING. clear_thread_flag() is defined as: clear_bit(flag, &ti->flags); and clear_bit() is a LOCK-ed atomic instruction on all x86 platforms: static inline void clear_bit(int nr, volatile unsigned long * addr) { __asm__ __volatile__( LOCK_PREFIX "btrl %1,%0" hence smp_mb__after_clear_bit() is defined as a simple compile barrier: #define smp_mb__after_clear_bit() barrier() but the explicit TS_POLLING clearing introduced by the patch: + current_thread_info()->status &= ~TS_POLLING; is not an atomic op! So the clearing of the TS_POLLING bit is freely reorderable with the reading of the NEED_RESCHED bit - and both now reside in different memory addresses. CPU idle wakeup very much depends on ordered memory ops, the clearing of the TS_POLLING flag must always be done before we test need_resched() and hit the idle instruction(s). [Symmetrically, the wakeup code needs to set NEED_RESCHED before it tests the TS_POLLING flag, so memory ordering is paramount.] Fernando's dual-core Athlon64 system has a sufficiently advanced memory ordering model so that it triggered this scenario very often. ( And it also turned out that the reason why these latencies never triggered on my testsystems is that i routinely use idle=poll, which was the only idle variant not affected by this bug. ) The fix is to change the smp_mb__after_clear_bit() to an smp_mb(), to act as an absolute barrier between the TS_POLLING write and the NEED_RESCHED read. This affects almost all idling methods (default, ACPI, APM), on all 3 x86 architectures: i386, x86_64, ia64. Signed-off-by: Ingo Molnar <mingo@elte.hu> Tested-by: Fernando Lopez-Lezcano <nando@ccrma.Stanford.EDU> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-12-22 02:11:56 -07:00
/*
* TS_POLLING-cleared state must be visible before we
* test NEED_RESCHED:
*/
smp_mb();
if (need_resched()) {
current_thread_info()->status |= TS_POLLING;
local_irq_enable();
return;
}
}
switch (cx->type) {
case ACPI_STATE_C1:
/*
* Invoke C1.
* Use the appropriate idle routine, the one that would
* be used without acpi C-states.
*/
if (pm_idle_save)
pm_idle_save();
else
[PATCH] sched: resched and cpu_idle rework Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce confusion, and make their semantics rigid. Improves efficiency of resched_task and some cpu_idle routines. * In resched_task: - TIF_NEED_RESCHED is only cleared with the task's runqueue lock held, and as we hold it during resched_task, then there is no need for an atomic test and set there. The only other time this should be set is when the task's quantum expires, in the timer interrupt - this is protected against because the rq lock is irq-safe. - If TIF_NEED_RESCHED is set, then we don't need to do anything. It won't get unset until the task get's schedule()d off. - If we are running on the same CPU as the task we resched, then set TIF_NEED_RESCHED and no further action is required. - If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set after TIF_NEED_RESCHED has been set, then we need to send an IPI. Using these rules, we are able to remove the test and set operation in resched_task, and make clear the previously vague semantics of POLLING_NRFLAG. * In idle routines: - Enter cpu_idle with preempt disabled. When the need_resched() condition becomes true, explicitly call schedule(). This makes things a bit clearer (IMO), but haven't updated all architectures yet. - Many do a test and clear of TIF_NEED_RESCHED for some reason. According to the resched_task rules, this isn't needed (and actually breaks the assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock held). So remove that. Generally one less locked memory op when switching to the idle thread. - Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner most polling idle loops. The above resched_task semantics allow it to be set until before the last time need_resched() is checked before going into a halt requiring interrupt wakeup. Many idle routines simply never enter such a halt, and so POLLING_NRFLAG can be always left set, completely eliminating resched IPIs when rescheduling the idle task. POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs. Signed-off-by: Nick Piggin <npiggin@suse.de> Cc: Ingo Molnar <mingo@elte.hu> Cc: Con Kolivas <kernel@kolivas.org> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-08 22:39:04 -07:00
acpi_safe_halt();
/*
* TBD: Can't get time duration while in C1, as resumes
* go to an ISR rather than here. Need to instrument
* base interrupt handler.
*/
sleep_ticks = 0xFFFFFFFF;
break;
case ACPI_STATE_C2:
/* Get start time (ticks) */
t1 = inl(acpi_gbl_FADT.xpm_timer_block.address);
/* Invoke C2 */
acpi_state_timer_broadcast(pr, cx, 1);
acpi_cstate_enter(cx);
/* Get end time (ticks) */
t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
#ifdef CONFIG_GENERIC_TIME
/* TSC halts in C2, so notify users */
mark_tsc_unstable();
#endif
/* Re-enable interrupts */
local_irq_enable();
current_thread_info()->status |= TS_POLLING;
/* Compute time (ticks) that we were actually asleep */
sleep_ticks =
ticks_elapsed(t1, t2) - cx->latency_ticks - C2_OVERHEAD;
acpi_state_timer_broadcast(pr, cx, 0);
break;
case ACPI_STATE_C3:
if (pr->flags.bm_check) {
if (atomic_inc_return(&c3_cpu_count) ==
num_online_cpus()) {
/*
* All CPUs are trying to go to C3
* Disable bus master arbitration
*/
acpi_set_register(ACPI_BITREG_ARB_DISABLE, 1);
}
} else {
/* SMP with no shared cache... Invalidate cache */
ACPI_FLUSH_CPU_CACHE();
}
/* Get start time (ticks) */
t1 = inl(acpi_gbl_FADT.xpm_timer_block.address);
/* Invoke C3 */
acpi_state_timer_broadcast(pr, cx, 1);
acpi_cstate_enter(cx);
/* Get end time (ticks) */
t2 = inl(acpi_gbl_FADT.xpm_timer_block.address);
if (pr->flags.bm_check) {
/* Enable bus master arbitration */
atomic_dec(&c3_cpu_count);
acpi_set_register(ACPI_BITREG_ARB_DISABLE, 0);
}
#ifdef CONFIG_GENERIC_TIME
/* TSC halts in C3, so notify users */
mark_tsc_unstable();
#endif
/* Re-enable interrupts */
local_irq_enable();
current_thread_info()->status |= TS_POLLING;
/* Compute time (ticks) that we were actually asleep */
sleep_ticks =
ticks_elapsed(t1, t2) - cx->latency_ticks - C3_OVERHEAD;
acpi_state_timer_broadcast(pr, cx, 0);
break;
default:
local_irq_enable();
return;
}
cx->usage++;
if ((cx->type != ACPI_STATE_C1) && (sleep_ticks > 0))
cx->time += sleep_ticks;
next_state = pr->power.state;
#ifdef CONFIG_HOTPLUG_CPU
/* Don't do promotion/demotion */
if ((cx->type == ACPI_STATE_C1) && (num_online_cpus() > 1) &&
!pr->flags.has_cst && !(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED)) {
next_state = cx;
goto end;
}
#endif
/*
* Promotion?
* ----------
* Track the number of longs (time asleep is greater than threshold)
* and promote when the count threshold is reached. Note that bus
* mastering activity may prevent promotions.
* Do not promote above max_cstate.
*/
if (cx->promotion.state &&
((cx->promotion.state - pr->power.states) <= max_cstate)) {
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
if (sleep_ticks > cx->promotion.threshold.ticks &&
cx->promotion.state->latency <= system_latency_constraint()) {
cx->promotion.count++;
cx->demotion.count = 0;
if (cx->promotion.count >=
cx->promotion.threshold.count) {
if (pr->flags.bm_check) {
if (!
(pr->power.bm_activity & cx->
promotion.threshold.bm)) {
next_state =
cx->promotion.state;
goto end;
}
} else {
next_state = cx->promotion.state;
goto end;
}
}
}
}
/*
* Demotion?
* ---------
* Track the number of shorts (time asleep is less than time threshold)
* and demote when the usage threshold is reached.
*/
if (cx->demotion.state) {
if (sleep_ticks < cx->demotion.threshold.ticks) {
cx->demotion.count++;
cx->promotion.count = 0;
if (cx->demotion.count >= cx->demotion.threshold.count) {
next_state = cx->demotion.state;
goto end;
}
}
}
end:
/*
* Demote if current state exceeds max_cstate
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
* or if the latency of the current state is unacceptable
*/
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
if ((pr->power.state - pr->power.states) > max_cstate ||
pr->power.state->latency > system_latency_constraint()) {
if (cx->demotion.state)
next_state = cx->demotion.state;
}
/*
* New Cx State?
* -------------
* If we're going to start using a new Cx state we must clean up
* from the previous and prepare to use the new.
*/
if (next_state != pr->power.state)
acpi_processor_power_activate(pr, next_state);
}
static int acpi_processor_set_power_policy(struct acpi_processor *pr)
{
unsigned int i;
unsigned int state_is_set = 0;
struct acpi_processor_cx *lower = NULL;
struct acpi_processor_cx *higher = NULL;
struct acpi_processor_cx *cx;
if (!pr)
return -EINVAL;
/*
* This function sets the default Cx state policy (OS idle handler).
* Our scheme is to promote quickly to C2 but more conservatively
* to C3. We're favoring C2 for its characteristics of low latency
* (quick response), good power savings, and ability to allow bus
* mastering activity. Note that the Cx state policy is completely
* customizable and can be altered dynamically.
*/
/* startup state */
for (i = 1; i < ACPI_PROCESSOR_MAX_POWER; i++) {
cx = &pr->power.states[i];
if (!cx->valid)
continue;
if (!state_is_set)
pr->power.state = cx;
state_is_set++;
break;
}
if (!state_is_set)
return -ENODEV;
/* demotion */
for (i = 1; i < ACPI_PROCESSOR_MAX_POWER; i++) {
cx = &pr->power.states[i];
if (!cx->valid)
continue;
if (lower) {
cx->demotion.state = lower;
cx->demotion.threshold.ticks = cx->latency_ticks;
cx->demotion.threshold.count = 1;
if (cx->type == ACPI_STATE_C3)
cx->demotion.threshold.bm = bm_history;
}
lower = cx;
}
/* promotion */
for (i = (ACPI_PROCESSOR_MAX_POWER - 1); i > 0; i--) {
cx = &pr->power.states[i];
if (!cx->valid)
continue;
if (higher) {
cx->promotion.state = higher;
cx->promotion.threshold.ticks = cx->latency_ticks;
if (cx->type >= ACPI_STATE_C2)
cx->promotion.threshold.count = 4;
else
cx->promotion.threshold.count = 10;
if (higher->type == ACPI_STATE_C3)
cx->promotion.threshold.bm = bm_history;
}
higher = cx;
}
return 0;
}
static int acpi_processor_get_power_info_fadt(struct acpi_processor *pr)
{
if (!pr)
return -EINVAL;
if (!pr->pblk)
return -ENODEV;
/* if info is obtained from pblk/fadt, type equals state */
pr->power.states[ACPI_STATE_C2].type = ACPI_STATE_C2;
pr->power.states[ACPI_STATE_C3].type = ACPI_STATE_C3;
#ifndef CONFIG_HOTPLUG_CPU
/*
* Check for P_LVL2_UP flag before entering C2 and above on
* an SMP system.
*/
if ((num_online_cpus() > 1) &&
!(acpi_gbl_FADT.flags & ACPI_FADT_C2_MP_SUPPORTED))
return -ENODEV;
#endif
/* determine C2 and C3 address from pblk */
pr->power.states[ACPI_STATE_C2].address = pr->pblk + 4;
pr->power.states[ACPI_STATE_C3].address = pr->pblk + 5;
/* determine latencies from FADT */
pr->power.states[ACPI_STATE_C2].latency = acpi_gbl_FADT.C2latency;
pr->power.states[ACPI_STATE_C3].latency = acpi_gbl_FADT.C3latency;
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"lvl2[0x%08x] lvl3[0x%08x]\n",
pr->power.states[ACPI_STATE_C2].address,
pr->power.states[ACPI_STATE_C3].address));
return 0;
}
static int acpi_processor_get_power_info_default(struct acpi_processor *pr)
{
if (!pr->power.states[ACPI_STATE_C1].valid) {
/* set the first C-State to C1 */
/* all processors need to support C1 */
pr->power.states[ACPI_STATE_C1].type = ACPI_STATE_C1;
pr->power.states[ACPI_STATE_C1].valid = 1;
}
/* the C0 state only exists as a filler in our array */
pr->power.states[ACPI_STATE_C0].valid = 1;
return 0;
}
static int acpi_processor_get_power_info_cst(struct acpi_processor *pr)
{
acpi_status status = 0;
acpi_integer count;
int current_count;
int i;
struct acpi_buffer buffer = { ACPI_ALLOCATE_BUFFER, NULL };
union acpi_object *cst;
if (nocst)
return -ENODEV;
current_count = 0;
status = acpi_evaluate_object(pr->handle, "_CST", NULL, &buffer);
if (ACPI_FAILURE(status)) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO, "No _CST, giving up\n"));
return -ENODEV;
}
cst = buffer.pointer;
/* There must be at least 2 elements */
if (!cst || (cst->type != ACPI_TYPE_PACKAGE) || cst->package.count < 2) {
printk(KERN_ERR PREFIX "not enough elements in _CST\n");
status = -EFAULT;
goto end;
}
count = cst->package.elements[0].integer.value;
/* Validate number of power states. */
if (count < 1 || count != cst->package.count - 1) {
printk(KERN_ERR PREFIX "count given by _CST is not valid\n");
status = -EFAULT;
goto end;
}
/* Tell driver that at least _CST is supported. */
pr->flags.has_cst = 1;
for (i = 1; i <= count; i++) {
union acpi_object *element;
union acpi_object *obj;
struct acpi_power_register *reg;
struct acpi_processor_cx cx;
memset(&cx, 0, sizeof(cx));
element = &(cst->package.elements[i]);
if (element->type != ACPI_TYPE_PACKAGE)
continue;
if (element->package.count != 4)
continue;
obj = &(element->package.elements[0]);
if (obj->type != ACPI_TYPE_BUFFER)
continue;
reg = (struct acpi_power_register *)obj->buffer.pointer;
if (reg->space_id != ACPI_ADR_SPACE_SYSTEM_IO &&
(reg->space_id != ACPI_ADR_SPACE_FIXED_HARDWARE))
continue;
/* There should be an easy way to extract an integer... */
obj = &(element->package.elements[1]);
if (obj->type != ACPI_TYPE_INTEGER)
continue;
cx.type = obj->integer.value;
/*
* Some buggy BIOSes won't list C1 in _CST -
* Let acpi_processor_get_power_info_default() handle them later
*/
if (i == 1 && cx.type != ACPI_STATE_C1)
current_count++;
cx.address = reg->address;
cx.index = current_count + 1;
cx.space_id = ACPI_CSTATE_SYSTEMIO;
if (reg->space_id == ACPI_ADR_SPACE_FIXED_HARDWARE) {
if (acpi_processor_ffh_cstate_probe
(pr->id, &cx, reg) == 0) {
cx.space_id = ACPI_CSTATE_FFH;
} else if (cx.type != ACPI_STATE_C1) {
/*
* C1 is a special case where FIXED_HARDWARE
* can be handled in non-MWAIT way as well.
* In that case, save this _CST entry info.
* That is, we retain space_id of SYSTEM_IO for
* halt based C1.
* Otherwise, ignore this info and continue.
*/
continue;
}
}
obj = &(element->package.elements[2]);
if (obj->type != ACPI_TYPE_INTEGER)
continue;
cx.latency = obj->integer.value;
obj = &(element->package.elements[3]);
if (obj->type != ACPI_TYPE_INTEGER)
continue;
cx.power = obj->integer.value;
current_count++;
memcpy(&(pr->power.states[current_count]), &cx, sizeof(cx));
/*
* We support total ACPI_PROCESSOR_MAX_POWER - 1
* (From 1 through ACPI_PROCESSOR_MAX_POWER - 1)
*/
if (current_count >= (ACPI_PROCESSOR_MAX_POWER - 1)) {
printk(KERN_WARNING
"Limiting number of power states to max (%d)\n",
ACPI_PROCESSOR_MAX_POWER);
printk(KERN_WARNING
"Please increase ACPI_PROCESSOR_MAX_POWER if needed.\n");
break;
}
}
ACPI_DEBUG_PRINT((ACPI_DB_INFO, "Found %d power states\n",
current_count));
/* Validate number of power states discovered */
if (current_count < 2)
status = -EFAULT;
end:
kfree(buffer.pointer);
return status;
}
static void acpi_processor_power_verify_c2(struct acpi_processor_cx *cx)
{
if (!cx->address)
return;
/*
* C2 latency must be less than or equal to 100
* microseconds.
*/
else if (cx->latency > ACPI_PROCESSOR_MAX_C2_LATENCY) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"latency too large [%d]\n", cx->latency));
return;
}
/*
* Otherwise we've met all of our C2 requirements.
* Normalize the C2 latency to expidite policy
*/
cx->valid = 1;
cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency);
return;
}
static void acpi_processor_power_verify_c3(struct acpi_processor *pr,
struct acpi_processor_cx *cx)
{
static int bm_check_flag;
if (!cx->address)
return;
/*
* C3 latency must be less than or equal to 1000
* microseconds.
*/
else if (cx->latency > ACPI_PROCESSOR_MAX_C3_LATENCY) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"latency too large [%d]\n", cx->latency));
return;
}
/*
* PIIX4 Erratum #18: We don't support C3 when Type-F (fast)
* DMA transfers are used by any ISA device to avoid livelock.
* Note that we could disable Type-F DMA (as recommended by
* the erratum), but this is known to disrupt certain ISA
* devices thus we take the conservative approach.
*/
else if (errata.piix4.fdma) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"C3 not supported on PIIX4 with Type-F DMA\n"));
return;
}
/* All the logic here assumes flags.bm_check is same across all CPUs */
if (!bm_check_flag) {
/* Determine whether bm_check is needed based on CPU */
acpi_processor_power_init_bm_check(&(pr->flags), pr->id);
bm_check_flag = pr->flags.bm_check;
} else {
pr->flags.bm_check = bm_check_flag;
}
if (pr->flags.bm_check) {
/* bus mastering control is necessary */
if (!pr->flags.bm_control) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"C3 support requires bus mastering control\n"));
return;
}
} else {
/*
* WBINVD should be set in fadt, for C3 state to be
* supported on when bm_check is not required.
*/
if (!(acpi_gbl_FADT.flags & ACPI_FADT_WBINVD)) {
ACPI_DEBUG_PRINT((ACPI_DB_INFO,
"Cache invalidation should work properly"
" for C3 to be enabled on SMP systems\n"));
return;
}
acpi_set_register(ACPI_BITREG_BUS_MASTER_RLD, 0);
}
/*
* Otherwise we've met all of our C3 requirements.
* Normalize the C3 latency to expidite policy. Enable
* checking of bus mastering status (bm_check) so we can
* use this in our C3 policy
*/
cx->valid = 1;
cx->latency_ticks = US_TO_PM_TIMER_TICKS(cx->latency);
return;
}
static int acpi_processor_power_verify(struct acpi_processor *pr)
{
unsigned int i;
unsigned int working = 0;
pr->power.timer_broadcast_on_state = INT_MAX;
for (i = 1; i < ACPI_PROCESSOR_MAX_POWER; i++) {
struct acpi_processor_cx *cx = &pr->power.states[i];
switch (cx->type) {
case ACPI_STATE_C1:
cx->valid = 1;
break;
case ACPI_STATE_C2:
acpi_processor_power_verify_c2(cx);
if (cx->valid)
acpi_timer_check_state(i, pr, cx);
break;
case ACPI_STATE_C3:
acpi_processor_power_verify_c3(pr, cx);
if (cx->valid)
acpi_timer_check_state(i, pr, cx);
break;
}
if (cx->valid)
working++;
}
acpi_propagate_timer_broadcast(pr);
return (working);
}
static int acpi_processor_get_power_info(struct acpi_processor *pr)
{
unsigned int i;
int result;
/* NOTE: the idle thread may not be running while calling
* this function */
/* Zero initialize all the C-states info. */
memset(pr->power.states, 0, sizeof(pr->power.states));
result = acpi_processor_get_power_info_cst(pr);
if (result == -ENODEV)
result = acpi_processor_get_power_info_fadt(pr);
if (result)
return result;
acpi_processor_get_power_info_default(pr);
pr->power.count = acpi_processor_power_verify(pr);
/*
* Set Default Policy
* ------------------
* Now that we know which states are supported, set the default
* policy. Note that this policy can be changed dynamically
* (e.g. encourage deeper sleeps to conserve battery life when
* not on AC).
*/
result = acpi_processor_set_power_policy(pr);
if (result)
return result;
/*
* if one state of type C2 or C3 is available, mark this
* CPU as being "idle manageable"
*/
for (i = 1; i < ACPI_PROCESSOR_MAX_POWER; i++) {
if (pr->power.states[i].valid) {
pr->power.count = i;
if (pr->power.states[i].type >= ACPI_STATE_C2)
pr->flags.power = 1;
}
}
return 0;
}
int acpi_processor_cst_has_changed(struct acpi_processor *pr)
{
int result = 0;
if (!pr)
return -EINVAL;
if (nocst) {
return -ENODEV;
}
if (!pr->flags.power_setup_done)
return -ENODEV;
/* Fall back to the default idle loop */
pm_idle = pm_idle_save;
synchronize_sched(); /* Relies on interrupts forcing exit from idle. */
pr->flags.power = 0;
result = acpi_processor_get_power_info(pr);
if ((pr->flags.power == 1) && (pr->flags.power_setup_done))
pm_idle = acpi_processor_idle;
return result;
}
/* proc interface */
static int acpi_processor_power_seq_show(struct seq_file *seq, void *offset)
{
struct acpi_processor *pr = seq->private;
unsigned int i;
if (!pr)
goto end;
seq_printf(seq, "active state: C%zd\n"
"max_cstate: C%d\n"
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
"bus master activity: %08x\n"
"maximum allowed latency: %d usec\n",
pr->power.state ? pr->power.state - pr->power.states : 0,
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
max_cstate, (unsigned)pr->power.bm_activity,
system_latency_constraint());
seq_puts(seq, "states:\n");
for (i = 1; i <= pr->power.count; i++) {
seq_printf(seq, " %cC%d: ",
(&pr->power.states[i] ==
pr->power.state ? '*' : ' '), i);
if (!pr->power.states[i].valid) {
seq_puts(seq, "<not supported>\n");
continue;
}
switch (pr->power.states[i].type) {
case ACPI_STATE_C1:
seq_printf(seq, "type[C1] ");
break;
case ACPI_STATE_C2:
seq_printf(seq, "type[C2] ");
break;
case ACPI_STATE_C3:
seq_printf(seq, "type[C3] ");
break;
default:
seq_printf(seq, "type[--] ");
break;
}
if (pr->power.states[i].promotion.state)
seq_printf(seq, "promotion[C%zd] ",
(pr->power.states[i].promotion.state -
pr->power.states));
else
seq_puts(seq, "promotion[--] ");
if (pr->power.states[i].demotion.state)
seq_printf(seq, "demotion[C%zd] ",
(pr->power.states[i].demotion.state -
pr->power.states));
else
seq_puts(seq, "demotion[--] ");
seq_printf(seq, "latency[%03d] usage[%08d] duration[%020llu]\n",
pr->power.states[i].latency,
pr->power.states[i].usage,
(unsigned long long)pr->power.states[i].time);
}
end:
return 0;
}
static int acpi_processor_power_open_fs(struct inode *inode, struct file *file)
{
return single_open(file, acpi_processor_power_seq_show,
PDE(inode)->data);
}
static const struct file_operations acpi_processor_power_fops = {
.open = acpi_processor_power_open_fs,
.read = seq_read,
.llseek = seq_lseek,
.release = single_release,
};
#ifdef CONFIG_SMP
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
static void smp_callback(void *v)
{
/* we already woke the CPU up, nothing more to do */
}
/*
* This function gets called when a part of the kernel has a new latency
* requirement. This means we need to get all processors out of their C-state,
* and then recalculate a new suitable C-state. Just do a cross-cpu IPI; that
* wakes them all right up.
*/
static int acpi_processor_latency_notify(struct notifier_block *b,
unsigned long l, void *v)
{
smp_call_function(smp_callback, NULL, 0, 1);
return NOTIFY_OK;
}
static struct notifier_block acpi_processor_latency_notifier = {
.notifier_call = acpi_processor_latency_notify,
};
#endif
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
int __cpuinit acpi_processor_power_init(struct acpi_processor *pr,
struct acpi_device *device)
{
acpi_status status = 0;
static int first_run;
struct proc_dir_entry *entry = NULL;
unsigned int i;
if (!first_run) {
dmi_check_system(processor_power_dmi_table);
if (max_cstate < ACPI_C_STATES_MAX)
printk(KERN_NOTICE
"ACPI: processor limited to max C-state %d\n",
max_cstate);
first_run++;
#ifdef CONFIG_SMP
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
register_latency_notifier(&acpi_processor_latency_notifier);
#endif
}
if (!pr)
return -EINVAL;
if (acpi_gbl_FADT.cst_control && !nocst) {
status =
acpi_os_write_port(acpi_gbl_FADT.smi_command, acpi_gbl_FADT.cst_control, 8);
if (ACPI_FAILURE(status)) {
ACPI_EXCEPTION((AE_INFO, status,
"Notifying BIOS of _CST ability failed"));
}
}
acpi_processor_get_power_info(pr);
/*
* Install the idle handler if processor power management is supported.
* Note that we use previously set idle handler will be used on
* platforms that only support C1.
*/
if ((pr->flags.power) && (!boot_option_idle_override)) {
printk(KERN_INFO PREFIX "CPU%d (power states:", pr->id);
for (i = 1; i <= pr->power.count; i++)
if (pr->power.states[i].valid)
printk(" C%d[C%d]", i,
pr->power.states[i].type);
printk(")\n");
if (pr->id == 0) {
pm_idle_save = pm_idle;
pm_idle = acpi_processor_idle;
}
}
/* 'power' [R] */
entry = create_proc_entry(ACPI_PROCESSOR_FILE_POWER,
S_IRUGO, acpi_device_dir(device));
if (!entry)
return -EIO;
else {
entry->proc_fops = &acpi_processor_power_fops;
entry->data = acpi_driver_data(device);
entry->owner = THIS_MODULE;
}
pr->flags.power_setup_done = 1;
return 0;
}
int acpi_processor_power_exit(struct acpi_processor *pr,
struct acpi_device *device)
{
pr->flags.power_setup_done = 0;
if (acpi_device_dir(device))
remove_proc_entry(ACPI_PROCESSOR_FILE_POWER,
acpi_device_dir(device));
/* Unregister the idle handler when processor #0 is removed. */
if (pr->id == 0) {
pm_idle = pm_idle_save;
/*
* We are about to unload the current idle thread pm callback
* (pm_idle), Wait for all processors to update cached/local
* copies of pm_idle before proceeding.
*/
cpu_idle_wait();
#ifdef CONFIG_SMP
[PATCH] maximum latency tracking infrastructure Add infrastructure to track "maximum allowable latency" for power saving policies. The reason for adding this infrastructure is that power management in the idle loop needs to make a tradeoff between latency and power savings (deeper power save modes have a longer latency to running code again). The code that today makes this tradeoff just does a rather simple algorithm; however this is not good enough: There are devices and use cases where a lower latency is required than that the higher power saving states provide. An example would be audio playback, but another example is the ipw2100 wireless driver that right now has a very direct and ugly acpi hook to disable some higher power states randomly when it gets certain types of error. The proposed solution is to have an interface where drivers can * announce the maximum latency (in microseconds) that they can deal with * modify this latency * give up their constraint and a function where the code that decides on power saving strategy can query the current global desired maximum. This patch has a user of each side: on the consumer side, ACPI is patched to use this, on the producer side the ipw2100 driver is patched. A generic maximum latency is also registered of 2 timer ticks (more and you lose accurate time tracking after all). While the existing users of the patch are x86 specific, the infrastructure is not. I'd like to ask the arch maintainers of other architectures if the infrastructure is generic enough for their use (assuming the architecture has such a tradeoff as concept at all), and the sound/multimedia driver owners to look at the driver facing API to see if this is something they can use. [akpm@osdl.org: cleanups] Signed-off-by: Arjan van de Ven <arjan@linux.intel.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Acked-by: Jesse Barnes <jesse.barnes@intel.com> Cc: "Brown, Len" <len.brown@intel.com> Signed-off-by: Andrew Morton <akpm@osdl.org> Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2006-09-30 23:27:17 -07:00
unregister_latency_notifier(&acpi_processor_latency_notifier);
#endif
}
return 0;
}