forked from Minki/linux
1aec1c558a
paravirtualize sched_clock. Signed-off-by: Isaku Yamahata <yamahata@valinux.co.jp> Signed-off-by: Tony Luck <tony.luck@intel.com>
262 lines
6.6 KiB
C
262 lines
6.6 KiB
C
/******************************************************************************
|
|
* arch/ia64/xen/time.c
|
|
*
|
|
* Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
|
|
* VA Linux Systems Japan K.K.
|
|
*
|
|
* 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/delay.h>
|
|
#include <linux/kernel_stat.h>
|
|
#include <linux/posix-timers.h>
|
|
#include <linux/irq.h>
|
|
#include <linux/clocksource.h>
|
|
|
|
#include <asm/timex.h>
|
|
|
|
#include <asm/xen/hypervisor.h>
|
|
|
|
#include <xen/interface/vcpu.h>
|
|
|
|
#include "../kernel/fsyscall_gtod_data.h"
|
|
|
|
DEFINE_PER_CPU(struct vcpu_runstate_info, runstate);
|
|
DEFINE_PER_CPU(unsigned long, processed_stolen_time);
|
|
DEFINE_PER_CPU(unsigned long, processed_blocked_time);
|
|
|
|
/* taken from i386/kernel/time-xen.c */
|
|
static void xen_init_missing_ticks_accounting(int cpu)
|
|
{
|
|
struct vcpu_register_runstate_memory_area area;
|
|
struct vcpu_runstate_info *runstate = &per_cpu(runstate, cpu);
|
|
int rc;
|
|
|
|
memset(runstate, 0, sizeof(*runstate));
|
|
|
|
area.addr.v = runstate;
|
|
rc = HYPERVISOR_vcpu_op(VCPUOP_register_runstate_memory_area, cpu,
|
|
&area);
|
|
WARN_ON(rc && rc != -ENOSYS);
|
|
|
|
per_cpu(processed_blocked_time, cpu) = runstate->time[RUNSTATE_blocked];
|
|
per_cpu(processed_stolen_time, cpu) = runstate->time[RUNSTATE_runnable]
|
|
+ runstate->time[RUNSTATE_offline];
|
|
}
|
|
|
|
/*
|
|
* Runstate accounting
|
|
*/
|
|
/* stolen from arch/x86/xen/time.c */
|
|
static void get_runstate_snapshot(struct vcpu_runstate_info *res)
|
|
{
|
|
u64 state_time;
|
|
struct vcpu_runstate_info *state;
|
|
|
|
BUG_ON(preemptible());
|
|
|
|
state = &__get_cpu_var(runstate);
|
|
|
|
/*
|
|
* The runstate info is always updated by the hypervisor on
|
|
* the current CPU, so there's no need to use anything
|
|
* stronger than a compiler barrier when fetching it.
|
|
*/
|
|
do {
|
|
state_time = state->state_entry_time;
|
|
rmb();
|
|
*res = *state;
|
|
rmb();
|
|
} while (state->state_entry_time != state_time);
|
|
}
|
|
|
|
#define NS_PER_TICK (1000000000LL/HZ)
|
|
|
|
static unsigned long
|
|
consider_steal_time(unsigned long new_itm)
|
|
{
|
|
unsigned long stolen, blocked;
|
|
unsigned long delta_itm = 0, stolentick = 0;
|
|
int cpu = smp_processor_id();
|
|
struct vcpu_runstate_info runstate;
|
|
struct task_struct *p = current;
|
|
|
|
get_runstate_snapshot(&runstate);
|
|
|
|
/*
|
|
* Check for vcpu migration effect
|
|
* In this case, itc value is reversed.
|
|
* This causes huge stolen value.
|
|
* This function just checks and reject this effect.
|
|
*/
|
|
if (!time_after_eq(runstate.time[RUNSTATE_blocked],
|
|
per_cpu(processed_blocked_time, cpu)))
|
|
blocked = 0;
|
|
|
|
if (!time_after_eq(runstate.time[RUNSTATE_runnable] +
|
|
runstate.time[RUNSTATE_offline],
|
|
per_cpu(processed_stolen_time, cpu)))
|
|
stolen = 0;
|
|
|
|
if (!time_after(delta_itm + new_itm, ia64_get_itc()))
|
|
stolentick = ia64_get_itc() - new_itm;
|
|
|
|
do_div(stolentick, NS_PER_TICK);
|
|
stolentick++;
|
|
|
|
do_div(stolen, NS_PER_TICK);
|
|
|
|
if (stolen > stolentick)
|
|
stolen = stolentick;
|
|
|
|
stolentick -= stolen;
|
|
do_div(blocked, NS_PER_TICK);
|
|
|
|
if (blocked > stolentick)
|
|
blocked = stolentick;
|
|
|
|
if (stolen > 0 || blocked > 0) {
|
|
account_steal_ticks(stolen);
|
|
account_idle_ticks(blocked);
|
|
run_local_timers();
|
|
|
|
if (rcu_pending(cpu))
|
|
rcu_check_callbacks(cpu, user_mode(get_irq_regs()));
|
|
|
|
scheduler_tick();
|
|
run_posix_cpu_timers(p);
|
|
delta_itm += local_cpu_data->itm_delta * (stolen + blocked);
|
|
|
|
if (cpu == time_keeper_id) {
|
|
write_seqlock(&xtime_lock);
|
|
do_timer(stolen + blocked);
|
|
local_cpu_data->itm_next = delta_itm + new_itm;
|
|
write_sequnlock(&xtime_lock);
|
|
} else {
|
|
local_cpu_data->itm_next = delta_itm + new_itm;
|
|
}
|
|
per_cpu(processed_stolen_time, cpu) += NS_PER_TICK * stolen;
|
|
per_cpu(processed_blocked_time, cpu) += NS_PER_TICK * blocked;
|
|
}
|
|
return delta_itm;
|
|
}
|
|
|
|
static int xen_do_steal_accounting(unsigned long *new_itm)
|
|
{
|
|
unsigned long delta_itm;
|
|
delta_itm = consider_steal_time(*new_itm);
|
|
*new_itm += delta_itm;
|
|
if (time_after(*new_itm, ia64_get_itc()) && delta_itm)
|
|
return 1;
|
|
|
|
return 0;
|
|
}
|
|
|
|
static void xen_itc_jitter_data_reset(void)
|
|
{
|
|
u64 lcycle, ret;
|
|
|
|
do {
|
|
lcycle = itc_jitter_data.itc_lastcycle;
|
|
ret = cmpxchg(&itc_jitter_data.itc_lastcycle, lcycle, 0);
|
|
} while (unlikely(ret != lcycle));
|
|
}
|
|
|
|
/* based on xen_sched_clock() in arch/x86/xen/time.c. */
|
|
/*
|
|
* This relies on HAVE_UNSTABLE_SCHED_CLOCK. If it can't be defined,
|
|
* something similar logic should be implemented here.
|
|
*/
|
|
/*
|
|
* Xen sched_clock implementation. Returns the number of unstolen
|
|
* nanoseconds, which is nanoseconds the VCPU spent in RUNNING+BLOCKED
|
|
* states.
|
|
*/
|
|
static unsigned long long xen_sched_clock(void)
|
|
{
|
|
struct vcpu_runstate_info runstate;
|
|
|
|
unsigned long long now;
|
|
unsigned long long offset;
|
|
unsigned long long ret;
|
|
|
|
/*
|
|
* Ideally sched_clock should be called on a per-cpu basis
|
|
* anyway, so preempt should already be disabled, but that's
|
|
* not current practice at the moment.
|
|
*/
|
|
preempt_disable();
|
|
|
|
/*
|
|
* both ia64_native_sched_clock() and xen's runstate are
|
|
* based on mAR.ITC. So difference of them makes sense.
|
|
*/
|
|
now = ia64_native_sched_clock();
|
|
|
|
get_runstate_snapshot(&runstate);
|
|
|
|
WARN_ON(runstate.state != RUNSTATE_running);
|
|
|
|
offset = 0;
|
|
if (now > runstate.state_entry_time)
|
|
offset = now - runstate.state_entry_time;
|
|
ret = runstate.time[RUNSTATE_blocked] +
|
|
runstate.time[RUNSTATE_running] +
|
|
offset;
|
|
|
|
preempt_enable();
|
|
|
|
return ret;
|
|
}
|
|
|
|
struct pv_time_ops xen_time_ops __initdata = {
|
|
.init_missing_ticks_accounting = xen_init_missing_ticks_accounting,
|
|
.do_steal_accounting = xen_do_steal_accounting,
|
|
.clocksource_resume = xen_itc_jitter_data_reset,
|
|
.sched_clock = xen_sched_clock,
|
|
};
|
|
|
|
/* Called after suspend, to resume time. */
|
|
static void xen_local_tick_resume(void)
|
|
{
|
|
/* Just trigger a tick. */
|
|
ia64_cpu_local_tick();
|
|
touch_softlockup_watchdog();
|
|
}
|
|
|
|
void
|
|
xen_timer_resume(void)
|
|
{
|
|
unsigned int cpu;
|
|
|
|
xen_local_tick_resume();
|
|
|
|
for_each_online_cpu(cpu)
|
|
xen_init_missing_ticks_accounting(cpu);
|
|
}
|
|
|
|
static void ia64_cpu_local_tick_fn(void *unused)
|
|
{
|
|
xen_local_tick_resume();
|
|
xen_init_missing_ticks_accounting(smp_processor_id());
|
|
}
|
|
|
|
void
|
|
xen_timer_resume_on_aps(void)
|
|
{
|
|
smp_call_function(&ia64_cpu_local_tick_fn, NULL, 1);
|
|
}
|