/**
* default_idle_call - Default CPU idle routine.
*
* To use when the cpuidle framework cannot be used.
*/
void default_idle_call(void)
{
if (current_clr_polling_and_test()) {
local_irq_enable();
} else {
stop_critical_timings();
arch_cpu_idle();
start_critical_timings();
}
}
/*
* Generic idle loop implementation
*/
static void cpu_idle_loop(void)
{
while (1) {
tick_nohz_idle_enter();
while (!need_resched()) {
check_pgt_cache();
rmb();
if (cpu_is_offline(smp_processor_id()))
arch_cpu_idle_dead();
local_irq_disable();
arch_cpu_idle_enter();
/*
* In poll mode we reenable interrupts and spin.
*
* Also if we detected in the wakeup from idle
* path that the tick broadcast device expired
* for us, we don't want to go deep idle as we
* know that the IPI is going to arrive right
* away
*/
if (cpu_idle_force_poll || tick_check_broadcast_expired()) {
cpu_idle_poll();
} else {
if (!current_clr_polling_and_test()) {
stop_critical_timings();
rcu_idle_enter();
arch_cpu_idle();
WARN_ON_ONCE(irqs_disabled());
rcu_idle_exit();
start_critical_timings();
} else {
local_irq_enable();
}
__current_set_polling();
}
arch_cpu_idle_exit();
}
/*
* Since we fell out of the loop above, we know
* TIF_NEED_RESCHED must be set, propagate it into
* PREEMPT_NEED_RESCHED.
*
* This is required because for polling idle loops we will
* not have had an IPI to fold the state for us.
*/
preempt_set_need_resched();
tick_nohz_idle_exit();
schedule_preempt_disabled();
}
}
static void default_idle_call(void)
{
/*
* We can't use the cpuidle framework, let's use the default idle
* routine.
*/
if (current_clr_polling_and_test())
local_irq_enable();
else
arch_cpu_idle();
}
/*
* Generic idle loop implementation
*/
static void cpu_idle_loop(void)
{
while (1) {
tick_nohz_idle_enter();
while (!need_resched()) {
check_pgt_cache();
rmb();
local_irq_disable();
arch_cpu_idle_enter();
/*
* In poll mode we reenable interrupts and spin.
*
* Also if we detected in the wakeup from idle
* path that the tick broadcast device expired
* for us, we don't want to go deep idle as we
* know that the IPI is going to arrive right
* away
*/
if (cpu_idle_force_poll ||
tick_check_broadcast_expired() ||
__get_cpu_var(idle_force_poll)) {
cpu_idle_poll();
} else {
if (!current_clr_polling_and_test()) {
stop_critical_timings();
rcu_idle_enter();
arch_cpu_idle();
WARN_ON_ONCE(irqs_disabled());
rcu_idle_exit();
start_critical_timings();
} else {
local_irq_enable();
}
__current_set_polling();
}
arch_cpu_idle_exit();
}
tick_nohz_idle_exit();
schedule_preempt_disabled();
if (cpu_is_offline(smp_processor_id()))
arch_cpu_idle_dead();
}
}
/**
* cpuidle_idle_call - the main idle function
*
* NOTE: no locks or semaphores should be used here
*/
static void cpuidle_idle_call(void)
{
struct cpuidle_device *dev = __this_cpu_read(cpuidle_devices);
struct cpuidle_driver *drv = cpuidle_get_cpu_driver(dev);
int next_state, entered_state;
bool broadcast;
/*
* Check if the idle task must be rescheduled. If it is the
* case, exit the function after re-enabling the local irq.
*/
if (need_resched()) {
local_irq_enable();
return;
}
/*
* During the idle period, stop measuring the disabled irqs
* critical sections latencies
*/
stop_critical_timings();
/*
* Tell the RCU framework we are entering an idle section,
* so no more rcu read side critical sections and one more
* step to the grace period
*/
rcu_idle_enter();
/*
* Ask the cpuidle framework to choose a convenient idle state.
* Fall back to the default arch idle method on errors.
*/
next_state = cpuidle_select(drv, dev);
if (next_state < 0) {
use_default:
/*
* We can't use the cpuidle framework, let's use the default
* idle routine.
*/
if (current_clr_polling_and_test())
local_irq_enable();
else
arch_cpu_idle();
goto exit_idle;
}
/*
* The idle task must be scheduled, it is pointless to
* go to idle, just update no idle residency and get
* out of this function
*/
if (current_clr_polling_and_test()) {
dev->last_residency = 0;
entered_state = next_state;
local_irq_enable();
goto exit_idle;
}
broadcast = !!(drv->states[next_state].flags & CPUIDLE_FLAG_TIMER_STOP);
/*
* Tell the time framework to switch to a broadcast timer
* because our local timer will be shutdown. If a local timer
* is used from another cpu as a broadcast timer, this call may
* fail if it is not available
*/
if (broadcast &&
clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_ENTER, &dev->cpu))
goto use_default;
trace_cpu_idle_rcuidle(next_state, dev->cpu);
/*
* Enter the idle state previously returned by the governor decision.
* This function will block until an interrupt occurs and will take
* care of re-enabling the local interrupts
*/
entered_state = cpuidle_enter(drv, dev, next_state);
trace_cpu_idle_rcuidle(PWR_EVENT_EXIT, dev->cpu);
if (broadcast)
clockevents_notify(CLOCK_EVT_NOTIFY_BROADCAST_EXIT, &dev->cpu);
/*
* Give the governor an opportunity to reflect on the outcome
*/
cpuidle_reflect(dev, entered_state);
exit_idle:
__current_set_polling();
/*
* It is up to the idle functions to reenable local interrupts
*/
if (WARN_ON_ONCE(irqs_disabled()))
local_irq_enable();
rcu_idle_exit();
start_critical_timings();
}
extern "C" int
_start(kernel_args *bootKernelArgs, int currentCPU)
{
if (bootKernelArgs->kernel_args_size != sizeof(kernel_args)
|| bootKernelArgs->version != CURRENT_KERNEL_ARGS_VERSION) {
// This is something we cannot handle right now - release kernels
// should always be able to handle the kernel_args of earlier
// released kernels.
debug_early_boot_message("Version mismatch between boot loader and "
"kernel!\n");
return -1;
}
smp_set_num_cpus(bootKernelArgs->num_cpus);
// wait for all the cpus to get here
smp_cpu_rendezvous(&sCpuRendezvous, currentCPU);
// the passed in kernel args are in a non-allocated range of memory
if (currentCPU == 0)
memcpy(&sKernelArgs, bootKernelArgs, sizeof(kernel_args));
smp_cpu_rendezvous(&sCpuRendezvous2, currentCPU);
// do any pre-booting cpu config
cpu_preboot_init_percpu(&sKernelArgs, currentCPU);
thread_preboot_init_percpu(&sKernelArgs, currentCPU);
// if we're not a boot cpu, spin here until someone wakes us up
if (smp_trap_non_boot_cpus(currentCPU, &sCpuRendezvous3)) {
// init platform
arch_platform_init(&sKernelArgs);
// setup debug output
debug_init(&sKernelArgs);
set_dprintf_enabled(true);
dprintf("Welcome to kernel debugger output!\n");
dprintf("Haiku revision: %lu\n", get_haiku_revision());
// init modules
TRACE("init CPU\n");
cpu_init(&sKernelArgs);
cpu_init_percpu(&sKernelArgs, currentCPU);
TRACE("init interrupts\n");
int_init(&sKernelArgs);
TRACE("init VM\n");
vm_init(&sKernelArgs);
// Before vm_init_post_sem() is called, we have to make sure that
// the boot loader allocated region is not used anymore
boot_item_init();
debug_init_post_vm(&sKernelArgs);
low_resource_manager_init();
// now we can use the heap and create areas
arch_platform_init_post_vm(&sKernelArgs);
lock_debug_init();
TRACE("init driver_settings\n");
driver_settings_init(&sKernelArgs);
debug_init_post_settings(&sKernelArgs);
TRACE("init notification services\n");
notifications_init();
TRACE("init teams\n");
team_init(&sKernelArgs);
TRACE("init ELF loader\n");
elf_init(&sKernelArgs);
TRACE("init modules\n");
module_init(&sKernelArgs);
TRACE("init semaphores\n");
haiku_sem_init(&sKernelArgs);
TRACE("init interrupts post vm\n");
int_init_post_vm(&sKernelArgs);
cpu_init_post_vm(&sKernelArgs);
commpage_init();
TRACE("init system info\n");
system_info_init(&sKernelArgs);
TRACE("init SMP\n");
smp_init(&sKernelArgs);
TRACE("init timer\n");
timer_init(&sKernelArgs);
TRACE("init real time clock\n");
rtc_init(&sKernelArgs);
TRACE("init condition variables\n");
condition_variable_init();
// now we can create and use semaphores
TRACE("init VM semaphores\n");
vm_init_post_sem(&sKernelArgs);
TRACE("init generic syscall\n");
generic_syscall_init();
smp_init_post_generic_syscalls();
TRACE("init scheduler\n");
scheduler_init();
TRACE("init threads\n");
thread_init(&sKernelArgs);
TRACE("init kernel daemons\n");
kernel_daemon_init();
arch_platform_init_post_thread(&sKernelArgs);
TRACE("init I/O interrupts\n");
int_init_io(&sKernelArgs);
TRACE("init VM threads\n");
vm_init_post_thread(&sKernelArgs);
low_resource_manager_init_post_thread();
TRACE("init VFS\n");
vfs_init(&sKernelArgs);
#if ENABLE_SWAP_SUPPORT
TRACE("init swap support\n");
swap_init();
#endif
TRACE("init POSIX semaphores\n");
realtime_sem_init();
xsi_sem_init();
xsi_msg_init();
// Start a thread to finish initializing the rest of the system. Note,
// it won't be scheduled before calling scheduler_start() (on any CPU).
TRACE("spawning main2 thread\n");
thread_id thread = spawn_kernel_thread(&main2, "main2",
B_NORMAL_PRIORITY, NULL);
resume_thread(thread);
// We're ready to start the scheduler and enable interrupts on all CPUs.
scheduler_enable_scheduling();
// bring up the AP cpus in a lock step fashion
TRACE("waking up AP cpus\n");
sCpuRendezvous = sCpuRendezvous2 = 0;
smp_wake_up_non_boot_cpus();
smp_cpu_rendezvous(&sCpuRendezvous, 0); // wait until they're booted
// exit the kernel startup phase (mutexes, etc work from now on out)
TRACE("exiting kernel startup\n");
gKernelStartup = false;
smp_cpu_rendezvous(&sCpuRendezvous2, 0);
// release the AP cpus to go enter the scheduler
TRACE("starting scheduler on cpu 0 and enabling interrupts\n");
scheduler_start();
enable_interrupts();
} else {
// lets make sure we're in sync with the main cpu
// the boot processor has probably been sending us
// tlb sync messages all along the way, but we've
// been ignoring them
arch_cpu_global_TLB_invalidate();
// this is run for each non boot processor after they've been set loose
cpu_init_percpu(&sKernelArgs, currentCPU);
smp_per_cpu_init(&sKernelArgs, currentCPU);
// wait for all other AP cpus to get to this point
smp_cpu_rendezvous(&sCpuRendezvous, currentCPU);
smp_cpu_rendezvous(&sCpuRendezvous2, currentCPU);
// welcome to the machine
scheduler_start();
enable_interrupts();
}
#ifdef TRACE_BOOT
// We disable interrupts for this dprintf(), since otherwise dprintf()
// would acquires a mutex, which is something we must not do in an idle
// thread, or otherwise the scheduler would be seriously unhappy.
disable_interrupts();
TRACE("main: done... begin idle loop on cpu %d\n", currentCPU);
enable_interrupts();
#endif
for (;;)
arch_cpu_idle();
return 0;
}
int mips_cpuidle_wait_enter(struct cpuidle_device *dev,
struct cpuidle_driver *drv, int index)
{
arch_cpu_idle();
return index;
}
/**
* cpuidle_idle_call - the main idle function
*
* NOTE: no locks or semaphores should be used here
*
* On archs that support TIF_POLLING_NRFLAG, is called with polling
* set, and it returns with polling set. If it ever stops polling, it
* must clear the polling bit.
*/
static void cpuidle_idle_call(void)
{
struct cpuidle_device *dev = __this_cpu_read(cpuidle_devices);
struct cpuidle_driver *drv = cpuidle_get_cpu_driver(dev);
int next_state, entered_state;
unsigned int broadcast;
bool reflect;
/*
* Check if the idle task must be rescheduled. If it is the
* case, exit the function after re-enabling the local irq.
*/
if (need_resched()) {
local_irq_enable();
return;
}
/*
* During the idle period, stop measuring the disabled irqs
* critical sections latencies
*/
stop_critical_timings();
/*
* Tell the RCU framework we are entering an idle section,
* so no more rcu read side critical sections and one more
* step to the grace period
*/
rcu_idle_enter();
if (cpuidle_not_available(drv, dev))
goto use_default;
/*
* Suspend-to-idle ("freeze") is a system state in which all user space
* has been frozen, all I/O devices have been suspended and the only
* activity happens here and in iterrupts (if any). In that case bypass
* the cpuidle governor and go stratight for the deepest idle state
* available. Possibly also suspend the local tick and the entire
* timekeeping to prevent timer interrupts from kicking us out of idle
* until a proper wakeup interrupt happens.
*/
if (idle_should_freeze()) {
entered_state = cpuidle_enter_freeze(drv, dev);
if (entered_state >= 0) {
local_irq_enable();
goto exit_idle;
}
reflect = false;
next_state = cpuidle_find_deepest_state(drv, dev);
} else {
reflect = true;
/*
* Ask the cpuidle framework to choose a convenient idle state.
*/
next_state = cpuidle_select(drv, dev);
}
/* Fall back to the default arch idle method on errors. */
if (next_state < 0)
goto use_default;
/*
* The idle task must be scheduled, it is pointless to
* go to idle, just update no idle residency and get
* out of this function
*/
if (current_clr_polling_and_test()) {
dev->last_residency = 0;
entered_state = next_state;
local_irq_enable();
goto exit_idle;
}
broadcast = drv->states[next_state].flags & CPUIDLE_FLAG_TIMER_STOP;
/*
* Tell the time framework to switch to a broadcast timer
* because our local timer will be shutdown. If a local timer
* is used from another cpu as a broadcast timer, this call may
* fail if it is not available
*/
if (broadcast && tick_broadcast_enter())
goto use_default;
/* Take note of the planned idle state. */
idle_set_state(this_rq(), &drv->states[next_state]);
/*
* Enter the idle state previously returned by the governor decision.
* This function will block until an interrupt occurs and will take
* care of re-enabling the local interrupts
*/
entered_state = cpuidle_enter(drv, dev, next_state);
/* The cpu is no longer idle or about to enter idle. */
idle_set_state(this_rq(), NULL);
if (broadcast)
tick_broadcast_exit();
/*
* Give the governor an opportunity to reflect on the outcome
*/
if (reflect)
cpuidle_reflect(dev, entered_state);
exit_idle:
__current_set_polling();
/*
* It is up to the idle functions to reenable local interrupts
*/
if (WARN_ON_ONCE(irqs_disabled()))
local_irq_enable();
rcu_idle_exit();
start_critical_timings();
return;
use_default:
/*
* We can't use the cpuidle framework, let's use the default
* idle routine.
*/
if (current_clr_polling_and_test())
local_irq_enable();
else
arch_cpu_idle();
goto exit_idle;
}
