perf_event.c

#undef DEBUG

/*
 * ARM performance counter support.
 *
 * Copyright (C) 2009 picoChip Designs, Ltd., Jamie Iles
 * Copyright (C) 2010 ARM Ltd., Will Deacon <will.deacon@arm.com>
 *
 * This code is based on the sparc64 perf event code, which is in turn based
 * on the x86 code. Callchain code is based on the ARM OProfile backtrace
 * code.
 */
#define pr_fmt(fmt) "hw perfevents: " fmt

#include <linux/kernel.h>
#include <linux/platform_device.h>
#include <linux/pm_runtime.h>
#include <linux/uaccess.h>

#include <asm/irq_regs.h>
#include <asm/pmu.h>
#include <asm/stacktrace.h>

static int
armpmu_map_cache_event(const unsigned (*cache_map)
				      [PERF_COUNT_HW_CACHE_MAX]
				      [PERF_COUNT_HW_CACHE_OP_MAX]
				      [PERF_COUNT_HW_CACHE_RESULT_MAX],
		       u64 config)
{
	unsigned int cache_type, cache_op, cache_result, ret;

	cache_type = (config >>  0) & 0xff;
	if (cache_type >= PERF_COUNT_HW_CACHE_MAX)
		return -EINVAL;

	cache_op = (config >>  8) & 0xff;
	if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX)
		return -EINVAL;

	cache_result = (config >> 16) & 0xff;
	if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
		return -EINVAL;

	ret = (int)(*cache_map)[cache_type][cache_op][cache_result];

	if (ret == CACHE_OP_UNSUPPORTED)
		return -ENOENT;

	return ret;
}

static int
armpmu_map_hw_event(const unsigned (*event_map)[PERF_COUNT_HW_MAX], u64 config)
{
	int mapping = (*event_map)[config];
	return mapping == HW_OP_UNSUPPORTED ? -ENOENT : mapping;
}

static int
armpmu_map_raw_event(u32 raw_event_mask, u64 config)
{
	return (int)(config & raw_event_mask);
}

int
armpmu_map_event(struct perf_event *event,
		 const unsigned (*event_map)[PERF_COUNT_HW_MAX],
		 const unsigned (*cache_map)
				[PERF_COUNT_HW_CACHE_MAX]
				[PERF_COUNT_HW_CACHE_OP_MAX]
				[PERF_COUNT_HW_CACHE_RESULT_MAX],
		 u32 raw_event_mask)
{
	u64 config = event->attr.config;

	switch (event->attr.type) {
	case PERF_TYPE_HARDWARE:
		return armpmu_map_hw_event(event_map, config);
	case PERF_TYPE_HW_CACHE:
		return armpmu_map_cache_event(cache_map, config);
	case PERF_TYPE_RAW:
		return armpmu_map_raw_event(raw_event_mask, config);
	}

	return -ENOENT;
}

int
armpmu_event_set_period(struct perf_event *event,
			struct hw_perf_event *hwc,
			int idx)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	s64 left = local64_read(&hwc->period_left);
	s64 period = hwc->sample_period;
	int ret = 0;

	/* The period may have been changed by PERF_EVENT_IOC_PERIOD */
	if (unlikely(period != hwc->last_period))
		left = period - (hwc->last_period - left);

	if (unlikely(left <= -period)) {
		left = period;
		local64_set(&hwc->period_left, left);
		hwc->last_period = period;
		ret = 1;
	}

	if (unlikely(left <= 0)) {
		left += period;
		local64_set(&hwc->period_left, left);
		hwc->last_period = period;
		ret = 1;
	}

	if (left > (s64)armpmu->max_period)
		left = armpmu->max_period;

	local64_set(&hwc->prev_count, (u64)-left);

	armpmu->write_counter(idx, (u64)(-left) & 0xffffffff);

	perf_event_update_userpage(event);

	return ret;
}

u64
armpmu_event_update(struct perf_event *event,
		    struct hw_perf_event *hwc,
		    int idx)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	u64 delta, prev_raw_count, new_raw_count;

again:
	prev_raw_count = local64_read(&hwc->prev_count);
	new_raw_count = armpmu->read_counter(idx);

	if (local64_cmpxchg(&hwc->prev_count, prev_raw_count,
			     new_raw_count) != prev_raw_count)
		goto again;

	delta = (new_raw_count - prev_raw_count) & armpmu->max_period;

	local64_add(delta, &event->count);
	local64_sub(delta, &hwc->period_left);

	return new_raw_count;
}

static void
armpmu_read(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;

	/* Don't read disabled counters! */
	if (hwc->idx < 0)
		return;

	armpmu_event_update(event, hwc, hwc->idx);
}

static void
armpmu_stop(struct perf_event *event, int flags)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct hw_perf_event *hwc = &event->hw;

	/*
	 * ARM pmu always has to update the counter, so ignore
	 * PERF_EF_UPDATE, see comments in armpmu_start().
	 */
	if (!(hwc->state & PERF_HES_STOPPED)) {
		armpmu->disable(hwc, hwc->idx);
		armpmu_event_update(event, hwc, hwc->idx);
		hwc->state |= PERF_HES_STOPPED | PERF_HES_UPTODATE;
	}
}

static void
armpmu_start(struct perf_event *event, int flags)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct hw_perf_event *hwc = &event->hw;

	/*
	 * ARM pmu always has to reprogram the period, so ignore
	 * PERF_EF_RELOAD, see the comment below.
	 */
	if (flags & PERF_EF_RELOAD)
		WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));

	hwc->state = 0;
	/*
	 * Set the period again. Some counters can't be stopped, so when we
	 * were stopped we simply disabled the IRQ source and the counter
	 * may have been left counting. If we don't do this step then we may
	 * get an interrupt too soon or *way* too late if the overflow has
	 * happened since disabling.
	 */
	armpmu_event_set_period(event, hwc, hwc->idx);
	armpmu->enable(hwc, hwc->idx);
}

static void
armpmu_del(struct perf_event *event, int flags)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct pmu_hw_events *hw_events = armpmu->get_hw_events();
	struct hw_perf_event *hwc = &event->hw;
	int idx = hwc->idx;

	WARN_ON(idx < 0);

	armpmu_stop(event, PERF_EF_UPDATE);
	hw_events->events[idx] = NULL;
	clear_bit(idx, hw_events->used_mask);

	perf_event_update_userpage(event);
}

static int
armpmu_add(struct perf_event *event, int flags)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct pmu_hw_events *hw_events = armpmu->get_hw_events();
	struct hw_perf_event *hwc = &event->hw;
	int idx;
	int err = 0;

	perf_pmu_disable(event->pmu);

	/* If we don't have a space for the counter then finish early. */
	idx = armpmu->get_event_idx(hw_events, hwc);
	if (idx < 0) {
		err = idx;
		goto out;
	}

	/*
	 * If there is an event in the counter we are going to use then make
	 * sure it is disabled.
	 */
	event->hw.idx = idx;
	armpmu->disable(hwc, idx);
	hw_events->events[idx] = event;

	hwc->state = PERF_HES_STOPPED | PERF_HES_UPTODATE;
	if (flags & PERF_EF_START)
		armpmu_start(event, PERF_EF_RELOAD);

	/* Propagate our changes to the userspace mapping. */
	perf_event_update_userpage(event);

out:
	perf_pmu_enable(event->pmu);
	return err;
}

static int
validate_event(struct pmu_hw_events *hw_events,
	       struct perf_event *event)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct hw_perf_event fake_event = event->hw;
	struct pmu *leader_pmu = event->group_leader->pmu;

	if (event->pmu != leader_pmu || event->state <= PERF_EVENT_STATE_OFF)
		return 1;

	return armpmu->get_event_idx(hw_events, &fake_event) >= 0;
}

static int
validate_group(struct perf_event *event)
{
	struct perf_event *sibling, *leader = event->group_leader;
	struct pmu_hw_events fake_pmu;
	DECLARE_BITMAP(fake_used_mask, ARMPMU_MAX_HWEVENTS);

	/*
	 * Initialise the fake PMU. We only need to populate the
	 * used_mask for the purposes of validation.
	 */
	memset(fake_used_mask, 0, sizeof(fake_used_mask));
	fake_pmu.used_mask = fake_used_mask;

	if (!validate_event(&fake_pmu, leader))
		return -EINVAL;

	list_for_each_entry(sibling, &leader->sibling_list, group_entry) {
		if (!validate_event(&fake_pmu, sibling))
			return -EINVAL;
	}

	if (!validate_event(&fake_pmu, event))
		return -EINVAL;

	return 0;
}

static irqreturn_t armpmu_dispatch_irq(int irq, void *dev)
{
	struct arm_pmu *armpmu = (struct arm_pmu *) dev;
	struct platform_device *plat_device = armpmu->plat_device;
	struct arm_pmu_platdata *plat = dev_get_platdata(&plat_device->dev);

	if (plat && plat->handle_irq)
		return plat->handle_irq(irq, dev, armpmu->handle_irq);
	else
		return armpmu->handle_irq(irq, dev);
}

static void
armpmu_release_hardware(struct arm_pmu *armpmu)
{
	armpmu->free_irq();
	pm_runtime_put_sync(&armpmu->plat_device->dev);
}

static int
armpmu_reserve_hardware(struct arm_pmu *armpmu)
{
	int err;
	struct platform_device *pmu_device = armpmu->plat_device;

	if (!pmu_device)
		return -ENODEV;

	pm_runtime_get_sync(&pmu_device->dev);
	err = armpmu->request_irq(armpmu_dispatch_irq);
	if (err) {
		armpmu_release_hardware(armpmu);
		return err;
	}

	return 0;
}

static void
hw_perf_event_destroy(struct perf_event *event)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	atomic_t *active_events	 = &armpmu->active_events;
	struct mutex *pmu_reserve_mutex = &armpmu->reserve_mutex;

	if (atomic_dec_and_mutex_lock(active_events, pmu_reserve_mutex)) {
		armpmu_release_hardware(armpmu);
		mutex_unlock(pmu_reserve_mutex);
	}
}

static int
event_requires_mode_exclusion(struct perf_event_attr *attr)
{
	return attr->exclude_idle || attr->exclude_user ||
	       attr->exclude_kernel || attr->exclude_hv;
}

static int
__hw_perf_event_init(struct perf_event *event)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	struct hw_perf_event *hwc = &event->hw;
	int mapping, err;

	mapping = armpmu->map_event(event);

	if (mapping < 0) {
		pr_debug("event %x:%llx not supported\n", event->attr.type,
			 event->attr.config);
		return mapping;
	}

	/*
	 * We don't assign an index until we actually place the event onto
	 * hardware. Use -1 to signify that we haven't decided where to put it
	 * yet. For SMP systems, each core has it's own PMU so we can't do any
	 * clever allocation or constraints checking at this point.
	 */
	hwc->idx		= -1;
	hwc->config_base	= 0;
	hwc->config		= 0;
	hwc->event_base		= 0;

	/*
	 * Check whether we need to exclude the counter from certain modes.
	 */
	if ((!armpmu->set_event_filter ||
	     armpmu->set_event_filter(hwc, &event->attr)) &&
	     event_requires_mode_exclusion(&event->attr)) {
		pr_debug("ARM performance counters do not support "
			 "mode exclusion\n");
		return -EOPNOTSUPP;
	}

	/*
	 * Store the event encoding into the config_base field.
	 */
	hwc->config_base	    |= (unsigned long)mapping;

	if (!hwc->sample_period) {
		/*
		 * For non-sampling runs, limit the sample_period to half
		 * of the counter width. That way, the new counter value
		 * is far less likely to overtake the previous one unless
		 * you have some serious IRQ latency issues.
		 */
		hwc->sample_period  = armpmu->max_period >> 1;
		hwc->last_period    = hwc->sample_period;
		local64_set(&hwc->period_left, hwc->sample_period);
	}

	err = 0;
	if (event->group_leader != event) {
		err = validate_group(event);
		if (err)
			return -EINVAL;
	}

	return err;
}

static int armpmu_event_init(struct perf_event *event)
{
	struct arm_pmu *armpmu = to_arm_pmu(event->pmu);
	int err = 0;
	atomic_t *active_events = &armpmu->active_events;

	/* does not support taken branch sampling */
	if (has_branch_stack(event))
		return -EOPNOTSUPP;

	if (armpmu->map_event(event) == -ENOENT)
		return -ENOENT;

	event->destroy = hw_perf_event_destroy;

	if (!atomic_inc_not_zero(active_events)) {
		mutex_lock(&armpmu->reserve_mutex);
		if (atomic_read(active_events) == 0)
			err = armpmu_reserve_hardware(armpmu);

		if (!err)
			atomic_inc(active_events);
		mutex_unlock(&armpmu->reserve_mutex);
	}

	if (err)
		return err;

	err = __hw_perf_event_init(event);
	if (err)
		hw_perf_event_destroy(event);

	return err;
}

static void armpmu_enable(struct pmu *pmu)
{
	struct arm_pmu *armpmu = to_arm_pmu(pmu);
	struct pmu_hw_events *hw_events = armpmu->get_hw_events();
	int enabled = bitmap_weight(hw_events->used_mask, armpmu->num_events);

	if (enabled)
		armpmu->start();
}

static void armpmu_disable(struct pmu *pmu)
{
	struct arm_pmu *armpmu = to_arm_pmu(pmu);
	armpmu->stop();
}

#ifdef CONFIG_PM_RUNTIME
static int armpmu_runtime_resume(struct device *dev)
{
	struct arm_pmu_platdata *plat = dev_get_platdata(dev);

	if (plat && plat->runtime_resume)
		return plat->runtime_resume(dev);

	return 0;
}

static int armpmu_runtime_suspend(struct device *dev)
{
	struct arm_pmu_platdata *plat = dev_get_platdata(dev);

	if (plat && plat->runtime_suspend)
		return plat->runtime_suspend(dev);

	return 0;
}
#endif

const struct dev_pm_ops armpmu_dev_pm_ops = {
	SET_RUNTIME_PM_OPS(armpmu_runtime_suspend, armpmu_runtime_resume, NULL)
};

static void __init armpmu_init(struct arm_pmu *armpmu)
{
	atomic_set(&armpmu->active_events, 0);
	mutex_init(&armpmu->reserve_mutex);

	armpmu->pmu = (struct pmu) {
		.pmu_enable	= armpmu_enable,
		.pmu_disable	= armpmu_disable,
		.event_init	= armpmu_event_init,
		.add		= armpmu_add,
		.del		= armpmu_del,
		.start		= armpmu_start,
		.stop		= armpmu_stop,
		.read		= armpmu_read,
	};
}

int armpmu_register(struct arm_pmu *armpmu, char *name, int type)
{
	armpmu_init(armpmu);
	pr_info("enabled with %s PMU driver, %d counters available\n",
			armpmu->name, armpmu->num_events);
	return perf_pmu_register(&armpmu->pmu, name, type);
}

/*
 * Callchain handling code.
 */

/*
 * The registers we're interested in are at the end of the variable
 * length saved register structure. The fp points at the end of this
 * structure so the address of this struct is:
 * (struct frame_tail *)(xxx->fp)-1
 *
 * This code has been adapted from the ARM OProfile support.
 */
struct frame_tail {
	struct frame_tail __user *fp;
	unsigned long sp;
	unsigned long lr;
} __attribute__((packed));

/*
 * Get the return address for a single stackframe and return a pointer to the
 * next frame tail.
 */
static struct frame_tail __user *
user_backtrace(struct frame_tail __user *tail,
	       struct perf_callchain_entry *entry)
{
	struct frame_tail buftail;

	/* Also check accessibility of one struct frame_tail beyond */
	if (!access_ok(VERIFY_READ, tail, sizeof(buftail)))
		return NULL;
	if (__copy_from_user_inatomic(&buftail, tail, sizeof(buftail)))
		return NULL;

	perf_callchain_store(entry, buftail.lr);

	/*
	 * Frame pointers should strictly progress back up the stack
	 * (towards higher addresses).
	 */
	if (tail + 1 >= buftail.fp)
		return NULL;

	return buftail.fp - 1;
}

void
perf_callchain_user(struct perf_callchain_entry *entry, struct pt_regs *regs)
{
	struct frame_tail __user *tail;


	tail = (struct frame_tail __user *)regs->ARM_fp - 1;

	while ((entry->nr < PERF_MAX_STACK_DEPTH) &&
	       tail && !((unsigned long)tail & 0x3))
		tail = user_backtrace(tail, entry);
}

/*
 * Gets called by walk_stackframe() for every stackframe. This will be called
 * whist unwinding the stackframe and is like a subroutine return so we use
 * the PC.
 */
static int
callchain_trace(struct stackframe *fr,
		void *data)
{
	struct perf_callchain_entry *entry = data;
	perf_callchain_store(entry, fr->pc);
	return 0;
}

void
perf_callchain_kernel(struct perf_callchain_entry *entry, struct pt_regs *regs)
{
	struct stackframe fr;

	fr.fp = regs->ARM_fp;
	fr.sp = regs->ARM_sp;
	fr.lr = regs->ARM_lr;
	fr.pc = regs->ARM_pc;
	walk_stackframe(&fr, callchain_trace, entry);
}