path: root/arch/arc/kernel/perf_event.c

// SPDX-License-Identifier: GPL-2.0+
//
// Linux performance counter support for ARC CPUs.
// This code is inspired by the perf support of various other architectures.
//
// Copyright (C) 2013-2018 Synopsys, Inc. (www.synopsys.com)

#include <linux/errno.h>
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/perf_event.h>
#include <linux/platform_device.h>
#include <asm/arcregs.h>
#include <asm/stacktrace.h>

/* HW holds 8 symbols + one for null terminator */
#define ARCPMU_EVENT_NAME_LEN	9

enum arc_pmu_attr_groups {
	ARCPMU_ATTR_GR_EVENTS,
	ARCPMU_ATTR_GR_FORMATS,
	ARCPMU_NR_ATTR_GR
};

struct arc_pmu_raw_event_entry {
	char name[ARCPMU_EVENT_NAME_LEN];
};

struct arc_pmu {
	struct pmu	pmu;
	unsigned int	irq;
	int		n_counters;
	int		n_events;
	u64		max_period;
	int		ev_hw_idx[PERF_COUNT_ARC_HW_MAX];

	struct arc_pmu_raw_event_entry	*raw_entry;
	struct attribute		**attrs;
	struct perf_pmu_events_attr	*attr;
	const struct attribute_group	*attr_groups[ARCPMU_NR_ATTR_GR + 1];
};

struct arc_pmu_cpu {
	/*
	 * A 1 bit for an index indicates that the counter is being used for
	 * an event. A 0 means that the counter can be used.
	 */
	unsigned long	used_mask[BITS_TO_LONGS(ARC_PERF_MAX_COUNTERS)];

	/*
	 * The events that are active on the PMU for the given index.
	 */
	struct perf_event *act_counter[ARC_PERF_MAX_COUNTERS];
};

struct arc_callchain_trace {
	int depth;
	void *perf_stuff;
};

static int callchain_trace(unsigned int addr, void *data)
{
	struct arc_callchain_trace *ctrl = data;
	struct perf_callchain_entry_ctx *entry = ctrl->perf_stuff;

	perf_callchain_store(entry, addr);

	if (ctrl->depth++ < 3)
		return 0;

	return -1;
}

void perf_callchain_kernel(struct perf_callchain_entry_ctx *entry,
			   struct pt_regs *regs)
{
	struct arc_callchain_trace ctrl = {
		.depth = 0,
		.perf_stuff = entry,
	};

	arc_unwind_core(NULL, regs, callchain_trace, &ctrl);
}

void perf_callchain_user(struct perf_callchain_entry_ctx *entry,
			 struct pt_regs *regs)
{
	/*
	 * User stack can't be unwound trivially with kernel dwarf unwinder
	 * So for now just record the user PC
	 */
	perf_callchain_store(entry, instruction_pointer(regs));
}

static struct arc_pmu *arc_pmu;
static DEFINE_PER_CPU(struct arc_pmu_cpu, arc_pmu_cpu);

/* read counter #idx; note that counter# != event# on ARC! */
static u64 arc_pmu_read_counter(int idx)
{
	u32 tmp;
	u64 result;

	/*
	 * ARC supports making 'snapshots' of the counters, so we don't
	 * need to care about counters wrapping to 0 underneath our feet
	 */
	write_aux_reg(ARC_REG_PCT_INDEX, idx);
	tmp = read_aux_reg(ARC_REG_PCT_CONTROL);
	write_aux_reg(ARC_REG_PCT_CONTROL, tmp | ARC_REG_PCT_CONTROL_SN);
	result = (u64) (read_aux_reg(ARC_REG_PCT_SNAPH)) << 32;
	result |= read_aux_reg(ARC_REG_PCT_SNAPL);

	return result;
}

static void arc_perf_event_update(struct perf_event *event,
				  struct hw_perf_event *hwc, int idx)
{
	u64 prev_raw_count = local64_read(&hwc->prev_count);
	u64 new_raw_count = arc_pmu_read_counter(idx);
	s64 delta = new_raw_count - prev_raw_count;

	/*
	 * We aren't afraid of hwc->prev_count changing beneath our feet
	 * because there's no way for us to re-enter this function anytime.
	 */
	local64_set(&hwc->prev_count, new_raw_count);
	local64_add(delta, &event->count);
	local64_sub(delta, &hwc->period_left);
}

static void arc_pmu_read(struct perf_event *event)
{
	arc_perf_event_update(event, &event->hw, event->hw.idx);
}

static int arc_pmu_cache_event(u64 config)
{
	unsigned int cache_type, cache_op, cache_result;
	int ret;

	cache_type	= (config >>  0) & 0xff;
	cache_op	= (config >>  8) & 0xff;
	cache_result	= (config >> 16) & 0xff;
	if (cache_type >= PERF_COUNT_HW_CACHE_MAX)
		return -EINVAL;
	if (cache_op >= PERF_COUNT_HW_CACHE_OP_MAX)
		return -EINVAL;
	if (cache_result >= PERF_COUNT_HW_CACHE_RESULT_MAX)
		return -EINVAL;

	ret = arc_pmu_cache_map[cache_type][cache_op][cache_result];

	if (ret == CACHE_OP_UNSUPPORTED)
		return -ENOENT;

	pr_debug("init cache event: type/op/result %d/%d/%d with h/w %d \'%s\'\n",
		 cache_type, cache_op, cache_result, ret,
		 arc_pmu_ev_hw_map[ret]);

	return ret;
}

/* initializes hw_perf_event structure if event is supported */
static int arc_pmu_event_init(struct perf_event *event)
{
	struct hw_perf_event *hwc = &event->hw;
	int ret;

	if (!is_sampling_event(event)) {
		hwc->sample_period = arc_pmu->max_period;
		hwc->last_period = hwc->sample_period;
		local64_set(&hwc->period_left, hwc->sample_period);
	}

	hwc->config = 0;

	if (is_isa_arcv2()) {
		/* "exclude user" means "count only kernel" */
		if (event->attr.exclude_user)
			hwc->config |= ARC_REG_PCT_CONFIG_KERN;

		/* "exclude kernel" means "count only user" */
		if (event->attr.exclude_kernel)
			hwc->config |= ARC_REG_PCT_CONFIG_USER;
	}

	switch (event->attr.type) {
	case PERF_TYPE_HARDWARE:
		if (event->attr.config >= PERF_COUNT_HW_MAX)
			return -ENOENT;
		if (arc_pmu->ev_hw_idx[event->attr.config] < 0)
			return -ENOENT;
		hwc->config |= arc_pmu->ev_hw_idx[event->attr.config];
		pr_debug("init event %d with h/w %08x \'%s\'\n",
			 (int)event->attr.config, (int)hwc->config,
			 arc_pmu_ev_hw_map[event->attr.config]);
		return 0;

	case PERF_TYPE_HW_CACHE:
		ret = arc_pmu_cache_event(event->attr.config);
		if (ret < 0)
			return ret;
		hwc->config |= arc_pmu->ev_hw_idx[ret];
		pr_debug("init cache event with h/w %08x \'%s\'\n",
			 (int)hwc->config, arc_pmu_ev_hw_map[ret]);
		return 0;

	case PERF_TYPE_RAW:
		if (event->attr.config >= arc_pmu->n_events)
			return -ENOENT;

		hwc->config |= event->attr.config;
		pr_debug("init raw event with idx %lld \'%s