/* Copyright (c) 2015, Linaro Limited
 * All rights reserved.
 *
 * SPDX-License-Identifier:	BSD-3-Clause
 */

/*
 * Program to read perf counter on ARMv8 from userspace using mmap way
 */
#define _GNU_SOURCE
#include <stdio.h>
#include <stdint.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/time.h>
#include <sys/syscall.h>
#include <linux/perf_event.h>
#include <linux/types.h>
#include <sys/mman.h>
#include <errno.h>
#include <string.h>
#define barrier() asm volatile("dmb ish" ::: "memory")
#define isb() asm volatile("isb" ::: "memory")
static int fddev = -1;
__attribute__((constructor)) static void
init(void)
{
	static struct perf_event_attr attr;
	attr.type = PERF_TYPE_HARDWARE;
	attr.config = PERF_COUNT_HW_CPU_CYCLES;
	fddev = syscall(__NR_perf_event_open, &attr, 0, -1, -1, 0);
}

__attribute__((destructor)) static void
fini(void)
{
	close(fddev);
}
#define ARMV8_PMCNTENSET_EL0_ENABLE (1<<31) /**< Enable Perf count reg */

static unsigned long rdpmc(unsigned int counter)
{
	unsigned long int ret;
	asm volatile("mrs %0, pmccntr_el0" : "=r" (ret));
	isb();
	return ret;
}

static unsigned long mmap_read_self(void *addr)
{
	struct perf_event_mmap_page *pc = addr;
	unsigned int seq, idx;
	unsigned long  count;
	signed long pmc = 0;
	do {
		count = 0;
		seq = pc->lock;
		barrier();
		idx = pc->index;
		if (idx)
			pmc = rdpmc(idx - 1);
		barrier();
	} while (pc->lock != seq);
	count = pmc;
	return count;
}

/* Simple loop body to keep things interested. Make sure it gets inlined. */
static inline int
loop(int* __restrict__ a, int* __restrict__ b, int n)
{
	unsigned sum = 0;
	int i = 0;
	for (i = 0; i < n; ++i)
		if (a[i] > b[i])
			sum += a[i] + 5;
	return sum;
}

static unsigned long page_size;

int main(int ac, char **av)
{
	long long time_start = 0;
	long long time_end   = 0;

	int *a  = NULL;
	int *b  = NULL;
	int len = 0;
	int i, sum = 0;
	void *addr; /* mmaped address */
	unsigned long start_count, stop_count;
	if (ac != 2)
		return -1;
	len = atoi(av[1]);
	printf("%s: len = %d\n", av[0], len);

	a = malloc(len*sizeof(*a));
	b = malloc(len*sizeof(*b));

	for (i = 0; i < len; ++i) {
		a[i] = i+128;
		b[i] = i+64;
	}

	page_size = sysconf(_SC_PAGESIZE);

	/*Allocate memory for the page size */
	addr = mmap(NULL, page_size, PROT_READ, MAP_SHARED, fddev, 0);
	if (addr == (void *)(-1)) {
		printf("Error: mmap() syscall returned with (%s)\n",
		       strerror(errno));
		return 0;
	}
	i = 0;
	start_count = stop_count = 0;
	/* --------------------Critical section-------------- */
	start_count = mmap_read_self(addr);
	for (i = 0; i < 1000; i++) {
		sum = loop(a, b, len);
		mmap_read_self(addr);
	}
	stop_count = mmap_read_self(addr);
	/* --------------------End Critical section-------------- */
	printf("sum=%d Avg count [ Loop + Read ] is = %ld\n", sum,
	       (stop_count-start_count)/1000);
	start_count = stop_count = 0;
	/* --------------------Critical section-------------- */
	start_count = mmap_read_self(addr);
	for (i = 0; i < 1000; i++)
		sum = loop(a, b, len);
	stop_count = mmap_read_self(addr);
	/* --------------------End Critical section-------------- */
	printf("sum=%d Avg count [ Loop ] \t is = %ld\n", sum,
	       (stop_count-start_count)/1000);
	munmap(addr, page_size);
	free(a);
	free(b);
	return 0;
}