path: root/tools/perf/builtin-kmem.c

#include "builtin.h"
#include "perf.h"

#include "util/util.h"
#include "util/cache.h"
#include "util/symbol.h"
#include "util/thread.h"
#include "util/header.h"
#include "util/session.h"

#include "util/parse-options.h"
#include "util/trace-event.h"

#include "util/debug.h"

#include <linux/rbtree.h>

struct alloc_stat;
typedef int (*sort_fn_t)(struct alloc_stat *, struct alloc_stat *);

static char const		*input_name = "perf.data";

static int			alloc_flag;
static int			caller_flag;

static int			alloc_lines = -1;
static int			caller_lines = -1;

static bool			raw_ip;

static char			default_sort_order[] = "frag,hit,bytes";

static int			*cpunode_map;
static int			max_cpu_num;

struct alloc_stat {
	u64	call_site;
	u64	ptr;
	u64	bytes_req;
	u64	bytes_alloc;
	u32	hit;
	u32	pingpong;

	short	alloc_cpu;

	struct rb_node node;
};

static struct rb_root root_alloc_stat;
static struct rb_root root_alloc_sorted;
static struct rb_root root_caller_stat;
static struct rb_root root_caller_sorted;

static unsigned long total_requested, total_allocated;
static unsigned long nr_allocs, nr_cross_allocs;

#define PATH_SYS_NODE	"/sys/devices/system/node"

static void init_cpunode_map(void)
{
	FILE *fp;
	int i;

	fp = fopen("/sys/devices/system/cpu/kernel_max", "r");
	if (!fp) {
		max_cpu_num = 4096;
		return;
	}

	if (fscanf(fp, "%d", &max_cpu_num) < 1)
		die("Failed to read 'kernel_max' from sysfs");
	max_cpu_num++;

	cpunode_map = calloc(max_cpu_num, sizeof(int));
	if (!cpunode_map)
		die("calloc");
	for (i = 0; i < max_cpu_num; i++)
		cpunode_map[i] = -1;
	fclose(fp);
}

static void setup_cpunode_map(void)
{
	struct dirent *dent1, *dent2;
	DIR *dir1, *dir2;
	unsigned int cpu, mem;
	char buf[PATH_MAX];

	init_cpunode_map();

	dir1 = opendir(PATH_SYS_NODE);
	if (!dir1)
		return;

	while ((dent1 = readdir(dir1)) != NULL) {
		if (dent1->d_type != DT_DIR ||
		    sscanf(dent1->d_name, "node%u", &mem) < 1)
			continue;

		snprintf(buf, PATH_MAX, "%s/%s", PATH_SYS_NODE, dent1->d_name);
		dir2 = opendir(buf);
		if (!dir2)
			continue;
		while ((dent2 = readdir(dir2)) != NULL) {
			if (dent2->d_type != DT_LNK ||
			    sscanf(dent2->d_name, "cpu%u", &cpu) < 1)
				continue;
			cpunode_map[cpu] = mem;
		}
	}
}

static void insert_alloc_stat(unsigned long call_site, unsigned long ptr,
			      int bytes_req, int bytes_alloc, int cpu)
{
	struct rb_node **node = &root_alloc_stat.rb_node;
	struct rb_node *parent = NULL;
	struct alloc_stat *data = NULL;

	while (*node) {
		parent = *node;
		data = rb_entry(*node, struct alloc_stat, node);

		if (ptr > data->ptr)
			node = &(*node)->rb_right;
		else if (ptr < data->ptr)
			node = &(*node)->rb_left;
		else
			break;
	}

	if (data && data->ptr == ptr) {
		data->hit++;
		data->bytes_req += bytes_req;
		data->bytes_alloc += bytes_alloc;
	} else {
		data = malloc(sizeof(*data));
		if (!data)
			die("malloc");
		data->ptr = ptr;
		data->pingpong = 0;
		data->hit = 1;
		data->bytes_req = bytes_req;
		data->bytes_alloc = bytes_alloc;

		rb_link_node(&data->node, parent, node);
		rb_insert_color(&data->node, &root_alloc_stat);
	}
	data->call_site = call_site;
	data->alloc_cpu = cpu;
}

static void insert_caller_stat(unsigned long call_site,
			      int bytes_req, int bytes_alloc)
{
	struct rb_node **node = &root_caller_stat.rb_node;
	struct rb_node *parent = NULL;
	struct alloc_stat *data = NULL;

	while (*node) {
		parent = *node;
		data = rb_entry(*node, struct alloc_stat, node);

		if (call_site > data->call_site)
			node = &(*node)->rb_right;
		else if (call_site < data->call_site)
			node = &(*node)->rb_left;
		else
			break;
	}

	if (data && data->call_site == call_site) {
		data->hit++;
		data->bytes_req += bytes_req;
		data->bytes_alloc += bytes_alloc;
	} else {
		data = malloc