[linux-2.6-block.git] / tools / accounting / getdelays.c

// SPDX-License-Identifier: GPL-2.0
/* getdelays.c
 *
 * Utility to get per-pid and per-tgid delay accounting statistics
 * Also illustrates usage of the taskstats interface
 *
 * Copyright (C) Shailabh Nagar, IBM Corp. 2005
 * Copyright (C) Balbir Singh, IBM Corp. 2006
 * Copyright (c) Jay Lan, SGI. 2006
 *
 * Compile with
 *	gcc -I/usr/src/linux/include getdelays.c -o getdelays
 */

#include <stdio.h>
#include <stdlib.h>
#include <errno.h>
#include <unistd.h>
#include <poll.h>
#include <string.h>
#include <fcntl.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/socket.h>
#include <sys/wait.h>
#include <signal.h>

#include <linux/genetlink.h>
#include <linux/taskstats.h>
#include <linux/cgroupstats.h>

/*
 * Generic macros for dealing with netlink sockets. Might be duplicated
 * elsewhere. It is recommended that commercial grade applications use
 * libnl or libnetlink and use the interfaces provided by the library
 */
#define GENLMSG_DATA(glh)	((void *)(NLMSG_DATA(glh) + GENL_HDRLEN))
#define GENLMSG_PAYLOAD(glh)	(NLMSG_PAYLOAD(glh, 0) - GENL_HDRLEN)
#define NLA_DATA(na)		((void *)((char*)(na) + NLA_HDRLEN))
#define NLA_PAYLOAD(len)	(len - NLA_HDRLEN)

#define err(code, fmt, arg...)			\
	do {					\
		fprintf(stderr, fmt, ##arg);	\
		exit(code);			\
	} while (0)

int done;
int rcvbufsz;
char name[100];
int dbg;
int print_delays;
int print_io_accounting;
int print_task_context_switch_counts;

#define PRINTF(fmt, arg...) {			\
	    if (dbg) {				\
		printf(fmt, ##arg);		\
	    }					\
	}

/* Maximum size of response requested or message sent */
#define MAX_MSG_SIZE	1024
/* Maximum number of cpus expected to be specified in a cpumask */
#define MAX_CPUS	32

struct msgtemplate {
	struct nlmsghdr n;
	struct genlmsghdr g;
	char buf[MAX_MSG_SIZE];
};

char cpumask[100+6*MAX_CPUS];

static void usage(void)
{
	fprintf(stderr, "getdelays [-dilv] [-w logfile] [-r bufsize] "
			"[-m cpumask] [-t tgid] [-p pid]\n");
	fprintf(stderr, "  -d: print delayacct stats\n");
	fprintf(stderr, "  -i: print IO accounting (works only with -p)\n");
	fprintf(stderr, "  -l: listen forever\n");
	fprintf(stderr, "  -v: debug on\n");
	fprintf(stderr, "  -C: container path\n");
}

/*
 * Create a raw netlink socket and bind
 */
static int create_nl_socket(int protocol)
{
	int fd;
	struct sockaddr_nl local;

	fd = socket(AF_NETLINK, SOCK_RAW, protocol);
	if (fd < 0)
		return -1;

	if (rcvbufsz)
		if (setsockopt(fd, SOL_SOCKET, SO_RCVBUF,
				&rcvbufsz, sizeof(rcvbufsz)) < 0) {
			fprintf(stderr, "Unable to set socket rcv buf size to %d\n",
				rcvbufsz);
			goto error;
		}

	memset(&local, 0, sizeof(local));
	local.nl_family = AF_NETLINK;

	if (bind(fd, (struct sockaddr *) &local, sizeof(local)) < 0)
		goto error;

	return fd;
error:
	close(fd);
	return -1;
}


static int send_cmd(int sd, __u16 nlmsg_type, __u32 nlmsg_pid,
	     __u8 genl_cmd, __u16 nla_type,
	     void *nla_data, int nla_len)
{
	struct nlattr *na;
	struct sockaddr_nl nladdr;
	int r, buflen;
	char *buf;

	struct msgtemplate msg;

	msg.n.nlmsg_len = NLMSG_LENGTH(GENL_HDRLEN);
	msg.n.nlmsg_type = nlmsg_type;
	msg.n.nlmsg_flags = NLM_F_REQUEST;
	msg.n.nlmsg_seq = 0;
	msg.n.nlmsg_pid = nlmsg_pid;
	msg.g.cmd = genl_cmd;
	msg.g.version = 0x1;
	na = (struct nlattr *) GENLMSG_DATA(&msg);
	na->nla_type = nla_type;
	na->nla_len = nla_len + 1 + NLA_HDRLEN;
	memcpy(NLA_DATA(na), nla_data, nla_len);
	msg.n.nlmsg_len += NLMSG_ALIGN(na->nla_len);

	buf = (char *) &msg;
	buflen = msg.n.nlmsg_len ;
	memset(&nladdr, 0, sizeof(nladdr));
	nladdr.nl_family = AF_NETLINK;
	while ((r = sendto(sd, buf, buflen, 0, (struct sockaddr *) &nladdr,
			   sizeof(nladdr))) < buflen) {
		if (r > 0) {
			buf += r;
			buflen -= r;
		} else if (errno != EAGAIN)
			return -1;
	}
	return 0;
}


/*
 * Probe the controller in genetlink to find the family id
 * for the TASKSTATS family
 */
static int get_family_id(int sd)
{
	struct {
		struct nlmsghdr n;
		struct genlmsghdr g;
		char buf[256];
	} ans;

	int id = 0, rc;
	struct nlattr *na;
	int rep_len;

	strcpy(name, TASKSTATS_GENL_NAME);
	rc = send_cmd(sd, GENL_ID_CTRL, getpid(), CTRL_CMD_GETFAMILY,
			CTRL_ATTR_FAMILY_NAME, (void *)name,
			strlen(TASKSTATS_GENL_NAME)+1);
	if (rc < 0)
		return 0;	/* sendto() failure? */

	rep_len = recv(sd, &ans, sizeof(ans), 0);
	if (ans.n.nlmsg_type == NLMSG_ERROR ||
	    (rep_len < 0) || !NLMSG_OK((&ans.n), rep_len))
		return 0;

	na = (struct nlattr *) GENLMSG_DATA(&ans);
	na = (struct nlattr *) ((char *) na + NLA_ALIGN(na->nla_len));
	if (na->nla_type == CTRL_ATTR_FAMILY_ID) {
		id = *(__u16 *) NLA_DATA(na);
	}
	return id;
}

#define average_ms(t, c) (t / 1000000ULL / (c ? c : 1))

static void print_delayacct(struct taskstats *t)
{
	printf("\n\nCPU   %15s%15s%15s%15s%15s\n"
	       "      %15llu%15llu%15llu%15llu%15.3fms\n"
	       "IO    %15s%15s%15s\n"
	       "      %15llu%15llu%15llums\n"
	       "SWAP  %15s%15s%15s\n"
	       "      %15llu%15llu%15llums\n"
	       "RECLAIM  %12s%15s%15s\n"
	       "      %15llu%15llu%15llums\n"
	       "THRASHING%12s%15s%15s\n"
	       "      %15llu%15llu%15llums\n",
	       "count", "real total", "virtual total",
	       "delay total", "delay average",
	       (unsigned long long)t->cpu_count,
	       (unsigned long long)t->cpu_run_real_total,
	       (unsigned long long)t->cpu_run_virtual_total,
	       (unsigned long long)t->cpu_delay_total,
	       average_ms((double)t->cpu_delay_total, t->cpu_count),
	       "count", "delay total", "delay average",
	       (unsigned long long)t->blkio_count,
	       (unsigned long long)t->blkio_delay_total,
	       average_ms(t->blkio_delay_total, t->blkio_count),
	       "count", "delay total", "delay average",
	       (unsigned long long)t->swapin_count,
	       (unsigned long long)t->swapin_delay_total,
	       average_ms(t->swapin_delay_total, t->swapin_count),
	       "count", "delay total", "delay average",
	       (unsigned long long)t->freepages_count,
	       (unsigned long long)t->freepages_delay_total,
	       average_ms(t->freepages_delay_total, t->freepages_count),
	       "count", "delay total", "delay average",
	       (unsigned long long)t->thrashing_count,
	       (unsigned long long)t->thrashing_delay_total,
	       average_ms(t->thrashing_delay_total, t->thrashing_count));
}

static void task_context_switch_counts(struct taskstats *t)
{
	printf("\n\nTask   %15s%15s\n"
	       "       %15llu%15llu\n",
	       "voluntary", "nonvoluntary",
	       (unsigned long long)t->nvcsw, (unsigned long long)t->nivcsw);
}

static void print_cgroupstats(struct cgroupstats *c)
{
	printf("sleeping %llu, blocked %llu, running %llu, stopped %llu, "
		"uninterruptible %llu\n", (unsigned long long)c->nr_sleeping,
		(unsigned long long)c->nr_io_wait,
		(unsigned long long)c->nr_running,
		(unsigned long long)c->nr_stopped,
		(unsigned long long)c->nr_uninterruptible);
}


static void print_ioacct(struct taskstats *t)
{
	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
		t->ac_comm,
		(unsigned long long)t->read_bytes,
		(unsigned long long)t->write_bytes,
		(unsigned long long)t->cancelled_write_bytes);
}

int main(int argc, char *argv[])
{
	int c, rc, rep_len, aggr_len, len2;
	int cmd_type = TASKSTATS_CMD_ATTR_UNSPEC;
	__u16 id;
	__u32 mypid;

	struct nlattr *na;
	int nl_sd = -1;
	int len = 0;
	pid_t tid = 0;
	pid_t rtid = 0;

	int fd = 0;
	int count = 0;
	int write_file = 0;
	int maskset = 0;
	char *logfile = NULL;
	int loop = 0;
	int containerset = 0;
	char *containerpath = NULL;
	int cfd = 0;
	int forking = 0;
	sigset_t sigset;

	struct msgtemplate msg;

	while (!forking) {
		c = getopt(argc, argv, "qdiw:r:m:t:p:vlC:c:");
		if (c < 0)
			break;

		switch (c) {
		case 'd':
			printf("print delayacct stats ON\n");
			print_delays = 1;
			break;
		case 'i':
			printf("printing IO accounting\n");
			print_io_accounting = 1;
			break;
		case 'q':
			printf("printing task/process context switch rates\n");
			print_task_context_switch_counts = 1;
			break;
		case 'C':
			containerset = 1;
			containerpath = optarg;
			break;
		case 'w':
			logfile = strdup(optarg);
			printf("write to file %s\n", logfile);
			write_file = 1;
			break;
		case 'r':
			rcvbufsz = atoi(optarg);
			printf("receive buf size %d\n", rcvbufsz);
			if (rcvbufsz < 0)
				err(1, "Invalid rcv buf size\n");
			break;
		case 'm':
			strncpy(cpumask, optarg, sizeof(cpumask));
			cpumask[sizeof(cpumask) - 1] = '\0';
			maskset = 1;
			printf("cpumask %s maskset %d\n", cpumask, maskset);
			break;
		case 't':
			tid = atoi(optarg);
			if (!tid)
				err(1, "Invalid tgid\n");
			cmd_type = TASKSTATS_CMD_ATTR_TGID;
			break;
		case 'p':
			tid = atoi(optarg);
			if (!tid)
				err(1, "Invalid pid\n");
			cmd_type = TASKSTATS_CMD_ATTR_PID;
			break;
		case 'c':

			/* Block SIGCHLD for sigwait() later */
			if (sigemptyset(&sigset) == -1)
				err(1, "Failed to empty sigset");
			if (sigaddset(&sigset, SIGCHLD))
				err(1, "Failed to set sigchld in sigset");
			sigprocmask(SIG_BLOCK, &sigset, NULL);

			/* fork/exec a child */
			tid = fork();
			if (tid < 0)
				err(1, "Fork failed\n");
			if (tid == 0)
				if (execvp(argv[optind - 1],
				    &argv[optind - 1]) < 0)
					exit(-1);

			/* Set the command type and avoid further processing */
			cmd_type = TASKSTATS_CMD_ATTR_PID;
			forking = 1;
			break;
		case 'v':
			printf("debug on\n");
			dbg = 1;
			break;
		case 'l':
			printf("listen forever\n");
			loop = 1;
			break;
		default:
			usage();
			exit(-1);
		}
	}

	if (write_file) {
		fd = open(logfile, O_WRONLY | O_CREAT | O_TRUNC,
			  S_IRUSR | S_IWUSR | S_IRGRP | S_IROTH);
		if (fd == -1) {
			perror("Cannot open output file\n");
			exit(1);
		}
	}

	nl_sd = create_nl_socket(NETLINK_GENERIC);
	if (nl_sd < 0)
		err(1, "error creating Netlink socket\n");


	mypid = getpid();
	id = get_family_id(nl_sd);
	if (!id) {
		fprintf(stderr, "Error getting family id, errno %d\n", errno);
		goto err;
	}
	PRINTF("family id %d\n", id);

	if (maskset) {
		rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
			      TASKSTATS_CMD_ATTR_REGISTER_CPUMASK,
			      &cpumask, strlen(cpumask) + 1);
		PRINTF("Sent register cpumask, retval %d\n", rc);
		if (rc < 0) {
			fprintf(stderr, "error sending register cpumask\n");
			goto err;
		}
	}

	if (tid && containerset) {
		fprintf(stderr, "Select either -t or -C, not both\n");
		goto err;
	}

	/*
	 * If we forked a child, wait for it to exit. Cannot use waitpid()
	 * as all the delicious data would be reaped as part of the wait
	 */
	if (tid && forking) {
		int sig_received;
		sigwait(&sigset, &sig_received);
	}

	if (tid) {
		rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
			      cmd_type, &tid, sizeof(__u32));
		PRINTF("Sent pid/tgid, retval %d\n", rc);
		if (rc < 0) {
			fprintf(stderr, "error sending tid/tgid cmd\n");
			goto done;
		}
	}

	if (containerset) {
		cfd = open(containerpath, O_RDONLY);
		if (cfd < 0) {
			perror("error opening container file");
			goto err;
		}
		rc = send_cmd(nl_sd, id, mypid, CGROUPSTATS_CMD_GET,
			      CGROUPSTATS_CMD_ATTR_FD, &cfd, sizeof(__u32));
		if (rc < 0) {
			perror("error sending cgroupstats command");
			goto err;
		}
	}
	if (!maskset && !tid && !containerset) {
		usage();
		goto err;
	}

	do {
		rep_len = recv(nl_sd, &msg, sizeof(msg), 0);
		PRINTF("received %d bytes\n", rep_len);

		if (rep_len < 0) {
			fprintf(stderr, "nonfatal reply error: errno %d\n",
				errno);
			continue;
		}
		if (msg.n.nlmsg_type == NLMSG_ERROR ||
		    !NLMSG_OK((&msg.n), rep_len)) {
			struct nlmsgerr *err = NLMSG_DATA(&msg);
			fprintf(stderr, "fatal reply error,  errno %d\n",
				err->error);
			goto done;
		}

		PRINTF("nlmsghdr size=%zu, nlmsg_len=%d, rep_len=%d\n",
		       sizeof(struct nlmsghdr), msg.n.nlmsg_len, rep_len);


		rep_len = GENLMSG_PAYLOAD(&msg.n);

		na = (struct nlattr *) GENLMSG_DATA(&msg);
		len = 0;
		while (len < rep_len) {
			len += NLA_ALIGN(na->nla_len);
			switch (na->nla_type) {
			case TASKSTATS_TYPE_AGGR_TGID:
				/* Fall through */
			case TASKSTATS_TYPE_AGGR_PID:
				aggr_len = NLA_PAYLOAD(na->nla_len);
				len2 = 0;
				/* For nested attributes, na follows */
				na = (struct nlattr *) NLA_DATA(na);
				done = 0;
				while (len2 < aggr_len) {
					switch (na->nla_type) {
					case TASKSTATS_TYPE_PID:
						rtid = *(int *) NLA_DATA(na);
						if (print_delays)
							printf("PID\t%d\n", rtid);
						break;
					case TASKSTATS_TYPE_TGID:
						rtid = *(int *) NLA_DATA(na);
						if (print_delays)
							printf("TGID\t%d\n", rtid);
						break;
					case TASKSTATS_TYPE_STATS:
						count++;
						if (print_delays)
							print_delayacct((struct taskstats *) NLA_DATA(na));
						if (print_io_accounting)
							print_ioacct((struct taskstats *) NLA_DATA(na));
						if (print_task_context_switch_counts)
							task_context_switch_counts((struct taskstats *) NLA_DATA(na));
						if (fd) {
							if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
								err(1,"write error\n");
							}
						}
						if (!loop)
							goto done;
						break;
					case TASKSTATS_TYPE_NULL:
						break;
					default:
						fprintf(stderr, "Unknown nested"
							" nla_type %d\n",
							na->nla_type);
						break;
					}
					len2 += NLA_ALIGN(na->nla_len);
					na = (struct nlattr *)((char *)na +
							       NLA_ALIGN(na->nla_len));
				}
				break;

			case CGROUPSTATS_TYPE_CGROUP_STATS:
				print_cgroupstats(NLA_DATA(na));
				break;
			default:
				fprintf(stderr, "Unknown nla_type %d\n",
					na->nla_type);
			case TASKSTATS_TYPE_NULL:
				break;
			}
			na = (struct nlattr *) (GENLMSG_DATA(&msg) + len);
		}
	} while (loop);
done:
	if (maskset) {
		rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
			      TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK,
			      &cpumask, strlen(cpumask) + 1);
		printf("Sent deregister mask, retval %d\n", rc);
		if (rc < 0)
			err(rc, "error sending deregister cpumask\n");
	}
err:
	close(nl_sd);
	if (fd)
		close(fd);
	if (cfd)
		close(cfd);
	return 0;
}
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
a3baf649 SN	2	/* getdelays.c
	3	*
	4	* Utility to get per-pid and per-tgid delay accounting statistics
	5	* Also illustrates usage of the taskstats interface
	6	*
	7	* Copyright (C) Shailabh Nagar, IBM Corp. 2005
	8	* Copyright (C) Balbir Singh, IBM Corp. 2006
9e06d3f9	9	* Copyright (c) Jay Lan, SGI. 2006
a3baf649	10	*
d2f7bf13 AM	11	* Compile with
d2f7bf13 AM	12	* gcc -I/usr/src/linux/include getdelays.c -o getdelays
a3baf649 SN	13	*/
	14
	15	#include <stdio.h>
	16	#include <stdlib.h>
	17	#include <errno.h>
	18	#include <unistd.h>
	19	#include <poll.h>
	20	#include <string.h>
	21	#include <fcntl.h>
	22	#include <sys/types.h>
	23	#include <sys/stat.h>
	24	#include <sys/socket.h>
db9e5679	25	#include <sys/wait.h>
a3baf649 SN	26	#include <signal.h>
	27
	28	#include <linux/genetlink.h>
	29	#include <linux/taskstats.h>
546040dc	30	#include <linux/cgroupstats.h>
a3baf649 SN	31
	32	/*
	33	* Generic macros for dealing with netlink sockets. Might be duplicated
	34	* elsewhere. It is recommended that commercial grade applications use
	35	* libnl or libnetlink and use the interfaces provided by the library
	36	*/
	37	#define GENLMSG_DATA(glh) ((void *)(NLMSG_DATA(glh) + GENL_HDRLEN))
	38	#define GENLMSG_PAYLOAD(glh) (NLMSG_PAYLOAD(glh, 0) - GENL_HDRLEN)
	39	#define NLA_DATA(na) ((void )((char)(na) + NLA_HDRLEN))
	40	#define NLA_PAYLOAD(len) (len - NLA_HDRLEN)
	41
d2f7bf13 AM	42	#define err(code, fmt, arg...) \
	43	do { \
	44	fprintf(stderr, fmt, ##arg); \
	45	exit(code); \
	46	} while (0)
	47
	48	int done;
	49	int rcvbufsz;
	50	char name[100];
	51	int dbg;
	52	int print_delays;
cf709844	53	int print_io_accounting;
b663a79c	54	int print_task_context_switch_counts;
d2f7bf13	55
9e06d3f9 SN	56	#define PRINTF(fmt, arg...) { \
	57	if (dbg) { \
	58	printf(fmt, ##arg); \
	59	} \
	60	}
	61
	62	/* Maximum size of response requested or message sent */
88040230	63	#define MAX_MSG_SIZE 1024
9e06d3f9 SN	64	/* Maximum number of cpus expected to be specified in a cpumask */
9e06d3f9 SN	65	#define MAX_CPUS 32
9e06d3f9 SN	66
	67	struct msgtemplate {
	68	struct nlmsghdr n;
	69	struct genlmsghdr g;
	70	char buf[MAX_MSG_SIZE];
	71	};
	72
	73	char cpumask[100+6*MAX_CPUS];
a3baf649	74
f16825bb RD	75	static void usage(void)
	76	{
	77	fprintf(stderr, "getdelays [-dilv] [-w logfile] [-r bufsize] "
	78	"[-m cpumask] [-t tgid] [-p pid]\n");
	79	fprintf(stderr, " -d: print delayacct stats\n");
	80	fprintf(stderr, " -i: print IO accounting (works only with -p)\n");
	81	fprintf(stderr, " -l: listen forever\n");
	82	fprintf(stderr, " -v: debug on\n");
546040dc	83	fprintf(stderr, " -C: container path\n");
f16825bb RD	84	}
f16825bb RD	85
a3baf649 SN	86	/*
	87	* Create a raw netlink socket and bind
	88	*/
9e06d3f9	89	static int create_nl_socket(int protocol)
a3baf649	90	{
9e06d3f9 SN	91	int fd;
	92	struct sockaddr_nl local;
	93
	94	fd = socket(AF_NETLINK, SOCK_RAW, protocol);
	95	if (fd < 0)
	96	return -1;
	97
	98	if (rcvbufsz)
	99	if (setsockopt(fd, SOL_SOCKET, SO_RCVBUF,
	100	&rcvbufsz, sizeof(rcvbufsz)) < 0) {
0de4b954	101	fprintf(stderr, "Unable to set socket rcv buf size to %d\n",
d2f7bf13	102	rcvbufsz);
0de4b954	103	goto error;
9e06d3f9	104	}
a3baf649	105
9e06d3f9 SN	106	memset(&local, 0, sizeof(local));
9e06d3f9 SN	107	local.nl_family = AF_NETLINK;
a3baf649	108
9e06d3f9 SN	109	if (bind(fd, (struct sockaddr *) &local, sizeof(local)) < 0)
9e06d3f9 SN	110	goto error;
a3baf649	111
9e06d3f9 SN	112	return fd;
	113	error:
	114	close(fd);
	115	return -1;
a3baf649 SN	116	}
a3baf649 SN	117
9e06d3f9	118
b7ed698c	119	static int send_cmd(int sd, __u16 nlmsg_type, __u32 nlmsg_pid,
9e06d3f9 SN	120	__u8 genl_cmd, __u16 nla_type,
9e06d3f9 SN	121	void *nla_data, int nla_len)
a3baf649	122	{
9e06d3f9 SN	123	struct nlattr *na;
	124	struct sockaddr_nl nladdr;
	125	int r, buflen;
	126	char *buf;
	127
	128	struct msgtemplate msg;
	129
	130	msg.n.nlmsg_len = NLMSG_LENGTH(GENL_HDRLEN);
	131	msg.n.nlmsg_type = nlmsg_type;
	132	msg.n.nlmsg_flags = NLM_F_REQUEST;
	133	msg.n.nlmsg_seq = 0;
	134	msg.n.nlmsg_pid = nlmsg_pid;
	135	msg.g.cmd = genl_cmd;
	136	msg.g.version = 0x1;
	137	na = (struct nlattr *) GENLMSG_DATA(&msg);
	138	na->nla_type = nla_type;
	139	na->nla_len = nla_len + 1 + NLA_HDRLEN;
	140	memcpy(NLA_DATA(na), nla_data, nla_len);
	141	msg.n.nlmsg_len += NLMSG_ALIGN(na->nla_len);
	142
	143	buf = (char *) &msg;
	144	buflen = msg.n.nlmsg_len ;
	145	memset(&nladdr, 0, sizeof(nladdr));
	146	nladdr.nl_family = AF_NETLINK;
	147	while ((r = sendto(sd, buf, buflen, 0, (struct sockaddr *) &nladdr,
	148	sizeof(nladdr))) < buflen) {
	149	if (r > 0) {
	150	buf += r;
	151	buflen -= r;
	152	} else if (errno != EAGAIN)
	153	return -1;
	154	}
	155	return 0;
a3baf649 SN	156	}
a3baf649 SN	157
9e06d3f9	158
a3baf649 SN	159	/*
	160	* Probe the controller in genetlink to find the family id
	161	* for the TASKSTATS family
	162	*/
b7ed698c	163	static int get_family_id(int sd)
a3baf649	164	{
9e06d3f9 SN	165	struct {
	166	struct nlmsghdr n;
	167	struct genlmsghdr g;
	168	char buf[256];
	169	} ans;
	170
10e6f32b	171	int id = 0, rc;
9e06d3f9 SN	172	struct nlattr *na;
	173	int rep_len;
	174
	175	strcpy(name, TASKSTATS_GENL_NAME);
	176	rc = send_cmd(sd, GENL_ID_CTRL, getpid(), CTRL_CMD_GETFAMILY,
	177	CTRL_ATTR_FAMILY_NAME, (void *)name,
	178	strlen(TASKSTATS_GENL_NAME)+1);
4ed960b1 AM	179	if (rc < 0)
4ed960b1 AM	180	return 0; /* sendto() failure? */
9e06d3f9 SN	181
	182	rep_len = recv(sd, &ans, sizeof(ans), 0);
	183	if (ans.n.nlmsg_type == NLMSG_ERROR \|\|
	184	(rep_len < 0) \|\| !NLMSG_OK((&ans.n), rep_len))
	185	return 0;
a3baf649	186
9e06d3f9 SN	187	na = (struct nlattr *) GENLMSG_DATA(&ans);
	188	na = (struct nlattr ) ((char ) na + NLA_ALIGN(na->nla_len));
	189	if (na->nla_type == CTRL_ATTR_FAMILY_ID) {
	190	id = (__u16 ) NLA_DATA(na);
	191	}
	192	return id;
a3baf649 SN	193	}
a3baf649 SN	194
02d54f09 WF	195	#define average_ms(t, c) (t / 1000000ULL / (c ? c : 1))
02d54f09 WF	196
b7ed698c	197	static void print_delayacct(struct taskstats *t)
a3baf649	198	{
02d54f09 WF	199	printf("\n\nCPU %15s%15s%15s%15s%15s\n"
	200	" %15llu%15llu%15llu%15llu%15.3fms\n"
	201	"IO %15s%15s%15s\n"
	202	" %15llu%15llu%15llums\n"
	203	"SWAP %15s%15s%15s\n"
	204	" %15llu%15llu%15llums\n"
	205	"RECLAIM %12s%15s%15s\n"
b1d29ba8 JW	206	" %15llu%15llu%15llums\n"
b1d29ba8 JW	207	"THRASHING%12s%15s%15s\n"
02d54f09 WF	208	" %15llu%15llu%15llums\n",
	209	"count", "real total", "virtual total",
	210	"delay total", "delay average",
66659313 RD	211	(unsigned long long)t->cpu_count,
	212	(unsigned long long)t->cpu_run_real_total,
	213	(unsigned long long)t->cpu_run_virtual_total,
	214	(unsigned long long)t->cpu_delay_total,
02d54f09 WF	215	average_ms((double)t->cpu_delay_total, t->cpu_count),
02d54f09 WF	216	"count", "delay total", "delay average",
66659313 RD	217	(unsigned long long)t->blkio_count,
66659313 RD	218	(unsigned long long)t->blkio_delay_total,
02d54f09 WF	219	average_ms(t->blkio_delay_total, t->blkio_count),
02d54f09 WF	220	"count", "delay total", "delay average",
66659313 RD	221	(unsigned long long)t->swapin_count,
66659313 RD	222	(unsigned long long)t->swapin_delay_total,
02d54f09 WF	223	average_ms(t->swapin_delay_total, t->swapin_count),
02d54f09 WF	224	"count", "delay total", "delay average",
66659313	225	(unsigned long long)t->freepages_count,
02d54f09	226	(unsigned long long)t->freepages_delay_total,
b1d29ba8 JW	227	average_ms(t->freepages_delay_total, t->freepages_count),
	228	"count", "delay total", "delay average",
	229	(unsigned long long)t->thrashing_count,
	230	(unsigned long long)t->thrashing_delay_total,
	231	average_ms(t->thrashing_delay_total, t->thrashing_count));
a3baf649 SN	232	}
a3baf649 SN	233
b7ed698c	234	static void task_context_switch_counts(struct taskstats *t)
b663a79c MU	235	{
b663a79c MU	236	printf("\n\nTask %15s%15s\n"
10e6f32b	237	" %15llu%15llu\n",
b663a79c	238	"voluntary", "nonvoluntary",
66659313	239	(unsigned long long)t->nvcsw, (unsigned long long)t->nivcsw);
b663a79c MU	240	}
b663a79c MU	241
b7ed698c	242	static void print_cgroupstats(struct cgroupstats *c)
546040dc BS	243	{
546040dc BS	244	printf("sleeping %llu, blocked %llu, running %llu, stopped %llu, "
66659313 RD	245	"uninterruptible %llu\n", (unsigned long long)c->nr_sleeping,
	246	(unsigned long long)c->nr_io_wait,
	247	(unsigned long long)c->nr_running,
	248	(unsigned long long)c->nr_stopped,
	249	(unsigned long long)c->nr_uninterruptible);
546040dc BS	250	}
	251
	252
b7ed698c	253	static void print_ioacct(struct taskstats *t)
cf709844 AM	254	{
	255	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
	256	t->ac_comm,
	257	(unsigned long long)t->read_bytes,
	258	(unsigned long long)t->write_bytes,
	259	(unsigned long long)t->cancelled_write_bytes);
	260	}
	261
a3baf649 SN	262	int main(int argc, char *argv[])
a3baf649 SN	263	{
b8d9a865 JSR	264	int c, rc, rep_len, aggr_len, len2;
b8d9a865 JSR	265	int cmd_type = TASKSTATS_CMD_ATTR_UNSPEC;
9e06d3f9 SN	266	__u16 id;
	267	__u32 mypid;
	268
	269	struct nlattr *na;
	270	int nl_sd = -1;
	271	int len = 0;
	272	pid_t tid = 0;
	273	pid_t rtid = 0;
	274
	275	int fd = 0;
	276	int count = 0;
	277	int write_file = 0;
	278	int maskset = 0;
7f76c403	279	char *logfile = NULL;
9e06d3f9	280	int loop = 0;
546040dc	281	int containerset = 0;
8da01af4	282	char *containerpath = NULL;
546040dc	283	int cfd = 0;
db9e5679 MG	284	int forking = 0;
db9e5679 MG	285	sigset_t sigset;
9e06d3f9 SN	286
	287	struct msgtemplate msg;
	288
db9e5679 MG	289	while (!forking) {
db9e5679 MG	290	c = getopt(argc, argv, "qdiw:r:m:t:p:vlC:c:");
9e06d3f9 SN	291	if (c < 0)
9e06d3f9 SN	292	break;
a3baf649	293
9e06d3f9 SN	294	switch (c) {
	295	case 'd':
	296	printf("print delayacct stats ON\n");
	297	print_delays = 1;
	298	break;
cf709844 AM	299	case 'i':
	300	printf("printing IO accounting\n");
	301	print_io_accounting = 1;
	302	break;
b663a79c MU	303	case 'q':
	304	printf("printing task/process context switch rates\n");
	305	print_task_context_switch_counts = 1;
	306	break;
546040dc BS	307	case 'C':
546040dc BS	308	containerset = 1;
8da01af4	309	containerpath = optarg;
546040dc	310	break;
9e06d3f9	311	case 'w':
7f76c403	312	logfile = strdup(optarg);
9e06d3f9 SN	313	printf("write to file %s\n", logfile);
	314	write_file = 1;
	315	break;
	316	case 'r':
	317	rcvbufsz = atoi(optarg);
	318	printf("receive buf size %d\n", rcvbufsz);
	319	if (rcvbufsz < 0)
	320	err(1, "Invalid rcv buf size\n");
	321	break;
	322	case 'm':
	323	strncpy(cpumask, optarg, sizeof(cpumask));
88e15ce4	324	cpumask[sizeof(cpumask) - 1] = '\0';
9e06d3f9 SN	325	maskset = 1;
	326	printf("cpumask %s maskset %d\n", cpumask, maskset);
	327	break;
	328	case 't':
	329	tid = atoi(optarg);
	330	if (!tid)
	331	err(1, "Invalid tgid\n");
	332	cmd_type = TASKSTATS_CMD_ATTR_TGID;
9e06d3f9 SN	333	break;
	334	case 'p':
	335	tid = atoi(optarg);
	336	if (!tid)
	337	err(1, "Invalid pid\n");
	338	cmd_type = TASKSTATS_CMD_ATTR_PID;
9e06d3f9	339	break;
db9e5679 MG	340	case 'c':
	341
	342	/* Block SIGCHLD for sigwait() later */
	343	if (sigemptyset(&sigset) == -1)
	344	err(1, "Failed to empty sigset");
	345	if (sigaddset(&sigset, SIGCHLD))
	346	err(1, "Failed to set sigchld in sigset");
	347	sigprocmask(SIG_BLOCK, &sigset, NULL);
	348
	349	/* fork/exec a child */
	350	tid = fork();
	351	if (tid < 0)
	352	err(1, "Fork failed\n");
	353	if (tid == 0)
	354	if (execvp(argv[optind - 1],
	355	&argv[optind - 1]) < 0)
	356	exit(-1);
	357
	358	/* Set the command type and avoid further processing */
	359	cmd_type = TASKSTATS_CMD_ATTR_PID;
	360	forking = 1;
	361	break;
9e06d3f9 SN	362	case 'v':
	363	printf("debug on\n");
	364	dbg = 1;
	365	break;
	366	case 'l':
	367	printf("listen forever\n");
	368	loop = 1;
	369	break;
	370	default:
f16825bb	371	usage();
9e06d3f9	372	exit(-1);
a3baf649	373	}
a3baf649	374	}
a3baf649	375
9e06d3f9 SN	376	if (write_file) {
	377	fd = open(logfile, O_WRONLY \| O_CREAT \| O_TRUNC,
	378	S_IRUSR \| S_IWUSR \| S_IRGRP \| S_IROTH);
	379	if (fd == -1) {
	380	perror("Cannot open output file\n");
	381	exit(1);
	382	}
	383	}
a3baf649	384
563c17cb ME	385	nl_sd = create_nl_socket(NETLINK_GENERIC);
563c17cb ME	386	if (nl_sd < 0)
9e06d3f9	387	err(1, "error creating Netlink socket\n");
a3baf649	388
a3baf649	389
9e06d3f9 SN	390	mypid = getpid();
	391	id = get_family_id(nl_sd);
	392	if (!id) {
d2f7bf13	393	fprintf(stderr, "Error getting family id, errno %d\n", errno);
9e06d3f9	394	goto err;
a3baf649	395	}
9e06d3f9 SN	396	PRINTF("family id %d\n", id);
	397
	398	if (maskset) {
	399	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	400	TASKSTATS_CMD_ATTR_REGISTER_CPUMASK,
7d1bdca9	401	&cpumask, strlen(cpumask) + 1);
9e06d3f9 SN	402	PRINTF("Sent register cpumask, retval %d\n", rc);
9e06d3f9 SN	403	if (rc < 0) {
d2f7bf13	404	fprintf(stderr, "error sending register cpumask\n");
9e06d3f9 SN	405	goto err;
9e06d3f9 SN	406	}
a3baf649 SN	407	}
a3baf649 SN	408
546040dc BS	409	if (tid && containerset) {
	410	fprintf(stderr, "Select either -t or -C, not both\n");
	411	goto err;
	412	}
	413
db9e5679 MG	414	/*
	415	* If we forked a child, wait for it to exit. Cannot use waitpid()
	416	* as all the delicious data would be reaped as part of the wait
	417	*/
	418	if (tid && forking) {
	419	int sig_received;
	420	sigwait(&sigset, &sig_received);
	421	}
	422
9e06d3f9 SN	423	if (tid) {
	424	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	425	cmd_type, &tid, sizeof(__u32));
	426	PRINTF("Sent pid/tgid, retval %d\n", rc);
	427	if (rc < 0) {
d2f7bf13	428	fprintf(stderr, "error sending tid/tgid cmd\n");
9e06d3f9 SN	429	goto done;
9e06d3f9 SN	430	}
a3baf649 SN	431	}
a3baf649 SN	432
546040dc BS	433	if (containerset) {
	434	cfd = open(containerpath, O_RDONLY);
	435	if (cfd < 0) {
	436	perror("error opening container file");
	437	goto err;
	438	}
	439	rc = send_cmd(nl_sd, id, mypid, CGROUPSTATS_CMD_GET,
	440	CGROUPSTATS_CMD_ATTR_FD, &cfd, sizeof(__u32));
	441	if (rc < 0) {
	442	perror("error sending cgroupstats command");
	443	goto err;
	444	}
	445	}
65a67bd2 MM	446	if (!maskset && !tid && !containerset) {
	447	usage();
	448	goto err;
	449	}
546040dc	450
9e06d3f9	451	do {
9e06d3f9 SN	452	rep_len = recv(nl_sd, &msg, sizeof(msg), 0);
9e06d3f9 SN	453	PRINTF("received %d bytes\n", rep_len);
a3baf649	454
9e06d3f9	455	if (rep_len < 0) {
d2f7bf13 AM	456	fprintf(stderr, "nonfatal reply error: errno %d\n",
d2f7bf13 AM	457	errno);
9e06d3f9 SN	458	continue;
	459	}
	460	if (msg.n.nlmsg_type == NLMSG_ERROR \|\|
	461	!NLMSG_OK((&msg.n), rep_len)) {
7d1bdca9	462	struct nlmsgerr *err = NLMSG_DATA(&msg);
d2f7bf13 AM	463	fprintf(stderr, "fatal reply error, errno %d\n",
d2f7bf13 AM	464	err->error);
9e06d3f9 SN	465	goto done;
	466	}
	467
10e6f32b	468	PRINTF("nlmsghdr size=%zu, nlmsg_len=%d, rep_len=%d\n",
9e06d3f9 SN	469	sizeof(struct nlmsghdr), msg.n.nlmsg_len, rep_len);
	470
	471
	472	rep_len = GENLMSG_PAYLOAD(&msg.n);
	473
	474	na = (struct nlattr *) GENLMSG_DATA(&msg);
	475	len = 0;
9e06d3f9 SN	476	while (len < rep_len) {
	477	len += NLA_ALIGN(na->nla_len);
	478	switch (na->nla_type) {
	479	case TASKSTATS_TYPE_AGGR_TGID:
	480	/* Fall through */
	481	case TASKSTATS_TYPE_AGGR_PID:
	482	aggr_len = NLA_PAYLOAD(na->nla_len);
	483	len2 = 0;
	484	/* For nested attributes, na follows */
	485	na = (struct nlattr *) NLA_DATA(na);
	486	done = 0;
	487	while (len2 < aggr_len) {
	488	switch (na->nla_type) {
	489	case TASKSTATS_TYPE_PID:
	490	rtid = (int ) NLA_DATA(na);
	491	if (print_delays)
	492	printf("PID\t%d\n", rtid);
	493	break;
	494	case TASKSTATS_TYPE_TGID:
	495	rtid = (int ) NLA_DATA(na);
	496	if (print_delays)
	497	printf("TGID\t%d\n", rtid);
	498	break;
	499	case TASKSTATS_TYPE_STATS:
	500	count++;
	501	if (print_delays)
	502	print_delayacct((struct taskstats *) NLA_DATA(na));
cf709844 AM	503	if (print_io_accounting)
cf709844 AM	504	print_ioacct((struct taskstats *) NLA_DATA(na));
b663a79c MU	505	if (print_task_context_switch_counts)
b663a79c MU	506	task_context_switch_counts((struct taskstats *) NLA_DATA(na));
9e06d3f9 SN	507	if (fd) {
	508	if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
	509	err(1,"write error\n");
	510	}
	511	}
	512	if (!loop)
	513	goto done;
	514	break;
570d8e93 ND	515	case TASKSTATS_TYPE_NULL:
570d8e93 ND	516	break;
9e06d3f9	517	default:
d2f7bf13 AM	518	fprintf(stderr, "Unknown nested"
	519	" nla_type %d\n",
	520	na->nla_type);
9e06d3f9 SN	521	break;
	522	}
	523	len2 += NLA_ALIGN(na->nla_len);
570d8e93 ND	524	na = (struct nlattr )((char )na +
570d8e93 ND	525	NLA_ALIGN(na->nla_len));
9e06d3f9 SN	526	}
	527	break;
	528
546040dc BS	529	case CGROUPSTATS_TYPE_CGROUP_STATS:
	530	print_cgroupstats(NLA_DATA(na));
	531	break;
9e06d3f9	532	default:
d2f7bf13 AM	533	fprintf(stderr, "Unknown nla_type %d\n",
d2f7bf13 AM	534	na->nla_type);
4be2c95d	535	case TASKSTATS_TYPE_NULL:
9e06d3f9	536	break;
a3baf649	537	}
9e06d3f9	538	na = (struct nlattr *) (GENLMSG_DATA(&msg) + len);
a3baf649	539	}
9e06d3f9 SN	540	} while (loop);
	541	done:
	542	if (maskset) {
	543	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	544	TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK,
7d1bdca9	545	&cpumask, strlen(cpumask) + 1);
9e06d3f9 SN	546	printf("Sent deregister mask, retval %d\n", rc);
	547	if (rc < 0)
	548	err(rc, "error sending deregister cpumask\n");
a3baf649	549	}
9e06d3f9 SN	550	err:
	551	close(nl_sd);
	552	if (fd)
	553	close(fd);
546040dc BS	554	if (cfd)
546040dc BS	555	close(cfd);
9e06d3f9	556	return 0;
a3baf649	557	}