[net-next-2.6.git] / Documentation / accounting / getdelays.c

/* 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;
__u64 stime, utime;

#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);
			return -1;
		}

	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);

	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;
}

static void print_delayacct(struct taskstats *t)
{
	printf("\n\nCPU   %15s%15s%15s%15s\n"
	       "      %15llu%15llu%15llu%15llu\n"
	       "IO    %15s%15s\n"
	       "      %15llu%15llu\n"
	       "SWAP  %15s%15s\n"
	       "      %15llu%15llu\n"
	       "RECLAIM  %12s%15s\n"
	       "      %15llu%15llu\n",
	       "count", "real total", "virtual total", "delay total",
	       (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,
	       "count", "delay total",
	       (unsigned long long)t->blkio_count,
	       (unsigned long long)t->blkio_delay_total,
	       "count", "delay total",
	       (unsigned long long)t->swapin_count,
	       (unsigned long long)t->swapin_delay_total,
	       "count", "delay total",
	       (unsigned long long)t->freepages_count,
	       (unsigned long long)t->freepages_delay_total);
}

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[1024];
	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;
			strncpy(containerpath, optarg, strlen(optarg) + 1);
			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));
			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);
		}
	}

	if ((nl_sd = create_nl_socket(NETLINK_GENERIC)) < 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 {
		int i;

		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;
		i = 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;
					default:
						fprintf(stderr, "Unknown nested"
							" nla_type %d\n",
							na->nla_type);
						break;
					}
					len2 += NLA_ALIGN(na->nla_len);
					na = (struct nlattr *) ((char *) na + len2);
				}
				break;

			case CGROUPSTATS_TYPE_CGROUP_STATS:
				print_cgroupstats(NLA_DATA(na));
				break;
			default:
				fprintf(stderr, "Unknown nla_type %d\n",
					na->nla_type);
				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
a3baf649 SN	1	/* getdelays.c
	2	*
	3	* Utility to get per-pid and per-tgid delay accounting statistics
	4	* Also illustrates usage of the taskstats interface
	5	*
	6	* Copyright (C) Shailabh Nagar, IBM Corp. 2005
	7	* Copyright (C) Balbir Singh, IBM Corp. 2006
9e06d3f9	8	* Copyright (c) Jay Lan, SGI. 2006
a3baf649	9	*
d2f7bf13 AM	10	* Compile with
d2f7bf13 AM	11	* gcc -I/usr/src/linux/include getdelays.c -o getdelays
a3baf649 SN	12	*/
	13
	14	#include <stdio.h>
	15	#include <stdlib.h>
	16	#include <errno.h>
	17	#include <unistd.h>
	18	#include <poll.h>
	19	#include <string.h>
	20	#include <fcntl.h>
	21	#include <sys/types.h>
	22	#include <sys/stat.h>
	23	#include <sys/socket.h>
db9e5679	24	#include <sys/wait.h>
a3baf649 SN	25	#include <signal.h>
	26
	27	#include <linux/genetlink.h>
	28	#include <linux/taskstats.h>
546040dc	29	#include <linux/cgroupstats.h>
a3baf649 SN	30
	31	/*
	32	* Generic macros for dealing with netlink sockets. Might be duplicated
	33	* elsewhere. It is recommended that commercial grade applications use
	34	* libnl or libnetlink and use the interfaces provided by the library
	35	*/
	36	#define GENLMSG_DATA(glh) ((void *)(NLMSG_DATA(glh) + GENL_HDRLEN))
	37	#define GENLMSG_PAYLOAD(glh) (NLMSG_PAYLOAD(glh, 0) - GENL_HDRLEN)
	38	#define NLA_DATA(na) ((void )((char)(na) + NLA_HDRLEN))
	39	#define NLA_PAYLOAD(len) (len - NLA_HDRLEN)
	40
d2f7bf13 AM	41	#define err(code, fmt, arg...) \
	42	do { \
	43	fprintf(stderr, fmt, ##arg); \
	44	exit(code); \
	45	} while (0)
	46
	47	int done;
	48	int rcvbufsz;
	49	char name[100];
	50	int dbg;
	51	int print_delays;
cf709844	52	int print_io_accounting;
b663a79c	53	int print_task_context_switch_counts;
9e06d3f9	54	__u64 stime, utime;
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) {
d2f7bf13 AM	101	fprintf(stderr, "Unable to set socket rcv buf size "
	102	"to %d\n",
	103	rcvbufsz);
9e06d3f9 SN	104	return -1;
9e06d3f9 SN	105	}
a3baf649	106
9e06d3f9 SN	107	memset(&local, 0, sizeof(local));
9e06d3f9 SN	108	local.nl_family = AF_NETLINK;
a3baf649	109
9e06d3f9 SN	110	if (bind(fd, (struct sockaddr *) &local, sizeof(local)) < 0)
9e06d3f9 SN	111	goto error;
a3baf649	112
9e06d3f9 SN	113	return fd;
	114	error:
	115	close(fd);
	116	return -1;
a3baf649 SN	117	}
a3baf649 SN	118
9e06d3f9	119
b7ed698c	120	static int send_cmd(int sd, __u16 nlmsg_type, __u32 nlmsg_pid,
9e06d3f9 SN	121	__u8 genl_cmd, __u16 nla_type,
9e06d3f9 SN	122	void *nla_data, int nla_len)
a3baf649	123	{
9e06d3f9 SN	124	struct nlattr *na;
	125	struct sockaddr_nl nladdr;
	126	int r, buflen;
	127	char *buf;
	128
	129	struct msgtemplate msg;
	130
	131	msg.n.nlmsg_len = NLMSG_LENGTH(GENL_HDRLEN);
	132	msg.n.nlmsg_type = nlmsg_type;
	133	msg.n.nlmsg_flags = NLM_F_REQUEST;
	134	msg.n.nlmsg_seq = 0;
	135	msg.n.nlmsg_pid = nlmsg_pid;
	136	msg.g.cmd = genl_cmd;
	137	msg.g.version = 0x1;
	138	na = (struct nlattr *) GENLMSG_DATA(&msg);
	139	na->nla_type = nla_type;
	140	na->nla_len = nla_len + 1 + NLA_HDRLEN;
	141	memcpy(NLA_DATA(na), nla_data, nla_len);
	142	msg.n.nlmsg_len += NLMSG_ALIGN(na->nla_len);
	143
	144	buf = (char *) &msg;
	145	buflen = msg.n.nlmsg_len ;
	146	memset(&nladdr, 0, sizeof(nladdr));
	147	nladdr.nl_family = AF_NETLINK;
	148	while ((r = sendto(sd, buf, buflen, 0, (struct sockaddr *) &nladdr,
	149	sizeof(nladdr))) < buflen) {
	150	if (r > 0) {
	151	buf += r;
	152	buflen -= r;
	153	} else if (errno != EAGAIN)
	154	return -1;
	155	}
	156	return 0;
a3baf649 SN	157	}
a3baf649 SN	158
9e06d3f9	159
a3baf649 SN	160	/*
	161	* Probe the controller in genetlink to find the family id
	162	* for the TASKSTATS family
	163	*/
b7ed698c	164	static int get_family_id(int sd)
a3baf649	165	{
9e06d3f9 SN	166	struct {
	167	struct nlmsghdr n;
	168	struct genlmsghdr g;
	169	char buf[256];
	170	} ans;
	171
10e6f32b	172	int id = 0, rc;
9e06d3f9 SN	173	struct nlattr *na;
	174	int rep_len;
	175
	176	strcpy(name, TASKSTATS_GENL_NAME);
	177	rc = send_cmd(sd, GENL_ID_CTRL, getpid(), CTRL_CMD_GETFAMILY,
	178	CTRL_ATTR_FAMILY_NAME, (void *)name,
	179	strlen(TASKSTATS_GENL_NAME)+1);
	180
	181	rep_len = recv(sd, &ans, sizeof(ans), 0);
	182	if (ans.n.nlmsg_type == NLMSG_ERROR \|\|
	183	(rep_len < 0) \|\| !NLMSG_OK((&ans.n), rep_len))
	184	return 0;
a3baf649	185
9e06d3f9 SN	186	na = (struct nlattr *) GENLMSG_DATA(&ans);
	187	na = (struct nlattr ) ((char ) na + NLA_ALIGN(na->nla_len));
	188	if (na->nla_type == CTRL_ATTR_FAMILY_ID) {
	189	id = (__u16 ) NLA_DATA(na);
	190	}
	191	return id;
a3baf649 SN	192	}
a3baf649 SN	193
b7ed698c	194	static void print_delayacct(struct taskstats *t)
a3baf649	195	{
9e06d3f9 SN	196	printf("\n\nCPU %15s%15s%15s%15s\n"
	197	" %15llu%15llu%15llu%15llu\n"
	198	"IO %15s%15s\n"
	199	" %15llu%15llu\n"
9b0975a2 KK	200	"SWAP %15s%15s\n"
	201	" %15llu%15llu\n"
	202	"RECLAIM %12s%15s\n"
efbee7f1	203	" %15llu%15llu\n",
9e06d3f9	204	"count", "real total", "virtual total", "delay total",
66659313 RD	205	(unsigned long long)t->cpu_count,
	206	(unsigned long long)t->cpu_run_real_total,
	207	(unsigned long long)t->cpu_run_virtual_total,
	208	(unsigned long long)t->cpu_delay_total,
9e06d3f9	209	"count", "delay total",
66659313 RD	210	(unsigned long long)t->blkio_count,
66659313 RD	211	(unsigned long long)t->blkio_delay_total,
9b0975a2	212	"count", "delay total",
66659313 RD	213	(unsigned long long)t->swapin_count,
	214	(unsigned long long)t->swapin_delay_total,
	215	"count", "delay total",
	216	(unsigned long long)t->freepages_count,
	217	(unsigned long long)t->freepages_delay_total);
a3baf649 SN	218	}
a3baf649 SN	219
b7ed698c	220	static void task_context_switch_counts(struct taskstats *t)
b663a79c MU	221	{
b663a79c MU	222	printf("\n\nTask %15s%15s\n"
10e6f32b	223	" %15llu%15llu\n",
b663a79c	224	"voluntary", "nonvoluntary",
66659313	225	(unsigned long long)t->nvcsw, (unsigned long long)t->nivcsw);
b663a79c MU	226	}
b663a79c MU	227
b7ed698c	228	static void print_cgroupstats(struct cgroupstats *c)
546040dc BS	229	{
546040dc BS	230	printf("sleeping %llu, blocked %llu, running %llu, stopped %llu, "
66659313 RD	231	"uninterruptible %llu\n", (unsigned long long)c->nr_sleeping,
	232	(unsigned long long)c->nr_io_wait,
	233	(unsigned long long)c->nr_running,
	234	(unsigned long long)c->nr_stopped,
	235	(unsigned long long)c->nr_uninterruptible);
546040dc BS	236	}
	237
	238
b7ed698c	239	static void print_ioacct(struct taskstats *t)
cf709844 AM	240	{
	241	printf("%s: read=%llu, write=%llu, cancelled_write=%llu\n",
	242	t->ac_comm,
	243	(unsigned long long)t->read_bytes,
	244	(unsigned long long)t->write_bytes,
	245	(unsigned long long)t->cancelled_write_bytes);
	246	}
	247
a3baf649 SN	248	int main(int argc, char *argv[])
a3baf649 SN	249	{
b8d9a865 JSR	250	int c, rc, rep_len, aggr_len, len2;
b8d9a865 JSR	251	int cmd_type = TASKSTATS_CMD_ATTR_UNSPEC;
9e06d3f9 SN	252	__u16 id;
	253	__u32 mypid;
	254
	255	struct nlattr *na;
	256	int nl_sd = -1;
	257	int len = 0;
	258	pid_t tid = 0;
	259	pid_t rtid = 0;
	260
	261	int fd = 0;
	262	int count = 0;
	263	int write_file = 0;
	264	int maskset = 0;
7f76c403	265	char *logfile = NULL;
9e06d3f9	266	int loop = 0;
546040dc BS	267	int containerset = 0;
	268	char containerpath[1024];
	269	int cfd = 0;
db9e5679 MG	270	int forking = 0;
db9e5679 MG	271	sigset_t sigset;
9e06d3f9 SN	272
	273	struct msgtemplate msg;
	274
db9e5679 MG	275	while (!forking) {
db9e5679 MG	276	c = getopt(argc, argv, "qdiw:r:m:t:p:vlC:c:");
9e06d3f9 SN	277	if (c < 0)
9e06d3f9 SN	278	break;
a3baf649	279
9e06d3f9 SN	280	switch (c) {
	281	case 'd':
	282	printf("print delayacct stats ON\n");
	283	print_delays = 1;
	284	break;
cf709844 AM	285	case 'i':
	286	printf("printing IO accounting\n");
	287	print_io_accounting = 1;
	288	break;
b663a79c MU	289	case 'q':
	290	printf("printing task/process context switch rates\n");
	291	print_task_context_switch_counts = 1;
	292	break;
546040dc BS	293	case 'C':
	294	containerset = 1;
	295	strncpy(containerpath, optarg, strlen(optarg) + 1);
	296	break;
9e06d3f9	297	case 'w':
7f76c403	298	logfile = strdup(optarg);
9e06d3f9 SN	299	printf("write to file %s\n", logfile);
	300	write_file = 1;
	301	break;
	302	case 'r':
	303	rcvbufsz = atoi(optarg);
	304	printf("receive buf size %d\n", rcvbufsz);
	305	if (rcvbufsz < 0)
	306	err(1, "Invalid rcv buf size\n");
	307	break;
	308	case 'm':
	309	strncpy(cpumask, optarg, sizeof(cpumask));
	310	maskset = 1;
	311	printf("cpumask %s maskset %d\n", cpumask, maskset);
	312	break;
	313	case 't':
	314	tid = atoi(optarg);
	315	if (!tid)
	316	err(1, "Invalid tgid\n");
	317	cmd_type = TASKSTATS_CMD_ATTR_TGID;
9e06d3f9 SN	318	break;
	319	case 'p':
	320	tid = atoi(optarg);
	321	if (!tid)
	322	err(1, "Invalid pid\n");
	323	cmd_type = TASKSTATS_CMD_ATTR_PID;
9e06d3f9	324	break;
db9e5679 MG	325	case 'c':
	326
	327	/* Block SIGCHLD for sigwait() later */
	328	if (sigemptyset(&sigset) == -1)
	329	err(1, "Failed to empty sigset");
	330	if (sigaddset(&sigset, SIGCHLD))
	331	err(1, "Failed to set sigchld in sigset");
	332	sigprocmask(SIG_BLOCK, &sigset, NULL);
	333
	334	/* fork/exec a child */
	335	tid = fork();
	336	if (tid < 0)
	337	err(1, "Fork failed\n");
	338	if (tid == 0)
	339	if (execvp(argv[optind - 1],
	340	&argv[optind - 1]) < 0)
	341	exit(-1);
	342
	343	/* Set the command type and avoid further processing */
	344	cmd_type = TASKSTATS_CMD_ATTR_PID;
	345	forking = 1;
	346	break;
9e06d3f9 SN	347	case 'v':
	348	printf("debug on\n");
	349	dbg = 1;
	350	break;
	351	case 'l':
	352	printf("listen forever\n");
	353	loop = 1;
	354	break;
	355	default:
f16825bb	356	usage();
9e06d3f9	357	exit(-1);
a3baf649	358	}
a3baf649	359	}
a3baf649	360
9e06d3f9 SN	361	if (write_file) {
	362	fd = open(logfile, O_WRONLY \| O_CREAT \| O_TRUNC,
	363	S_IRUSR \| S_IWUSR \| S_IRGRP \| S_IROTH);
	364	if (fd == -1) {
	365	perror("Cannot open output file\n");
	366	exit(1);
	367	}
	368	}
a3baf649	369
9e06d3f9 SN	370	if ((nl_sd = create_nl_socket(NETLINK_GENERIC)) < 0)
9e06d3f9 SN	371	err(1, "error creating Netlink socket\n");
a3baf649	372
a3baf649	373
9e06d3f9 SN	374	mypid = getpid();
	375	id = get_family_id(nl_sd);
	376	if (!id) {
d2f7bf13	377	fprintf(stderr, "Error getting family id, errno %d\n", errno);
9e06d3f9	378	goto err;
a3baf649	379	}
9e06d3f9 SN	380	PRINTF("family id %d\n", id);
	381
	382	if (maskset) {
	383	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	384	TASKSTATS_CMD_ATTR_REGISTER_CPUMASK,
7d1bdca9	385	&cpumask, strlen(cpumask) + 1);
9e06d3f9 SN	386	PRINTF("Sent register cpumask, retval %d\n", rc);
9e06d3f9 SN	387	if (rc < 0) {
d2f7bf13	388	fprintf(stderr, "error sending register cpumask\n");
9e06d3f9 SN	389	goto err;
9e06d3f9 SN	390	}
a3baf649 SN	391	}
a3baf649 SN	392
546040dc BS	393	if (tid && containerset) {
	394	fprintf(stderr, "Select either -t or -C, not both\n");
	395	goto err;
	396	}
	397
db9e5679 MG	398	/*
	399	* If we forked a child, wait for it to exit. Cannot use waitpid()
	400	* as all the delicious data would be reaped as part of the wait
	401	*/
	402	if (tid && forking) {
	403	int sig_received;
	404	sigwait(&sigset, &sig_received);
	405	}
	406
9e06d3f9 SN	407	if (tid) {
	408	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	409	cmd_type, &tid, sizeof(__u32));
	410	PRINTF("Sent pid/tgid, retval %d\n", rc);
	411	if (rc < 0) {
d2f7bf13	412	fprintf(stderr, "error sending tid/tgid cmd\n");
9e06d3f9 SN	413	goto done;
9e06d3f9 SN	414	}
a3baf649 SN	415	}
a3baf649 SN	416
546040dc BS	417	if (containerset) {
	418	cfd = open(containerpath, O_RDONLY);
	419	if (cfd < 0) {
	420	perror("error opening container file");
	421	goto err;
	422	}
	423	rc = send_cmd(nl_sd, id, mypid, CGROUPSTATS_CMD_GET,
	424	CGROUPSTATS_CMD_ATTR_FD, &cfd, sizeof(__u32));
	425	if (rc < 0) {
	426	perror("error sending cgroupstats command");
	427	goto err;
	428	}
	429	}
65a67bd2 MM	430	if (!maskset && !tid && !containerset) {
	431	usage();
	432	goto err;
	433	}
546040dc	434
9e06d3f9 SN	435	do {
9e06d3f9 SN	436	int i;
a3baf649	437
9e06d3f9 SN	438	rep_len = recv(nl_sd, &msg, sizeof(msg), 0);
9e06d3f9 SN	439	PRINTF("received %d bytes\n", rep_len);
a3baf649	440
9e06d3f9	441	if (rep_len < 0) {
d2f7bf13 AM	442	fprintf(stderr, "nonfatal reply error: errno %d\n",
d2f7bf13 AM	443	errno);
9e06d3f9 SN	444	continue;
	445	}
	446	if (msg.n.nlmsg_type == NLMSG_ERROR \|\|
	447	!NLMSG_OK((&msg.n), rep_len)) {
7d1bdca9	448	struct nlmsgerr *err = NLMSG_DATA(&msg);
d2f7bf13 AM	449	fprintf(stderr, "fatal reply error, errno %d\n",
d2f7bf13 AM	450	err->error);
9e06d3f9 SN	451	goto done;
	452	}
	453
10e6f32b	454	PRINTF("nlmsghdr size=%zu, nlmsg_len=%d, rep_len=%d\n",
9e06d3f9 SN	455	sizeof(struct nlmsghdr), msg.n.nlmsg_len, rep_len);
	456
	457
	458	rep_len = GENLMSG_PAYLOAD(&msg.n);
	459
	460	na = (struct nlattr *) GENLMSG_DATA(&msg);
	461	len = 0;
	462	i = 0;
	463	while (len < rep_len) {
	464	len += NLA_ALIGN(na->nla_len);
	465	switch (na->nla_type) {
	466	case TASKSTATS_TYPE_AGGR_TGID:
	467	/* Fall through */
	468	case TASKSTATS_TYPE_AGGR_PID:
	469	aggr_len = NLA_PAYLOAD(na->nla_len);
	470	len2 = 0;
	471	/* For nested attributes, na follows */
	472	na = (struct nlattr *) NLA_DATA(na);
	473	done = 0;
	474	while (len2 < aggr_len) {
	475	switch (na->nla_type) {
	476	case TASKSTATS_TYPE_PID:
	477	rtid = (int ) NLA_DATA(na);
	478	if (print_delays)
	479	printf("PID\t%d\n", rtid);
	480	break;
	481	case TASKSTATS_TYPE_TGID:
	482	rtid = (int ) NLA_DATA(na);
	483	if (print_delays)
	484	printf("TGID\t%d\n", rtid);
	485	break;
	486	case TASKSTATS_TYPE_STATS:
	487	count++;
	488	if (print_delays)
	489	print_delayacct((struct taskstats *) NLA_DATA(na));
cf709844 AM	490	if (print_io_accounting)
cf709844 AM	491	print_ioacct((struct taskstats *) NLA_DATA(na));
b663a79c MU	492	if (print_task_context_switch_counts)
b663a79c MU	493	task_context_switch_counts((struct taskstats *) NLA_DATA(na));
9e06d3f9 SN	494	if (fd) {
	495	if (write(fd, NLA_DATA(na), na->nla_len) < 0) {
	496	err(1,"write error\n");
	497	}
	498	}
	499	if (!loop)
	500	goto done;
	501	break;
	502	default:
d2f7bf13 AM	503	fprintf(stderr, "Unknown nested"
	504	" nla_type %d\n",
	505	na->nla_type);
9e06d3f9 SN	506	break;
	507	}
	508	len2 += NLA_ALIGN(na->nla_len);
	509	na = (struct nlattr ) ((char ) na + len2);
	510	}
	511	break;
	512
546040dc BS	513	case CGROUPSTATS_TYPE_CGROUP_STATS:
	514	print_cgroupstats(NLA_DATA(na));
	515	break;
9e06d3f9	516	default:
d2f7bf13 AM	517	fprintf(stderr, "Unknown nla_type %d\n",
d2f7bf13 AM	518	na->nla_type);
9e06d3f9	519	break;
a3baf649	520	}
9e06d3f9	521	na = (struct nlattr *) (GENLMSG_DATA(&msg) + len);
a3baf649	522	}
9e06d3f9 SN	523	} while (loop);
	524	done:
	525	if (maskset) {
	526	rc = send_cmd(nl_sd, id, mypid, TASKSTATS_CMD_GET,
	527	TASKSTATS_CMD_ATTR_DEREGISTER_CPUMASK,
7d1bdca9	528	&cpumask, strlen(cpumask) + 1);
9e06d3f9 SN	529	printf("Sent deregister mask, retval %d\n", rc);
	530	if (rc < 0)
	531	err(rc, "error sending deregister cpumask\n");
a3baf649	532	}
9e06d3f9 SN	533	err:
	534	close(nl_sd);
	535	if (fd)
	536	close(fd);
546040dc BS	537	if (cfd)
546040dc BS	538	close(cfd);
9e06d3f9	539	return 0;
a3baf649	540	}