Merging upstream version 6.7.7.

Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>

3 files changed, 193 insertions, 42 deletions

diff --git a/tools/perf/bench/sched-messaging.c b/tools/perf/bench/sched-messaging.c
index fa1f8f9988..93dcd9dba3 100644
--- a/tools/perf/bench/sched-messaging.c
+++ b/tools/perf/bench/sched-messaging.c
@@ -36,6 +36,7 @@ static bool use_pipes = false;
 static unsigned int nr_loops = 100;
 static bool thread_mode = false;
 static unsigned int num_groups = 10;
+static unsigned int total_children = 0;
 static struct list_head sender_contexts = LIST_HEAD_INIT(sender_contexts);
 static struct list_head receiver_contexts = LIST_HEAD_INIT(receiver_contexts);
 
@@ -55,6 +56,13 @@ struct receiver_context {
 	int wakefd;
 };
 
+union messaging_worker {
+	pthread_t thread;
+	pid_t pid;
+};
+
+static union messaging_worker *worker_tab;
+
 static void fdpair(int fds[2])
 {
 	if (use_pipes) {
@@ -98,7 +106,7 @@ static void *sender(struct sender_context *ctx)
 
 again:
 			ret = write(ctx->out_fds[j], data + done,
-				    sizeof(data)-done);
+				    sizeof(data) - done);
 			if (ret < 0)
 				err(EXIT_FAILURE, "SENDER: write");
 			done += ret;
@@ -139,30 +147,12 @@ again:
 	return NULL;
 }
 
-static pthread_t create_worker(void *ctx, void *(*func)(void *))
+static void create_thread_worker(union messaging_worker *worker,
+				 void *ctx, void *(*func)(void *))
 {
 	pthread_attr_t attr;
-	pthread_t childid;
 	int ret;
 
-	if (!thread_mode) {
-		/* process mode */
-		/* Fork the receiver. */
-		switch (fork()) {
-		case -1:
-			err(EXIT_FAILURE, "fork()");
-			break;
-		case 0:
-			(*func) (ctx);
-			exit(0);
-			break;
-		default:
-			break;
-		}
-
-		return (pthread_t)0;
-	}
-
 	if (pthread_attr_init(&attr) != 0)
 		err(EXIT_FAILURE, "pthread_attr_init:");
 
@@ -171,15 +161,37 @@ static pthread_t create_worker(void *ctx, void *(*func)(void *))
 		err(EXIT_FAILURE, "pthread_attr_setstacksize");
 #endif
 
-	ret = pthread_create(&childid, &attr, func, ctx);
+	ret = pthread_create(&worker->thread, &attr, func, ctx);
 	if (ret != 0)
 		err(EXIT_FAILURE, "pthread_create failed");
 
 	pthread_attr_destroy(&attr);
-	return childid;
 }
 
-static void reap_worker(pthread_t id)
+static void create_process_worker(union messaging_worker *worker,
+				  void *ctx, void *(*func)(void *))
+{
+	/* Fork the receiver. */
+	worker->pid = fork();
+
+	if (worker->pid == -1) {
+		err(EXIT_FAILURE, "fork()");
+	} else if (worker->pid == 0) {
+		(*func) (ctx);
+		exit(0);
+	}
+}
+
+static void create_worker(union messaging_worker *worker,
+			  void *ctx, void *(*func)(void *))
+{
+	if (!thread_mode)
+		return create_process_worker(worker, ctx, func);
+	else
+		return create_thread_worker(worker, ctx, func);
+}
+
+static void reap_worker(union messaging_worker *worker)
 {
 	int proc_status;
 	void *thread_status;
@@ -190,19 +202,19 @@ static void reap_worker(pthread_t id)
 		if (!WIFEXITED(proc_status))
 			exit(1);
 	} else {
-		pthread_join(id, &thread_status);
+		pthread_join(worker->thread, &thread_status);
 	}
 }
 
 /* One group of senders and receivers */
-static unsigned int group(pthread_t *pth,
+static unsigned int group(union messaging_worker *worker,
 		unsigned int num_fds,
 		int ready_out,
 		int wakefd)
 {
 	unsigned int i;
-	struct sender_context *snd_ctx = malloc(sizeof(struct sender_context)
-			+ num_fds * sizeof(int));
+	struct sender_context *snd_ctx = malloc(sizeof(struct sender_context) +
+						num_fds * sizeof(int));
 
 	if (!snd_ctx)
 		err(EXIT_FAILURE, "malloc()");
@@ -226,7 +238,7 @@ static unsigned int group(pthread_t *pth,
 		ctx->ready_out = ready_out;
 		ctx->wakefd = wakefd;
 
-		pth[i] = create_worker(ctx, (void *)receiver);
+		create_worker(worker + i, ctx, (void *)receiver);
 
 		snd_ctx->out_fds[i] = fds[1];
 		if (!thread_mode)
@@ -239,7 +251,7 @@ static unsigned int group(pthread_t *pth,
 		snd_ctx->wakefd = wakefd;
 		snd_ctx->num_fds = num_fds;
 
-		pth[num_fds+i] = create_worker(snd_ctx, (void *)sender);
+		create_worker(worker + num_fds + i, snd_ctx, (void *)sender);
 	}
 
 	/* Close the fds we have left */
@@ -251,6 +263,17 @@ static unsigned int group(pthread_t *pth,
 	return num_fds * 2;
 }
 
+static void sig_handler(int sig __maybe_unused)
+{
+	unsigned int i;
+
+	/*
+	 * When exit abnormally, kill all forked child processes.
+	 */
+	for (i = 0; i < total_children; i++)
+		kill(worker_tab[i].pid, SIGKILL);
+}
+
 static const struct option options[] = {
 	OPT_BOOLEAN('p', "pipe", &use_pipes,
 		    "Use pipe() instead of socketpair()"),
@@ -268,27 +291,30 @@ static const char * const bench_sched_message_usage[] = {
 
 int bench_sched_messaging(int argc, const char **argv)
 {
-	unsigned int i, total_children;
+	unsigned int i;
 	struct timeval start, stop, diff;
 	unsigned int num_fds = 20;
 	int readyfds[2], wakefds[2];
 	char dummy;
-	pthread_t *pth_tab;
 	struct sender_context *pos, *n;
 
 	argc = parse_options(argc, argv, options,
 			     bench_sched_message_usage, 0);
 
-	pth_tab = malloc(num_fds * 2 * num_groups * sizeof(pthread_t));
-	if (!pth_tab)
+	worker_tab = malloc(num_fds * 2 * num_groups * sizeof(union messaging_worker));
+	if (!worker_tab)
 		err(EXIT_FAILURE, "main:malloc()");
 
 	fdpair(readyfds);
 	fdpair(wakefds);
 
-	total_children = 0;
+	if (!thread_mode) {
+		signal(SIGINT, sig_handler);
+		signal(SIGTERM, sig_handler);
+	}
+
 	for (i = 0; i < num_groups; i++)
-		total_children += group(pth_tab+total_children, num_fds,
+		total_children += group(worker_tab + total_children, num_fds,
 					readyfds[1], wakefds[0]);
 
 	/* Wait for everyone to be ready */
@@ -304,7 +330,7 @@ int bench_sched_messaging(int argc, const char **argv)
 
 	/* Reap them all */
 	for (i = 0; i < total_children; i++)
-		reap_worker(pth_tab[i]);
+		reap_worker(worker_tab + i);
 
 	gettimeofday(&stop, NULL);
 
@@ -332,7 +358,7 @@ int bench_sched_messaging(int argc, const char **argv)
 		break;
 	}
 
-	free(pth_tab);
+	free(worker_tab);
 	list_for_each_entry_safe(pos, n, &sender_contexts, list) {
 		list_del_init(&pos->list);
 		free(pos);
diff --git a/tools/perf/bench/sched-pipe.c b/tools/perf/bench/sched-pipe.c
index a960e7a93a..3af6d3c55a 100644
--- a/tools/perf/bench/sched-pipe.c
+++ b/tools/perf/bench/sched-pipe.c
@@ -10,7 +10,9 @@
  * Ported to perf by Hitoshi Mitake <mitake@dcl.info.waseda.ac.jp>
  */
 #include <subcmd/parse-options.h>
+#include <api/fs/fs.h>
 #include "bench.h"
+#include "util/cgroup.h"
 
 #include <unistd.h>
 #include <stdio.h>
@@ -19,6 +21,7 @@
 #include <sys/wait.h>
 #include <string.h>
 #include <errno.h>
+#include <fcntl.h>
 #include <assert.h>
 #include <sys/time.h>
 #include <sys/types.h>
@@ -31,6 +34,7 @@ struct thread_data {
 	int			nr;
 	int			pipe_read;
 	int			pipe_write;
+	bool			cgroup_failed;
 	pthread_t		pthread;
 };
 
@@ -40,9 +44,48 @@ static	int			loops = LOOPS_DEFAULT;
 /* Use processes by default: */
 static bool			threaded;
 
+static char			*cgrp_names[2];
+static struct cgroup		*cgrps[2];
+
+static int parse_two_cgroups(const struct option *opt __maybe_unused,
+			     const char *str, int unset __maybe_unused)
+{
+	char *p = strdup(str);
+	char *q;
+	int ret = -1;
+
+	if (p == NULL) {
+		fprintf(stderr, "memory allocation failure\n");
+		return -1;
+	}
+
+	q = strchr(p, ',');
+	if (q == NULL) {
+		fprintf(stderr, "it should have two cgroup names: %s\n", p);
+		goto out;
+	}
+	*q = '\0';
+
+	cgrp_names[0] = strdup(p);
+	cgrp_names[1] = strdup(q + 1);
+
+	if (cgrp_names[0] == NULL || cgrp_names[1] == NULL) {
+		fprintf(stderr, "memory allocation failure\n");
+		goto out;
+	}
+	ret = 0;
+
+out:
+	free(p);
+	return ret;
+}
+
 static const struct option options[] = {
 	OPT_INTEGER('l', "loop",	&loops,		"Specify number of loops"),
 	OPT_BOOLEAN('T', "threaded",	&threaded,	"Specify threads/process based task setup"),
+	OPT_CALLBACK('G', "cgroups", NULL, "SEND,RECV",
+		     "Put sender and receivers in given cgroups",
+		     parse_two_cgroups),
 	OPT_END()
 };
 
@@ -51,12 +94,89 @@ static const char * const bench_sched_pipe_usage[] = {
 	NULL
 };
 
+static int enter_cgroup(int nr)
+{
+	char buf[32];
+	int fd, len, ret;
+	int saved_errno;
+	struct cgroup *cgrp;
+	pid_t pid;
+
+	if (cgrp_names[nr] == NULL)
+		return 0;
+
+	if (cgrps[nr] == NULL) {
+		cgrps[nr] = cgroup__new(cgrp_names[nr], /*do_open=*/true);
+		if (cgrps[nr] == NULL)
+			goto err;
+	}
+	cgrp = cgrps[nr];
+
+	if (threaded)
+		pid = syscall(__NR_gettid);
+	else
+		pid = getpid();
+
+	snprintf(buf, sizeof(buf), "%d\n", pid);
+	len = strlen(buf);
+
+	/* try cgroup v2 interface first */
+	if (threaded)
+		fd = openat(cgrp->fd, "cgroup.threads", O_WRONLY);
+	else
+		fd = openat(cgrp->fd, "cgroup.procs", O_WRONLY);
+
+	/* try cgroup v1 if failed */
+	if (fd < 0 && errno == ENOENT)
+		fd = openat(cgrp->fd, "tasks", O_WRONLY);
+
+	if (fd < 0)
+		goto err;
+
+	ret = write(fd, buf, len);
+	close(fd);
+
+	if (ret != len) {
+		printf("Cannot enter to cgroup: %s\n", cgrp->name);
+		return -1;
+	}
+	return 0;
+
+err:
+	saved_errno = errno;
+	printf("Failed to open cgroup file in %s\n", cgrp_names[nr]);
+
+	if (saved_errno == ENOENT) {
+		char mnt[PATH_MAX];
+
+		if (cgroupfs_find_mountpoint(mnt, sizeof(mnt), "perf_event") == 0)
+			printf(" Hint: create the cgroup first, like 'mkdir %s/%s'\n",
+			       mnt, cgrp_names[nr]);
+	} else if (saved_errno == EACCES && geteuid() > 0) {
+		printf(" Hint: try to run as root\n");
+	}
+
+	return -1;
+}
+
+static void exit_cgroup(int nr)
+{
+	cgroup__put(cgrps[nr]);
+	free(cgrp_names[nr]);
+}
+
 static void *worker_thread(void *__tdata)
 {
 	struct thread_data *td = __tdata;
 	int m = 0, i;
 	int ret;
 
+	ret = enter_cgroup(td->nr);
+	if (ret < 0) {
+		td->cgroup_failed = true;
+		return NULL;
+	}
+
 	for (i = 0; i < loops; i++) {
 		if (!td->nr) {
 			ret = read(td->pipe_read, &m, sizeof(int));
@@ -76,7 +196,8 @@ static void *worker_thread(void *__tdata)
 
 int bench_sched_pipe(int argc, const char **argv)
 {
-	struct thread_data threads[2], *td;
+	struct thread_data threads[2] = {};
+	struct thread_data *td;
 	int pipe_1[2], pipe_2[2];
 	struct timeval start, stop, diff;
 	unsigned long long result_usec = 0;
@@ -112,9 +233,7 @@ int bench_sched_pipe(int argc, const char **argv)
 		}
 	}
 
-
 	if (threaded) {
-
 		for (t = 0; t < nr_threads; t++) {
 			td = threads + t;
 
@@ -128,7 +247,6 @@ int bench_sched_pipe(int argc, const char **argv)
 			ret = pthread_join(td->pthread, NULL);
 			BUG_ON(ret);
 		}
-
 	} else {
 		pid = fork();
 		assert(pid >= 0);
@@ -147,6 +265,12 @@ int bench_sched_pipe(int argc, const char **argv)
 	gettimeofday(&stop, NULL);
 	timersub(&stop, &start, &diff);
 
+	exit_cgroup(0);
+	exit_cgroup(1);
+
+	if (threads[0].cgroup_failed || threads[1].cgroup_failed)
+		return 0;
+
 	switch (bench_format) {
 	case BENCH_FORMAT_DEFAULT:
 		printf("# Executed %d pipe operations between two %s\n\n",
diff --git a/tools/perf/bench/uprobe.c b/tools/perf/bench/uprobe.c
index 914c0817fe..5c71fdc419 100644
--- a/tools/perf/bench/uprobe.c
+++ b/tools/perf/bench/uprobe.c
@@ -89,6 +89,7 @@ static int bench_uprobe__setup_bpf_skel(enum bench_uprobe bench)
 	return err;
 cleanup:
 	bench_uprobe_bpf__destroy(skel);
+	skel = NULL;
 	return err;
 }