Merge branch 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip

* 'perf-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip: (28 commits)
  perf session: Fix infinite loop in __perf_session__process_events
  perf evsel: Support perf_evsel__open(cpus > 1 && threads > 1)
  perf sched: Use PTHREAD_STACK_MIN to avoid pthread_attr_setstacksize() fail
  perf tools: Emit clearer message for sys_perf_event_open ENOENT return
  perf stat: better error message for unsupported events
  perf sched: Fix allocation result check
  perf, x86: P4 PMU - Fix unflagged overflows handling
  dynamic debug: Fix build issue with older gcc
  tracing: Fix TRACE_EVENT power tracepoint creation
  tracing: Fix preempt count leak
  tracepoint: Add __rcu annotation
  tracing: remove duplicate null-pointer check in skb tracepoint
  tracing/trivial: Add missing comma in TRACE_EVENT comment
  tracing: Include module.h in define_trace.h
  x86: Save rbp in pt_regs on irq entry
  x86, dumpstack: Fix unused variable warning
  x86, NMI: Clean-up default_do_nmi()
  x86, NMI: Allow NMI reason io port (0x61) to be processed on any CPU
  x86, NMI: Remove DIE_NMI_IPI
  x86, NMI: Add priorities to handlers
  ...

tools/perf/Makefile

@@ -227,7 +227,7 @@ ifndef PERF_DEBUG
   CFLAGS_OPTIMIZE = -O6
 endif
 
-CFLAGS = -ggdb3 -Wall -Wextra -std=gnu99 -Werror $(CFLAGS_OPTIMIZE) -D_FORTIFY_SOURCE=2 $(EXTRA_WARNINGS) $(EXTRA_CFLAGS)
+CFLAGS = -fno-omit-frame-pointer -ggdb3 -Wall -Wextra -std=gnu99 -Werror $(CFLAGS_OPTIMIZE) -D_FORTIFY_SOURCE=2 $(EXTRA_WARNINGS) $(EXTRA_CFLAGS)
 EXTLIBS = -lpthread -lrt -lelf -lm
 ALL_CFLAGS = $(CFLAGS) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64
 ALL_LDFLAGS = $(LDFLAGS)

tools/perf/builtin-record.c

@@ -331,6 +331,9 @@ try_again:
 			else if (err ==  ENODEV && cpu_list) {
 				die("No such device - did you specify"
 					" an out-of-range profile CPU?\n");
+			} else if (err == ENOENT) {
+				die("%s event is not supported. ",
+				     event_name(evsel));
 			} else if (err == EINVAL && sample_id_all_avail) {
 				/*
 				 * Old kernel, no attr->sample_id_type_all field

tools/perf/builtin-sched.c

@@ -489,7 +489,8 @@ static void create_tasks(void)
 
 	err = pthread_attr_init(&attr);
 	BUG_ON(err);
-	err = pthread_attr_setstacksize(&attr, (size_t)(16*1024));
+	err = pthread_attr_setstacksize(&attr,
+			(size_t) max(16 * 1024, PTHREAD_STACK_MIN));
 	BUG_ON(err);
 	err = pthread_mutex_lock(&start_work_mutex);
 	BUG_ON(err);
@@ -1861,7 +1862,7 @@ static int __cmd_record(int argc, const char **argv)
 	rec_argc = ARRAY_SIZE(record_args) + argc - 1;
 	rec_argv = calloc(rec_argc + 1, sizeof(char *));
 
-	if (rec_argv)
+	if (rec_argv == NULL)
 		return -ENOMEM;
 
 	for (i = 0; i < ARRAY_SIZE(record_args); i++)

tools/perf/builtin-stat.c

@@ -316,6 +316,8 @@ static int run_perf_stat(int argc __used, const char **argv)
 				      "\t Consider tweaking"
 				      " /proc/sys/kernel/perf_event_paranoid or running as root.",
 				      system_wide ? "system-wide " : "");
+			} else if (errno == ENOENT) {
+				error("%s event is not supported. ", event_name(counter));
 			} else {
 				error("open_counter returned with %d (%s). "
 				      "/bin/dmesg may provide additional information.\n",
@@ -683,8 +685,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 		nr_counters = ARRAY_SIZE(default_attrs);
 
 		for (c = 0; c < ARRAY_SIZE(default_attrs); ++c) {
-			pos = perf_evsel__new(default_attrs[c].type,
-					      default_attrs[c].config,
+			pos = perf_evsel__new(&default_attrs[c],
 					      nr_counters);
 			if (pos == NULL)
 				goto out;

tools/perf/builtin-test.c

@@ -234,6 +234,7 @@ out:
 	return err;
 }
 
+#include "util/cpumap.h"
 #include "util/evsel.h"
 #include <sys/types.h>
 
@@ -264,6 +265,7 @@ static int test__open_syscall_event(void)
 	int err = -1, fd;
 	struct thread_map *threads;
 	struct perf_evsel *evsel;
+	struct perf_event_attr attr;
 	unsigned int nr_open_calls = 111, i;
 	int id = trace_event__id("sys_enter_open");
 
@@ -278,7 +280,10 @@ static int test__open_syscall_event(void)
 		return -1;
 	}
 
-	evsel = perf_evsel__new(PERF_TYPE_TRACEPOINT, id, 0);
+	memset(&attr, 0, sizeof(attr));
+	attr.type = PERF_TYPE_TRACEPOINT;
+	attr.config = id;
+	evsel = perf_evsel__new(&attr, 0);
 	if (evsel == NULL) {
 		pr_debug("perf_evsel__new\n");
 		goto out_thread_map_delete;
@@ -317,6 +322,111 @@ out_thread_map_delete:
 	return err;
 }
 
+#include <sched.h>
+
+static int test__open_syscall_event_on_all_cpus(void)
+{
+	int err = -1, fd, cpu;
+	struct thread_map *threads;
+	struct cpu_map *cpus;
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr;
+	unsigned int nr_open_calls = 111, i;
+	cpu_set_t *cpu_set;
+	size_t cpu_set_size;
+	int id = trace_event__id("sys_enter_open");
+
+	if (id < 0) {
+		pr_debug("is debugfs mounted on /sys/kernel/debug?\n");
+		return -1;
+	}
+
+	threads = thread_map__new(-1, getpid());
+	if (threads == NULL) {
+		pr_debug("thread_map__new\n");
+		return -1;
+	}
+
+	cpus = cpu_map__new(NULL);
+	if (threads == NULL) {
+		pr_debug("thread_map__new\n");
+		return -1;
+	}
+
+	cpu_set = CPU_ALLOC(cpus->nr);
+
+	if (cpu_set == NULL)
+		goto out_thread_map_delete;
+
+	cpu_set_size = CPU_ALLOC_SIZE(cpus->nr);
+	CPU_ZERO_S(cpu_set_size, cpu_set);
+
+	memset(&attr, 0, sizeof(attr));
+	attr.type = PERF_TYPE_TRACEPOINT;
+	attr.config = id;
+	evsel = perf_evsel__new(&attr, 0);
+	if (evsel == NULL) {
+		pr_debug("perf_evsel__new\n");
+		goto out_cpu_free;
+	}
+
+	if (perf_evsel__open(evsel, cpus, threads) < 0) {
+		pr_debug("failed to open counter: %s, "
+			 "tweak /proc/sys/kernel/perf_event_paranoid?\n",
+			 strerror(errno));
+		goto out_evsel_delete;
+	}
+
+	for (cpu = 0; cpu < cpus->nr; ++cpu) {
+		unsigned int ncalls = nr_open_calls + cpu;
+
+		CPU_SET(cpu, cpu_set);
+		sched_setaffinity(0, cpu_set_size, cpu_set);
+		for (i = 0; i < ncalls; ++i) {
+			fd = open("/etc/passwd", O_RDONLY);
+			close(fd);
+		}
+		CPU_CLR(cpu, cpu_set);
+	}
+
+	/*
+	 * Here we need to explicitely preallocate the counts, as if
+	 * we use the auto allocation it will allocate just for 1 cpu,
+	 * as we start by cpu 0.
+	 */
+	if (perf_evsel__alloc_counts(evsel, cpus->nr) < 0) {
+		pr_debug("perf_evsel__alloc_counts(ncpus=%d)\n", cpus->nr);
+		goto out_close_fd;
+	}
+
+	for (cpu = 0; cpu < cpus->nr; ++cpu) {
+		unsigned int expected;
+
+		if (perf_evsel__read_on_cpu(evsel, cpu, 0) < 0) {
+			pr_debug("perf_evsel__open_read_on_cpu\n");
+			goto out_close_fd;
+		}
+
+		expected = nr_open_calls + cpu;
+		if (evsel->counts->cpu[cpu].val != expected) {
+			pr_debug("perf_evsel__read_on_cpu: expected to intercept %d calls on cpu %d, got %Ld\n",
+				 expected, cpu, evsel->counts->cpu[cpu].val);
+			goto out_close_fd;
+		}
+	}
+
+	err = 0;
+out_close_fd:
+	perf_evsel__close_fd(evsel, 1, threads->nr);
+out_evsel_delete:
+	perf_evsel__delete(evsel);
+out_cpu_free:
+	CPU_FREE(cpu_set);
+out_thread_map_delete:
+	thread_map__delete(threads);
+	return err;
+}
+
 static struct test {
 	const char *desc;
 	int (*func)(void);
@@ -329,6 +439,10 @@ static struct test {
 		.desc = "detect open syscall event",
 		.func = test__open_syscall_event,
 	},
+	{
+		.desc = "detect open syscall event on all cpus",
+		.func = test__open_syscall_event_on_all_cpus,
+	},
 	{
 		.func = NULL,
 	},

tools/perf/builtin-top.c

@@ -1247,6 +1247,8 @@ try_again:
 				die("Permission error - are you root?\n"
 					"\t Consider tweaking"
 					" /proc/sys/kernel/perf_event_paranoid.\n");
+			if (err == ENOENT)
+				die("%s event is not supported. ", event_name(evsel));
 			/*
 			 * If it's cycles then fall back to hrtimer
 			 * based cpu-clock-tick sw counter, which

tools/perf/util/evsel.c

@@ -6,14 +6,13 @@
 
 #define FD(e, x, y) (*(int *)xyarray__entry(e->fd, x, y))
 
-struct perf_evsel *perf_evsel__new(u32 type, u64 config, int idx)
+struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx)
 {
 	struct perf_evsel *evsel = zalloc(sizeof(*evsel));
 
 	if (evsel != NULL) {
 		evsel->idx	   = idx;
-		evsel->attr.type   = type;
-		evsel->attr.config = config;
+		evsel->attr	   = *attr;
 		INIT_LIST_HEAD(&evsel->node);
 	}
 
@@ -128,59 +127,75 @@ int __perf_evsel__read(struct perf_evsel *evsel,
 	return 0;
 }
 
-int perf_evsel__open_per_cpu(struct perf_evsel *evsel, struct cpu_map *cpus)
+static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
+			      struct thread_map *threads)
 {
-	int cpu;
+	int cpu, thread;
 
-	if (evsel->fd == NULL && perf_evsel__alloc_fd(evsel, cpus->nr, 1) < 0)
+	if (evsel->fd == NULL &&
+	    perf_evsel__alloc_fd(evsel, cpus->nr, threads->nr) < 0)
 		return -1;
 
 	for (cpu = 0; cpu < cpus->nr; cpu++) {
-		FD(evsel, cpu, 0) = sys_perf_event_open(&evsel->attr, -1,
-							cpus->map[cpu], -1, 0);
-		if (FD(evsel, cpu, 0) < 0)
-			goto out_close;
+		for (thread = 0; thread < threads->nr; thread++) {
+			FD(evsel, cpu, thread) = sys_perf_event_open(&evsel->attr,
+								     threads->map[thread],
+								     cpus->map[cpu], -1, 0);
+			if (FD(evsel, cpu, thread) < 0)
+				goto out_close;
+		}
 	}
 
 	return 0;
 
 out_close:
-	while (--cpu >= 0) {
-		close(FD(evsel, cpu, 0));
-		FD(evsel, cpu, 0) = -1;
-	}
+	do {
+		while (--thread >= 0) {
+			close(FD(evsel, cpu, thread));
+			FD(evsel, cpu, thread) = -1;
+		}
+		thread = threads->nr;
+	} while (--cpu >= 0);
 	return -1;
 }
 
+static struct {
+	struct cpu_map map;
+	int cpus[1];
+} empty_cpu_map = {
+	.map.nr	= 1,
+	.cpus	= { -1, },
+};
+
+static struct {
+	struct thread_map map;
+	int threads[1];
+} empty_thread_map = {
+	.map.nr	 = 1,
+	.threads = { -1, },
+};
+
+int perf_evsel__open(struct perf_evsel *evsel,
+		     struct cpu_map *cpus, struct thread_map *threads)
+{
+
+	if (cpus == NULL) {
+		/* Work around old compiler warnings about strict aliasing */
+		cpus = &empty_cpu_map.map;
+	}
+
+	if (threads == NULL)
+		threads = &empty_thread_map.map;
+
+	return __perf_evsel__open(evsel, cpus, threads);
+}
+
+int perf_evsel__open_per_cpu(struct perf_evsel *evsel, struct cpu_map *cpus)
+{
+	return __perf_evsel__open(evsel, cpus, &empty_thread_map.map);
+}
+
 int perf_evsel__open_per_thread(struct perf_evsel *evsel, struct thread_map *threads)
 {
-	int thread;
-
-	if (evsel->fd == NULL && perf_evsel__alloc_fd(evsel, 1, threads->nr))
-		return -1;
-
-	for (thread = 0; thread < threads->nr; thread++) {
-		FD(evsel, 0, thread) = sys_perf_event_open(&evsel->attr,
-							   threads->map[thread], -1, -1, 0);
-		if (FD(evsel, 0, thread) < 0)
-			goto out_close;
-	}
-
-	return 0;
-
-out_close:
-	while (--thread >= 0) {
-		close(FD(evsel, 0, thread));
-		FD(evsel, 0, thread) = -1;
-	}
-	return -1;
-}
-
-int perf_evsel__open(struct perf_evsel *evsel, 
-		     struct cpu_map *cpus, struct thread_map *threads)
-{
-	if (threads == NULL)
-		return perf_evsel__open_per_cpu(evsel, cpus);
-
-	return perf_evsel__open_per_thread(evsel, threads);
+	return __perf_evsel__open(evsel, &empty_cpu_map.map, threads);
 }

tools/perf/util/evsel.h

@@ -37,7 +37,7 @@ struct perf_evsel {
 struct cpu_map;
 struct thread_map;
 
-struct perf_evsel *perf_evsel__new(u32 type, u64 config, int idx);
+struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr, int idx);
 void perf_evsel__delete(struct perf_evsel *evsel);
 
 int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);

tools/perf/util/parse-events.c

@@ -490,6 +490,31 @@ parse_multiple_tracepoint_event(char *sys_name, const char *evt_exp,
 	return EVT_HANDLED_ALL;
 }
 
+static int store_event_type(const char *orgname)
+{
+	char filename[PATH_MAX], *c;
+	FILE *file;
+	int id, n;
+
+	sprintf(filename, "%s/", debugfs_path);
+	strncat(filename, orgname, strlen(orgname));
+	strcat(filename, "/id");
+
+	c = strchr(filename, ':');
+	if (c)
+		*c = '/';
+
+	file = fopen(filename, "r");
+	if (!file)
+		return 0;
+	n = fscanf(file, "%i", &id);
+	fclose(file);
+	if (n < 1) {
+		pr_err("cannot store event ID\n");
+		return -EINVAL;
+	}
+	return perf_header__push_event(id, orgname);
+}
 
 static enum event_result parse_tracepoint_event(const char **strp,
 				    struct perf_event_attr *attr)
@@ -533,9 +558,13 @@ static enum event_result parse_tracepoint_event(const char **strp,
 		*strp += strlen(sys_name) + evt_length;
 		return parse_multiple_tracepoint_event(sys_name, evt_name,
 						       flags);
-	} else
+	} else {
+		if (store_event_type(evt_name) < 0)
+			return EVT_FAILED;
+
 		return parse_single_tracepoint_event(sys_name, evt_name,
 						     evt_length, attr, strp);
+	}
 }
 
 static enum event_result
@@ -778,41 +807,11 @@ modifier:
 	return ret;
 }
 
-static int store_event_type(const char *orgname)
-{
-	char filename[PATH_MAX], *c;
-	FILE *file;
-	int id, n;
-
-	sprintf(filename, "%s/", debugfs_path);
-	strncat(filename, orgname, strlen(orgname));
-	strcat(filename, "/id");
-
-	c = strchr(filename, ':');
-	if (c)
-		*c = '/';
-
-	file = fopen(filename, "r");
-	if (!file)
-		return 0;
-	n = fscanf(file, "%i", &id);
-	fclose(file);
-	if (n < 1) {
-		pr_err("cannot store event ID\n");
-		return -EINVAL;
-	}
-	return perf_header__push_event(id, orgname);
-}
-
 int parse_events(const struct option *opt __used, const char *str, int unset __used)
 {
 	struct perf_event_attr attr;
 	enum event_result ret;
 
-	if (strchr(str, ':'))
-		if (store_event_type(str) < 0)
-			return -1;
-
 	for (;;) {
 		memset(&attr, 0, sizeof(attr));
 		ret = parse_event_symbols(&str, &attr);
@@ -824,7 +823,7 @@ int parse_events(const struct option *opt __used, const char *str, int unset __u
 
 		if (ret != EVT_HANDLED_ALL) {
 			struct perf_evsel *evsel;
-			evsel = perf_evsel__new(attr.type, attr.config,
+			evsel = perf_evsel__new(&attr,
 						nr_counters);
 			if (evsel == NULL)
 				return -1;
@@ -1014,8 +1013,15 @@ void print_events(void)
 
 int perf_evsel_list__create_default(void)
 {
-	struct perf_evsel *evsel = perf_evsel__new(PERF_TYPE_HARDWARE,
-						   PERF_COUNT_HW_CPU_CYCLES, 0);
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr;
+
+	memset(&attr, 0, sizeof(attr));
+	attr.type = PERF_TYPE_HARDWARE;
+	attr.config = PERF_COUNT_HW_CPU_CYCLES;
+
+	evsel = perf_evsel__new(&attr, 0);
+
 	if (evsel == NULL)
 		return -ENOMEM;
 

tools/perf/util/session.c

@@ -1007,7 +1007,7 @@ more:
 	if (size == 0)
 		size = 8;
 
-	if (head + event->header.size >= mmap_size) {
+	if (head + event->header.size > mmap_size) {
 		if (mmaps[map_idx]) {
 			munmap(mmaps[map_idx], mmap_size);
 			mmaps[map_idx] = NULL;