Merge branch 'master' into for-next

Fast-forwarded to current state of Linus' tree as there are patches to be
applied for files that didn't exist on the old branch.

41 files changed, 545 insertions, 218 deletions

diff --git a/tools/perf/Makefile b/tools/perf/Makefile
index 158c30e8210..207dee5c5b1 100644
--- a/tools/perf/Makefile
+++ b/tools/perf/Makefile
@@ -165,8 +165,12 @@ grep-libs = $(filter -l%,$(1))
 strip-libs = $(filter-out -l%,$(1))
 
 $(OUTPUT)python/perf.so: $(PYRF_OBJS)
-	$(QUIET_GEN)python util/setup.py --quiet  build_ext --build-lib='$(OUTPUT)python' \
-						--build-temp='$(OUTPUT)python/temp'
+	$(QUIET_GEN)( \
+		export CFLAGS="$(BASIC_CFLAGS)"; \
+		python util/setup.py --quiet  build_ext --build-lib='$(OUTPUT)python' \
+			--build-temp='$(OUTPUT)python/temp' \
+	)
+
 #
 # No Perl scripts right now:
 #
diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index 695de4b5ae6..e18eb7ed30a 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -42,9 +42,9 @@ static const char *sym_hist_filter;
 
 static int perf_evlist__add_sample(struct perf_evlist *evlist,
 				   struct perf_sample *sample,
+				   struct perf_evsel *evsel,
 				   struct addr_location *al)
 {
-	struct perf_evsel *evsel;
 	struct hist_entry *he;
 	int ret;
 
@@ -59,18 +59,6 @@ static int perf_evlist__add_sample(struct perf_evlist *evlist,
 		return 0;
 	}
 
-	evsel = perf_evlist__id2evsel(evlist, sample->id);
-	if (evsel == NULL) {
-		/*
-		 * FIXME: Propagate this back, but at least we're in a builtin,
-		 * where exit() is allowed. ;-)
-		 */
-		ui__warning("Invalid %s file, contains samples with id not in "
-			    "its header!\n", input_name);
-		exit_browser(0);
-		exit(1);
-	}
-
 	he = __hists__add_entry(&evsel->hists, al, NULL, 1);
 	if (he == NULL)
 		return -ENOMEM;
@@ -92,6 +80,7 @@ static int perf_evlist__add_sample(struct perf_evlist *evlist,
 
 static int process_sample_event(union perf_event *event,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel,
 				struct perf_session *session)
 {
 	struct addr_location al;
@@ -103,7 +92,8 @@ static int process_sample_event(union perf_event *event,
 		return -1;
 	}
 
-	if (!al.filtered && perf_evlist__add_sample(session->evlist, sample, &al)) {
+	if (!al.filtered &&
+	    perf_evlist__add_sample(session->evlist, sample, evsel, &al)) {
 		pr_warning("problem incrementing symbol count, "
 			   "skipping event\n");
 		return -1;
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 6b7d91160ec..e8219990f8b 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -32,6 +32,7 @@ static int hists__add_entry(struct hists *self,
 
 static int diff__process_sample_event(union perf_event *event,
 				      struct perf_sample *sample,
+				      struct perf_evsel *evsel __used,
 				      struct perf_session *session)
 {
 	struct addr_location al;
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index e29f04ed339..8dfc12bb119 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -43,6 +43,14 @@ static int perf_event__repipe(union perf_event *event,
 	return perf_event__repipe_synth(event, session);
 }
 
+static int perf_event__repipe_sample(union perf_event *event,
+			      struct perf_sample *sample __used,
+			      struct perf_evsel *evsel __used,
+			      struct perf_session *session)
+{
+	return perf_event__repipe_synth(event, session);
+}
+
 static int perf_event__repipe_mmap(union perf_event *event,
 				   struct perf_sample *sample,
 				   struct perf_session *session)
@@ -124,6 +132,7 @@ static int dso__inject_build_id(struct dso *self, struct perf_session *session)
 
 static int perf_event__inject_buildid(union perf_event *event,
 				      struct perf_sample *sample,
+				      struct perf_evsel *evsel __used,
 				      struct perf_session *session)
 {
 	struct addr_location al;
@@ -164,7 +173,7 @@ repipe:
 }
 
 struct perf_event_ops inject_ops = {
-	.sample		= perf_event__repipe,
+	.sample		= perf_event__repipe_sample,
 	.mmap		= perf_event__repipe,
 	.comm		= perf_event__repipe,
 	.fork		= perf_event__repipe,
diff --git a/tools/perf/builtin-kmem.c b/tools/perf/builtin-kmem.c
index 7f618f4e7b7..225e963df10 100644
--- a/tools/perf/builtin-kmem.c
+++ b/tools/perf/builtin-kmem.c
@@ -305,6 +305,7 @@ static void process_raw_event(union perf_event *raw_event __used, void *data,
 
 static int process_sample_event(union perf_event *event,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel __used,
 				struct perf_session *session)
 {
 	struct thread *thread = perf_session__findnew(session, event->ip.pid);
diff --git a/tools/perf/builtin-lock.c b/tools/perf/builtin-lock.c
index 7a2a79d2cf2..9ac05aafd9b 100644
--- a/tools/perf/builtin-lock.c
+++ b/tools/perf/builtin-lock.c
@@ -845,7 +845,9 @@ static void dump_info(void)
 		die("Unknown type of information\n");
 }
 
-static int process_sample_event(union perf_event *event, struct perf_sample *sample,
+static int process_sample_event(union perf_event *event,
+				struct perf_sample *sample,
+				struct perf_evsel *evsel __used,
 				struct perf_session *s)
 {
 	struct thread *thread = perf_session__findnew(s, sample->tid);
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 6febcc168a8..416538248a4 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -41,7 +41,7 @@ static u64			user_interval			= ULLONG_MAX;
 static u64			default_interval		=      0;
 
 static unsigned int		page_size;
-static unsigned int		mmap_pages			=    128;
+static unsigned int		mmap_pages			= UINT_MAX;
 static unsigned int		user_freq 			= UINT_MAX;
 static int			freq				=   1000;
 static int			output;
@@ -163,6 +163,7 @@ static void config_attr(struct perf_evsel *evsel, struct perf_evlist *evlist)
 	struct perf_event_attr *attr = &evsel->attr;
 	int track = !evsel->idx; /* only the first counter needs these */
 
+	attr->inherit		= !no_inherit;
 	attr->read_format	= PERF_FORMAT_TOTAL_TIME_ENABLED |
 				  PERF_FORMAT_TOTAL_TIME_RUNNING |
 				  PERF_FORMAT_ID;
@@ -251,6 +252,9 @@ static void open_counters(struct perf_evlist *evlist)
 {
 	struct perf_evsel *pos;
 
+	if (evlist->cpus->map[0] < 0)
+		no_inherit = true;
+
 	list_for_each_entry(pos, &evlist->entries, node) {
 		struct perf_event_attr *attr = &pos->attr;
 		/*
@@ -271,15 +275,13 @@ static void open_counters(struct perf_evlist *evlist)
 retry_sample_id:
 		attr->sample_id_all = sample_id_all_avail ? 1 : 0;
 try_again:
-		if (perf_evsel__open(pos, evlist->cpus, evlist->threads, group,
-				     !no_inherit) < 0) {
+		if (perf_evsel__open(pos, evlist->cpus, evlist->threads, group) < 0) {
 			int err = errno;
 
-			if (err == EPERM || err == EACCES)
-				die("Permission error - are you root?\n"
-					"\t Consider tweaking"
-					" /proc/sys/kernel/perf_event_paranoid.\n");
-			else if (err ==  ENODEV && cpu_list) {
+			if (err == EPERM || err == EACCES) {
+				ui__warning_paranoid();
+				exit(EXIT_FAILURE);
+			} else if (err ==  ENODEV && cpu_list) {
 				die("No such device - did you specify"
 					" an out-of-range profile CPU?\n");
 			} else if (err == EINVAL && sample_id_all_avail) {
@@ -302,11 +304,19 @@ try_again:
 					&& attr->config == PERF_COUNT_HW_CPU_CYCLES) {
 
 				if (verbose)
-					warning(" ... trying to fall back to cpu-clock-ticks\n");
+					ui__warning("The cycles event is not supported, "
+						    "trying to fall back to cpu-clock-ticks\n");
 				attr->type = PERF_TYPE_SOFTWARE;
 				attr->config = PERF_COUNT_SW_CPU_CLOCK;
 				goto try_again;
 			}
+
+			if (err == ENOENT) {
+				ui__warning("The %s event is not supported.\n",
+					    event_name(pos));
+				exit(EXIT_FAILURE);
+			}
+
 			printf("\n");
 			error("sys_perf_event_open() syscall returned with %d (%s).  /bin/dmesg may provide additional information.\n",
 			      err, strerror(err));
@@ -506,6 +516,10 @@ static int __cmd_record(int argc, const char **argv)
 	if (have_tracepoints(&evsel_list->entries))
 		perf_header__set_feat(&session->header, HEADER_TRACE_INFO);
 
+	/* 512 kiB: default amount of unprivileged mlocked memory */
+	if (mmap_pages == UINT_MAX)
+		mmap_pages = (512 * 1024) / page_size;
+
 	if (forks) {
 		child_pid = fork();
 		if (child_pid < 0) {
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index b1b82009ab9..498c6f70a74 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -50,12 +50,12 @@ static symbol_filter_t	annotate_init;
 
 static int perf_session__add_hist_entry(struct perf_session *session,
 					struct addr_location *al,
-					struct perf_sample *sample)
+					struct perf_sample *sample,
+					struct perf_evsel *evsel)
 {
 	struct symbol *parent = NULL;
 	int err = 0;
 	struct hist_entry *he;
-	struct perf_evsel *evsel;
 
 	if ((sort__has_parent || symbol_conf.use_callchain) && sample->callchain) {
 		err = perf_session__resolve_callchain(session, al->thread,
@@ -64,18 +64,6 @@ static int perf_session__add_hist_entry(struct perf_session *session,
 			return err;
 	}
 
-	evsel = perf_evlist__id2evsel(session->evlist, sample->id);
-	if (evsel == NULL) {
-		/*
-		 * FIXME: Propagate this back, but at least we're in a builtin,
-		 * where exit() is allowed. ;-)
-		 */
-		ui__warning("Invalid %s file, contains samples with id %" PRIu64 " not in "
-			    "its header!\n", input_name, sample->id);
-		exit_browser(0);
-		exit(1);
-	}
-
 	he = __hists__add_entry(&evsel->hists, al, parent, sample->period);
 	if (he == NULL)
 		return -ENOMEM;
@@ -113,6 +101,7 @@ out:
 
 static int process_sample_event(union perf_event *event,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel,
 				struct perf_session *session)
 {
 	struct addr_location al;
@@ -127,7 +116,7 @@ static int process_sample_event(union perf_event *event,
 	if (al.filtered || (hide_unresolved && al.sym == NULL))
 		return 0;
 
-	if (perf_session__add_hist_entry(session, &al, sample)) {
+	if (perf_session__add_hist_entry(session, &al, sample, evsel)) {
 		pr_debug("problem incrementing symbol period, skipping event\n");
 		return -1;
 	}
diff --git a/tools/perf/builtin-sched.c b/tools/perf/builtin-sched.c
index a32f411faea..dcfe8873c9a 100644
--- a/tools/perf/builtin-sched.c
+++ b/tools/perf/builtin-sched.c
@@ -1603,6 +1603,7 @@ static void process_raw_event(union perf_event *raw_event __used,
 
 static int process_sample_event(union perf_event *event,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel __used,
 				struct perf_session *session)
 {
 	struct thread *thread;
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 9f5fc549214..ac574ea2391 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -162,19 +162,11 @@ static void print_sample_start(struct perf_sample *sample,
 
 static void process_event(union perf_event *event __unused,
 			  struct perf_sample *sample,
+			  struct perf_evsel *evsel,
 			  struct perf_session *session,
 			  struct thread *thread)
 {
-	struct perf_event_attr *attr;
-	struct perf_evsel *evsel;
-
-	evsel = perf_evlist__id2evsel(session->evlist, sample->id);
-	if (evsel == NULL) {
-		pr_err("Invalid data. Contains samples with id not in "
-		       "its header!\n");
-		return;
-	}
-	attr = &evsel->attr;
+	struct perf_event_attr *attr = &evsel->attr;
 
 	if (output_fields[attr->type] == 0)
 		return;
@@ -244,6 +236,7 @@ static char const		*input_name = "perf.data";
 
 static int process_sample_event(union perf_event *event,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel,
 				struct perf_session *session)
 {
 	struct thread *thread = perf_session__findnew(session, event->ip.pid);
@@ -264,7 +257,7 @@ static int process_sample_event(union perf_event *event,
 		last_timestamp = sample->time;
 		return 0;
 	}
-	scripting_ops->process_event(event, sample, session, thread);
+	scripting_ops->process_event(event, sample, evsel, session, thread);
 
 	session->hists.stats.total_period += sample->period;
 	return 0;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index e2109f9b43e..03f0e45f147 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -167,16 +167,17 @@ static int create_perf_stat_counter(struct perf_evsel *evsel)
 		attr->read_format = PERF_FORMAT_TOTAL_TIME_ENABLED |
 				    PERF_FORMAT_TOTAL_TIME_RUNNING;
 
+	attr->inherit = !no_inherit;
+
 	if (system_wide)
-		return perf_evsel__open_per_cpu(evsel, evsel_list->cpus, false, false);
+		return perf_evsel__open_per_cpu(evsel, evsel_list->cpus, false);
 
-	attr->inherit = !no_inherit;
 	if (target_pid == -1 && target_tid == -1) {
 		attr->disabled = 1;
 		attr->enable_on_exec = 1;
 	}
 
-	return perf_evsel__open_per_thread(evsel, evsel_list->threads, false, false);
+	return perf_evsel__open_per_thread(evsel, evsel_list->threads, false);
 }
 
 /*
diff --git a/tools/perf/builtin-test.c b/tools/perf/builtin-test.c
index 1b2106c58f6..11e3c845836 100644
--- a/tools/perf/builtin-test.c
+++ b/tools/perf/builtin-test.c
@@ -290,7 +290,7 @@ static int test__open_syscall_event(void)
 		goto out_thread_map_delete;
 	}
 
-	if (perf_evsel__open_per_thread(evsel, threads, false, false) < 0) {
+	if (perf_evsel__open_per_thread(evsel, threads, false) < 0) {
 		pr_debug("failed to open counter: %s, "
 			 "tweak /proc/sys/kernel/perf_event_paranoid?\n",
 			 strerror(errno));
@@ -303,7 +303,7 @@ static int test__open_syscall_event(void)
 	}
 
 	if (perf_evsel__read_on_cpu(evsel, 0, 0) < 0) {
-		pr_debug("perf_evsel__open_read_on_cpu\n");
+		pr_debug("perf_evsel__read_on_cpu\n");
 		goto out_close_fd;
 	}
 
@@ -365,7 +365,7 @@ static int test__open_syscall_event_on_all_cpus(void)
 		goto out_thread_map_delete;
 	}
 
-	if (perf_evsel__open(evsel, cpus, threads, false, false) < 0) {
+	if (perf_evsel__open(evsel, cpus, threads, false) < 0) {
 		pr_debug("failed to open counter: %s, "
 			 "tweak /proc/sys/kernel/perf_event_paranoid?\n",
 			 strerror(errno));
@@ -418,7 +418,7 @@ static int test__open_syscall_event_on_all_cpus(void)
 			continue;
 
 		if (perf_evsel__read_on_cpu(evsel, cpu, 0) < 0) {
-			pr_debug("perf_evsel__open_read_on_cpu\n");
+			pr_debug("perf_evsel__read_on_cpu\n");
 			err = -1;
 			break;
 		}
@@ -529,7 +529,7 @@ static int test__basic_mmap(void)
 
 		perf_evlist__add(evlist, evsels[i]);
 
-		if (perf_evsel__open(evsels[i], cpus, threads, false, false) < 0) {
+		if (perf_evsel__open(evsels[i], cpus, threads, false) < 0) {
 			pr_debug("failed to open counter: %s, "
 				 "tweak /proc/sys/kernel/perf_event_paranoid?\n",
 				 strerror(errno));
diff --git a/tools/perf/builtin-timechart.c b/tools/perf/builtin-timechart.c
index 67c0459dc32..aa26f4d66d1 100644
--- a/tools/perf/builtin-timechart.c
+++ b/tools/perf/builtin-timechart.c
@@ -488,6 +488,7 @@ static void sched_switch(int cpu, u64 timestamp, struct trace_entry *te)
 
 static int process_sample_event(union perf_event *event __used,
 				struct perf_sample *sample,
+				struct perf_evsel *evsel __used,
 				struct perf_session *session)
 {
 	struct trace_entry *te;
@@ -506,6 +507,16 @@ static int process_sample_event(union perf_event *event __used,
 		struct power_entry_old *peo;
 		peo = (void *)te;
 #endif
+		/*
+		 * FIXME: use evsel, its already mapped from id to perf_evsel,
+		 * remove perf_header__find_event infrastructure bits.
+		 * Mapping all these "power:cpu_idle" strings to the tracepoint
+		 * ID and then just comparing against evsel->attr.config.
+		 *
+		 * e.g.:
+		 *
+		 * if (evsel->attr.config == power_cpu_idle_id)
+		 */
 		event_str = perf_header__find_event(te->type);
 
 		if (!event_str)
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 70f1075cc5b..7e3d6e310bf 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -515,7 +515,9 @@ static void handle_keypress(struct perf_session *session, int c)
 			break;
 		case 'E':
 			if (top.evlist->nr_entries > 1) {
-				int counter;
+				/* Select 0 as the default event: */
+				int counter = 0;
+
 				fprintf(stderr, "\nAvailable events:");
 
 				list_for_each_entry(top.sym_evsel, &top.evlist->entries, node)
@@ -843,15 +845,16 @@ static void start_counters(struct perf_evlist *evlist)
 		}
 
 		attr->mmap = 1;
+		attr->inherit = inherit;
 try_again:
 		if (perf_evsel__open(counter, top.evlist->cpus,
-				     top.evlist->threads, group, inherit) < 0) {
+				     top.evlist->threads, group) < 0) {
 			int err = errno;
 
-			if (err == EPERM || err == EACCES)
-				die("Permission error - are you root?\n"
-					"\t Consider tweaking"
-					" /proc/sys/kernel/perf_event_paranoid.\n");
+			if (err == EPERM || err == EACCES) {
+				ui__warning_paranoid();
+				goto out_err;
+			}
 			/*
 			 * If it's cycles then fall back to hrtimer
 			 * based cpu-clock-tick sw counter, which
@@ -859,25 +862,41 @@ try_again:
 			 */
 			if (attr->type == PERF_TYPE_HARDWARE &&
 			    attr->config == PERF_COUNT_HW_CPU_CYCLES) {
-
 				if (verbose)
-					warning(" ... trying to fall back to cpu-clock-ticks\n");
+					ui__warning("Cycles event not supported,\n"
+						    "trying to fall back to cpu-clock-ticks\n");
 
 				attr->type = PERF_TYPE_SOFTWARE;
 				attr->config = PERF_COUNT_SW_CPU_CLOCK;
 				goto try_again;
 			}
-			printf("\n");
-			error("sys_perf_event_open() syscall returned with %d "
-			      "(%s).  /bin/dmesg may provide additional information.\n",
-			      err, strerror(err));
-			die("No CONFIG_PERF_EVENTS=y kernel support configured?\n");
-			exit(-1);
+
+			if (err == ENOENT) {
+				ui__warning("The %s event is not supported.\n",
+					    event_name(counter));
+				goto out_err;
+			}
+
+			ui__warning("The sys_perf_event_open() syscall "
+				    "returned with %d (%s).  /bin/dmesg "
+				    "may provide additional information.\n"
+				    "No CONFIG_PERF_EVENTS=y kernel support "
+				    "configured?\n", err, strerror(err));
+			goto out_err;
 		}
 	}
 
-	if (perf_evlist__mmap(evlist, mmap_pages, false) < 0)
-		die("failed to mmap with %d (%s)\n", errno, strerror(errno));
+	if (perf_evlist__mmap(evlist, mmap_pages, false) < 0) {
+		ui__warning("Failed to mmap with %d (%s)\n",
+			    errno, strerror(errno));
+		goto out_err;
+	}
+
+	return;
+
+out_err:
+	exit_browser(0);
+	exit(0);
 }
 
 static int __cmd_top(void)
diff --git a/tools/perf/util/build-id.c b/tools/perf/util/build-id.c
index 31f934af986..a91cd99f26e 100644
--- a/tools/perf/util/build-id.c
+++ b/tools/perf/util/build-id.c
@@ -16,6 +16,7 @@
 
 static int build_id__mark_dso_hit(union perf_event *event,
 				  struct perf_sample *sample __used,
+				  struct perf_evsel *evsel __used,
 				  struct perf_session *session)
 {
 	struct addr_location al;
diff --git a/tools/perf/util/cgroup.c b/tools/perf/util/cgroup.c
index 9fea7553522..96bee5c4600 100644
--- a/tools/perf/util/cgroup.c
+++ b/tools/perf/util/cgroup.c
@@ -13,7 +13,7 @@ cgroupfs_find_mountpoint(char *buf, size_t maxlen)
 {
 	FILE *fp;
 	char mountpoint[MAX_PATH+1], tokens[MAX_PATH+1], type[MAX_PATH+1];
-	char *token, *saved_ptr;
+	char *token, *saved_ptr = NULL;
 	int found = 0;
 
 	fp = fopen("/proc/mounts", "r");
diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c
index d4536a9e0d8..155749d7435 100644
--- a/tools/perf/util/debug.c
+++ b/tools/perf/util/debug.c
@@ -57,6 +57,16 @@ void ui__warning(const char *format, ...)
 }
 #endif
 
+void ui__warning_paranoid(void)
+{
+	ui__warning("Permission error - are you root?\n"
+		    "Consider tweaking /proc/sys/kernel/perf_event_paranoid:\n"
+		    " -1 - Not paranoid at all\n"
+		    "  0 - Disallow raw tracepoint access for unpriv\n"
+		    "  1 - Disallow cpu events for unpriv\n"
+		    "  2 - Disallow kernel profiling for unpriv\n");
+}
+
 void trace_event(union perf_event *event)
 {
 	unsigned char *raw_event = (void *)event;
diff --git a/tools/perf/util/debug.h b/tools/perf/util/debug.h
index 93516cf4682..fd53db47e3d 100644
--- a/tools/perf/util/debug.h
+++ b/tools/perf/util/debug.h
@@ -36,5 +36,6 @@ int ui_helpline__show_help(const char *format, va_list ap);
 #endif
 
 void ui__warning(const char *format, ...) __attribute__((format(printf, 1, 2)));
+void ui__warning_paranoid(void);
 
 #endif	/* __PERF_DEBUG_H */
diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index 2b15c362ef5..1023f67633a 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -710,7 +710,7 @@ try_again:
 		 * in the whole kernel symbol list.
 		 */
 		if ((long long)al->addr < 0 &&
-		    cpumode == PERF_RECORD_MISC_KERNEL &&
+		    cpumode == PERF_RECORD_MISC_USER &&
 		    machine && mg != &machine->kmaps) {
 			mg = &machine->kmaps;
 			goto try_again;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index d852cefa20d..45da8d186b4 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -12,6 +12,7 @@
 #include "evlist.h"
 #include "evsel.h"
 #include "util.h"
+#include "debug.h"
 
 #include <sys/mman.h>
 
@@ -250,15 +251,19 @@ int perf_evlist__alloc_mmap(struct perf_evlist *evlist)
 	return evlist->mmap != NULL ? 0 : -ENOMEM;
 }
 
-static int __perf_evlist__mmap(struct perf_evlist *evlist, int cpu, int prot,
-			       int mask, int fd)
+static int __perf_evlist__mmap(struct perf_evlist *evlist, struct perf_evsel *evsel,
+			       int cpu, int prot, int mask, int fd)
 {
 	evlist->mmap[cpu].prev = 0;
 	evlist->mmap[cpu].mask = mask;
 	evlist->mmap[cpu].base = mmap(NULL, evlist->mmap_len, prot,
 				      MAP_SHARED, fd, 0);
-	if (evlist->mmap[cpu].base == MAP_FAILED)
+	if (evlist->mmap[cpu].base == MAP_FAILED) {
+		if (evlist->cpus->map[cpu] == -1 && evsel->attr.inherit)
+			ui__warning("Inherit is not allowed on per-task "
+				    "events using mmap.\n");
 		return -1;
+	}
 
 	perf_evlist__add_pollfd(evlist, fd);
 	return 0;
@@ -312,7 +317,8 @@ int perf_evlist__mmap(struct perf_evlist *evlist, int pages, bool overwrite)
 					if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT,
 						  FD(first_evsel, cpu, 0)) != 0)
 						goto out_unmap;
-				} else if (__perf_evlist__mmap(evlist, cpu, prot, mask, fd) < 0)
+				} else if (__perf_evlist__mmap(evlist, evsel, cpu,
+							       prot, mask, fd) < 0)
 					goto out_unmap;
 
 				if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index 662596afd7f..d6fd59beb86 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -175,7 +175,7 @@ int __perf_evsel__read(struct perf_evsel *evsel,
 }
 
 static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
-			      struct thread_map *threads, bool group, bool inherit)
+			      struct thread_map *threads, bool group)
 {
 	int cpu, thread;
 	unsigned long flags = 0;
@@ -192,19 +192,6 @@ static int __perf_evsel__open(struct perf_evsel *evsel, struct cpu_map *cpus,
 
 	for (cpu = 0; cpu < cpus->nr; cpu++) {
 		int group_fd = -1;
-		/*
-		 * Don't allow mmap() of inherited per-task counters. This
-		 * would create a performance issue due to all children writing
-		 * to the same buffer.
-		 *
-		 * FIXME:
-		 * Proper fix is not to pas