21 files changed, 1005 insertions, 30 deletions

diff --git a/tools/perf/lib/Build b/tools/perf/lib/Build
index c31f1c111f8f..2ef9a4ec6d99 100644
--- a/tools/perf/lib/Build
+++ b/tools/perf/lib/Build
@@ -3,6 +3,7 @@ libperf-y += cpumap.o
 libperf-y += threadmap.o
 libperf-y += evsel.o
 libperf-y += evlist.o
+libperf-y += mmap.o
 libperf-y += zalloc.o
 libperf-y += xyarray.o
 libperf-y += lib.o
diff --git a/tools/perf/lib/Makefile b/tools/perf/lib/Makefile
index 85ccb8c439a4..0f233638ef1f 100644
--- a/tools/perf/lib/Makefile
+++ b/tools/perf/lib/Makefile
@@ -107,6 +107,7 @@ else
 endif
 
 LIBAPI = $(API_PATH)libapi.a
+export LIBAPI
 
 $(LIBAPI): FORCE
 	$(Q)$(MAKE) -C $(LIB_DIR) O=$(OUTPUT) $(OUTPUT)libapi.a
@@ -172,8 +173,9 @@ install_headers:
 		$(call do_install,include/perf/cpumap.h,$(prefix)/include/perf,644); \
 		$(call do_install,include/perf/threadmap.h,$(prefix)/include/perf,644); \
 		$(call do_install,include/perf/evlist.h,$(prefix)/include/perf,644); \
-		$(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644);
-		$(call do_install,include/perf/event.h,$(prefix)/include/perf,644);
+		$(call do_install,include/perf/evsel.h,$(prefix)/include/perf,644); \
+		$(call do_install,include/perf/event.h,$(prefix)/include/perf,644); \
+		$(call do_install,include/perf/mmap.h,$(prefix)/include/perf,644);
 
 install_pkgconfig: $(LIBPERF_PC)
 	$(call QUIET_INSTALL, $(LIBPERF_PC)) \
diff --git a/tools/perf/lib/core.c b/tools/perf/lib/core.c
index d0b9ae422b9f..58fc894b76c5 100644
--- a/tools/perf/lib/core.c
+++ b/tools/perf/lib/core.c
@@ -5,11 +5,12 @@
 #include <stdio.h>
 #include <stdarg.h>
 #include <unistd.h>
+#include <linux/compiler.h>
 #include <perf/core.h>
 #include <internal/lib.h>
 #include "internal.h"
 
-static int __base_pr(enum libperf_print_level level, const char *format,
+static int __base_pr(enum libperf_print_level level __maybe_unused, const char *format,
 		     va_list args)
 {
 	return vfprintf(stderr, format, args);
diff --git a/tools/perf/lib/evlist.c b/tools/perf/lib/evlist.c
index d1496fee810c..205ddbb80bc1 100644
--- a/tools/perf/lib/evlist.c
+++ b/tools/perf/lib/evlist.c
@@ -8,13 +8,20 @@
 #include <internal/evlist.h>
 #include <internal/evsel.h>
 #include <internal/xyarray.h>
+#include <internal/mmap.h>
+#include <internal/cpumap.h>
+#include <internal/threadmap.h>
+#include <internal/xyarray.h>
+#include <internal/lib.h>
 #include <linux/zalloc.h>
+#include <sys/ioctl.h>
 #include <stdlib.h>
 #include <errno.h>
 #include <unistd.h>
 #include <fcntl.h>
 #include <signal.h>
 #include <poll.h>
+#include <sys/mman.h>
 #include <perf/cpumap.h>
 #include <perf/threadmap.h>
 #include <api/fd/array.h>
@@ -27,6 +34,7 @@ void perf_evlist__init(struct perf_evlist *evlist)
 		INIT_HLIST_HEAD(&evlist->heads[i]);
 	INIT_LIST_HEAD(&evlist->entries);
 	evlist->nr_entries = 0;
+	fdarray__init(&evlist->pollfd, 64);
 }
 
 static void __perf_evlist__propagate_maps(struct perf_evlist *evlist,
@@ -101,8 +109,36 @@ perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)
 	return next;
 }
 
+static void perf_evlist__purge(struct perf_evlist *evlist)
+{
+	struct perf_evsel *pos, *n;
+
+	perf_evlist__for_each_entry_safe(evlist, n, pos) {
+		list_del_init(&pos->node);
+		perf_evsel__delete(pos);
+	}
+
+	evlist->nr_entries = 0;
+}
+
+void perf_evlist__exit(struct perf_evlist *evlist)
+{
+	perf_cpu_map__put(evlist->cpus);
+	perf_thread_map__put(evlist->threads);
+	evlist->cpus = NULL;
+	evlist->threads = NULL;
+	fdarray__exit(&evlist->pollfd);
+}
+
 void perf_evlist__delete(struct perf_evlist *evlist)
 {
+	if (evlist == NULL)
+		return;
+
+	perf_evlist__munmap(evlist);
+	perf_evlist__close(evlist);
+	perf_evlist__purge(evlist);
+	perf_evlist__exit(evlist);
 	free(evlist);
 }
 
@@ -277,7 +313,328 @@ int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,
 	return pos;
 }
 
+static void perf_evlist__munmap_filtered(struct fdarray *fda, int fd,
+					 void *arg __maybe_unused)
+{
+	struct perf_mmap *map = fda->priv[fd].ptr;
+
+	if (map)
+		perf_mmap__put(map);
+}
+
+int perf_evlist__filter_pollfd(struct perf_evlist *evlist, short revents_and_mask)
+{
+	return fdarray__filter(&evlist->pollfd, revents_and_mask,
+			       perf_evlist__munmap_filtered, NULL);
+}
+
 int perf_evlist__poll(struct perf_evlist *evlist, int timeout)
 {
 	return fdarray__poll(&evlist->pollfd, timeout);
 }
+
+static struct perf_mmap* perf_evlist__alloc_mmap(struct perf_evlist *evlist, bool overwrite)
+{
+	int i;
+	struct perf_mmap *map;
+
+	map = zalloc(evlist->nr_mmaps * sizeof(struct perf_mmap));
+	if (!map)
+		return NULL;
+
+	for (i = 0; i < evlist->nr_mmaps; i++) {
+		struct perf_mmap *prev = i ? &map[i - 1] : NULL;
+
+		/*
+		 * When the perf_mmap() call is made we grab one refcount, plus
+		 * one extra to let perf_mmap__consume() get the last
+		 * events after all real references (perf_mmap__get()) are
+		 * dropped.
+		 *
+		 * Each PERF_EVENT_IOC_SET_OUTPUT points to this mmap and
+		 * thus does perf_mmap__get() on it.
+		 */
+		perf_mmap__init(&map[i], prev, overwrite, NULL);
+	}
+
+	return map;
+}
+
+static void perf_evlist__set_sid_idx(struct perf_evlist *evlist,
+				     struct perf_evsel *evsel, int idx, int cpu,
+				     int thread)
+{
+	struct perf_sample_id *sid = SID(evsel, cpu, thread);
+
+	sid->idx = idx;
+	if (evlist->cpus && cpu >= 0)
+		sid->cpu = evlist->cpus->map[cpu];
+	else
+		sid->cpu = -1;
+	if (!evsel->system_wide && evlist->threads && thread >= 0)
+		sid->tid = perf_thread_map__pid(evlist->threads, thread);
+	else
+		sid->tid = -1;
+}
+
+static struct perf_mmap*
+perf_evlist__mmap_cb_get(struct perf_evlist *evlist, bool overwrite, int idx)
+{
+	struct perf_mmap *maps;
+
+	maps = overwrite ? evlist->mmap_ovw : evlist->mmap;
+
+	if (!maps) {
+		maps = perf_evlist__alloc_mmap(evlist, overwrite);
+		if (!maps)
+			return NULL;
+
+		if (overwrite)
+			evlist->mmap_ovw = maps;
+		else
+			evlist->mmap = maps;
+	}
+
+	return &maps[idx];
+}
+
+#define FD(e, x, y) (*(int *) xyarray__entry(e->fd, x, y))
+
+static int
+perf_evlist__mmap_cb_mmap(struct perf_mmap *map, struct perf_mmap_param *mp,
+			  int output, int cpu)
+{
+	return perf_mmap__mmap(map, mp, output, cpu);
+}
+
+static void perf_evlist__set_mmap_first(struct perf_evlist *evlist, struct perf_mmap *map,
+					bool overwrite)
+{
+	if (overwrite)
+		evlist->mmap_ovw_first = map;
+	else
+		evlist->mmap_first = map;
+}
+
+static int
+mmap_per_evsel(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,
+	       int idx, struct perf_mmap_param *mp, int cpu_idx,
+	       int thread, int *_output, int *_output_overwrite)
+{
+	int evlist_cpu = perf_cpu_map__cpu(evlist->cpus, cpu_idx);
+	struct perf_evsel *evsel;
+	int revent;
+
+	perf_evlist__for_each_entry(evlist, evsel) {
+		bool overwrite = evsel->attr.write_backward;
+		struct perf_mmap *map;
+		int *output, fd, cpu;
+
+		if (evsel->system_wide && thread)
+			continue;
+
+		cpu = perf_cpu_map__idx(evsel->cpus, evlist_cpu);
+		if (cpu == -1)
+			continue;
+
+		map = ops->get(evlist, overwrite, idx);
+		if (map == NULL)
+			return -ENOMEM;
+
+		if (overwrite) {
+			mp->prot = PROT_READ;
+			output   = _output_overwrite;
+		} else {
+			mp->prot = PROT_READ | PROT_WRITE;
+			output   = _output;
+		}
+
+		fd = FD(evsel, cpu, thread);
+
+		if (*output == -1) {
+			*output = fd;
+
+			/*
+			 * The last one will be done at perf_mmap__consume(), so that we
+			 * make sure we don't prevent tools from consuming every last event in
+			 * the ring buffer.
+			 *
+			 * I.e. we can get the POLLHUP meaning that the fd doesn't exist
+			 * anymore, but the last events for it are still in the ring buffer,
+			 * waiting to be consumed.
+			 *
+			 * Tools can chose to ignore this at their own discretion, but the
+			 * evlist layer can't just drop it when filtering events in
+			 * perf_evlist__filter_pollfd().
+			 */
+			refcount_set(&map->refcnt, 2);
+
+			if (ops->mmap(map, mp, *output, evlist_cpu) < 0)
+				return -1;
+
+			if (!idx)
+				perf_evlist__set_mmap_first(evlist, map, overwrite);
+		} else {
+			if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, *output) != 0)
+				return -1;
+
+			perf_mmap__get(map);
+		}
+
+		revent = !overwrite ? POLLIN : 0;
+
+		if (!evsel->system_wide &&
+		    perf_evlist__add_pollfd(evlist, fd, map, revent) < 0) {
+			perf_mmap__put(map);
+			return -1;
+		}
+
+		if (evsel->attr.read_format & PERF_FORMAT_ID) {
+			if (perf_evlist__id_add_fd(evlist, evsel, cpu, thread,
+						   fd) < 0)
+				return -1;
+			perf_evlist__set_sid_idx(evlist, evsel, idx, cpu,
+						 thread);
+		}
+	}
+
+	return 0;
+}
+
+static int
+mmap_per_thread(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,
+		struct perf_mmap_param *mp)
+{
+	int thread;
+	int nr_threads = perf_thread_map__nr(evlist->threads);
+
+	for (thread = 0; thread < nr_threads; thread++) {
+		int output = -1;
+		int output_overwrite = -1;
+
+		if (ops->idx)
+			ops->idx(evlist, mp, thread, false);
+
+		if (mmap_per_evsel(evlist, ops, thread, mp, 0, thread,
+				   &output, &output_overwrite))
+			goto out_unmap;
+	}
+
+	return 0;
+
+out_unmap:
+	perf_evlist__munmap(evlist);
+	return -1;
+}
+
+static int
+mmap_per_cpu(struct perf_evlist *evlist, struct perf_evlist_mmap_ops *ops,
+	     struct perf_mmap_param *mp)
+{
+	int nr_threads = perf_thread_map__nr(evlist->threads);
+	int nr_cpus    = perf_cpu_map__nr(evlist->cpus);
+	int cpu, thread;
+
+	for (cpu = 0; cpu < nr_cpus; cpu++) {
+		int output = -1;
+		int output_overwrite = -1;
+
+		if (ops->idx)
+			ops->idx(evlist, mp, cpu, true);
+
+		for (thread = 0; thread < nr_threads; thread++) {
+			if (mmap_per_evsel(evlist, ops, cpu, mp, cpu,
+					   thread, &output, &output_overwrite))
+				goto out_unmap;
+		}
+	}
+
+	return 0;
+
+out_unmap:
+	perf_evlist__munmap(evlist);
+	return -1;
+}
+
+static int perf_evlist__nr_mmaps(struct perf_evlist *evlist)
+{
+	int nr_mmaps;
+
+	nr_mmaps = perf_cpu_map__nr(evlist->cpus);
+	if (perf_cpu_map__empty(evlist->cpus))
+		nr_mmaps = perf_thread_map__nr(evlist->threads);
+
+	return nr_mmaps;
+}
+
+int perf_evlist__mmap_ops(struct perf_evlist *evlist,
+			  struct perf_evlist_mmap_ops *ops,
+			  struct perf_mmap_param *mp)
+{
+	struct perf_evsel *evsel;
+	const struct perf_cpu_map *cpus = evlist->cpus;
+	const struct perf_thread_map *threads = evlist->threads;
+
+	if (!ops || !ops->get || !ops->mmap)
+		return -EINVAL;
+
+	mp->mask = evlist->mmap_len - page_size - 1;
+
+	evlist->nr_mmaps = perf_evlist__nr_mmaps(evlist);
+
+	perf_evlist__for_each_entry(evlist, evsel) {
+		if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
+		    evsel->sample_id == NULL &&
+		    perf_evsel__alloc_id(evsel, perf_cpu_map__nr(cpus), threads->nr) < 0)
+			return -ENOMEM;
+	}
+
+	if (evlist->pollfd.entries == NULL && perf_evlist__alloc_pollfd(evlist) < 0)
+		return -ENOMEM;
+
+	if (perf_cpu_map__empty(cpus))
+		return mmap_per_thread(evlist, ops, mp);
+
+	return mmap_per_cpu(evlist, ops, mp);
+}
+
+int perf_evlist__mmap(struct perf_evlist *evlist, int pages)
+{
+	struct perf_mmap_param mp;
+	struct perf_evlist_mmap_ops ops = {
+		.get  = perf_evlist__mmap_cb_get,
+		.mmap = perf_evlist__mmap_cb_mmap,
+	};
+
+	evlist->mmap_len = (pages + 1) * page_size;
+
+	return perf_evlist__mmap_ops(evlist, &ops, &mp);
+}
+
+void perf_evlist__munmap(struct perf_evlist *evlist)
+{
+	int i;
+
+	if (evlist->mmap) {
+		for (i = 0; i < evlist->nr_mmaps; i++)
+			perf_mmap__munmap(&evlist->mmap[i]);
+	}
+
+	if (evlist->mmap_ovw) {
+		for (i = 0; i < evlist->nr_mmaps; i++)
+			perf_mmap__munmap(&evlist->mmap_ovw[i]);
+	}
+
+	zfree(&evlist->mmap);
+	zfree(&evlist->mmap_ovw);
+}
+
+struct perf_mmap*
+perf_evlist__next_mmap(struct perf_evlist *evlist, struct perf_mmap *map,
+		       bool overwrite)
+{
+	if (map)
+		return map->next;
+
+	return overwrite ? evlist->mmap_ovw_first : evlist->mmap_first;
+}
diff --git a/tools/perf/lib/evsel.c b/tools/perf/lib/evsel.c
index a8cb582e2721..5a89857b0381 100644
--- a/tools/perf/lib/evsel.c
+++ b/tools/perf/lib/evsel.c
@@ -120,7 +120,8 @@ void perf_evsel__close_fd(struct perf_evsel *evsel)
 
 	for (cpu = 0; cpu < xyarray__max_x(evsel->fd); cpu++)
 		for (thread = 0; thread < xyarray__max_y(evsel->fd); ++thread) {
-			close(FD(evsel, cpu, thread));
+			if (FD(evsel, cpu, thread) >= 0)
+				close(FD(evsel, cpu, thread));
 			FD(evsel, cpu, thread) = -1;
 		}
 }
diff --git a/tools/perf/lib/include/internal/evlist.h b/tools/perf/lib/include/internal/evlist.h
index 9f440ab12b76..a2fbccf1922f 100644
--- a/tools/perf/lib/include/internal/evlist.h
+++ b/tools/perf/lib/include/internal/evlist.h
@@ -11,6 +11,7 @@
 
 struct perf_cpu_map;
 struct perf_thread_map;
+struct perf_mmap_param;
 
 struct perf_evlist {
 	struct list_head	 entries;
@@ -22,12 +23,36 @@ struct perf_evlist {
 	size_t			 mmap_len;
 	struct fdarray		 pollfd;
 	struct hlist_head	 heads[PERF_EVLIST__HLIST_SIZE];
+	struct perf_mmap	*mmap;
+	struct perf_mmap	*mmap_ovw;
+	struct perf_mmap	*mmap_first;
+	struct perf_mmap	*mmap_ovw_first;
+};
+
+typedef void
+(*perf_evlist_mmap__cb_idx_t)(struct perf_evlist*, struct perf_mmap_param*, int, bool);
+typedef struct perf_mmap*
+(*perf_evlist_mmap__cb_get_t)(struct perf_evlist*, bool, int);
+typedef int
+(*perf_evlist_mmap__cb_mmap_t)(struct perf_mmap*, struct perf_mmap_param*, int, int);
+
+struct perf_evlist_mmap_ops {
+	perf_evlist_mmap__cb_idx_t	idx;
+	perf_evlist_mmap__cb_get_t	get;
+	perf_evlist_mmap__cb_mmap_t	mmap;
 };
 
 int perf_evlist__alloc_pollfd(struct perf_evlist *evlist);
 int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,
 			    void *ptr, short revent);
 
+int perf_evlist__mmap_ops(struct perf_evlist *evlist,
+			  struct perf_evlist_mmap_ops *ops,
+			  struct perf_mmap_param *mp);
+
+void perf_evlist__init(struct perf_evlist *evlist);
+void perf_evlist__exit(struct perf_evlist *evlist);
+
 /**
  * __perf_evlist__for_each_entry - iterate thru all the evsels
  * @list: list_head instance to iterate
@@ -60,6 +85,24 @@ int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,
 #define perf_evlist__for_each_entry_reverse(evlist, evsel) \
 	__perf_evlist__for_each_entry_reverse(&(evlist)->entries, evsel)
 
+/**
+ * __perf_evlist__for_each_entry_safe - safely iterate thru all the evsels
+ * @list: list_head instance to iterate
+ * @tmp: struct evsel temp iterator
+ * @evsel: struct evsel iterator
+ */
+#define __perf_evlist__for_each_entry_safe(list, tmp, evsel) \
+	list_for_each_entry_safe(evsel, tmp, list, node)
+
+/**
+ * perf_evlist__for_each_entry_safe - safely iterate thru all the evsels
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ * @tmp: struct evsel temp iterator
+ */
+#define perf_evlist__for_each_entry_safe(evlist, tmp, evsel) \
+	__perf_evlist__for_each_entry_safe(&(evlist)->entries, tmp, evsel)
+
 static inline struct perf_evsel *perf_evlist__first(struct perf_evlist *evlist)
 {
 	return list_entry(evlist->entries.next, struct perf_evsel, node);
diff --git a/tools/perf/lib/include/internal/evsel.h b/tools/perf/lib/include/internal/evsel.h
index a69b8299c36f..1ffd083b235e 100644
--- a/tools/perf/lib/include/internal/evsel.h
+++ b/tools/perf/lib/include/internal/evsel.h
@@ -50,6 +50,7 @@ struct perf_evsel {
 	bool			 system_wide;
 };
 
+void perf_evsel__init(struct perf_evsel *evsel, struct perf_event_attr *attr);
 int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads);
 void perf_evsel__close_fd(struct perf_evsel *evsel);
 void perf_evsel__free_fd(struct perf_evsel *evsel);
diff --git a/tools/perf/lib/include/internal/mmap.h b/tools/perf/lib/include/internal/mmap.h
index ba1e519c15b9..be7556e0a2b2 100644
--- a/tools/perf/lib/include/internal/mmap.h
+++ b/tools/perf/lib/include/internal/mmap.h
@@ -10,23 +10,46 @@
 /* perf sample has 16 bits size limit */
 #define PERF_SAMPLE_MAX_SIZE (1 << 16)
 
+struct perf_mmap;
+
+typedef void (*libperf_unmap_cb_t)(struct perf_mmap *map);
+
 /**
  * struct perf_mmap - perf's ring buffer mmap details
  *
  * @refcnt - e.g. code using PERF_EVENT_IOC_SET_OUTPUT to share this
  */
 struct perf_mmap {
-	void		*base;
-	int		 mask;
-	int		 fd;
-	int		 cpu;
-	refcount_t	 refcnt;
-	u64		 prev;
-	u64		 start;
-	u64		 end;
-	bool		 overwrite;
-	u64		 flush;
-	char		 event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8);
+	void			*base;
+	int			 mask;
+	int			 fd;
+	int			 cpu;
+	refcount_t		 refcnt;
+	u64			 prev;
+	u64			 start;
+	u64			 end;
+	bool			 overwrite;
+	u64			 flush;
+	libperf_unmap_cb_t	 unmap_cb;
+	char			 event_copy[PERF_SAMPLE_MAX_SIZE] __aligned(8);
+	struct perf_mmap	*next;
+};
+
+struct perf_mmap_param {
+	int	prot;
+	int	mask;
 };
 
+size_t perf_mmap__mmap_len(struct perf_mmap *map);
+
+void perf_mmap__init(struct perf_mmap *map, struct perf_mmap *prev,
+		     bool overwrite, libperf_unmap_cb_t unmap_cb);
+int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp,
+		    int fd, int cpu);
+void perf_mmap__munmap(struct perf_mmap *map);
+void perf_mmap__get(struct perf_mmap *map);
+void perf_mmap__put(struct perf_mmap *map);
+
+u64 perf_mmap__read_head(struct perf_mmap *map);
+
 #endif /* __LIBPERF_INTERNAL_MMAP_H */
diff --git a/tools/perf/lib/include/internal/tests.h b/tools/perf/lib/include/internal/tests.h
index b7a20cd24ee1..2093e8868a67 100644
--- a/tools/perf/lib/include/internal/tests.h
+++ b/tools/perf/lib/include/internal/tests.h
@@ -4,14 +4,28 @@
 
 #include <stdio.h>
 
-#define __T_START fprintf(stdout, "- running %s...", __FILE__)
-#define __T_OK    fprintf(stdout, "OK\n")
-#define __T_FAIL  fprintf(stdout, "FAIL\n")
+int tests_failed;
+
+#define __T_START					\
+do {							\
+	fprintf(stdout, "- running %s...", __FILE__);	\
+	fflush(NULL);					\
+	tests_failed = 0;				\
+} while (0)
+
+#define __T_END								\
+do {									\
+	if (tests_failed)						\
+		fprintf(stdout, "  FAILED (%d)\n", tests_failed);	\
+	else								\
+		fprintf(stdout, "OK\n");				\
+} while (0)
 
 #define __T(text, cond)                                                          \
 do {                                                                             \
 	if (!(cond)) {                                                           \
 		fprintf(stderr, "FAILED %s:%d %s\n", __FILE__, __LINE__, text);  \
+		tests_failed++;                                                  \
 		return -1;                                                       \
 	}                                                                        \
 } while (0)
diff --git a/tools/perf/lib/include/perf/core.h b/tools/perf/lib/include/perf/core.h
index cfd70e720c1c..a3f6d68edad7 100644
--- a/tools/perf/lib/include/perf/core.h
+++ b/tools/perf/lib/include/perf/core.h
@@ -9,9 +9,12 @@
 #endif
 
 enum libperf_print_level {
+	LIBPERF_ERR,
 	LIBPERF_WARN,
 	LIBPERF_INFO,
 	LIBPERF_DEBUG,
+	LIBPERF_DEBUG2,
+	LIBPERF_DEBUG3,
 };
 
 typedef int (*libperf_print_fn_t)(enum libperf_print_level level,
diff --git a/tools/perf/lib/include/perf/evlist.h b/tools/perf/lib/include/perf/evlist.h
index 8a2ce0757ab2..0a7479dc13bf 100644
--- a/tools/perf/lib/include/perf/evlist.h
+++ b/tools/perf/lib/include/perf/evlist.h
@@ -3,13 +3,13 @@
 #define __LIBPERF_EVLIST_H
 
 #include <perf/core.h>
+#include <stdbool.h>
 
 struct perf_evlist;
 struct perf_evsel;
 struct perf_cpu_map;
 struct perf_thread_map;
 
-LIBPERF_API void perf_evlist__init(struct perf_evlist *evlist);
 LIBPERF_API void perf_evlist__add(struct perf_evlist *evlist,
 				  struct perf_evsel *evsel);
 LIBPERF_API void perf_evlist__remove(struct perf_evlist *evlist,
@@ -32,5 +32,18 @@ LIBPERF_API void perf_evlist__set_maps(struct perf_evlist *evlist,
 				       struct perf_cpu_map *cpus,
 				       struct perf_thread_map *threads);
 LIBPERF_API int perf_evlist__poll(struct perf_evlist *evlist, int timeout);
+LIBPERF_API int perf_evlist__filter_pollfd(struct perf_evlist *evlist,
+					   short revents_and_mask);
+
+LIBPERF_API int perf_evlist__mmap(struct perf_evlist *evlist, int pages);
+LIBPERF_API void perf_evlist__munmap(struct perf_evlist *evlist);
+
+LIBPERF_API struct perf_mmap *perf_evlist__next_mmap(struct perf_evlist *evlist,
+						     struct perf_mmap *map,
+						     bool overwrite);
+#define perf_evlist__for_each_mmap(evlist, pos, overwrite)		\
+	for ((pos) = perf_evlist__next_mmap((evlist), NULL, overwrite);	\
+	     (pos) != NULL;						\
+	     (pos) = perf_evlist__next_mmap((evlist), (pos), overwrite))
 
 #endif /* __LIBPERF_EVLIST_H */
diff --git a/tools/perf/lib/include/perf/evsel.h b/tools/perf/lib/include/perf/evsel.h
index 4388667f265c..557f5815a9c9 100644
--- a/tools/perf/lib/include/perf/evsel.h
+++ b/tools/perf/lib/include/perf/evsel.h
@@ -21,8 +21,6 @@ struct perf_counts_values {
 	};
 };
 
-LIBPERF_API void perf_evsel__init(struct perf_evsel *evsel,
-				  struct perf_event_attr *attr);
 LIBPERF_API struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr);
 LIBPERF_API void perf_evsel__delete(struct perf_evsel *evsel);
 LIBPERF_API int perf_evsel__open(struct perf_evsel *evsel, struct perf_cpu_map *cpus,
diff --git a/tools/perf/lib/include/perf/mmap.h b/tools/perf/lib/include/perf/mmap.h
new file mode 100644
index 000000000000..9508ad90d8b9
--- /dev/null
+++ b/tools/perf/lib/include/perf/mmap.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_MMAP_H
+#define __LIBPERF_MMAP_H
+
+#include <perf/core.h>
+
+struct perf_mmap;
+union perf_event;
+
+LIBPERF_API void perf_mmap__consume(struct perf_mmap *map);
+LIBPERF_API int perf_mmap__read_init(struct perf_mmap *map);
+LIBPERF_API void perf_mmap__read_done(struct perf_mmap *map);
+LIBPERF_API union perf_event *perf_mmap__read_event(struct perf_mmap *map);
+
+#endif /* __LIBPERF_MMAP_H */
diff --git a/tools/perf/lib/internal.h b/tools/perf/lib/internal.h
index dc92f241732e..2c27e158de6b 100644
--- a/tools/perf/lib/internal.h
+++ b/tools/perf/lib/internal.h
@@ -2,6 +2,8 @@
 #ifndef __LIBPERF_INTERNAL_H
 #define __LIBPERF_INTERNAL_H
 
+#include <perf/core.h>
+
 void libperf_print(enum libperf_print_level level,
 		   const char *format, ...)
 	__attribute__((format(printf, 2, 3)));
@@ -11,8 +13,11 @@ do {                            \
 	libperf_print(level, "libperf: " fmt, ##__VA_ARGS__);     \
 } while (0)
 
+#define pr_err(fmt, ...)        __pr(LIBPERF_ERR, fmt, ##__VA_ARGS__)
 #define pr_warning(fmt, ...)    __pr(LIBPERF_WARN, fmt, ##__VA_ARGS__)
 #define pr_info(fmt, ...)       __pr(LIBPERF_INFO, fmt, ##__VA_ARGS__)
 #define pr_debug(fmt, ...)      __pr(LIBPERF_DEBUG, fmt, ##__VA_ARGS__)
+#define pr_debug2(fmt, ...)     __pr(LIBPERF_DEBUG2, fmt, ##__VA_ARGS__)
+#define pr_debug3(fmt, ...)     __pr(LIBPERF_DEBUG3, fmt, ##__VA_ARGS__)
 
 #endif /* __LIBPERF_INTERNAL_H */
diff --git a/tools/perf/lib/libperf.map b/tools/perf/lib/libperf.map
index ab8dbde1136c..7be1af8a546c 100644
--- a/tools/perf/lib/libperf.map
+++ b/tools/perf/lib/libperf.map
@@ -21,7 +21,6 @@ LIBPERF_0.0.1 {
 		perf_evsel__delete;
 		perf_evsel__enable;
 		perf_evsel__disable;
-		perf_evsel__init;
 		perf_evsel__open;
 		perf_evsel__close;
 		perf_evsel__read;
@@ -34,12 +33,19 @@ LIBPERF_0.0.1 {
 		perf_evlist__close;
 		perf_evlist__enable;
 		perf_evlist__disable;
-		perf_evlist__init;
 		perf_evlist__add;
 		perf_evlist__remove;
 		perf_evlist__next;
 		perf_evlist__set_maps;
 		perf_evlist__poll;
+		perf_evlist__mmap;
+		perf_evlist__munmap;
+		perf_evlist__filter_pollfd;
+		perf_evlist__next_mmap;
+		perf_mmap__consume;
+		perf_mmap__read_init;
+		perf_mmap__read_done;
+		perf_mmap__read_event;
 	local:
 		*;
 };
diff --git a/tools/perf/lib/mmap.c b/tools/perf/lib/mmap.c
new file mode 100644
index 000000000000..79d5ed6c38cc
--- /dev/null
+++ b/tools/perf/lib/mmap.c
@@ -0,0 +1,275 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <sys/mman.h>
+#include <inttypes.h>
+#include <asm/bug.h>
+#include <errno.h>
+#include <string.h>
+#include <linux/ring_buffer.h>
+#include <linux/perf_event.h>
+#include <perf/mmap.h>
+#include <perf/event.h>
+#include <internal/mmap.h>
+#include <internal/lib.h>
+#include <linux/kernel.h>
+#include "internal.h"
+
+void perf_mmap__init(struct perf_mmap *map, struct perf_mmap *prev,
+		     bool overwrite, libperf_unmap_cb_t unmap_cb)
+{
+	map->fd = -1;
+	map->overwrite = overwrite;
+	map->unmap_cb  = unmap_cb;
+	refcount_set(&map->refcnt, 0);
+	if (prev)
+		prev->next = map;
+}
+
+size_t perf_mmap__mmap_len(struct perf_mmap *map)
+{
+	return map->mask + 1 + page_size;
+}
+
+int perf_mmap__mmap(struct perf_mmap *map, struct perf_mmap_param *mp,
+		    int fd, int cpu)
+{
+	map->prev = 0;
+	map->mask = mp->mask;
+	map->base = mmap(NULL, perf_mmap__mmap_len(map), mp->prot,
+			 MAP_SHARED, fd, 0);
+	if (map->base == MAP_FAILED) {
+		map->base = NULL;
+		return -1;
+	}
+
+	map->fd  = fd;
+	map->cpu = cpu;
+	return 0;
+}
+
+void perf_mmap__munmap(struct perf_mmap *map)
+{
+	if (map && map->base != NULL) {
+		munmap(map->base, perf_mmap__mmap_len(map));
+		map->base = NULL;
+		map->fd = -1;
+		refcount_set(&map->refcnt, 0);
+	}
+	if (map && map->unmap_cb)
+		map->unmap_cb(map);
+}
+
+void perf_mmap__get(struct perf_mmap *map)
+{
+	refcount_inc(&map->refcnt);
+}
+
+void perf_mmap__put(struct perf_mmap *map)
+{
+	BUG_ON(map->base && refcount_read(&map->refcnt) == 0);
+
+	if (refcount_dec_and_test(&map->refcnt))
+		perf_mmap__munmap(map);
+}
+
+static inline void perf_mmap__write_tail(struct perf_mmap *md, u64 tail)
+{
+	ring_buffer_write_tail(md->base, tail);
+}
+
+u64 perf_mmap__read_head(struct perf_mmap *map)
+{
+	return ring_buffer_read_head(map->base);
+}
+
+static bool perf_mmap__empty(struct perf_mmap *map)
+{
+	struct perf_event_mmap_page *pc = map->base;
+
+	return perf_mmap__read_head(map) == map->prev && !pc->aux_size;
+}
+
+void perf_mmap__consume(struct perf_mmap *map)
+{
+	if (!map->overwrite) {
+		u64 old = map->prev;
+
+		perf_mmap__write_tail(map, old);
+	}
+
+	if (refcount_read(&map->refcnt) == 1 && perf_mmap__empty(map))
+		perf_mmap__put(map);
+}
+
+static int overwrite_rb_find_range(void *buf, int mask, u64 *start, u64 *end)
+{
+	struct perf_event_header *pheader;
+	u64 evt_head = *start;
+	int size = mask + 1;
+
+	pr_debug2("%s: buf=%p, start=%"PRIx64"\n", __func__, buf, *start);
+	pheader = (struct perf_event_header *)(buf + (*start & mask));
+	while (true) {
+		if (evt_head - *start >= (unsigned int)size) {
+			pr_debug("Finished reading overwrite ring buffer: rewind\n");
+			if (evt_head - *start > (unsigned int)size)
+				evt_head -= pheader->size;
+			*end = evt_head;
+			return 0;
+		}
+
+		pheader = (struct perf_event_header *)(buf + (evt_head & mask));
+
+		if (pheader->size == 0) {
+			pr_debug("Finished reading overwrite ring buffer: get start\n");
+			*end = evt_head;
+			return 0;
+		}
+
+		evt_head += pheader->size;
+		pr_debug3("move evt_head: %"PRIx64"\n", evt_head);
+	}
+	WARN_ONCE(1, "Shouldn't get here\n");
+	return -1;
+}
+
+/*
+ * Report the start and end of the available data in ringbuffer
+ */
+static int __perf_mmap__read_init(struct perf_mmap *md)
+{
+	u64 head = perf_mmap__read_head(md);
+	u64 old = md->prev;
+	unsigned char *data = md->base + page_size;
+	unsigned long size;
+
+	md->start = md->overwrite ? head : old;
+	md->end = md->overwrite ? old : head;
+
+	if ((md->end - md->start) < md->flush)
+		return -EAGAIN;
+
+	size = md->end - md->start;
+	if (size > (unsigned long)(md->mask) + 1) {
+		if (!md->overwrite) {
+			WARN_ONCE(1, "failed to keep up with mmap data. (warn only once)\n");
+
+			md->prev = head;
+			perf_mmap__consume(md);
+			return -EAGAIN;
+		}
+
+		/*
+		 * Backward ring buffer is full. We still have a chance to read
+		 * most of data