Adding upstream version 6.6.15.upstream/6.6.15

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

49 files changed, 6605 insertions, 0 deletions

diff --git a/tools/lib/perf/Build b/tools/lib/perf/Build
new file mode 100644
index 000000000..e8f5b7fb9
--- /dev/null
+++ b/tools/lib/perf/Build
@@ -0,0 +1,15 @@
+libperf-y += core.o
+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
+
+$(OUTPUT)zalloc.o: ../../lib/zalloc.c FORCE
+	$(call rule_mkdir)
+	$(call if_changed_dep,cc_o_c)
+
+tests-y += tests/
diff --git a/tools/lib/perf/Documentation/Makefile b/tools/lib/perf/Documentation/Makefile
new file mode 100644
index 000000000..972754082
--- /dev/null
+++ b/tools/lib/perf/Documentation/Makefile
@@ -0,0 +1,156 @@
+# SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
+# Most of this file is copied from tools/perf/Documentation/Makefile
+
+include ../../../scripts/Makefile.include
+include ../../../scripts/utilities.mak
+
+MAN3_TXT  = libperf.txt
+MAN7_TXT  = libperf-counting.txt libperf-sampling.txt
+MAN_EX    = examples/*.c
+
+MAN_TXT   = $(MAN3_TXT) $(MAN7_TXT)
+
+_MAN_XML  = $(patsubst %.txt,%.xml,$(MAN_TXT))
+_MAN_HTML = $(patsubst %.txt,%.html,$(MAN_TXT))
+_MAN_3    = $(patsubst %.txt,%.3,$(MAN3_TXT))
+_MAN_7    = $(patsubst %.txt,%.7,$(MAN7_TXT))
+
+MAN_XML   = $(addprefix $(OUTPUT),$(_MAN_XML))
+MAN_HTML  = $(addprefix $(OUTPUT),$(_MAN_HTML))
+MAN_3     = $(addprefix $(OUTPUT),$(_MAN_3))
+MAN_7     = $(addprefix $(OUTPUT),$(_MAN_7))
+MAN_X     = $(MAN_3) $(MAN_7)
+
+# Make the path relative to DESTDIR, not prefix
+ifndef DESTDIR
+  prefix ?=$(HOME)
+endif
+
+mandir  ?= $(prefix)/share/man
+man3dir  = $(mandir)/man3
+man7dir  = $(mandir)/man7
+
+docdir  ?= $(prefix)/share/doc/libperf
+htmldir  = $(docdir)/html
+exdir    = $(docdir)/examples
+
+ASCIIDOC        = asciidoc
+ASCIIDOC_EXTRA  = --unsafe -f asciidoc.conf
+ASCIIDOC_HTML   = xhtml11
+MANPAGE_XSL     = manpage-normal.xsl
+XMLTO_EXTRA     =
+XMLTO           =xmlto
+
+INSTALL ?= install
+RM      ?= rm -f
+
+# For asciidoc ...
+#	-7.1.2,	no extra settings are needed.
+#	8.0-,	set ASCIIDOC8.
+#
+
+# For docbook-xsl ...
+#	-1.68.1,	set ASCIIDOC_NO_ROFF? (based on changelog from 1.73.0)
+#	1.69.0,		no extra settings are needed?
+#	1.69.1-1.71.0,	set DOCBOOK_SUPPRESS_SP?
+#	1.71.1,		no extra settings are needed?
+#	1.72.0,		set DOCBOOK_XSL_172.
+#	1.73.0-,	set ASCIIDOC_NO_ROFF
+
+# If you had been using DOCBOOK_XSL_172 in an attempt to get rid
+# of 'the ".ft C" problem' in your generated manpages, and you
+# instead ended up with weird characters around callouts, try
+# using ASCIIDOC_NO_ROFF instead (it works fine with ASCIIDOC8).
+
+ifdef ASCIIDOC8
+  ASCIIDOC_EXTRA += -a asciidoc7compatible
+endif
+ifdef DOCBOOK_XSL_172
+  ASCIIDOC_EXTRA += -a libperf-asciidoc-no-roff
+  MANPAGE_XSL = manpage-1.72.xsl
+else
+  ifdef ASCIIDOC_NO_ROFF
+    # docbook-xsl after 1.72 needs the regular XSL, but will not
+    # pass-thru raw roff codes from asciidoc.conf, so turn them off.
+    ASCIIDOC_EXTRA += -a libperf-asciidoc-no-roff
+  endif
+endif
+ifdef MAN_BOLD_LITERAL
+  XMLTO_EXTRA += -m manpage-bold-literal.xsl
+endif
+ifdef DOCBOOK_SUPPRESS_SP
+  XMLTO_EXTRA += -m manpage-suppress-sp.xsl
+endif
+
+DESTDIR ?=
+DESTDIR_SQ = '$(subst ','\'',$(DESTDIR))'
+
+export DESTDIR DESTDIR_SQ
+
+# Please note that there is a minor bug in asciidoc.
+# The version after 6.0.3 _will_ include the patch found here:
+#   http://marc.theaimsgroup.com/?l=libtraceevent&m=111558757202243&w=2
+#
+# Until that version is released you may have to apply the patch
+# yourself - yes, all 6 characters of it!
+
+QUIET_SUBDIR0  = +$(MAKE) -C # space to separate -C and subdir
+QUIET_SUBDIR1  =
+
+ifneq ($(findstring $(MAKEFLAGS),w),w)
+  PRINT_DIR = --no-print-directory
+else # "make -w"
+  NO_SUBDIR = :
+endif
+
+ifneq ($(findstring $(MAKEFLAGS),s),s)
+  ifneq ($(V),1)
+    QUIET_ASCIIDOC = @echo '  ASCIIDOC '$@;
+    QUIET_XMLTO    = @echo '  XMLTO    '$@;
+  endif
+endif
+
+all: $(MAN_X) $(MAN_HTML)
+
+$(MAN_HTML) $(MAN_X): asciidoc.conf
+
+install-man: all
+	$(call QUIET_INSTALL, man) \
+		$(INSTALL) -d -m 755 $(DESTDIR)$(man3dir); \
+		$(INSTALL) -m 644 $(MAN_3) $(DESTDIR)$(man3dir); \
+		$(INSTALL) -d -m 755 $(DESTDIR)$(man7dir); \
+		$(INSTALL) -m 644 $(MAN_7) $(DESTDIR)$(man7dir);
+
+install-html:
+	$(call QUIET_INSTALL, html) \
+		$(INSTALL) -d -m 755 $(DESTDIR)$(htmldir); \
+		$(INSTALL) -m 644 $(MAN_HTML) $(DESTDIR)$(htmldir); \
+
+install-examples:
+	$(call QUIET_INSTALL, examples) \
+		$(INSTALL) -d -m 755 $(DESTDIR)$(exdir); \
+		$(INSTALL) -m 644 $(MAN_EX) $(DESTDIR)$(exdir); \
+
+CLEAN_FILES =					\
+	$(MAN_XML) $(addsuffix +,$(MAN_XML))	\
+	$(MAN_HTML) $(addsuffix +,$(MAN_HTML))	\
+	$(MAN_X)
+
+clean:
+	$(call QUIET_CLEAN, Documentation) $(RM) $(CLEAN_FILES)
+
+$(MAN_3): $(OUTPUT)%.3: %.xml
+	$(QUIET_XMLTO)$(XMLTO) -o $(OUTPUT). -m $(MANPAGE_XSL) $(XMLTO_EXTRA) man $<
+
+$(MAN_7): $(OUTPUT)%.7: %.xml
+	$(QUIET_XMLTO)$(XMLTO) -o $(OUTPUT). -m $(MANPAGE_XSL) $(XMLTO_EXTRA) man $<
+
+$(MAN_XML): $(OUTPUT)%.xml: %.txt
+	$(QUIET_ASCIIDOC)$(ASCIIDOC) -b docbook -d manpage \
+		$(ASCIIDOC_EXTRA) -alibperf_version=$(EVENT_PARSE_VERSION) -o $@+ $< && \
+	mv $@+ $@
+
+$(MAN_HTML): $(OUTPUT)%.html: %.txt
+	$(QUIET_ASCIIDOC)$(ASCIIDOC) -b $(ASCIIDOC_HTML) -d manpage \
+	$(ASCIIDOC_EXTRA) -aperf_version=$(EVENT_PARSE_VERSION) -o $@+ $< && \
+	mv $@+ $@
diff --git a/tools/lib/perf/Documentation/asciidoc.conf b/tools/lib/perf/Documentation/asciidoc.conf
new file mode 100644
index 000000000..9d5a5a5ee
--- /dev/null
+++ b/tools/lib/perf/Documentation/asciidoc.conf
@@ -0,0 +1,120 @@
+## linktep: macro
+#
+# Usage: linktep:command[manpage-section]
+#
+# Note, {0} is the manpage section, while {target} is the command.
+#
+# Show TEP link as: <command>(<section>); if section is defined, else just show
+# the command.
+
+[macros]
+(?su)[\\]?(?P<name>linktep):(?P<target>\S*?)\[(?P<attrlist>.*?)\]=
+
+[attributes]
+asterisk=&#42;
+plus=&#43;
+caret=&#94;
+startsb=&#91;
+endsb=&#93;
+tilde=&#126;
+
+ifdef::backend-docbook[]
+[linktep-inlinemacro]
+{0%{target}}
+{0#<citerefentry>}
+{0#<refentrytitle>{target}</refentrytitle><manvolnum>{0}</manvolnum>}
+{0#</citerefentry>}
+endif::backend-docbook[]
+
+ifdef::backend-docbook[]
+ifndef::tep-asciidoc-no-roff[]
+# "unbreak" docbook-xsl v1.68 for manpages. v1.69 works with or without this.
+# v1.72 breaks with this because it replaces dots not in roff requests.
+[listingblock]
+<example><title>{title}</title>
+<literallayout>
+ifdef::doctype-manpage[]
+&#10;.ft C&#10;
+endif::doctype-manpage[]
+|
+ifdef::doctype-manpage[]
+&#10;.ft&#10;
+endif::doctype-manpage[]
+</literallayout>
+{title#}</example>
+endif::tep-asciidoc-no-roff[]
+
+ifdef::tep-asciidoc-no-roff[]
+ifdef::doctype-manpage[]
+# The following two small workarounds insert a simple paragraph after screen
+[listingblock]
+<example><title>{title}</title>
+<literallayout>
+|
+</literallayout><simpara></simpara>
+{title#}</example>
+
+[verseblock]
+<formalpara{id? id="{id}"}><title>{title}</title><para>
+{title%}<literallayout{id? id="{id}"}>
+{title#}<literallayout>
+|
+</literallayout>
+{title#}</para></formalpara>
+{title%}<simpara></simpara>
+endif::doctype-manpage[]
+endif::tep-asciidoc-no-roff[]
+endif::backend-docbook[]
+
+ifdef::doctype-manpage[]
+ifdef::backend-docbook[]
+[header]
+template::[header-declarations]
+<refentry>
+<refmeta>
+<refentrytitle>{mantitle}</refentrytitle>
+<manvolnum>{manvolnum}</manvolnum>
+<refmiscinfo class="source">libperf</refmiscinfo>
+<refmiscinfo class="version">{libperf_version}</refmiscinfo>
+<refmiscinfo class="manual">libperf Manual</refmiscinfo>
+</refmeta>
+<refnamediv>
+  <refname>{manname1}</refname>
+  <refname>{manname2}</refname>
+  <refname>{manname3}</refname>
+  <refname>{manname4}</refname>
+  <refname>{manname5}</refname>
+  <refname>{manname6}</refname>
+  <refname>{manname7}</refname>
+  <refname>{manname8}</refname>
+  <refname>{manname9}</refname>
+  <refname>{manname10}</refname>
+  <refname>{manname11}</refname>
+  <refname>{manname12}</refname>
+  <refname>{manname13}</refname>
+  <refname>{manname14}</refname>
+  <refname>{manname15}</refname>
+  <refname>{manname16}</refname>
+  <refname>{manname17}</refname>
+  <refname>{manname18}</refname>
+  <refname>{manname19}</refname>
+  <refname>{manname20}</refname>
+  <refname>{manname21}</refname>
+  <refname>{manname22}</refname>
+  <refname>{manname23}</refname>
+  <refname>{manname24}</refname>
+  <refname>{manname25}</refname>
+  <refname>{manname26}</refname>
+  <refname>{manname27}</refname>
+  <refname>{manname28}</refname>
+  <refname>{manname29}</refname>
+  <refname>{manname30}</refname>
+  <refpurpose>{manpurpose}</refpurpose>
+</refnamediv>
+endif::backend-docbook[]
+endif::doctype-manpage[]
+
+ifdef::backend-xhtml11[]
+[linktep-inlinemacro]
+<a href="{target}.html">{target}{0?({0})}</a>
+endif::backend-xhtml11[]
diff --git a/tools/lib/perf/Documentation/examples/counting.c b/tools/lib/perf/Documentation/examples/counting.c
new file mode 100644
index 000000000..608569357
--- /dev/null
+++ b/tools/lib/perf/Documentation/examples/counting.c
@@ -0,0 +1,83 @@
+#include <linux/perf_event.h>
+#include <perf/evlist.h>
+#include <perf/evsel.h>
+#include <perf/cpumap.h>
+#include <perf/threadmap.h>
+#include <perf/mmap.h>
+#include <perf/core.h>
+#include <perf/event.h>
+#include <stdio.h>
+#include <unistd.h>
+
+static int libperf_print(enum libperf_print_level level,
+                         const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+int main(int argc, char **argv)
+{
+	int count = 100000, err = 0;
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel;
+	struct perf_thread_map *threads;
+	struct perf_counts_values counts;
+
+	struct perf_event_attr attr1 = {
+		.type        = PERF_TYPE_SOFTWARE,
+		.config      = PERF_COUNT_SW_CPU_CLOCK,
+		.read_format = PERF_FORMAT_TOTAL_TIME_ENABLED|PERF_FORMAT_TOTAL_TIME_RUNNING,
+		.disabled    = 1,
+	};
+	struct perf_event_attr attr2 = {
+		.type        = PERF_TYPE_SOFTWARE,
+		.config      = PERF_COUNT_SW_TASK_CLOCK,
+		.read_format = PERF_FORMAT_TOTAL_TIME_ENABLED|PERF_FORMAT_TOTAL_TIME_RUNNING,
+		.disabled    = 1,
+	};
+
+	libperf_init(libperf_print);
+	threads = perf_thread_map__new_dummy();
+	if (!threads) {
+		fprintf(stderr, "failed to create threads\n");
+		return -1;
+	}
+	perf_thread_map__set_pid(threads, 0, 0);
+	evlist = perf_evlist__new();
+	if (!evlist) {
+		fprintf(stderr, "failed to create evlist\n");
+		goto out_threads;
+	}
+	evsel = perf_evsel__new(&attr1);
+	if (!evsel) {
+		fprintf(stderr, "failed to create evsel1\n");
+		goto out_evlist;
+	}
+	perf_evlist__add(evlist, evsel);
+	evsel = perf_evsel__new(&attr2);
+	if (!evsel) {
+		fprintf(stderr, "failed to create evsel2\n");
+		goto out_evlist;
+	}
+	perf_evlist__add(evlist, evsel);
+	perf_evlist__set_maps(evlist, NULL, threads);
+	err = perf_evlist__open(evlist);
+	if (err) {
+		fprintf(stderr, "failed to open evsel\n");
+		goto out_evlist;
+	}
+	perf_evlist__enable(evlist);
+	while (count--);
+	perf_evlist__disable(evlist);
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		perf_evsel__read(evsel, 0, 0, &counts);
+		fprintf(stdout, "count %llu, enabled %llu, run %llu\n",
+				counts.val, counts.ena, counts.run);
+	}
+	perf_evlist__close(evlist);
+out_evlist:
+	perf_evlist__delete(evlist);
+out_threads:
+	perf_thread_map__put(threads);
+	return err;
+}
diff --git a/tools/lib/perf/Documentation/examples/sampling.c b/tools/lib/perf/Documentation/examples/sampling.c
new file mode 100644
index 000000000..8e1a926a9
--- /dev/null
+++ b/tools/lib/perf/Documentation/examples/sampling.c
@@ -0,0 +1,119 @@
+#include <linux/perf_event.h>
+#include <perf/evlist.h>
+#include <perf/evsel.h>
+#include <perf/cpumap.h>
+#include <perf/threadmap.h>
+#include <perf/mmap.h>
+#include <perf/core.h>
+#include <perf/event.h>
+#include <stdio.h>
+#include <unistd.h>
+
+static int libperf_print(enum libperf_print_level level,
+                         const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+union u64_swap {
+	__u64 val64;
+	__u32 val32[2];
+};
+
+int main(int argc, char **argv)
+{
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel;
+	struct perf_mmap *map;
+	struct perf_cpu_map *cpus;
+	struct perf_event_attr attr = {
+		.type        = PERF_TYPE_HARDWARE,
+		.config      = PERF_COUNT_HW_CPU_CYCLES,
+		.disabled    = 1,
+		.freq        = 1,
+		.sample_freq = 10,
+		.sample_type = PERF_SAMPLE_IP|PERF_SAMPLE_TID|PERF_SAMPLE_CPU|PERF_SAMPLE_PERIOD,
+	};
+	int err = -1;
+	union perf_event *event;
+
+	libperf_init(libperf_print);
+
+	cpus = perf_cpu_map__new(NULL);
+	if (!cpus) {
+		fprintf(stderr, "failed to create cpus\n");
+		return -1;
+	}
+
+	evlist = perf_evlist__new();
+	if (!evlist) {
+		fprintf(stderr, "failed to create evlist\n");
+		goto out_cpus;
+	}
+
+	evsel = perf_evsel__new(&attr);
+	if (!evsel) {
+		fprintf(stderr, "failed to create cycles\n");
+		goto out_cpus;
+	}
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_maps(evlist, cpus, NULL);
+
+	err = perf_evlist__open(evlist);
+	if (err) {
+		fprintf(stderr, "failed to open evlist\n");
+		goto out_evlist;
+	}
+
+	err = perf_evlist__mmap(evlist, 4);
+	if (err) {
+		fprintf(stderr, "failed to mmap evlist\n");
+		goto out_evlist;
+	}
+
+	perf_evlist__enable(evlist);
+	sleep(3);
+	perf_evlist__disable(evlist);
+
+	perf_evlist__for_each_mmap(evlist, map, false) {
+		if (perf_mmap__read_init(map) < 0)
+			continue;
+
+		while ((event = perf_mmap__read_event(map)) != NULL) {
+			int cpu, pid, tid;
+			__u64 ip, period, *array;
+			union u64_swap u;
+
+			array = event->sample.array;
+
+			ip = *array;
+			array++;
+
+			u.val64 = *array;
+			pid = u.val32[0];
+			tid = u.val32[1];
+			array++;
+
+			u.val64 = *array;
+			cpu = u.val32[0];
+			array++;
+
+			period = *array;
+
+			fprintf(stdout, "cpu %3d, pid %6d, tid %6d, ip %20llx, period %20llu\n",
+				cpu, pid, tid, ip, period);
+
+			perf_mmap__consume(map);
+		}
+
+		perf_mmap__read_done(map);
+	}
+
+out_evlist:
+	perf_evlist__delete(evlist);
+out_cpus:
+	perf_cpu_map__put(cpus);
+	return err;
+}
diff --git a/tools/lib/perf/Documentation/libperf-counting.txt b/tools/lib/perf/Documentation/libperf-counting.txt
new file mode 100644
index 000000000..8b75efcd6
--- /dev/null
+++ b/tools/lib/perf/Documentation/libperf-counting.txt
@@ -0,0 +1,213 @@
+libperf-counting(7)
+===================
+
+NAME
+----
+libperf-counting - counting interface
+
+DESCRIPTION
+-----------
+The counting interface provides API to measure and get count for specific perf events.
+
+The following test tries to explain count on `counting.c` example.
+
+It is by no means complete guide to counting, but shows libperf basic API for counting.
+
+The `counting.c` comes with libperf package and can be compiled and run like:
+
+[source,bash]
+--
+$ gcc -o counting counting.c -lperf
+$ sudo ./counting
+count 176792, enabled 176944, run 176944
+count 176242, enabled 176242, run 176242
+--
+
+It requires root access, because of the `PERF_COUNT_SW_CPU_CLOCK` event,
+which is available only for root.
+
+The `counting.c` example monitors two events on the current process and displays
+their count, in a nutshell it:
+
+* creates events
+* adds them to the event list
+* opens and enables events through the event list
+* does some workload
+* disables events
+* reads and displays event counts
+* destroys the event list
+
+The first thing you need to do before using libperf is to call init function:
+
+[source,c]
+--
+  8 static int libperf_print(enum libperf_print_level level,
+  9                          const char *fmt, va_list ap)
+ 10 {
+ 11         return vfprintf(stderr, fmt, ap);
+ 12 }
+
+ 14 int main(int argc, char **argv)
+ 15 {
+ ...
+ 35         libperf_init(libperf_print);
+--
+
+It will setup the library and sets function for debug output from library.
+
+The `libperf_print` callback will receive any message with its debug level,
+defined as:
+
+[source,c]
+--
+enum libperf_print_level {
+        LIBPERF_ERR,
+        LIBPERF_WARN,
+        LIBPERF_INFO,
+        LIBPERF_DEBUG,
+        LIBPERF_DEBUG2,
+        LIBPERF_DEBUG3,
+};
+--
+
+Once the setup is complete we start by defining specific events using the `struct perf_event_attr`.
+
+We create software events for cpu and task:
+
+[source,c]
+--
+ 20         struct perf_event_attr attr1 = {
+ 21                 .type        = PERF_TYPE_SOFTWARE,
+ 22                 .config      = PERF_COUNT_SW_CPU_CLOCK,
+ 23                 .read_format = PERF_FORMAT_TOTAL_TIME_ENABLED|PERF_FORMAT_TOTAL_TIME_RUNNING,
+ 24                 .disabled    = 1,
+ 25         };
+ 26         struct perf_event_attr attr2 = {
+ 27                 .type        = PERF_TYPE_SOFTWARE,
+ 28                 .config      = PERF_COUNT_SW_TASK_CLOCK,
+ 29                 .read_format = PERF_FORMAT_TOTAL_TIME_ENABLED|PERF_FORMAT_TOTAL_TIME_RUNNING,
+ 30                 .disabled    = 1,
+ 31         };
+--
+
+The `read_format` setup tells perf to include timing details together with each count.
+
+Next step is to prepare threads map.
+
+In this case we will monitor current process, so we create threads map with single pid (0):
+
+[source,c]
+--
+ 37         threads = perf_thread_map__new_dummy();
+ 38         if (!threads) {
+ 39                 fprintf(stderr, "failed to create threads\n");
+ 40                 return -1;
+ 41         }
+ 42
+ 43         perf_thread_map__set_pid(threads, 0, 0);
+--
+
+Now we create libperf's event list, which will serve as holder for the events we want:
+
+[source,c]
+--
+ 45         evlist = perf_evlist__new();
+ 46         if (!evlist) {
+ 47                 fprintf(stderr, "failed to create evlist\n");
+ 48                 goto out_threads;
+ 49         }
+--
+
+We create libperf's events for the attributes we defined earlier and add them to the list:
+
+[source,c]
+--
+ 51         evsel = perf_evsel__new(&attr1);
+ 52         if (!evsel) {
+ 53                 fprintf(stderr, "failed to create evsel1\n");
+ 54                 goto out_evlist;
+ 55         }
+ 56
+ 57         perf_evlist__add(evlist, evsel);
+ 58
+ 59         evsel = perf_evsel__new(&attr2);
+ 60         if (!evsel) {
+ 61                 fprintf(stderr, "failed to create evsel2\n");
+ 62                 goto out_evlist;
+ 63         }
+ 64
+ 65         perf_evlist__add(evlist, evsel);
+--
+
+Configure event list with the thread map and open events:
+
+[source,c]
+--
+ 67         perf_evlist__set_maps(evlist, NULL, threads);
+ 68
+ 69         err = perf_evlist__open(evlist);
+ 70         if (err) {
+ 71                 fprintf(stderr, "failed to open evsel\n");
+ 72                 goto out_evlist;
+ 73         }
+--
+
+Both events are created as disabled (note the `disabled = 1` assignment above),
+so we need to enable the whole list explicitly (both events).
+
+From this moment events are counting and we can do our workload.
+
+When we are done we disable the events list.
+
+[source,c]
+--
+ 75         perf_evlist__enable(evlist);
+ 76
+ 77         while (count--);
+ 78
+ 79         perf_evlist__disable(evlist);
+--
+
+Now we need to get the counts from events, following code iterates through the
+events list and read counts:
+
+[source,c]
+--
+ 81         perf_evlist__for_each_evsel(evlist, evsel) {
+ 82                 perf_evsel__read(evsel, 0, 0, &counts);
+ 83                 fprintf(stdout, "count %llu, enabled %llu, run %llu\n",
+ 84                         counts.val, counts.ena, counts.run);
+ 85         }
+--
+
+And finally cleanup.
+
+We close the whole events list (both events) and remove it together with the threads map:
+
+[source,c]
+--
+ 87         perf_evlist__close(evlist);
+ 88
+ 89 out_evlist:
+ 90         perf_evlist__delete(evlist);
+ 91 out_threads:
+ 92         perf_thread_map__put(threads);
+ 93         return err;
+ 94 }
+--
+
+REPORTING BUGS
+--------------
+Report bugs to <linux-perf-users@vger.kernel.org>.
+
+LICENSE
+-------
+libperf is Free Software licensed under the GNU LGPL 2.1
+
+RESOURCES
+---------
+https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
+
+SEE ALSO
+--------
+libperf(3), libperf-sampling(7)
diff --git a/tools/lib/perf/Documentation/libperf-sampling.txt b/tools/lib/perf/Documentation/libperf-sampling.txt
new file mode 100644
index 000000000..d6ca24f6e
--- /dev/null
+++ b/tools/lib/perf/Documentation/libperf-sampling.txt
@@ -0,0 +1,244 @@
+libperf-sampling(7)
+===================
+
+NAME
+----
+libperf-sampling - sampling interface
+
+
+DESCRIPTION
+-----------
+The sampling interface provides API to measure and get count for specific perf events.
+
+The following test tries to explain count on `sampling.c` example.
+
+It is by no means complete guide to sampling, but shows libperf basic API for sampling.
+
+The `sampling.c` comes with libperf package and can be compiled and run like:
+
+[source,bash]
+--
+$ gcc -o sampling sampling.c -lperf
+$ sudo ./sampling
+cpu   0, pid      0, tid      0, ip     ffffffffad06c4e6, period                    1
+cpu   0, pid   4465, tid   4469, ip     ffffffffad118748, period             18322959
+cpu   0, pid      0, tid      0, ip     ffffffffad115722, period             33544846
+cpu   0, pid   4465, tid   4470, ip         7f84fe0cdad6, period             23687474
+cpu   0, pid      0, tid      0, ip     ffffffffad9e0349, period             34255790
+cpu   0, pid   4465, tid   4469, ip     ffffffffad136581, period             38664069
+cpu   0, pid      0, tid      0, ip     ffffffffad9e55e2, period             21922384
+cpu   0, pid   4465, tid   4470, ip         7f84fe0ebebf, period             17655175
+...
+--
+
+It requires root access, because it uses hardware cycles event.
+
+The `sampling.c` example profiles/samples all CPUs with hardware cycles, in a
+nutshell it:
+
+- creates events
+- adds them to the event list
+- opens and enables events through the event list
+- sleeps for 3 seconds
+- disables events
+- reads and displays recorded samples
+- destroys the event list
+
+The first thing you need to do before using libperf is to call init function:
+
+[source,c]
+--
+ 12 static int libperf_print(enum libperf_print_level level,
+ 13                          const char *fmt, va_list ap)
+ 14 {
+ 15         return vfprintf(stderr, fmt, ap);
+ 16 }
+
+ 23 int main(int argc, char **argv)
+ 24 {
+ ...
+ 40         libperf_init(libperf_print);
+--
+
+It will setup the library and sets function for debug output from library.
+
+The `libperf_print` callback will receive any message with its debug level,
+defined as:
+
+[source,c]
+--
+enum libperf_print_level {
+        LIBPERF_ERR,
+        LIBPERF_WARN,
+        LIBPERF_INFO,
+        LIBPERF_DEBUG,
+        LIBPERF_DEBUG2,
+        LIBPERF_DEBUG3,
+};
+--
+
+Once the setup is complete we start by defining cycles event using the `struct perf_event_attr`:
+
+[source,c]
+--
+ 29         struct perf_event_attr attr = {
+ 30                 .type        = PERF_TYPE_HARDWARE,
+ 31                 .config      = PERF_COUNT_HW_CPU_CYCLES,
+ 32                 .disabled    = 1,
+ 33                 .freq        = 1,
+ 34                 .sample_freq = 10,
+ 35                 .sample_type = PERF_SAMPLE_IP|PERF_SAMPLE_TID|PERF_SAMPLE_CPU|PERF_SAMPLE_PERIOD,
+ 36         };
+--
+
+Next step is to prepare CPUs map.
+
+In this case we will monitor all the available CPUs:
+
+[source,c]
+--
+ 42         cpus = perf_cpu_map__new(NULL);
+ 43         if (!cpus) {
+ 44                 fprintf(stderr, "failed to create cpus\n");
+ 45                 return -1;
+ 46         }
+--
+
+Now we create libperf's event list, which will serve as holder for the cycles event:
+
+[source,c]
+--
+ 48         evlist = perf_evlist__new();
+ 49         if (!evlist) {
+ 50                 fprintf(stderr, "failed to create evlist\n");
+ 51                 goto out_cpus;
+ 52         }
+--
+
+We create libperf's event for the cycles attribute we defined earlier and add it to the list:
+
+[source,c]
+--
+ 54         evsel = perf_evsel__new(&attr);
+ 55         if (!evsel) {
+ 56                 fprintf(stderr, "failed to create cycles\n");
+ 57                 goto out_cpus;
+ 58         }
+ 59
+ 60         perf_evlist__add(evlist, evsel);
+--
+
+Configure event list with the cpus map and open event:
+
+[source,c]
+--
+ 62         perf_evlist__set_maps(evlist, cpus, NULL);
+ 63
+ 64         err = perf_evlist__open(evlist);
+ 65         if (err) {
+ 66                 fprintf(stderr, "failed to open evlist\n");
+ 67                 goto out_evlist;
+ 68         }
+--
+
+Once the events list is open, we can create memory maps AKA perf ring buffers:
+
+[source,c]
+--
+ 70         err = perf_evlist__mmap(evlist, 4);
+ 71         if (err) {
+ 72                 fprintf(stderr, "failed to mmap evlist\n");
+ 73                 goto out_evlist;
+ 74         }
+--
+
+The event is created as disabled (note the `disabled = 1` assignment above),
+so we need to enable the events list explicitly.
+
+From this moment the cycles event is sampling.
+
+We will sleep for 3 seconds while the ring buffers get data from all CPUs, then we disable the events list.
+
+[source,c]
+--
+ 76         perf_evlist__enable(evlist);
+ 77         sleep(3);
+ 78         perf_evlist__disable(evlist);
+--
+
+Following code walks through the ring buffers and reads stored events/samples:
+
+[source,c]
+--
+ 80         perf_evlist__for_each_mmap(evlist, map, false) {
+ 81                 if (perf_mmap__read_init(map) < 0)
+ 82                         continue;
+ 83
+ 84                 while ((event = perf_mmap__read_event(map)) != NULL) {
+
+                            /* process event */
+
+108                         perf_mmap__consume(map);
+109                 }
+110                 perf_mmap__read_done(map);
+111         }
+
+--
+
+Each sample needs to get parsed:
+
+[source,c]
+--
+ 85                         int cpu, pid, tid;
+ 86                         __u64 ip, period, *array;
+ 87                         union u64_swap u;
+ 88
+ 89                         array = event->sample.array;
+ 90
+ 91                         ip = *array;
+ 92                         array++;
+ 93
+ 94                         u.val64 = *array;
+ 95                         pid = u.val32[0];
+ 96                         tid = u.val32[1];
+ 97                         array++;
+ 98
+ 99                         u.val64 = *array;
+100                         cpu = u.val32[0];
+101                         array++;
+102
+103                         period = *array;
+104
+105                         fprintf(stdout, "cpu %3d, pid %6d, tid %6d, ip %20llx, period %20llu\n",
+106                                 cpu, pid, tid, ip, period);
+--
+
+And finally cleanup.
+
+We close the whole events list (both events) and remove it together with the threads map:
+
+[source,c]
+--
+113 out_evlist:
+114         perf_evlist__delete(evlist);
+115 out_cpus:
+116         perf_cpu_map__put(cpus);
+117         return err;
+118 }
+--
+
+REPORTING BUGS
+--------------
+Report bugs to <linux-perf-users@vger.kernel.org>.
+
+LICENSE
+-------
+libperf is Free Software licensed under the GNU LGPL 2.1
+
+RESOURCES
+---------
+https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
+
+SEE ALSO
+--------
+libperf(3), libperf-counting(7)
diff --git a/tools/lib/perf/Documentation/libperf.txt b/tools/lib/perf/Documentation/libperf.txt
new file mode 100644
index 000000000..a8f1a2379
--- /dev/null
+++ b/tools/lib/perf/Documentation/libperf.txt
@@ -0,0 +1,251 @@
+libperf(3)
+==========
+
+NAME
+----
+libperf - Linux kernel perf event library
+
+
+SYNOPSIS
+--------
+*Generic API:*
+
+[source,c]
+--
+  #include <perf/core.h>
+
+  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,
+                                    const char *, va_list ap);
+
+  void libperf_init(libperf_print_fn_t fn);
+--
+
+*API to handle CPU maps:*
+
+[source,c]
+--
+  #include <perf/cpumap.h>
+
+  struct perf_cpu_map;
+
+  struct perf_cpu_map *perf_cpu_map__dummy_new(void);
+  struct perf_cpu_map *perf_cpu_map__new(const char *cpu_list);
+  struct perf_cpu_map *perf_cpu_map__read(FILE *file);
+  struct perf_cpu_map *perf_cpu_map__get(struct perf_cpu_map *map);
+  struct perf_cpu_map *perf_cpu_map__merge(struct perf_cpu_map *orig,
+                                           struct perf_cpu_map *other);
+  void perf_cpu_map__put(struct perf_cpu_map *map);
+  int perf_cpu_map__cpu(const struct perf_cpu_map *cpus, int idx);
+  int perf_cpu_map__nr(const struct perf_cpu_map *cpus);
+  bool perf_cpu_map__empty(const struct perf_cpu_map *map);
+  int perf_cpu_map__max(struct perf_cpu_map *map);
+  bool perf_cpu_map__has(const struct perf_cpu_map *map, int cpu);
+
+  #define perf_cpu_map__for_each_cpu(cpu, idx, cpus)
+--
+
+*API to handle thread maps:*
+
+[source,c]
+--
+  #include <perf/threadmap.h>
+
+  struct perf_thread_map;
+
+  struct perf_thread_map *perf_thread_map__new_dummy(void);
+  struct perf_thread_map *perf_thread_map__new_array(int nr_threads, pid_t *array);
+
+  void perf_thread_map__set_pid(struct perf_thread_map *map, int idx, pid_t pid);
+  char *perf_thread_map__comm(struct perf_thread_map *map, int idx);
+  int perf_thread_map__nr(struct perf_thread_map *threads);
+  pid_t perf_thread_map__pid(struct perf_thread_map *map, int idx);
+
+  struct perf_thread_map *perf_thread_map__get(struct perf_thread_map *map);
+  void perf_thread_map__put(struct perf_thread_map *map);
+--
+
+*API to handle event lists:*
+
+[source,c]
+--
+  #include <perf/evlist.h>
+
+  struct perf_evlist;
+
+  void perf_evlist__add(struct perf_evlist *evlist,
+                        struct perf_evsel *evsel);
+  void perf_evlist__remove(struct perf_evlist *evlist,
+                           struct perf_evsel *evsel);
+  struct perf_evlist *perf_evlist__new(void);
+  void perf_evlist__delete(struct perf_evlist *evlist);
+  struct perf_evsel* perf_evlist__next(struct perf_evlist *evlist,
+                                       struct perf_evsel *evsel);
+  int perf_evlist__open(struct perf_evlist *evlist);
+  void perf_evlist__close(struct perf_evlist *evlist);
+  void perf_evlist__enable(struct perf_evlist *evlist);
+  void perf_evlist__disable(struct perf_evlist *evlist);
+
+  #define perf_evlist__for_each_evsel(evlist, pos)
+
+  void perf_evlist__set_maps(struct perf_evlist *evlist,
+                             struct perf_cpu_map *cpus,
+                             struct perf_thread_map *threads);
+  int perf_evlist__poll(struct perf_evlist *evlist, int timeout);
+  int perf_evlist__filter_pollfd(struct perf_evlist *evlist,
+                                 short revents_and_mask);
+
+  int perf_evlist__mmap(struct perf_evlist *evlist, int pages);
+  void perf_evlist__munmap(struct perf_evlist *evlist);
+
+  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)
+--
+
+*API to handle events:*
+
+[source,c]
+--
+  #include <perf/evsel.h>*
+
+  struct perf_evsel;
+
+  struct perf_counts_values {
+          union {
+                  struct {
+                          uint64_t val;
+                          uint64_t ena;
+                          uint64_t run;
+                  };
+                  uint64_t values[3];
+          };
+  };
+
+  struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr);
+  void perf_evsel__delete(struct perf_evsel *evsel);
+  int perf_evsel__open(struct perf_evsel *evsel, struct perf_cpu_map *cpus,
+                       struct perf_thread_map *threads);
+  void perf_evsel__close(struct perf_evsel *evsel);
+  void perf_evsel__close_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+  int perf_evsel__mmap(struct perf_evsel *evsel, int pages);
+  void perf_evsel__munmap(struct perf_evsel *evsel);
+  void *perf_evsel__mmap_base(struct perf_evsel *evsel, int cpu_map_idx, int thread);
+  int perf_evsel__read(struct perf_evsel *evsel, int cpu_map_idx, int thread,
+                       struct perf_counts_values *count);
+  int perf_evsel__enable(struct perf_evsel *evsel);
+  int perf_evsel__enable_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+  int perf_evsel__disable(struct perf_evsel *evsel);
+  int perf_evsel__disable_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+  struct perf_cpu_map *perf_evsel__cpus(struct perf_evsel *evsel);
+  struct perf_thread_map *perf_evsel__threads(struct perf_evsel *evsel);
+  struct perf_event_attr *perf_evsel__attr(struct perf_evsel *evsel);
+--
+
+*API to handle maps (perf ring buffers):*
+
+[source,c]
+--
+  #include <perf/mmap.h>
+
+  struct perf_mmap;
+
+  void perf_mmap__consume(struct perf_mmap *map);
+  int perf_mmap__read_init(struct perf_mmap *map);
+  void perf_mmap__read_done(struct perf_mmap *map);
+  union perf_event *perf_mmap__read_event(struct perf_mmap *map);
+--
+
+*Structures to access perf API events:*
+
+[source,c]
+--
+  #include <perf/event.h>
+
+  struct perf_record_mmap;
+  struct perf_record_mmap2;
+  struct perf_record_comm;
+  struct perf_record_namespaces;
+  struct perf_record_fork;
+  struct perf_record_lost;
+  struct perf_record_lost_samples;
+  struct perf_record_read;
+  struct perf_record_throttle;
+  struct perf_record_ksymbol;
+  struct perf_record_bpf_event;
+  struct perf_record_sample;
+  struct perf_record_switch;
+  struct perf_record_header_attr;
+  struct perf_record_record_cpu_map;
+  struct perf_record_cpu_map_data;
+  struct perf_record_cpu_map;
+  struct perf_record_event_update_cpus;
+  struct perf_record_event_update_scale;
+  struct perf_record_event_update;
+  struct perf_trace_event_type;
+  struct perf_record_header_event_type;
+  struct perf_record_header_tracing_data;
+  struct perf_record_header_build_id;
+  struct perf_record_id_index;
+  struct perf_record_auxtrace_info;
+  struct perf_record_auxtrace;
+  struct perf_record_auxtrace_error;
+  struct perf_record_aux;
+  struct perf_record_itrace_start;
+  struct perf_record_thread_map_entry;
+  struct perf_record_thread_map;
+  struct perf_record_stat_config_entry;
+  struct perf_record_stat_config;
+  struct perf_record_stat;
+  struct perf_record_stat_round;
+  struct perf_record_time_conv;
+  struct perf_record_header_feature;
+  struct perf_record_compressed;
+--
+
+DESCRIPTION
+-----------
+The libperf library provides an API to access the linux kernel perf
+events subsystem.
+
+Following objects are key to the libperf interface:
+
+[horizontal]
+
+struct perf_cpu_map:: Provides a CPU list abstraction.
+
+struct perf_thread_map:: Provides a thread list abstraction.
+
+struct perf_evsel:: Provides an abstraction for single a perf event.
+
+struct perf_evlist:: Gathers several struct perf_evsel object and performs functions on all of them.
+
+struct perf_mmap:: Provides an abstraction for accessing perf ring buffer.
+
+The exported API functions bind these objects together.
+
+REPORTING BUGS
+--------------
+Report bugs to <linux-perf-users@vger.kernel.org>.
+
+LICENSE
+-------
+libperf is Free Software licensed under the GNU LGPL 2.1
+
+RESOURCES
+---------
+https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
+
+SEE ALSO
+--------
+libperf-sampling(7), libperf-counting(7)
diff --git a/tools/lib/perf/Documentation/manpage-1.72.xsl b/tools/lib/perf/Documentation/manpage-1.72.xsl
new file mode 100644
index 000000000..b4d315cb8
--- /dev/null
+++ b/tools/lib/perf/Documentation/manpage-1.72.xsl
@@ -0,0 +1,14 @@
+<!-- manpage-1.72.xsl:
+     special settings for manpages rendered from asciidoc+docbook
+     handles peculiarities in docbook-xsl 1.72.0 -->
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+		version="1.0">
+
+<xsl:import href="manpage-base.xsl"/>
+
+<!-- these are the special values for the roff control characters
+     needed for docbook-xsl 1.72.0 -->
+<xsl:param name="git.docbook.backslash">&#x2593;</xsl:param>
+<xsl:param name="git.docbook.dot"      >&#x2302;</xsl:param>
+
+</xsl:stylesheet>
diff --git a/tools/lib/perf/Documentation/manpage-base.xsl b/tools/lib/perf/Documentation/manpage-base.xsl
new file mode 100644
index 000000000..a264fa616
--- /dev/null
+++ b/tools/lib/perf/Documentation/manpage-base.xsl
@@ -0,0 +1,35 @@
+<!-- manpage-base.xsl:
+     special formatting for manpages rendered from asciidoc+docbook -->
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+		version="1.0">
+
+<!-- these params silence some output from xmlto -->
+<xsl:param name="man.output.quietly" select="1"/>
+<xsl:param name="refentry.meta.get.quietly" select="1"/>
+
+<!-- convert asciidoc callouts to man page format;
+     git.docbook.backslash and git.docbook.dot params
+     must be supplied by another XSL file or other means -->
+<xsl:template match="co">
+	<xsl:value-of select="concat(
+			      $git.docbook.backslash,'fB(',
+			      substring-after(@id,'-'),')',
+			      $git.docbook.backslash,'fR')"/>
+</xsl:template>
+<xsl:template match="calloutlist">
+	<xsl:value-of select="$git.docbook.dot"/>
+	<xsl:text>sp&#10;</xsl:text>
+	<xsl:apply-templates/>
+	<xsl:text>&#10;</xsl:text>
+</xsl:template>
+<xsl:template match="callout">
+	<xsl:value-of select="concat(
+			      $git.docbook.backslash,'fB',
+			      substring-after(@arearefs,'-'),
+			      '. ',$git.docbook.backslash,'fR')"/>
+	<xsl:apply-templates/>
+	<xsl:value-of select="$git.docbook.dot"/>
+	<xsl:text>br&#10;</xsl:text>
+</xsl:template>
+
+</xsl:stylesheet>
diff --git a/tools/lib/perf/Documentation/manpage-bold-literal.xsl b/tools/lib/perf/Documentation/manpage-bold-literal.xsl
new file mode 100644
index 000000000..608eb5df6
--- /dev/null
+++ b/tools/lib/perf/Documentation/manpage-bold-literal.xsl
@@ -0,0 +1,17 @@
+<!-- manpage-bold-literal.xsl:
+     special formatting for manpages rendered from asciidoc+docbook -->
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+		version="1.0">
+
+<!-- render literal text as bold (instead of plain or monospace);
+     this makes literal text easier to distinguish in manpages
+     viewed on a tty -->
+<xsl:template match="literal">
+	<xsl:value-of select="$git.docbook.backslash"/>
+	<xsl:text>fB</xsl:text>
+	<xsl:apply-templates/>
+	<xsl:value-of select="$git.docbook.backslash"/>
+	<xsl:text>fR</xsl:text>
+</xsl:template>
+
+</xsl:stylesheet>
diff --git a/tools/lib/perf/Documentation/manpage-normal.xsl b/tools/lib/perf/Documentation/manpage-normal.xsl
new file mode 100644
index 000000000..a48f5b11f
--- /dev/null
+++ b/tools/lib/perf/Documentation/manpage-normal.xsl
@@ -0,0 +1,13 @@
+<!-- manpage-normal.xsl:
+     special settings for manpages rendered from asciidoc+docbook
+     handles anything we want to keep away from docbook-xsl 1.72.0 -->
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+		version="1.0">
+
+<xsl:import href="manpage-base.xsl"/>
+
+<!-- these are the normal values for the roff control characters -->
+<xsl:param name="git.docbook.backslash">\</xsl:param>
+<xsl:param name="git.docbook.dot"	>.</xsl:param>
+
+</xsl:stylesheet>
diff --git a/tools/lib/perf/Documentation/manpage-suppress-sp.xsl b/tools/lib/perf/Documentation/manpage-suppress-sp.xsl
new file mode 100644
index 000000000..a63c7632a
--- /dev/null
+++ b/tools/lib/perf/Documentation/manpage-suppress-sp.xsl
@@ -0,0 +1,21 @@
+<!-- manpage-suppress-sp.xsl:
+     special settings for manpages rendered from asciidoc+docbook
+     handles erroneous, inline .sp in manpage output of some
+     versions of docbook-xsl -->
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform"
+		version="1.0">
+
+<!-- attempt to work around spurious .sp at the tail of the line
+     that some versions of docbook stylesheets seem to add -->
+<xsl:template match="simpara">
+  <xsl:variable name="content">
+    <xsl:apply-templates/>
+  </xsl:variable>
+  <xsl:value-of select="normalize-space($content)"/>
+  <xsl:if test="not(ancestor::authorblurb) and
+                not(ancestor::personblurb)">
+    <xsl:text>&#10;&#10;</xsl:text>
+  </xsl:if>
+</xsl:template>
+
+</xsl:stylesheet>
diff --git a/tools/lib/perf/Makefile b/tools/lib/perf/Makefile
new file mode 100644
index 000000000..3a9b2140a
--- /dev/null
+++ b/tools/lib/perf/Makefile
@@ -0,0 +1,220 @@
+# SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
+# Most of this file is copied from tools/lib/bpf/Makefile
+
+LIBPERF_VERSION = 0
+LIBPERF_PATCHLEVEL = 0
+LIBPERF_EXTRAVERSION = 1
+
+MAKEFLAGS += --no-print-directory
+
+ifeq ($(srctree),)
+srctree := $(patsubst %/,%,$(dir $(CURDIR)))
+srctree := $(patsubst %/,%,$(dir $(srctree)))
+srctree := $(patsubst %/,%,$(dir $(srctree)))
+#$(info Determined 'srctree' to be $(srctree))
+endif
+
+INSTALL = install
+
+# Use DESTDIR for installing into a different root directory.
+# This is useful for building a package. The program will be
+# installed in this directory as if it was the root directory.
+# Then the build tool can move it later.
+DESTDIR ?=
+DESTDIR_SQ = '$(subst ','\'',$(DESTDIR))'
+
+include $(srctree)/tools/scripts/Makefile.include
+include $(srctree)/tools/scripts/Makefile.arch
+
+ifeq ($(LP64), 1)
+  libdir_relative = lib64
+else
+  libdir_relative = lib
+endif
+
+prefix ?=
+libdir = $(prefix)/$(libdir_relative)
+
+# Shell quotes
+libdir_SQ = $(subst ','\'',$(libdir))
+libdir_relative_SQ = $(subst ','\'',$(libdir_relative))
+
+ifeq ("$(origin V)", "command line")
+  VERBOSE = $(V)
+endif
+ifndef VERBOSE
+  VERBOSE = 0
+endif
+
+ifeq ($(VERBOSE),1)
+  Q =
+else
+  Q = @
+endif
+
+TEST_ARGS := $(if $(V),-v)
+
+# Set compile option CFLAGS
+ifdef EXTRA_CFLAGS
+  CFLAGS := $(EXTRA_CFLAGS)
+else
+  CFLAGS := -g -Wall
+endif
+
+INCLUDES = \
+-I$(srctree)/tools/lib/perf/include \
+-I$(srctree)/tools/lib/ \
+-I$(srctree)/tools/include \
+-I$(srctree)/tools/arch/$(SRCARCH)/include/ \
+-I$(srctree)/tools/arch/$(SRCARCH)/include/uapi \
+-I$(srctree)/tools/include/uapi
+
+# Append required CFLAGS
+override CFLAGS += $(EXTRA_WARNINGS)
+override CFLAGS += -Werror -Wall
+override CFLAGS += -fPIC
+override CFLAGS += $(INCLUDES)
+override CFLAGS += -fvisibility=hidden
+
+all:
+
+export srctree OUTPUT CC LD CFLAGS V
+export DESTDIR DESTDIR_SQ
+
+include $(srctree)/tools/build/Makefile.include
+
+VERSION_SCRIPT := libperf.map
+
+PATCHLEVEL    = $(LIBPERF_PATCHLEVEL)
+EXTRAVERSION  = $(LIBPERF_EXTRAVERSION)
+VERSION       = $(LIBPERF_VERSION).$(LIBPERF_PATCHLEVEL).$(LIBPERF_EXTRAVERSION)
+
+LIBPERF_SO := $(OUTPUT)libperf.so.$(VERSION)
+LIBPERF_A  := $(OUTPUT)libperf.a
+LIBPERF_IN := $(OUTPUT)libperf-in.o
+LIBPERF_PC := $(OUTPUT)libperf.pc
+
+LIBPERF_ALL := $(LIBPERF_A) $(OUTPUT)libperf.so*
+
+LIB_DIR := $(srctree)/tools/lib/api/
+
+ifneq ($(OUTPUT),)
+ifneq ($(subdir),)
+  API_PATH=$(OUTPUT)/../lib/api/
+else
+  API_PATH=$(OUTPUT)
+endif
+else
+  API_PATH=$(LIB_DIR)
+endif
+
+LIBAPI = $(API_PATH)libapi.a
+export LIBAPI
+
+$(LIBAPI): FORCE
+	$(Q)$(MAKE) -C $(LIB_DIR) O=$(OUTPUT) $(OUTPUT)libapi.a
+
+$(LIBAPI)-clean:
+	$(call QUIET_CLEAN, libapi)
+	$(Q)$(MAKE) -C $(LIB_DIR) O=$(OUTPUT) clean >/dev/null
+
+$(LIBPERF_IN): FORCE
+	$(Q)$(MAKE) $(build)=libperf
+
+$(LIBPERF_A): $(LIBPERF_IN)
+	$(QUIET_AR)$(RM) $@ && $(AR) rcs $@ $(LIBPERF_IN)
+
+$(LIBPERF_SO): $(LIBPERF_IN) $(LIBAPI)
+	$(QUIET_LINK)$(CC) --shared -Wl,-soname,libperf.so \
+                                    -Wl,--version-script=$(VERSION_SCRIPT) $^ -o $@
+	@ln -sf $(@F) $(OUTPUT)libperf.so
+	@ln -sf $(@F) $(OUTPUT)libperf.so.$(LIBPERF_VERSION)
+
+
+libs: $(LIBPERF_A) $(LIBPERF_SO) $(LIBPERF_PC)
+
+all: fixdep
+	$(Q)$(MAKE) libs
+
+clean: $(LIBAPI)-clean
+	$(call QUIET_CLEAN, libperf) $(RM) $(LIBPERF_A) \
+                *.o *~ *.a *.so *.so.$(VERSION) *.so.$(LIBPERF_VERSION) .*.d .*.cmd tests/*.o LIBPERF-CFLAGS $(LIBPERF_PC) \
+                $(TESTS_STATIC) $(TESTS_SHARED)
+
+TESTS_IN = tests-in.o
+
+TESTS_STATIC = $(OUTPUT)tests-static
+TESTS_SHARED = $(OUTPUT)tests-shared
+
+$(TESTS_IN): FORCE
+	$(Q)$(MAKE) $(build)=tests
+
+$(TESTS_STATIC): $(TESTS_IN) $(LIBPERF_A) $(LIBAPI)
+	$(QUIET_LINK)$(CC) -o $@ $^
+
+$(TESTS_SHARED): $(TESTS_IN) $(LIBAPI)
+	$(QUIET_LINK)$(CC) -o $@ -L$(or $(OUTPUT),.) $^ -lperf
+
+make-tests: libs $(TESTS_SHARED) $(TESTS_STATIC)
+
+tests: make-tests
+	@echo "running static:"
+	@./$(TESTS_STATIC) $(TEST_ARGS)
+	@echo "running dynamic:"
+	@LD_LIBRARY_PATH=. ./$(TESTS_SHARED) $(TEST_ARGS)
+
+$(LIBPERF_PC):
+	$(QUIET_GEN)sed -e "s|@PREFIX@|$(prefix)|" \
+		-e "s|@LIBDIR@|$(libdir_SQ)|" \
+		-e "s|@VERSION@|$(VERSION)|" \
+		< libperf.pc.template > $@
+
+define do_install_mkdir
+	if [ ! -d '$(DESTDIR_SQ)$1' ]; then             \
+		$(INSTALL) -d -m 755 '$(DESTDIR_SQ)$1'; \
+	fi
+endef
+
+define do_install
+	if [ ! -d '$2' ]; then             \
+		$(INSTALL) -d -m 755 '$2'; \
+	fi;                                \
+	$(INSTALL) $1 $(if $3,-m $3,) '$2'
+endef
+
+install_lib: libs
+	$(call QUIET_INSTALL, $(LIBPERF_ALL)) \
+		$(call do_install_mkdir,$(libdir_SQ)); \
+		cp -fpR $(LIBPERF_ALL) $(DESTDIR)$(libdir_SQ)
+
+HDRS := bpf_perf.h core.h cpumap.h threadmap.h evlist.h evsel.h event.h mmap.h
+INTERNAL_HDRS := cpumap.h evlist.h evsel.h lib.h mmap.h rc_check.h threadmap.h xyarray.h
+
+INSTALL_HDRS_PFX := $(DESTDIR)$(prefix)/include/perf
+INSTALL_HDRS := $(addprefix $(INSTALL_HDRS_PFX)/, $(HDRS))
+INSTALL_INTERNAL_HDRS_PFX := $(DESTDIR)$(prefix)/include/internal
+INSTALL_INTERNAL_HDRS := $(addprefix $(INSTALL_INTERNAL_HDRS_PFX)/, $(INTERNAL_HDRS))
+
+$(INSTALL_HDRS): $(INSTALL_HDRS_PFX)/%.h: include/perf/%.h
+	$(call QUIET_INSTALL, $@) \
+		$(call do_install,$<,$(INSTALL_HDRS_PFX)/,644)
+
+$(INSTALL_INTERNAL_HDRS): $(INSTALL_INTERNAL_HDRS_PFX)/%.h: include/internal/%.h
+	$(call QUIET_INSTALL, $@) \
+		$(call do_install,$<,$(INSTALL_INTERNAL_HDRS_PFX)/,644)
+
+install_headers: $(INSTALL_HDRS) $(INSTALL_INTERNAL_HDRS)
+	$(call QUIET_INSTALL, libperf_headers)
+
+install_pkgconfig: $(LIBPERF_PC)
+	$(call QUIET_INSTALL, $(LIBPERF_PC)) \
+		$(call do_install,$(LIBPERF_PC),$(DESTDIR_SQ)$(libdir_SQ)/pkgconfig,644)
+
+install_doc:
+	$(Q)$(MAKE) -C Documentation install-man install-html install-examples
+
+install: install_lib install_headers install_pkgconfig install_doc
+
+FORCE:
+
+.PHONY: all install clean tests FORCE
diff --git a/tools/lib/perf/core.c b/tools/lib/perf/core.c
new file mode 100644
index 000000000..58fc894b7
--- /dev/null
+++ b/tools/lib/perf/core.c
@@ -0,0 +1,38 @@
+// SPDX-License-Identifier: GPL-2.0-only
+
+#define __printf(a, b)  __attribute__((format(printf, a, b)))
+
+#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 __maybe_unused, const char *format,
+		     va_list args)
+{
+	return vfprintf(stderr, format, args);
+}
+
+static libperf_print_fn_t __libperf_pr = __base_pr;
+
+__printf(2, 3)
+void libperf_print(enum libperf_print_level level, const char *format, ...)
+{
+	va_list args;
+
+	if (!__libperf_pr)
+		return;
+
+	va_start(args, format);
+	__libperf_pr(level, format, args);
+	va_end(args);
+}
+
+void libperf_init(libperf_print_fn_t fn)
+{
+	page_size = sysconf(_SC_PAGE_SIZE);
+	__libperf_pr = fn;
+}
diff --git a/tools/lib/perf/cpumap.c b/tools/lib/perf/cpumap.c
new file mode 100644
index 000000000..2a5a29217
--- /dev/null
+++ b/tools/lib/perf/cpumap.c
@@ -0,0 +1,481 @@
+// SPDX-License-Identifier: GPL-2.0-only
+#include <perf/cpumap.h>
+#include <stdlib.h>
+#include <linux/refcount.h>
+#include <internal/cpumap.h>
+#include <asm/bug.h>
+#include <stdio.h>
+#include <string.h>
+#include <unistd.h>
+#include <ctype.h>
+#include <limits.h>
+
+void perf_cpu_map__set_nr(struct perf_cpu_map *map, int nr_cpus)
+{
+	RC_CHK_ACCESS(map)->nr = nr_cpus;
+}
+
+struct perf_cpu_map *perf_cpu_map__alloc(int nr_cpus)
+{
+	RC_STRUCT(perf_cpu_map) *cpus = malloc(sizeof(*cpus) + sizeof(struct perf_cpu) * nr_cpus);
+	struct perf_cpu_map *result;
+
+	if (ADD_RC_CHK(result, cpus)) {
+		cpus->nr = nr_cpus;
+		refcount_set(&cpus->refcnt, 1);
+	}
+	return result;
+}
+
+struct perf_cpu_map *perf_cpu_map__dummy_new(void)
+{
+	struct perf_cpu_map *cpus = perf_cpu_map__alloc(1);
+
+	if (cpus)
+		RC_CHK_ACCESS(cpus)->map[0].cpu = -1;
+
+	return cpus;
+}
+
+static void cpu_map__delete(struct perf_cpu_map *map)
+{
+	if (map) {
+		WARN_ONCE(refcount_read(perf_cpu_map__refcnt(map)) != 0,
+			  "cpu_map refcnt unbalanced\n");
+		RC_CHK_FREE(map);
+	}
+}
+
+struct perf_cpu_map *perf_cpu_map__get(struct perf_cpu_map *map)
+{
+	struct perf_cpu_map *result;
+
+	if (RC_CHK_GET(result, map))
+		refcount_inc(perf_cpu_map__refcnt(map));
+
+	return result;
+}
+
+void perf_cpu_map__put(struct perf_cpu_map *map)
+{
+	if (map) {
+		if (refcount_dec_and_test(perf_cpu_map__refcnt(map)))
+			cpu_map__delete(map);
+		else
+			RC_CHK_PUT(map);
+	}
+}
+
+static struct perf_cpu_map *cpu_map__default_new(void)
+{
+	struct perf_cpu_map *cpus;
+	int nr_cpus;
+
+	nr_cpus = sysconf(_SC_NPROCESSORS_ONLN);
+	if (nr_cpus < 0)
+		return NULL;
+
+	cpus = perf_cpu_map__alloc(nr_cpus);
+	if (cpus != NULL) {
+		int i;
+
+		for (i = 0; i < nr_cpus; ++i)
+			RC_CHK_ACCESS(cpus)->map[i].cpu = i;
+	}
+
+	return cpus;
+}
+
+struct perf_cpu_map *perf_cpu_map__default_new(void)
+{
+	return cpu_map__default_new();
+}
+
+
+static int cmp_cpu(const void *a, const void *b)
+{
+	const struct perf_cpu *cpu_a = a, *cpu_b = b;
+
+	return cpu_a->cpu - cpu_b->cpu;
+}
+
+static struct perf_cpu __perf_cpu_map__cpu(const struct perf_cpu_map *cpus, int idx)
+{
+	return RC_CHK_ACCESS(cpus)->map[idx];
+}
+
+static struct perf_cpu_map *cpu_map__trim_new(int nr_cpus, const struct perf_cpu *tmp_cpus)
+{
+	size_t payload_size = nr_cpus * sizeof(struct perf_cpu);
+	struct perf_cpu_map *cpus = perf_cpu_map__alloc(nr_cpus);
+	int i, j;
+
+	if (cpus != NULL) {
+		memcpy(RC_CHK_ACCESS(cpus)->map, tmp_cpus, payload_size);
+		qsort(RC_CHK_ACCESS(cpus)->map, nr_cpus, sizeof(struct perf_cpu), cmp_cpu);
+		/* Remove dups */
+		j = 0;
+		for (i = 0; i < nr_cpus; i++) {
+			if (i == 0 ||
+			    __perf_cpu_map__cpu(cpus, i).cpu !=
+			    __perf_cpu_map__cpu(cpus, i - 1).cpu) {
+				RC_CHK_ACCESS(cpus)->map[j++].cpu =
+					__perf_cpu_map__cpu(cpus, i).cpu;
+			}
+		}
+		perf_cpu_map__set_nr(cpus, j);
+		assert(j <= nr_cpus);
+	}
+	return cpus;
+}
+
+struct perf_cpu_map *perf_cpu_map__read(FILE *file)
+{
+	struct perf_cpu_map *cpus = NULL;
+	int nr_cpus = 0;
+	struct perf_cpu *tmp_cpus = NULL, *tmp;
+	int max_entries = 0;
+	int n, cpu, prev;
+	char sep;
+
+	sep = 0;
+	prev = -1;
+	for (;;) {
+		n = fscanf(file, "%u%c", &cpu, &sep);
+		if (n <= 0)
+			break;
+		if (prev >= 0) {
+			int new_max = nr_cpus + cpu - prev - 1;
+
+			WARN_ONCE(new_max >= MAX_NR_CPUS, "Perf can support %d CPUs. "
+							  "Consider raising MAX_NR_CPUS\n", MAX_NR_CPUS);
+
+			if (new_max >= max_entries) {
+				max_entries = new_max + MAX_NR_CPUS / 2;
+				tmp = realloc(tmp_cpus, max_entries * sizeof(struct perf_cpu));
+				if (tmp == NULL)
+					goto out_free_tmp;
+				tmp_cpus = tmp;
+			}
+
+			while (++prev < cpu)
+				tmp_cpus[nr_cpus++].cpu = prev;
+		}
+		if (nr_cpus == max_entries) {
+			max_entries += MAX_NR_CPUS;
+			tmp = realloc(tmp_cpus, max_entries * sizeof(struct perf_cpu));
+			if (tmp == NULL)
+				goto out_free_tmp;
+			tmp_cpus = tmp;
+		}
+
+		tmp_cpus[nr_cpus++].cpu = cpu;
+		if (n == 2 && sep == '-')
+			prev = cpu;
+		else
+			prev = -1;
+		if (n == 1 || sep == '\n')
+			break;
+	}
+
+	if (nr_cpus > 0)
+		cpus = cpu_map__trim_new(nr_cpus, tmp_cpus);
+	else
+		cpus = cpu_map__default_new();
+out_free_tmp:
+	free(tmp_cpus);
+	return cpus;
+}
+
+static struct perf_cpu_map *cpu_map__read_all_cpu_map(void)
+{
+	struct perf_cpu_map *cpus = NULL;
+	FILE *onlnf;
+
+	onlnf = fopen("/sys/devices/system/cpu/online", "r");
+	if (!onlnf)
+		return cpu_map__default_new();
+
+	cpus = perf_cpu_map__read(onlnf);
+	fclose(onlnf);
+	return cpus;
+}
+
+struct perf_cpu_map *perf_cpu_map__new(const char *cpu_list)
+{
+	struct perf_cpu_map *cpus = NULL;
+	unsigned long start_cpu, end_cpu = 0;
+	char *p = NULL;
+	int i, nr_cpus = 0;
+	struct perf_cpu *tmp_cpus = NULL, *tmp;
+	int max_entries = 0;
+
+	if (!cpu_list)
+		return cpu_map__read_all_cpu_map();
+
+	/*
+	 * must handle the case of empty cpumap to cover
+	 * TOPOLOGY header for NUMA nodes with no CPU
+	 * ( e.g., because of CPU hotplug)
+	 */
+	if (!isdigit(*cpu_list) && *cpu_list != '\0')
+		goto out;
+
+	while (isdigit(*cpu_list)) {
+		p = NULL;
+		start_cpu = strtoul(cpu_list, &p, 0);
+		if (start_cpu >= INT_MAX
+		    || (*p != '\0' && *p != ',' && *p != '-'))
+			goto invalid;
+
+		if (*p == '-') {
+			cpu_list = ++p;
+			p = NULL;
+			end_cpu = strtoul(cpu_list, &p, 0);
+
+			if (end_cpu >= INT_MAX || (*p != '\0' && *p != ','))
+				goto invalid;
+
+			if (end_cpu < start_cpu)
+				goto invalid;
+		} else {
+			end_cpu = start_cpu;
+		}
+
+		WARN_ONCE(end_cpu >= MAX_NR_CPUS, "Perf can support %d CPUs. "
+						  "Consider raising MAX_NR_CPUS\n", MAX_NR_CPUS);
+
+		for (; start_cpu <= end_cpu; start_cpu++) {
+			/* check for duplicates */
+			for (i = 0; i < nr_cpus; i++)
+				if (tmp_cpus[i].cpu == (int)start_cpu)
+					goto invalid;
+
+			if (nr_cpus == max_entries) {
+				max_entries += MAX_NR_CPUS;
+				tmp = realloc(tmp_cpus, max_entries * sizeof(struct perf_cpu));
+				if (tmp == NULL)
+					goto invalid;
+				tmp_cpus = tmp;
+			}
+			tmp_cpus[nr_cpus++].cpu = (int)start_cpu;
+		}
+		if (*p)
+			++p;
+
+		cpu_list = p;
+	}
+
+	if (nr_cpus > 0)
+		cpus = cpu_map__trim_new(nr_cpus, tmp_cpus);
+	else if (*cpu_list != '\0')
+		cpus = cpu_map__default_new();
+	else
+		cpus = perf_cpu_map__dummy_new();
+invalid:
+	free(tmp_cpus);
+out:
+	return cpus;
+}
+
+static int __perf_cpu_map__nr(const struct perf_cpu_map *cpus)
+{
+	return RC_CHK_ACCESS(cpus)->nr;
+}
+
+struct perf_cpu perf_cpu_map__cpu(const struct perf_cpu_map *cpus, int idx)
+{
+	struct perf_cpu result = {
+		.cpu = -1
+	};
+
+	if (cpus && idx < __perf_cpu_map__nr(cpus))
+		return __perf_cpu_map__cpu(cpus, idx);
+
+	return result;
+}
+
+int perf_cpu_map__nr(const struct perf_cpu_map *cpus)
+{
+	return cpus ? __perf_cpu_map__nr(cpus) : 1;
+}
+
+bool perf_cpu_map__empty(const struct perf_cpu_map *map)
+{
+	return map ? __perf_cpu_map__cpu(map, 0).cpu == -1 : true;
+}
+
+int perf_cpu_map__idx(const struct perf_cpu_map *cpus, struct perf_cpu cpu)
+{
+	int low, high;
+
+	if (!cpus)
+		return -1;
+
+	low = 0;
+	high = __perf_cpu_map__nr(cpus);
+	while (low < high) {
+		int idx = (low + high) / 2;
+		struct perf_cpu cpu_at_idx = __perf_cpu_map__cpu(cpus, idx);
+
+		if (cpu_at_idx.cpu == cpu.cpu)
+			return idx;
+
+		if (cpu_at_idx.cpu > cpu.cpu)
+			high = idx;
+		else
+			low = idx + 1;
+	}
+
+	return -1;
+}
+
+bool perf_cpu_map__has(const struct perf_cpu_map *cpus, struct perf_cpu cpu)
+{
+	return perf_cpu_map__idx(cpus, cpu) != -1;
+}
+
+bool perf_cpu_map__equal(const struct perf_cpu_map *lhs, const struct perf_cpu_map *rhs)
+{
+	int nr;
+
+	if (lhs == rhs)
+		return true;
+
+	if (!lhs || !rhs)
+		return false;
+
+	nr = __perf_cpu_map__nr(lhs);
+	if (nr != __perf_cpu_map__nr(rhs))
+		return false;
+
+	for (int idx = 0; idx < nr; idx++) {
+		if (__perf_cpu_map__cpu(lhs, idx).cpu != __perf_cpu_map__cpu(rhs, idx).cpu)
+			return false;
+	}
+	return true;
+}
+
+bool perf_cpu_map__has_any_cpu(const struct perf_cpu_map *map)
+{
+	return map && __perf_cpu_map__cpu(map, 0).cpu == -1;
+}
+
+struct perf_cpu perf_cpu_map__max(const struct perf_cpu_map *map)
+{
+	struct perf_cpu result = {
+		.cpu = -1
+	};
+
+	// cpu_map__trim_new() qsort()s it, cpu_map__default_new() sorts it as well.
+	return __perf_cpu_map__nr(map) > 0
+		? __perf_cpu_map__cpu(map, __perf_cpu_map__nr(map) - 1)
+		: result;
+}
+
+/** Is 'b' a subset of 'a'. */
+bool perf_cpu_map__is_subset(const struct perf_cpu_map *a, const struct perf_cpu_map *b)
+{
+	if (a == b || !b)
+		return true;
+	if (!a || __perf_cpu_map__nr(b) > __perf_cpu_map__nr(a))
+		return false;
+
+	for (int i = 0, j = 0; i < __perf_cpu_map__nr(a); i++) {
+		if (__perf_cpu_map__cpu(a, i).cpu > __perf_cpu_map__cpu(b, j).cpu)
+			return false;
+		if (__perf_cpu_map__cpu(a, i).cpu == __perf_cpu_map__cpu(b, j).cpu) {
+			j++;
+			if (j == __perf_cpu_map__nr(b))
+				return true;
+		}
+	}
+	return false;
+}
+
+/*
+ * Merge two cpumaps
+ *
+ * orig either gets freed and replaced with a new map, or reused
+ * with no reference count change (similar to "realloc")
+ * other has its reference count increased.
+ */
+
+struct perf_cpu_map *perf_cpu_map__merge(struct perf_cpu_map *orig,
+					 struct perf_cpu_map *other)
+{
+	struct perf_cpu *tmp_cpus;
+	int tmp_len;
+	int i, j, k;
+	struct perf_cpu_map *merged;
+
+	if (perf_cpu_map__is_subset(orig, other))
+		return orig;
+	if (perf_cpu_map__is_subset(other, orig)) {
+		perf_cpu_map__put(orig);
+		return perf_cpu_map__get(other);
+	}
+
+	tmp_len = __perf_cpu_map__nr(orig) + __perf_cpu_map__nr(other);
+	tmp_cpus = malloc(tmp_len * sizeof(struct perf_cpu));
+	if (!tmp_cpus)
+		return NULL;
+
+	/* Standard merge algorithm from wikipedia */
+	i = j = k = 0;
+	while (i < __perf_cpu_map__nr(orig) && j < __perf_cpu_map__nr(other)) {
+		if (__perf_cpu_map__cpu(orig, i).cpu <= __perf_cpu_map__cpu(other, j).cpu) {
+			if (__perf_cpu_map__cpu(orig, i).cpu == __perf_cpu_map__cpu(other, j).cpu)
+				j++;
+			tmp_cpus[k++] = __perf_cpu_map__cpu(orig, i++);
+		} else
+			tmp_cpus[k++] = __perf_cpu_map__cpu(other, j++);
+	}
+
+	while (i < __perf_cpu_map__nr(orig))
+		tmp_cpus[k++] = __perf_cpu_map__cpu(orig, i++);
+
+	while (j < __perf_cpu_map__nr(other))
+		tmp_cpus[k++] = __perf_cpu_map__cpu(other, j++);
+	assert(k <= tmp_len);
+
+	merged = cpu_map__trim_new(k, tmp_cpus);
+	free(tmp_cpus);
+	perf_cpu_map__put(orig);
+	return merged;
+}
+
+struct perf_cpu_map *perf_cpu_map__intersect(struct perf_cpu_map *orig,
+					     struct perf_cpu_map *other)
+{
+	struct perf_cpu *tmp_cpus;
+	int tmp_len;
+	int i, j, k;
+	struct perf_cpu_map *merged = NULL;
+
+	if (perf_cpu_map__is_subset(other, orig))
+		return perf_cpu_map__get(orig);
+	if (perf_cpu_map__is_subset(orig, other))
+		return perf_cpu_map__get(other);
+
+	tmp_len = max(__perf_cpu_map__nr(orig), __perf_cpu_map__nr(other));
+	tmp_cpus = malloc(tmp_len * sizeof(struct perf_cpu));
+	if (!tmp_cpus)
+		return NULL;
+
+	i = j = k = 0;
+	while (i < __perf_cpu_map__nr(orig) && j < __perf_cpu_map__nr(other)) {
+		if (__perf_cpu_map__cpu(orig, i).cpu < __perf_cpu_map__cpu(other, j).cpu)
+			i++;
+		else if (__perf_cpu_map__cpu(orig, i).cpu > __perf_cpu_map__cpu(other, j).cpu)
+			j++;
+		else {
+			j++;
+			tmp_cpus[k++] = __perf_cpu_map__cpu(orig, i++);
+		}
+	}
+	if (k)
+		merged = cpu_map__trim_new(k, tmp_cpus);
+	free(tmp_cpus);
+	return merged;
+}
diff --git a/tools/lib/perf/evlist.c b/tools/lib/perf/evlist.c
new file mode 100644
index 000000000..b8b066d0d
--- /dev/null
+++ b/tools/lib/perf/evlist.c
@@ -0,0 +1,740 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <perf/evlist.h>
+#include <perf/evsel.h>
+#include <linux/bitops.h>
+#include <linux/list.h>
+#include <linux/hash.h>
+#include <sys/ioctl.h>
+#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/lib.h>
+#include <linux/zalloc.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>
+#include "internal.h"
+
+void perf_evlist__init(struct perf_evlist *evlist)
+{
+	INIT_LIST_HEAD(&evlist->entries);
+	evlist->nr_entries = 0;
+	fdarray__init(&evlist->pollfd, 64);
+	perf_evlist__reset_id_hash(evlist);
+}
+
+static void __perf_evlist__propagate_maps(struct perf_evlist *evlist,
+					  struct perf_evsel *evsel)
+{
+	if (evsel->system_wide) {
+		/* System wide: set the cpu map of the evsel to all online CPUs. */
+		perf_cpu_map__put(evsel->cpus);
+		evsel->cpus = perf_cpu_map__new(NULL);
+	} else if (evlist->has_user_cpus && evsel->is_pmu_core) {
+		/*
+		 * User requested CPUs on a core PMU, ensure the requested CPUs
+		 * are valid by intersecting with those of the PMU.
+		 */
+		perf_cpu_map__put(evsel->cpus);
+		evsel->cpus = perf_cpu_map__intersect(evlist->user_requested_cpus, evsel->own_cpus);
+	} else if (!evsel->own_cpus || evlist->has_user_cpus ||
+		(!evsel->requires_cpu && perf_cpu_map__has_any_cpu(evlist->user_requested_cpus))) {
+		/*
+		 * The PMU didn't specify a default cpu map, this isn't a core
+		 * event and the user requested CPUs or the evlist user
+		 * requested CPUs have the "any CPU" (aka dummy) CPU value. In
+		 * which case use the user requested CPUs rather than the PMU
+		 * ones.
+		 */
+		perf_cpu_map__put(evsel->cpus);
+		evsel->cpus = perf_cpu_map__get(evlist->user_requested_cpus);
+	} else if (evsel->cpus != evsel->own_cpus) {
+		/*
+		 * No user requested cpu map but the PMU cpu map doesn't match
+		 * the evsel's. Reset it back to the PMU cpu map.
+		 */
+		perf_cpu_map__put(evsel->cpus);
+		evsel->cpus = perf_cpu_map__get(evsel->own_cpus);
+	}
+
+	if (evsel->system_wide) {
+		perf_thread_map__put(evsel->threads);
+		evsel->threads = perf_thread_map__new_dummy();
+	} else {
+		perf_thread_map__put(evsel->threads);
+		evsel->threads = perf_thread_map__get(evlist->threads);
+	}
+
+	evlist->all_cpus = perf_cpu_map__merge(evlist->all_cpus, evsel->cpus);
+}
+
+static void perf_evlist__propagate_maps(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	evlist->needs_map_propagation = true;
+
+	perf_evlist__for_each_evsel(evlist, evsel)
+		__perf_evlist__propagate_maps(evlist, evsel);
+}
+
+void perf_evlist__add(struct perf_evlist *evlist,
+		      struct perf_evsel *evsel)
+{
+	evsel->idx = evlist->nr_entries;
+	list_add_tail(&evsel->node, &evlist->entries);
+	evlist->nr_entries += 1;
+
+	if (evlist->needs_map_propagation)
+		__perf_evlist__propagate_maps(evlist, evsel);
+}
+
+void perf_evlist__remove(struct perf_evlist *evlist,
+			 struct perf_evsel *evsel)
+{
+	list_del_init(&evsel->node);
+	evlist->nr_entries -= 1;
+}
+
+struct perf_evlist *perf_evlist__new(void)
+{
+	struct perf_evlist *evlist = zalloc(sizeof(*evlist));
+
+	if (evlist != NULL)
+		perf_evlist__init(evlist);
+
+	return evlist;
+}
+
+struct perf_evsel *
+perf_evlist__next(struct perf_evlist *evlist, struct perf_evsel *prev)
+{
+	struct perf_evsel *next;
+
+	if (!prev) {
+		next = list_first_entry(&evlist->entries,
+					struct perf_evsel,
+					node);
+	} else {
+		next = list_next_entry(prev, node);
+	}
+
+	/* Empty list is noticed here so don't need checking on entry. */
+	if (&next->node == &evlist->entries)
+		return NULL;
+
+	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->user_requested_cpus);
+	perf_cpu_map__put(evlist->all_cpus);
+	perf_thread_map__put(evlist->threads);
+	evlist->user_requested_cpus = NULL;
+	evlist->all_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);
+}
+
+void perf_evlist__set_maps(struct perf_evlist *evlist,
+			   struct perf_cpu_map *cpus,
+			   struct perf_thread_map *threads)
+{
+	/*
+	 * Allow for the possibility that one or another of the maps isn't being
+	 * changed i.e. don't put it.  Note we are assuming the maps that are
+	 * being applied are brand new and evlist is taking ownership of the
+	 * original reference count of 1.  If that is not the case it is up to
+	 * the caller to increase the reference count.
+	 */
+	if (cpus != evlist->user_requested_cpus) {
+		perf_cpu_map__put(evlist->user_requested_cpus);
+		evlist->user_requested_cpus = perf_cpu_map__get(cpus);
+	}
+
+	if (threads != evlist->threads) {
+		perf_thread_map__put(evlist->threads);
+		evlist->threads = perf_thread_map__get(threads);
+	}
+
+	perf_evlist__propagate_maps(evlist);
+}
+
+int perf_evlist__open(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+	int err;
+
+	perf_evlist__for_each_entry(evlist, evsel) {
+		err = perf_evsel__open(evsel, evsel->cpus, evsel->threads);
+		if (err < 0)
+			goto out_err;
+	}
+
+	return 0;
+
+out_err:
+	perf_evlist__close(evlist);
+	return err;
+}
+
+void perf_evlist__close(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	perf_evlist__for_each_entry_reverse(evlist, evsel)
+		perf_evsel__close(evsel);
+}
+
+void perf_evlist__enable(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	perf_evlist__for_each_entry(evlist, evsel)
+		perf_evsel__enable(evsel);
+}
+
+void perf_evlist__disable(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+
+	perf_evlist__for_each_entry(evlist, evsel)
+		perf_evsel__disable(evsel);
+}
+
+u64 perf_evlist__read_format(struct perf_evlist *evlist)
+{
+	struct perf_evsel *first = perf_evlist__first(evlist);
+
+	return first->attr.read_format;
+}
+
+#define SID(e, x, y) xyarray__entry(e->sample_id, x, y)
+
+static void perf_evlist__id_hash(struct perf_evlist *evlist,
+				 struct perf_evsel *evsel,
+				 int cpu, int thread, u64 id)
+{
+	int hash;
+	struct perf_sample_id *sid = SID(evsel, cpu, thread);
+
+	sid->id = id;
+	sid->evsel = evsel;
+	hash = hash_64(sid->id, PERF_EVLIST__HLIST_BITS);
+	hlist_add_head(&sid->node, &evlist->heads[hash]);
+}
+
+void perf_evlist__reset_id_hash(struct perf_evlist *evlist)
+{
+	int i;
+
+	for (i = 0; i < PERF_EVLIST__HLIST_SIZE; ++i)
+		INIT_HLIST_HEAD(&evlist->heads[i]);
+}
+
+void perf_evlist__id_add(struct perf_evlist *evlist,
+			 struct perf_evsel *evsel,
+			 int cpu, int thread, u64 id)
+{
+	perf_evlist__id_hash(evlist, evsel, cpu, thread, id);
+	evsel->id[evsel->ids++] = id;
+}
+
+int perf_evlist__id_add_fd(struct perf_evlist *evlist,
+			   struct perf_evsel *evsel,
+			   int cpu, int thread, int fd)
+{
+	u64 read_data[4] = { 0, };
+	int id_idx = 1; /* The first entry is the counter value */
+	u64 id;
+	int ret;
+
+	ret = ioctl(fd, PERF_EVENT_IOC_ID, &id);
+	if (!ret)
+		goto add;
+
+	if (errno != ENOTTY)
+		return -1;
+
+	/* Legacy way to get event id.. All hail to old kernels! */
+
+	/*
+	 * This way does not work with group format read, so bail
+	 * out in that case.
+	 */
+	if (perf_evlist__read_format(evlist) & PERF_FORMAT_GROUP)
+		return -1;
+
+	if (!(evsel->attr.read_format & PERF_FORMAT_ID) ||
+	    read(fd, &read_data, sizeof(read_data)) == -1)
+		return -1;
+
+	if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		++id_idx;
+	if (evsel->attr.read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		++id_idx;
+
+	id = read_data[id_idx];
+
+add:
+	perf_evlist__id_add(evlist, evsel, cpu, thread, id);
+	return 0;
+}
+
+int perf_evlist__alloc_pollfd(struct perf_evlist *evlist)
+{
+	int nr_cpus = perf_cpu_map__nr(evlist->all_cpus);
+	int nr_threads = perf_thread_map__nr(evlist->threads);
+	int nfds = 0;
+	struct perf_evsel *evsel;
+
+	perf_evlist__for_each_entry(evlist, evsel) {
+		if (evsel->system_wide)
+			nfds += nr_cpus;
+		else
+			nfds += nr_cpus * nr_threads;
+	}
+
+	if (fdarray__available_entries(&evlist->pollfd) < nfds &&
+	    fdarray__grow(&evlist->pollfd, nfds) < 0)
+		return -ENOMEM;
+
+	return 0;
+}
+
+int perf_evlist__add_pollfd(struct perf_evlist *evlist, int fd,
+			    void *ptr, short revent, enum fdarray_flags flags)
+{
+	int pos = fdarray__add(&evlist->pollfd, fd, revent | POLLERR | POLLHUP, flags);
+
+	if (pos >= 0) {
+		evlist->pollfd.priv[pos].ptr = ptr;
+		fcntl(fd, F_SETFL, O_NONBLOCK);
+	}
+
+	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_evsel__set_sid_idx(struct perf_evsel *evsel, int idx, int cpu, int thread)
+{
+	struct perf_sample_id *sid = SID(evsel, cpu, thread);
+
+	sid->idx = idx;
+	sid->cpu = perf_cpu_map__cpu(evsel->cpus, cpu);
+	sid->tid = perf_thread_map__pid(evsel->threads, thread);
+}
+
+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, struct perf_cpu 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 *nr_mmaps)
+{
+	struct perf_cpu evlist_cpu = perf_cpu_map__cpu(evlist->all_cpus, cpu_idx);
+	struct perf_evsel *evsel;
+	int revent;
+
+	perf_evlist__for_each_entry(evlist, evsel) {
+		bool overwrite = evsel->attr.write_backward;
+		enum fdarray_flags flgs;
+		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->idx)
+				ops->idx(evlist, evsel, mp, idx);
+
+			/* Debug message used by test scripts */
+			pr_debug("idx %d: mmapping fd %d\n", idx, *output);
+			if (ops->mmap(map, mp, *output, evlist_cpu) < 0)
+				return -1;
+
+			*nr_mmaps += 1;
+
+			if (!idx)
+				perf_evlist__set_mmap_first(evlist, map, overwrite);
+		} else {
+			/* Debug message used by test scripts */
+			pr_debug("idx %d: set output fd %d -> %d\n", idx, fd, *output);
+			if (ioctl(fd, PERF_EVENT_IOC_SET_OUTPUT, *output) != 0)
+				return -1;
+
+			perf_mmap__get(map);
+		}
+
+		revent = !overwrite ? POLLIN : 0;
+
+		flgs = evsel->system_wide ? fdarray_flag__nonfilterable : fdarray_flag__default;
+		if (perf_evlist__add_pollfd(evlist, fd, map, revent, flgs) < 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_evsel__set_sid_idx(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 nr_threads = perf_thread_map__nr(evlist->threads);
+	int nr_cpus    = perf_cpu_map__nr(evlist->all_cpus);
+	int cpu, thread, idx = 0;
+	int nr_mmaps = 0;
+
+	pr_debug("%s: nr cpu values (may include -1) %d nr threads %d\n",
+		 __func__, nr_cpus, nr_threads);
+
+	/* per-thread mmaps */
+	for (thread = 0; thread < nr_threads; thread++, idx++) {
+		int output = -1;
+		int output_overwrite = -1;
+
+		if (mmap_per_evsel(evlist, ops, idx, mp, 0, thread, &output,
+				   &output_overwrite, &nr_mmaps))
+			goto out_unmap;
+	}
+
+	/* system-wide mmaps i.e. per-cpu */
+	for (cpu = 1; cpu < nr_cpus; cpu++, idx++) {
+		int output = -1;
+		int output_overwrite = -1;
+
+		if (mmap_per_evsel(evlist, ops, idx, mp, cpu, 0, &output,
+				   &output_overwrite, &nr_mmaps))
+			goto out_unmap;
+	}
+
+	if (nr_mmaps != evlist->nr_mmaps)
+		pr_err("Miscounted nr_mmaps %d vs %d\n", nr_mmaps, evlist->nr_mmaps);
+
+	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->all_cpus);
+	int nr_mmaps = 0;
+	int cpu, thread;
+
+	pr_debug("%s: nr cpu values %d nr threads %d\n", __func__, nr_cpus, nr_threads);
+
+	for (cpu = 0; cpu < nr_cpus; cpu++) {
+		int output = -1;
+		int output_overwrite = -1;
+
+		for (thread = 0; thread < nr_threads; thread++) {
+			if (mmap_per_evsel(evlist, ops, cpu, mp, cpu,
+					   thread, &output, &output_overwrite, &nr_mmaps))
+				goto out_unmap;
+		}
+	}
+
+	if (nr_mmaps != evlist->nr_mmaps)
+		pr_err("Miscounted nr_mmaps %d vs %d\n", nr_mmaps, evlist->nr_mmaps);
+
+	return 0;
+
+out_unmap:
+	perf_evlist__munmap(evlist);
+	return -1;
+}
+
+static int perf_evlist__nr_mmaps(struct perf_evlist *evlist)
+{
+	int nr_mmaps;
+
+	/* One for each CPU */
+	nr_mmaps = perf_cpu_map__nr(evlist->all_cpus);
+	if (perf_cpu_map__empty(evlist->all_cpus)) {
+		/* Plus one for each thread */
+		nr_mmaps += perf_thread_map__nr(evlist->threads);
+		/* Minus the per-thread CPU (-1) */
+		nr_mmaps -= 1;
+	}
+
+	return nr_mmaps;
+}
+
+int perf_evlist__mmap_ops(struct perf_evlist *evlist,
+			  struct perf_evlist_mmap_ops *ops,
+			  struct perf_mmap_param *mp)
+{
+	const struct perf_cpu_map *cpus = evlist->all_cpus;
+	struct perf_evsel *evsel;
+
+	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, evsel->fd->max_x, evsel->fd->max_y) < 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;
+}
+
+void __perf_evlist__set_leader(struct list_head *list, struct perf_evsel *leader)
+{
+	struct perf_evsel *evsel;
+	int n = 0;
+
+	__perf_evlist__for_each_entry(list, evsel) {
+		evsel->leader = leader;
+		n++;
+	}
+	leader->nr_members = n;
+}
+
+void perf_evlist__set_leader(struct perf_evlist *evlist)
+{
+	if (evlist->nr_entries) {
+		struct perf_evsel *first = list_entry(evlist->entries.next,
+						struct perf_evsel, node);
+
+		__perf_evlist__set_leader(&evlist->entries, first);
+	}
+}
+
+int perf_evlist__nr_groups(struct perf_evlist *evlist)
+{
+	struct perf_evsel *evsel;
+	int nr_groups = 0;
+
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		/*
+		 * evsels by default have a nr_members of 1, and they are their
+		 * own leader. If the nr_members is >1 then this is an
+		 * indication of a group.
+		 */
+		if (evsel->leader == evsel && evsel->nr_members > 1)
+			nr_groups++;
+	}
+	return nr_groups;
+}
diff --git a/tools/lib/perf/evsel.c b/tools/lib/perf/evsel.c
new file mode 100644
index 000000000..8b51b008a
--- /dev/null
+++ b/tools/lib/perf/evsel.c
@@ -0,0 +1,557 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <errno.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+#include <perf/evsel.h>
+#include <perf/cpumap.h>
+#include <perf/threadmap.h>
+#include <linux/list.h>
+#include <internal/evsel.h>
+#include <linux/zalloc.h>
+#include <stdlib.h>
+#include <internal/xyarray.h>
+#include <internal/cpumap.h>
+#include <internal/mmap.h>
+#include <internal/threadmap.h>
+#include <internal/lib.h>
+#include <linux/string.h>
+#include <sys/ioctl.h>
+#include <sys/mman.h>
+#include <asm/bug.h>
+
+void perf_evsel__init(struct perf_evsel *evsel, struct perf_event_attr *attr,
+		      int idx)
+{
+	INIT_LIST_HEAD(&evsel->node);
+	evsel->attr = *attr;
+	evsel->idx  = idx;
+	evsel->leader = evsel;
+}
+
+struct perf_evsel *perf_evsel__new(struct perf_event_attr *attr)
+{
+	struct perf_evsel *evsel = zalloc(sizeof(*evsel));
+
+	if (evsel != NULL)
+		perf_evsel__init(evsel, attr, 0);
+
+	return evsel;
+}
+
+void perf_evsel__delete(struct perf_evsel *evsel)
+{
+	free(evsel);
+}
+
+#define FD(_evsel, _cpu_map_idx, _thread)				\
+	((int *)xyarray__entry(_evsel->fd, _cpu_map_idx, _thread))
+#define MMAP(_evsel, _cpu_map_idx, _thread)				\
+	(_evsel->mmap ? ((struct perf_mmap *) xyarray__entry(_evsel->mmap, _cpu_map_idx, _thread)) \
+		      : NULL)
+
+int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)
+{
+	evsel->fd = xyarray__new(ncpus, nthreads, sizeof(int));
+
+	if (evsel->fd) {
+		int idx, thread;
+
+		for (idx = 0; idx < ncpus; idx++) {
+			for (thread = 0; thread < nthreads; thread++) {
+				int *fd = FD(evsel, idx, thread);
+
+				if (fd)
+					*fd = -1;
+			}
+		}
+	}
+
+	return evsel->fd != NULL ? 0 : -ENOMEM;
+}
+
+static int perf_evsel__alloc_mmap(struct perf_evsel *evsel, int ncpus, int nthreads)
+{
+	evsel->mmap = xyarray__new(ncpus, nthreads, sizeof(struct perf_mmap));
+
+	return evsel->mmap != NULL ? 0 : -ENOMEM;
+}
+
+static int
+sys_perf_event_open(struct perf_event_attr *attr,
+		    pid_t pid, struct perf_cpu cpu, int group_fd,
+		    unsigned long flags)
+{
+	return syscall(__NR_perf_event_open, attr, pid, cpu.cpu, group_fd, flags);
+}
+
+static int get_group_fd(struct perf_evsel *evsel, int cpu_map_idx, int thread, int *group_fd)
+{
+	struct perf_evsel *leader = evsel->leader;
+	int *fd;
+
+	if (evsel == leader) {
+		*group_fd = -1;
+		return 0;
+	}
+
+	/*
+	 * Leader must be already processed/open,
+	 * if not it's a bug.
+	 */
+	if (!leader->fd)
+		return -ENOTCONN;
+
+	fd = FD(leader, cpu_map_idx, thread);
+	if (fd == NULL || *fd == -1)
+		return -EBADF;
+
+	*group_fd = *fd;
+
+	return 0;
+}
+
+int perf_evsel__open(struct perf_evsel *evsel, struct perf_cpu_map *cpus,
+		     struct perf_thread_map *threads)
+{
+	struct perf_cpu cpu;
+	int idx, thread, err = 0;
+
+	if (cpus == NULL) {
+		static struct perf_cpu_map *empty_cpu_map;
+
+		if (empty_cpu_map == NULL) {
+			empty_cpu_map = perf_cpu_map__dummy_new();
+			if (empty_cpu_map == NULL)
+				return -ENOMEM;
+		}
+
+		cpus = empty_cpu_map;
+	}
+
+	if (threads == NULL) {
+		static struct perf_thread_map *empty_thread_map;
+
+		if (empty_thread_map == NULL) {
+			empty_thread_map = perf_thread_map__new_dummy();
+			if (empty_thread_map == NULL)
+				return -ENOMEM;
+		}
+
+		threads = empty_thread_map;
+	}
+
+	if (evsel->fd == NULL &&
+	    perf_evsel__alloc_fd(evsel, perf_cpu_map__nr(cpus), threads->nr) < 0)
+		return -ENOMEM;
+
+	perf_cpu_map__for_each_cpu(cpu, idx, cpus) {
+		for (thread = 0; thread < threads->nr; thread++) {
+			int fd, group_fd, *evsel_fd;
+
+			evsel_fd = FD(evsel, idx, thread);
+			if (evsel_fd == NULL) {
+				err = -EINVAL;
+				goto out;
+			}
+
+			err = get_group_fd(evsel, idx, thread, &group_fd);
+			if (err < 0)
+				goto out;
+
+			fd = sys_perf_event_open(&evsel->attr,
+						 threads->map[thread].pid,
+						 cpu, group_fd, 0);
+
+			if (fd < 0) {
+				err = -errno;
+				goto out;
+			}
+
+			*evsel_fd = fd;
+		}
+	}
+out:
+	if (err)
+		perf_evsel__close(evsel);
+
+	return err;
+}
+
+static void perf_evsel__close_fd_cpu(struct perf_evsel *evsel, int cpu_map_idx)
+{
+	int thread;
+
+	for (thread = 0; thread < xyarray__max_y(evsel->fd); ++thread) {
+		int *fd = FD(evsel, cpu_map_idx, thread);
+
+		if (fd && *fd >= 0) {
+			close(*fd);
+			*fd = -1;
+		}
+	}
+}
+
+void perf_evsel__close_fd(struct perf_evsel *evsel)
+{
+	for (int idx = 0; idx < xyarray__max_x(evsel->fd); idx++)
+		perf_evsel__close_fd_cpu(evsel, idx);
+}
+
+void perf_evsel__free_fd(struct perf_evsel *evsel)
+{
+	xyarray__delete(evsel->fd);
+	evsel->fd = NULL;
+}
+
+void perf_evsel__close(struct perf_evsel *evsel)
+{
+	if (evsel->fd == NULL)
+		return;
+
+	perf_evsel__close_fd(evsel);
+	perf_evsel__free_fd(evsel);
+}
+
+void perf_evsel__close_cpu(struct perf_evsel *evsel, int cpu_map_idx)
+{
+	if (evsel->fd == NULL)
+		return;
+
+	perf_evsel__close_fd_cpu(evsel, cpu_map_idx);
+}
+
+void perf_evsel__munmap(struct perf_evsel *evsel)
+{
+	int idx, thread;
+
+	if (evsel->fd == NULL || evsel->mmap == NULL)
+		return;
+
+	for (idx = 0; idx < xyarray__max_x(evsel->fd); idx++) {
+		for (thread = 0; thread < xyarray__max_y(evsel->fd); thread++) {
+			int *fd = FD(evsel, idx, thread);
+
+			if (fd == NULL || *fd < 0)
+				continue;
+
+			perf_mmap__munmap(MMAP(evsel, idx, thread));
+		}
+	}
+
+	xyarray__delete(evsel->mmap);
+	evsel->mmap = NULL;
+}
+
+int perf_evsel__mmap(struct perf_evsel *evsel, int pages)
+{
+	int ret, idx, thread;
+	struct perf_mmap_param mp = {
+		.prot = PROT_READ | PROT_WRITE,
+		.mask = (pages * page_size) - 1,
+	};
+
+	if (evsel->fd == NULL || evsel->mmap)
+		return -EINVAL;
+
+	if (perf_evsel__alloc_mmap(evsel, xyarray__max_x(evsel->fd), xyarray__max_y(evsel->fd)) < 0)
+		return -ENOMEM;
+
+	for (idx = 0; idx < xyarray__max_x(evsel->fd); idx++) {
+		for (thread = 0; thread < xyarray__max_y(evsel->fd); thread++) {
+			int *fd = FD(evsel, idx, thread);
+			struct perf_mmap *map;
+			struct perf_cpu cpu = perf_cpu_map__cpu(evsel->cpus, idx);
+
+			if (fd == NULL || *fd < 0)
+				continue;
+
+			map = MMAP(evsel, idx, thread);
+			perf_mmap__init(map, NULL, false, NULL);
+
+			ret = perf_mmap__mmap(map, &mp, *fd, cpu);
+			if (ret) {
+				perf_evsel__munmap(evsel);
+				return ret;
+			}
+		}
+	}
+
+	return 0;
+}
+
+void *perf_evsel__mmap_base(struct perf_evsel *evsel, int cpu_map_idx, int thread)
+{
+	int *fd = FD(evsel, cpu_map_idx, thread);
+
+	if (fd == NULL || *fd < 0 || MMAP(evsel, cpu_map_idx, thread) == NULL)
+		return NULL;
+
+	return MMAP(evsel, cpu_map_idx, thread)->base;
+}
+
+int perf_evsel__read_size(struct perf_evsel *evsel)
+{
+	u64 read_format = evsel->attr.read_format;
+	int entry = sizeof(u64); /* value */
+	int size = 0;
+	int nr = 1;
+
+	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		size += sizeof(u64);
+
+	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		size += sizeof(u64);
+
+	if (read_format & PERF_FORMAT_ID)
+		entry += sizeof(u64);
+
+	if (read_format & PERF_FORMAT_LOST)
+		entry += sizeof(u64);
+
+	if (read_format & PERF_FORMAT_GROUP) {
+		nr = evsel->nr_members;
+		size += sizeof(u64);
+	}
+
+	size += entry * nr;
+	return size;
+}
+
+/* This only reads values for the leader */
+static int perf_evsel__read_group(struct perf_evsel *evsel, int cpu_map_idx,
+				  int thread, struct perf_counts_values *count)
+{
+	size_t size = perf_evsel__read_size(evsel);
+	int *fd = FD(evsel, cpu_map_idx, thread);
+	u64 read_format = evsel->attr.read_format;
+	u64 *data;
+	int idx = 1;
+
+	if (fd == NULL || *fd < 0)
+		return -EINVAL;
+
+	data = calloc(1, size);
+	if (data == NULL)
+		return -ENOMEM;
+
+	if (readn(*fd, data, size) <= 0) {
+		free(data);
+		return -errno;
+	}
+
+	/*
+	 * This reads only the leader event intentionally since we don't have
+	 * perf counts values for sibling events.
+	 */
+	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		count->ena = data[idx++];
+	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		count->run = data[idx++];
+
+	/* value is always available */
+	count->val = data[idx++];
+	if (read_format & PERF_FORMAT_ID)
+		count->id = data[idx++];
+	if (read_format & PERF_FORMAT_LOST)
+		count->lost = data[idx++];
+
+	free(data);
+	return 0;
+}
+
+/*
+ * The perf read format is very flexible.  It needs to set the proper
+ * values according to the read format.
+ */
+static void perf_evsel__adjust_values(struct perf_evsel *evsel, u64 *buf,
+				      struct perf_counts_values *count)
+{
+	u64 read_format = evsel->attr.read_format;
+	int n = 0;
+
+	count->val = buf[n++];
+
+	if (read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		count->ena = buf[n++];
+
+	if (read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		count->run = buf[n++];
+
+	if (read_format & PERF_FORMAT_ID)
+		count->id = buf[n++];
+
+	if (read_format & PERF_FORMAT_LOST)
+		count->lost = buf[n++];
+}
+
+int perf_evsel__read(struct perf_evsel *evsel, int cpu_map_idx, int thread,
+		     struct perf_counts_values *count)
+{
+	size_t size = perf_evsel__read_size(evsel);
+	int *fd = FD(evsel, cpu_map_idx, thread);
+	u64 read_format = evsel->attr.read_format;
+	struct perf_counts_values buf;
+
+	memset(count, 0, sizeof(*count));
+
+	if (fd == NULL || *fd < 0)
+		return -EINVAL;
+
+	if (read_format & PERF_FORMAT_GROUP)
+		return perf_evsel__read_group(evsel, cpu_map_idx, thread, count);
+
+	if (MMAP(evsel, cpu_map_idx, thread) &&
+	    !(read_format & (PERF_FORMAT_ID | PERF_FORMAT_LOST)) &&
+	    !perf_mmap__read_self(MMAP(evsel, cpu_map_idx, thread), count))
+		return 0;
+
+	if (readn(*fd, buf.values, size) <= 0)
+		return -errno;
+
+	perf_evsel__adjust_values(evsel, buf.values, count);
+	return 0;
+}
+
+static int perf_evsel__ioctl(struct perf_evsel *evsel, int ioc, void *arg,
+			     int cpu_map_idx, int thread)
+{
+	int *fd = FD(evsel, cpu_map_idx, thread);
+
+	if (fd == NULL || *fd < 0)
+		return -1;
+
+	return ioctl(*fd, ioc, arg);
+}
+
+static int perf_evsel__run_ioctl(struct perf_evsel *evsel,
+				 int ioc,  void *arg,
+				 int cpu_map_idx)
+{
+	int thread;
+
+	for (thread = 0; thread < xyarray__max_y(evsel->fd); thread++) {
+		int err = perf_evsel__ioctl(evsel, ioc, arg, cpu_map_idx, thread);
+
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int perf_evsel__enable_cpu(struct perf_evsel *evsel, int cpu_map_idx)
+{
+	return perf_evsel__run_ioctl(evsel, PERF_EVENT_IOC_ENABLE, NULL, cpu_map_idx);
+}
+
+int perf_evsel__enable_thread(struct perf_evsel *evsel, int thread)
+{
+	struct perf_cpu cpu __maybe_unused;
+	int idx;
+	int err;
+
+	perf_cpu_map__for_each_cpu(cpu, idx, evsel->cpus) {
+		err = perf_evsel__ioctl(evsel, PERF_EVENT_IOC_ENABLE, NULL, idx, thread);
+		if (err)
+			return err;
+	}
+
+	return 0;
+}
+
+int perf_evsel__enable(struct perf_evsel *evsel)
+{
+	int i;
+	int err = 0;
+
+	for (i = 0; i < xyarray__max_x(evsel->fd) && !err; i++)
+		err = perf_evsel__run_ioctl(evsel, PERF_EVENT_IOC_ENABLE, NULL, i);
+	return err;
+}
+
+int perf_evsel__disable_cpu(struct perf_evsel *evsel, int cpu_map_idx)
+{
+	return perf_evsel__run_ioctl(evsel, PERF_EVENT_IOC_DISABLE, NULL, cpu_map_idx);
+}
+
+int perf_evsel__disable(struct perf_evsel *evsel)
+{
+	int i;
+	int err = 0;
+
+	for (i = 0; i < xyarray__max_x(evsel->fd) && !err; i++)
+		err = perf_evsel__run_ioctl(evsel, PERF_EVENT_IOC_DISABLE, NULL, i);
+	return err;
+}
+
+int perf_evsel__apply_filter(struct perf_evsel *evsel, const char *filter)
+{
+	int err = 0, i;
+
+	for (i = 0; i < perf_cpu_map__nr(evsel->cpus) && !err; i++)
+		err = perf_evsel__run_ioctl(evsel,
+				     PERF_EVENT_IOC_SET_FILTER,
+				     (void *)filter, i);
+	return err;
+}
+
+struct perf_cpu_map *perf_evsel__cpus(struct perf_evsel *evsel)
+{
+	return evsel->cpus;
+}
+
+struct perf_thread_map *perf_evsel__threads(struct perf_evsel *evsel)
+{
+	return evsel->threads;
+}
+
+struct perf_event_attr *perf_evsel__attr(struct perf_evsel *evsel)
+{
+	return &evsel->attr;
+}
+
+int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads)
+{
+	if (ncpus == 0 || nthreads == 0)
+		return 0;
+
+	evsel->sample_id = xyarray__new(ncpus, nthreads, sizeof(struct perf_sample_id));
+	if (evsel->sample_id == NULL)
+		return -ENOMEM;
+
+	evsel->id = zalloc(ncpus * nthreads * sizeof(u64));
+	if (evsel->id == NULL) {
+		xyarray__delete(evsel->sample_id);
+		evsel->sample_id = NULL;
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+void perf_evsel__free_id(struct perf_evsel *evsel)
+{
+	xyarray__delete(evsel->sample_id);
+	evsel->sample_id = NULL;
+	zfree(&evsel->id);
+	evsel->ids = 0;
+}
+
+void perf_counts_values__scale(struct perf_counts_values *count,
+			       bool scale, __s8 *pscaled)
+{
+	s8 scaled = 0;
+
+	if (scale) {
+		if (count->run == 0) {
+			scaled = -1;
+			count->val = 0;
+		} else if (count->run < count->ena) {
+			scaled = 1;
+			count->val = (u64)((double)count->val * count->ena / count->run);
+		}
+	}
+
+	if (pscaled)
+		*pscaled = scaled;
+}
diff --git a/tools/lib/perf/include/internal/cpumap.h b/tools/lib/perf/include/internal/cpumap.h
new file mode 100644
index 000000000..49649eb51
--- /dev/null
+++ b/tools/lib/perf/include/internal/cpumap.h
@@ -0,0 +1,38 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_CPUMAP_H
+#define __LIBPERF_INTERNAL_CPUMAP_H
+
+#include <linux/refcount.h>
+#include <perf/cpumap.h>
+#include <internal/rc_check.h>
+
+/**
+ * A sized, reference counted, sorted array of integers representing CPU
+ * numbers. This is commonly used to capture which CPUs a PMU is associated
+ * with. The indices into the cpumap are frequently used as they avoid having
+ * gaps if CPU numbers were used. For events associated with a pid, rather than
+ * a CPU, a single dummy map with an entry of -1 is used.
+ */
+DECLARE_RC_STRUCT(perf_cpu_map) {
+	refcount_t	refcnt;
+	/** Length of the map array. */
+	int		nr;
+	/** The CPU values. */
+	struct perf_cpu	map[];
+};
+
+#ifndef MAX_NR_CPUS
+#define MAX_NR_CPUS	2048
+#endif
+
+struct perf_cpu_map *perf_cpu_map__alloc(int nr_cpus);
+int perf_cpu_map__idx(const struct perf_cpu_map *cpus, struct perf_cpu cpu);
+bool perf_cpu_map__is_subset(const struct perf_cpu_map *a, const struct perf_cpu_map *b);
+
+void perf_cpu_map__set_nr(struct perf_cpu_map *map, int nr_cpus);
+
+static inline refcount_t *perf_cpu_map__refcnt(struct perf_cpu_map *map)
+{
+	return &RC_CHK_ACCESS(map)->refcnt;
+}
+#endif /* __LIBPERF_INTERNAL_CPUMAP_H */
diff --git a/tools/lib/perf/include/internal/evlist.h b/tools/lib/perf/include/internal/evlist.h
new file mode 100644
index 000000000..3339bc2f1
--- /dev/null
+++ b/tools/lib/perf/include/internal/evlist.h
@@ -0,0 +1,138 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_EVLIST_H
+#define __LIBPERF_INTERNAL_EVLIST_H
+
+#include <linux/list.h>
+#include <api/fd/array.h>
+#include <internal/cpumap.h>
+#include <internal/evsel.h>
+
+#define PERF_EVLIST__HLIST_BITS 8
+#define PERF_EVLIST__HLIST_SIZE (1 << PERF_EVLIST__HLIST_BITS)
+
+struct perf_cpu_map;
+struct perf_thread_map;
+struct perf_mmap_param;
+
+struct perf_evlist {
+	struct list_head	 entries;
+	int			 nr_entries;
+	bool			 has_user_cpus;
+	bool			 needs_map_propagation;
+	/**
+	 * The cpus passed from the command line or all online CPUs by
+	 * default.
+	 */
+	struct perf_cpu_map	*user_requested_cpus;
+	/** The union of all evsel cpu maps. */
+	struct perf_cpu_map	*all_cpus;
+	struct perf_thread_map	*threads;
+	int			 nr_mmaps;
+	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_evsel*,
+			      struct perf_mmap_param*, int);
+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, struct perf_cpu);
+
+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, enum fdarray_flags flags);
+
+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
+ * @evsel: struct perf_evsel iterator
+ */
+#define __perf_evlist__for_each_entry(list, evsel) \
+	list_for_each_entry(evsel, list, node)
+
+/**
+ * evlist__for_each_entry - iterate thru all the evsels
+ * @evlist: perf_evlist instance to iterate
+ * @evsel: struct perf_evsel iterator
+ */
+#define perf_evlist__for_each_entry(evlist, evsel) \
+	__perf_evlist__for_each_entry(&(evlist)->entries, evsel)
+
+/**
+ * __perf_evlist__for_each_entry_reverse - iterate thru all the evsels in reverse order
+ * @list: list_head instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define __perf_evlist__for_each_entry_reverse(list, evsel) \
+	list_for_each_entry_reverse(evsel, list, node)
+
+/**
+ * perf_evlist__for_each_entry_reverse - iterate thru all the evsels in reverse order
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#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);
+}
+
+static inline struct perf_evsel *perf_evlist__last(struct perf_evlist *evlist)
+{
+	return list_entry(evlist->entries.prev, struct perf_evsel, node);
+}
+
+u64 perf_evlist__read_format(struct perf_evlist *evlist);
+
+void perf_evlist__id_add(struct perf_evlist *evlist,
+			 struct perf_evsel *evsel,
+			 int cpu, int thread, u64 id);
+
+int perf_evlist__id_add_fd(struct perf_evlist *evlist,
+			   struct perf_evsel *evsel,
+			   int cpu, int thread, int fd);
+
+void perf_evlist__reset_id_hash(struct perf_evlist *evlist);
+
+void __perf_evlist__set_leader(struct list_head *list, struct perf_evsel *leader);
+#endif /* __LIBPERF_INTERNAL_EVLIST_H */
diff --git a/tools/lib/perf/include/internal/evsel.h b/tools/lib/perf/include/internal/evsel.h
new file mode 100644
index 000000000..5cd220a61
--- /dev/null
+++ b/tools/lib/perf/include/internal/evsel.h
@@ -0,0 +1,91 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_EVSEL_H
+#define __LIBPERF_INTERNAL_EVSEL_H
+
+#include <linux/types.h>
+#include <linux/perf_event.h>
+#include <stdbool.h>
+#include <sys/types.h>
+#include <internal/cpumap.h>
+
+struct perf_thread_map;
+struct xyarray;
+
+/*
+ * Per fd, to map back from PERF_SAMPLE_ID to evsel, only used when there are
+ * more than one entry in the evlist.
+ */
+struct perf_sample_id {
+	struct hlist_node	 node;
+	u64			 id;
+	struct perf_evsel	*evsel;
+       /*
+	* 'idx' will be used for AUX area sampling. A sample will have AUX area
+	* data that will be queued for decoding, where there are separate
+	* queues for each CPU (per-cpu tracing) or task (per-thread tracing).
+	* The sample ID can be used to lookup 'idx' which is effectively the
+	* queue number.
+	*/
+	int			 idx;
+	struct perf_cpu		 cpu;
+	pid_t			 tid;
+
+	/* Guest machine pid and VCPU, valid only if machine_pid is non-zero */
+	pid_t			 machine_pid;
+	struct perf_cpu		 vcpu;
+
+	/* Holds total ID period value for PERF_SAMPLE_READ processing. */
+	u64			 period;
+};
+
+struct perf_evsel {
+	struct list_head	 node;
+	struct perf_event_attr	 attr;
+	/** The commonly used cpu map of CPUs the event should be opened upon, etc. */
+	struct perf_cpu_map	*cpus;
+	/**
+	 * The cpu map read from the PMU. For core PMUs this is the list of all
+	 * CPUs the event can be opened upon. For other PMUs this is the default
+	 * cpu map for opening the event on, for example, the first CPU on a
+	 * socket for an uncore event.
+	 */
+	struct perf_cpu_map	*own_cpus;
+	struct perf_thread_map	*threads;
+	struct xyarray		*fd;
+	struct xyarray		*mmap;
+	struct xyarray		*sample_id;
+	u64			*id;
+	u32			 ids;
+	struct perf_evsel	*leader;
+
+	/* parse modifier helper */
+	int			 nr_members;
+	/*
+	 * system_wide is for events that need to be on every CPU, irrespective
+	 * of user requested CPUs or threads. Tha main example of this is the
+	 * dummy event. Map propagation will set cpus for this event to all CPUs
+	 * as software PMU events like dummy, have a CPU map that is empty.
+	 */
+	bool			 system_wide;
+	/*
+	 * Some events, for example uncore events, require a CPU.
+	 * i.e. it cannot be the 'any CPU' value of -1.
+	 */
+	bool			 requires_cpu;
+	/** Is the PMU for the event a core one? Effects the handling of own_cpus. */
+	bool			 is_pmu_core;
+	int			 idx;
+};
+
+void perf_evsel__init(struct perf_evsel *evsel, struct perf_event_attr *attr,
+		      int idx);
+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);
+int perf_evsel__read_size(struct perf_evsel *evsel);
+int perf_evsel__apply_filter(struct perf_evsel *evsel, const char *filter);
+
+int perf_evsel__alloc_id(struct perf_evsel *evsel, int ncpus, int nthreads);
+void perf_evsel__free_id(struct perf_evsel *evsel);
+
+#endif /* __LIBPERF_INTERNAL_EVSEL_H */
diff --git a/tools/lib/perf/include/internal/lib.h b/tools/lib/perf/include/internal/lib.h
new file mode 100644
index 000000000..85471a4b9
--- /dev/null
+++ b/tools/lib/perf/include/internal/lib.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_LIB_H
+#define __LIBPERF_INTERNAL_LIB_H
+
+#include <sys/types.h>
+
+extern unsigned int page_size;
+
+ssize_t readn(int fd, void *buf, size_t n);
+ssize_t writen(int fd, const void *buf, size_t n);
+
+ssize_t preadn(int fd, void *buf, size_t n, off_t offs);
+
+#endif /* __LIBPERF_INTERNAL_CPUMAP_H */
diff --git a/tools/lib/perf/include/internal/mmap.h b/tools/lib/perf/include/internal/mmap.h
new file mode 100644
index 000000000..5a062af8e
--- /dev/null
+++ b/tools/lib/perf/include/internal/mmap.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_MMAP_H
+#define __LIBPERF_INTERNAL_MMAP_H
+
+#include <linux/compiler.h>
+#include <linux/refcount.h>
+#include <linux/types.h>
+#include <stdbool.h>
+#include <internal/cpumap.h>
+
+/* perf sample has 16 bits size limit */
+#define PERF_SAMPLE_MAX_SIZE (1 << 16)
+
+struct perf_mmap;
+struct perf_counts_values;
+
+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;
+	struct perf_cpu		 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, struct perf_cpu 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);
+
+int perf_mmap__read_self(struct perf_mmap *map, struct perf_counts_values *count);
+
+#endif /* __LIBPERF_INTERNAL_MMAP_H */
diff --git a/tools/lib/perf/include/internal/rc_check.h b/tools/lib/perf/include/internal/rc_check.h
new file mode 100644
index 000000000..e88a6d8a0
--- /dev/null
+++ b/tools/lib/perf/include/internal/rc_check.h
@@ -0,0 +1,106 @@
+/* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */
+#ifndef __LIBPERF_INTERNAL_RC_CHECK_H
+#define __LIBPERF_INTERNAL_RC_CHECK_H
+
+#include <stdlib.h>
+#include <linux/zalloc.h>
+
+/*
+ * Enable reference count checking implicitly with leak checking, which is
+ * integrated into address sanitizer.
+ */
+#if defined(__SANITIZE_ADDRESS__) || defined(LEAK_SANITIZER) || defined(ADDRESS_SANITIZER)
+#define REFCNT_CHECKING 1
+#elif defined(__has_feature)
+#if __has_feature(address_sanitizer) || __has_feature(leak_sanitizer)
+#define REFCNT_CHECKING 1
+#endif
+#endif
+
+/*
+ * Shared reference count checking macros.
+ *
+ * Reference count checking is an approach to sanitizing the use of reference
+ * counted structs. It leverages address and leak sanitizers to make sure gets
+ * are paired with a put. Reference count checking adds a malloc-ed layer of
+ * indirection on a get, and frees it on a put. A missed put will be reported as
+ * a memory leak. A double put will be reported as a double free. Accessing
+ * after a put will cause a use-after-free and/or a segfault.
+ */
+
+#ifndef REFCNT_CHECKING
+/* Replaces "struct foo" so that the pointer may be interposed. */
+#define DECLARE_RC_STRUCT(struct_name)		\
+	struct struct_name
+
+/* Declare a reference counted struct variable. */
+#define RC_STRUCT(struct_name) struct struct_name
+
+/*
+ * Interpose the indirection. Result will hold the indirection and object is the
+ * reference counted struct.
+ */
+#define ADD_RC_CHK(result, object) (result = object, object)
+
+/* Strip the indirection layer. */
+#define RC_CHK_ACCESS(object) object
+
+/* Frees the object and the indirection layer. */
+#define RC_CHK_FREE(object) free(object)
+
+/* A get operation adding the indirection layer. */
+#define RC_CHK_GET(result, object) ADD_RC_CHK(result, object)
+
+/* A put operation removing the indirection layer. */
+#define RC_CHK_PUT(object) {}
+
+#else
+
+/* Replaces "struct foo" so that the pointer may be interposed. */
+#define DECLARE_RC_STRUCT(struct_name)			\
+	struct original_##struct_name;			\
+	struct struct_name {				\
+		struct original_##struct_name *orig;	\
+	};						\
+	struct original_##struct_name
+
+/* Declare a reference counted struct variable. */
+#define RC_STRUCT(struct_name) struct original_##struct_name
+
+/*
+ * Interpose the indirection. Result will hold the indirection and object is the
+ * reference counted struct.
+ */
+#define ADD_RC_CHK(result, object)					\
+	(								\
+		object ? (result = malloc(sizeof(*result)),		\
+			result ? (result->orig = object, result)	\
+			: (result = NULL, NULL))			\
+		: (result = NULL, NULL)					\
+		)
+
+/* Strip the indirection layer. */
+#define RC_CHK_ACCESS(object) object->orig
+
+/* Frees the object and the indirection layer. */
+#define RC_CHK_FREE(object)			\
+	do {					\
+		zfree(&object->orig);		\
+		free(object);			\
+	} while(0)
+
+/* A get operation adding the indirection layer. */
+#define RC_CHK_GET(result, object) ADD_RC_CHK(result, (object ? object->orig : NULL))
+
+/* A put operation removing the indirection layer. */
+#define RC_CHK_PUT(object)			\
+	do {					\
+		if (object) {			\
+			object->orig = NULL;	\
+			free(object);		\
+		}				\
+	} while(0)
+
+#endif
+
+#endif /* __LIBPERF_INTERNAL_RC_CHECK_H */
diff --git a/tools/lib/perf/include/internal/tests.h b/tools/lib/perf/include/internal/tests.h
new file mode 100644
index 000000000..b130a6663
--- /dev/null
+++ b/tools/lib/perf/include/internal/tests.h
@@ -0,0 +1,67 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_TESTS_H
+#define __LIBPERF_INTERNAL_TESTS_H
+
+#include <stdio.h>
+#include <unistd.h>
+
+extern int tests_failed;
+extern int tests_verbose;
+
+static inline int get_verbose(char **argv, int argc)
+{
+	int c;
+	int verbose = 0;
+
+	while ((c = getopt(argc, argv, "v")) != -1) {
+		switch (c)
+		{
+		case 'v':
+			verbose = 1;
+			break;
+		default:
+			break;
+		}
+	}
+	optind = 1;
+
+	return verbose;
+}
+
+#define __T_START					\
+do {							\
+	tests_verbose = get_verbose(argv, argc);	\
+	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)
+
+#define __T_VERBOSE(...)						\
+do {									\
+	if (tests_verbose) {						\
+		if (tests_verbose == 1) {				\
+			fputc('\n', stderr);				\
+			tests_verbose++;				\
+		}							\
+		fprintf(stderr, ##__VA_ARGS__);				\
+	}								\
+} while (0)
+
+#endif /* __LIBPERF_INTERNAL_TESTS_H */
diff --git a/tools/lib/perf/include/internal/threadmap.h b/tools/lib/perf/include/internal/threadmap.h
new file mode 100644
index 000000000..df748baf9
--- /dev/null
+++ b/tools/lib/perf/include/internal/threadmap.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_THREADMAP_H
+#define __LIBPERF_INTERNAL_THREADMAP_H
+
+#include <linux/refcount.h>
+#include <sys/types.h>
+#include <unistd.h>
+
+struct thread_map_data {
+	pid_t	 pid;
+	char	*comm;
+};
+
+struct perf_thread_map {
+	refcount_t	refcnt;
+	int		nr;
+	int		err_thread;
+	struct thread_map_data map[];
+};
+
+struct perf_thread_map *perf_thread_map__realloc(struct perf_thread_map *map, int nr);
+
+#endif /* __LIBPERF_INTERNAL_THREADMAP_H */
diff --git a/tools/lib/perf/include/internal/xyarray.h b/tools/lib/perf/include/internal/xyarray.h
new file mode 100644
index 000000000..f10af3da7
--- /dev/null
+++ b/tools/lib/perf/include/internal/xyarray.h
@@ -0,0 +1,43 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_INTERNAL_XYARRAY_H
+#define __LIBPERF_INTERNAL_XYARRAY_H
+
+#include <linux/compiler.h>
+#include <sys/types.h>
+
+struct xyarray {
+	size_t row_size;
+	size_t entry_size;
+	size_t entries;
+	size_t max_x;
+	size_t max_y;
+	char contents[] __aligned(8);
+};
+
+struct xyarray *xyarray__new(int xlen, int ylen, size_t entry_size);
+void xyarray__delete(struct xyarray *xy);
+void xyarray__reset(struct xyarray *xy);
+
+static inline void *__xyarray__entry(struct xyarray *xy, int x, int y)
+{
+	return &xy->contents[x * xy->row_size + y * xy->entry_size];
+}
+
+static inline void *xyarray__entry(struct xyarray *xy, size_t x, size_t y)
+{
+	if (x >= xy->max_x || y >= xy->max_y)
+		return NULL;
+	return __xyarray__entry(xy, x, y);
+}
+
+static inline int xyarray__max_y(struct xyarray *xy)
+{
+	return xy->max_y;
+}
+
+static inline int xyarray__max_x(struct xyarray *xy)
+{
+	return xy->max_x;
+}
+
+#endif /* __LIBPERF_INTERNAL_XYARRAY_H */
diff --git a/tools/lib/perf/include/perf/bpf_perf.h b/tools/lib/perf/include/perf/bpf_perf.h
new file mode 100644
index 000000000..e7cf6ba7b
--- /dev/null
+++ b/tools/lib/perf/include/perf/bpf_perf.h
@@ -0,0 +1,31 @@
+/* SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause) */
+#ifndef __LIBPERF_BPF_PERF_H
+#define __LIBPERF_BPF_PERF_H
+
+#include <linux/types.h>  /* for __u32 */
+
+/*
+ * bpf_perf uses a hashmap, the attr_map, to track all the leader programs.
+ * The hashmap is pinned in bpffs. flock() on this file is used to ensure
+ * no concurrent access to the attr_map.  The key of attr_map is struct
+ * perf_event_attr, and the value is struct perf_event_attr_map_entry.
+ *
+ * struct perf_event_attr_map_entry contains two __u32 IDs, bpf_link of the
+ * leader prog, and the diff_map. Each perf-stat session holds a reference
+ * to the bpf_link to make sure the leader prog is attached to sched_switch
+ * tracepoint.
+ *
+ * Since the hashmap only contains IDs of the bpf_link and diff_map, it
+ * does not hold any references to the leader program. Once all perf-stat
+ * sessions of these events exit, the leader prog, its maps, and the
+ * perf_events will be freed.
+ */
+struct perf_event_attr_map_entry {
+	__u32 link_id;
+	__u32 diff_map_id;
+};
+
+/* default attr_map name */
+#define BPF_PERF_DEFAULT_ATTR_MAP_PATH "perf_attr_map"
+
+#endif /* __LIBPERF_BPF_PERF_H */
diff --git a/tools/lib/perf/include/perf/core.h b/tools/lib/perf/include/perf/core.h
new file mode 100644
index 000000000..a3f6d68ed
--- /dev/null
+++ b/tools/lib/perf/include/perf/core.h
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_CORE_H
+#define __LIBPERF_CORE_H
+
+#include <stdarg.h>
+
+#ifndef LIBPERF_API
+#define LIBPERF_API __attribute__((visibility("default")))
+#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,
+				  const char *, va_list ap);
+
+LIBPERF_API void libperf_init(libperf_print_fn_t fn);
+
+#endif /* __LIBPERF_CORE_H */
diff --git a/tools/lib/perf/include/perf/cpumap.h b/tools/lib/perf/include/perf/cpumap.h
new file mode 100644
index 000000000..e38d859a3
--- /dev/null
+++ b/tools/lib/perf/include/perf/cpumap.h
@@ -0,0 +1,57 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_CPUMAP_H
+#define __LIBPERF_CPUMAP_H
+
+#include <perf/core.h>
+#include <stdio.h>
+#include <stdbool.h>
+
+/** A wrapper around a CPU to avoid confusion with the perf_cpu_map's map's indices. */
+struct perf_cpu {
+	int cpu;
+};
+
+struct perf_cache {
+	int cache_lvl;
+	int cache;
+};
+
+struct perf_cpu_map;
+
+/**
+ * perf_cpu_map__dummy_new - a map with a singular "any CPU"/dummy -1 value.
+ */
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__dummy_new(void);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__default_new(void);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__new(const char *cpu_list);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__read(FILE *file);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__get(struct perf_cpu_map *map);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__merge(struct perf_cpu_map *orig,
+						     struct perf_cpu_map *other);
+LIBPERF_API struct perf_cpu_map *perf_cpu_map__intersect(struct perf_cpu_map *orig,
+							 struct perf_cpu_map *other);
+LIBPERF_API void perf_cpu_map__put(struct perf_cpu_map *map);
+LIBPERF_API struct perf_cpu perf_cpu_map__cpu(const struct perf_cpu_map *cpus, int idx);
+LIBPERF_API int perf_cpu_map__nr(const struct perf_cpu_map *cpus);
+/**
+ * perf_cpu_map__empty - is map either empty or the "any CPU"/dummy value.
+ */
+LIBPERF_API bool perf_cpu_map__empty(const struct perf_cpu_map *map);
+LIBPERF_API struct perf_cpu perf_cpu_map__max(const struct perf_cpu_map *map);
+LIBPERF_API bool perf_cpu_map__has(const struct perf_cpu_map *map, struct perf_cpu cpu);
+LIBPERF_API bool perf_cpu_map__equal(const struct perf_cpu_map *lhs,
+				     const struct perf_cpu_map *rhs);
+/**
+ * perf_cpu_map__any_cpu - Does the map contain the "any CPU"/dummy -1 value?
+ */
+LIBPERF_API bool perf_cpu_map__has_any_cpu(const struct perf_cpu_map *map);
+
+#define perf_cpu_map__for_each_cpu(cpu, idx, cpus)		\
+	for ((idx) = 0, (cpu) = perf_cpu_map__cpu(cpus, idx);	\
+	     (idx) < perf_cpu_map__nr(cpus);			\
+	     (idx)++, (cpu) = perf_cpu_map__cpu(cpus, idx))
+
+#define perf_cpu_map__for_each_idx(idx, cpus)				\
+	for ((idx) = 0; (idx) < perf_cpu_map__nr(cpus); (idx)++)
+
+#endif /* __LIBPERF_CPUMAP_H */
diff --git a/tools/lib/perf/include/perf/event.h b/tools/lib/perf/include/perf/event.h
new file mode 100644
index 000000000..ae6409018
--- /dev/null
+++ b/tools/lib/perf/include/perf/event.h
@@ -0,0 +1,517 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_EVENT_H
+#define __LIBPERF_EVENT_H
+
+#include <linux/perf_event.h>
+#include <linux/types.h>
+#include <linux/limits.h>
+#include <linux/bpf.h>
+#include <sys/types.h> /* pid_t */
+
+#define event_contains(obj, mem) ((obj).header.size > offsetof(typeof(obj), mem))
+
+struct perf_record_mmap {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 start;
+	__u64			 len;
+	__u64			 pgoff;
+	char			 filename[PATH_MAX];
+};
+
+struct perf_record_mmap2 {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 start;
+	__u64			 len;
+	__u64			 pgoff;
+	union {
+		struct {
+			__u32	 maj;
+			__u32	 min;
+			__u64	 ino;
+			__u64	 ino_generation;
+		};
+		struct {
+			__u8	 build_id_size;
+			__u8	 __reserved_1;
+			__u16	 __reserved_2;
+			__u8	 build_id[20];
+		};
+	};
+	__u32			 prot;
+	__u32			 flags;
+	char			 filename[PATH_MAX];
+};
+
+struct perf_record_comm {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	char			 comm[16];
+};
+
+struct perf_record_namespaces {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 nr_namespaces;
+	struct perf_ns_link_info link_info[];
+};
+
+struct perf_record_fork {
+	struct perf_event_header header;
+	__u32			 pid, ppid;
+	__u32			 tid, ptid;
+	__u64			 time;
+};
+
+struct perf_record_lost {
+	struct perf_event_header header;
+	__u64			 id;
+	__u64			 lost;
+};
+
+#define PERF_RECORD_MISC_LOST_SAMPLES_BPF (1 << 15)
+
+struct perf_record_lost_samples {
+	struct perf_event_header header;
+	__u64			 lost;
+};
+
+/*
+ * PERF_FORMAT_ENABLED | PERF_FORMAT_RUNNING | PERF_FORMAT_ID | PERF_FORMAT_LOST
+ */
+struct perf_record_read {
+	struct perf_event_header header;
+	__u32			 pid, tid;
+	__u64			 value;
+	__u64			 time_enabled;
+	__u64			 time_running;
+	__u64			 id;
+	__u64			 lost;
+};
+
+struct perf_record_throttle {
+	struct perf_event_header header;
+	__u64			 time;
+	__u64			 id;
+	__u64			 stream_id;
+};
+
+#ifndef KSYM_NAME_LEN
+#define KSYM_NAME_LEN 512
+#endif
+
+struct perf_record_ksymbol {
+	struct perf_event_header header;
+	__u64			 addr;
+	__u32			 len;
+	__u16			 ksym_type;
+	__u16			 flags;
+	char			 name[KSYM_NAME_LEN];
+};
+
+struct perf_record_bpf_event {
+	struct perf_event_header header;
+	__u16			 type;
+	__u16			 flags;
+	__u32			 id;
+
+	/* for bpf_prog types */
+	__u8			 tag[BPF_TAG_SIZE];  // prog tag
+};
+
+struct perf_record_cgroup {
+	struct perf_event_header header;
+	__u64			 id;
+	char			 path[PATH_MAX];
+};
+
+struct perf_record_text_poke_event {
+	struct perf_event_header header;
+	__u64			addr;
+	__u16			old_len;
+	__u16			new_len;
+	__u8			bytes[];
+};
+
+struct perf_record_sample {
+	struct perf_event_header header;
+	__u64			 array[];
+};
+
+struct perf_record_switch {
+	struct perf_event_header header;
+	__u32			 next_prev_pid;
+	__u32			 next_prev_tid;
+};
+
+struct perf_record_header_attr {
+	struct perf_event_header header;
+	struct perf_event_attr	 attr;
+	/*
+	 * Array of u64 id follows here but we cannot use a flexible array
+	 * because size of attr in the data can be different then current
+	 * version.  Please use perf_record_header_attr_id() below.
+	 *
+	 * __u64		 id[];  // do not use this
+	 */
+};
+
+/* Returns the pointer to id array based on the actual attr size. */
+#define perf_record_header_attr_id(evt)			\
+	((void *)&(evt)->attr.attr + (evt)->attr.attr.size)
+
+enum {
+	PERF_CPU_MAP__CPUS = 0,
+	PERF_CPU_MAP__MASK = 1,
+	PERF_CPU_MAP__RANGE_CPUS = 2,
+};
+
+/*
+ * Array encoding of a perf_cpu_map where nr is the number of entries in cpu[]
+ * and each entry is a value for a CPU in the map.
+ */
+struct cpu_map_entries {
+	__u16			 nr;
+	__u16			 cpu[];
+};
+
+/* Bitmap encoding of a perf_cpu_map where bitmap entries are 32-bit. */
+struct perf_record_mask_cpu_map32 {
+	/* Number of mask values. */
+	__u16			 nr;
+	/* Constant 4. */
+	__u16			 long_size;
+	/* Bitmap data. */
+	__u32			 mask[];
+};
+
+/* Bitmap encoding of a perf_cpu_map where bitmap entries are 64-bit. */
+struct perf_record_mask_cpu_map64 {
+	/* Number of mask values. */
+	__u16			 nr;
+	/* Constant 8. */
+	__u16			 long_size;
+	/* Legacy padding. */
+	char                     __pad[4];
+	/* Bitmap data. */
+	__u64			 mask[];
+};
+
+/*
+ * 'struct perf_record_cpu_map_data' is packed as unfortunately an earlier
+ * version had unaligned data and we wish to retain file format compatibility.
+ * -irogers
+ */
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wpacked"
+#pragma GCC diagnostic ignored "-Wattributes"
+
+/*
+ * An encoding of a CPU map for a range starting at start_cpu through to
+ * end_cpu. If any_cpu is 1, an any CPU (-1) value (aka dummy value) is present.
+ */
+struct perf_record_range_cpu_map {
+	__u8 any_cpu;
+	__u8 __pad;
+	__u16 start_cpu;
+	__u16 end_cpu;
+};
+
+struct perf_record_cpu_map_data {
+	__u16			 type;
+	union {
+		/* Used when type == PERF_CPU_MAP__CPUS. */
+		struct cpu_map_entries cpus_data;
+		/* Used when type == PERF_CPU_MAP__MASK and long_size == 4. */
+		struct perf_record_mask_cpu_map32 mask32_data;
+		/* Used when type == PERF_CPU_MAP__MASK and long_size == 8. */
+		struct perf_record_mask_cpu_map64 mask64_data;
+		/* Used when type == PERF_CPU_MAP__RANGE_CPUS. */
+		struct perf_record_range_cpu_map range_cpu_data;
+	};
+} __attribute__((packed));
+
+#pragma GCC diagnostic pop
+
+struct perf_record_cpu_map {
+	struct perf_event_header	 header;
+	struct perf_record_cpu_map_data	 data;
+};
+
+enum {
+	PERF_EVENT_UPDATE__UNIT  = 0,
+	PERF_EVENT_UPDATE__SCALE = 1,
+	PERF_EVENT_UPDATE__NAME  = 2,
+	PERF_EVENT_UPDATE__CPUS  = 3,
+};
+
+struct perf_record_event_update_cpus {
+	struct perf_record_cpu_map_data	 cpus;
+};
+
+struct perf_record_event_update_scale {
+	double			 scale;
+};
+
+struct perf_record_event_update {
+	struct perf_event_header header;
+	__u64			 type;
+	__u64			 id;
+	union {
+		/* Used when type == PERF_EVENT_UPDATE__SCALE. */
+		struct perf_record_event_update_scale scale;
+		/* Used when type == PERF_EVENT_UPDATE__UNIT. */
+		char unit[0];
+		/* Used when type == PERF_EVENT_UPDATE__NAME. */
+		char name[0];
+		/* Used when type == PERF_EVENT_UPDATE__CPUS. */
+		struct perf_record_event_update_cpus cpus;
+	};
+};
+
+#define MAX_EVENT_NAME 64
+
+struct perf_trace_event_type {
+	__u64			 event_id;
+	char			 name[MAX_EVENT_NAME];
+};
+
+struct perf_record_header_event_type {
+	struct perf_event_header	 header;
+	struct perf_trace_event_type	 event_type;
+};
+
+struct perf_record_header_tracing_data {
+	struct perf_event_header header;
+	__u32			 size;
+};
+
+#define PERF_RECORD_MISC_BUILD_ID_SIZE (1 << 15)
+
+struct perf_record_header_build_id {
+	struct perf_event_header header;
+	pid_t			 pid;
+	union {
+		__u8		 build_id[24];
+		struct {
+			__u8	 data[20];
+			__u8	 size;
+			__u8	 reserved1__;
+			__u16	 reserved2__;
+		};
+	};
+	char			 filename[];
+};
+
+struct id_index_entry {
+	__u64			 id;
+	__u64			 idx;
+	__u64			 cpu;
+	__u64			 tid;
+};
+
+struct id_index_entry_2 {
+	__u64			 machine_pid;
+	__u64			 vcpu;
+};
+
+struct perf_record_id_index {
+	struct perf_event_header header;
+	__u64			 nr;
+	struct id_index_entry	 entries[];
+};
+
+struct perf_record_auxtrace_info {
+	struct perf_event_header header;
+	__u32			 type;
+	__u32			 reserved__; /* For alignment */
+	__u64			 priv[];
+};
+
+struct perf_record_auxtrace {
+	struct perf_event_header header;
+	__u64			 size;
+	__u64			 offset;
+	__u64			 reference;
+	__u32			 idx;
+	__u32			 tid;
+	__u32			 cpu;
+	__u32			 reserved__; /* For alignment */
+};
+
+#define MAX_AUXTRACE_ERROR_MSG 64
+
+struct perf_record_auxtrace_error {
+	struct perf_event_header header;
+	__u32			 type;
+	__u32			 code;
+	__u32			 cpu;
+	__u32			 pid;
+	__u32			 tid;
+	__u32			 fmt;
+	__u64			 ip;
+	__u64			 time;
+	char			 msg[MAX_AUXTRACE_ERROR_MSG];
+	__u32			 machine_pid;
+	__u32			 vcpu;
+};
+
+struct perf_record_aux {
+	struct perf_event_header header;
+	__u64			 aux_offset;
+	__u64			 aux_size;
+	__u64			 flags;
+};
+
+struct perf_record_itrace_start {
+	struct perf_event_header header;
+	__u32			 pid;
+	__u32			 tid;
+};
+
+struct perf_record_aux_output_hw_id {
+	struct perf_event_header header;
+	__u64			hw_id;
+};
+
+struct perf_record_thread_map_entry {
+	__u64			 pid;
+	char			 comm[16];
+};
+
+struct perf_record_thread_map {
+	struct perf_event_header		 header;
+	__u64					 nr;
+	struct perf_record_thread_map_entry	 entries[];
+};
+
+enum {
+	PERF_STAT_CONFIG_TERM__AGGR_MODE	= 0,
+	PERF_STAT_CONFIG_TERM__INTERVAL		= 1,
+	PERF_STAT_CONFIG_TERM__SCALE		= 2,
+	PERF_STAT_CONFIG_TERM__AGGR_LEVEL	= 3,
+	PERF_STAT_CONFIG_TERM__MAX		= 4,
+};
+
+struct perf_record_stat_config_entry {
+	__u64			 tag;
+	__u64			 val;
+};
+
+struct perf_record_stat_config {
+	struct perf_event_header		 header;
+	__u64					 nr;
+	struct perf_record_stat_config_entry	 data[];
+};
+
+struct perf_record_stat {
+	struct perf_event_header header;
+
+	__u64			 id;
+	__u32			 cpu;
+	__u32			 thread;
+
+	union {
+		struct {
+			__u64	 val;
+			__u64	 ena;
+			__u64	 run;
+		};
+		__u64		 values[3];
+	};
+};
+
+struct perf_record_stat_round {
+	struct perf_event_header header;
+	__u64			 type;
+	__u64			 time;
+};
+
+struct perf_record_time_conv {
+	struct perf_event_header header;
+	__u64			 time_shift;
+	__u64			 time_mult;
+	__u64			 time_zero;
+	__u64			 time_cycles;
+	__u64			 time_mask;
+	__u8			 cap_user_time_zero;
+	__u8			 cap_user_time_short;
+	__u8			 reserved[6];	/* For alignment */
+};
+
+struct perf_record_header_feature {
+	struct perf_event_header header;
+	__u64			 feat_id;
+	char			 data[];
+};
+
+struct perf_record_compressed {
+	struct perf_event_header header;
+	char			 data[];
+};
+
+enum perf_user_event_type { /* above any possible kernel type */
+	PERF_RECORD_USER_TYPE_START		= 64,
+	PERF_RECORD_HEADER_ATTR			= 64,
+	PERF_RECORD_HEADER_EVENT_TYPE		= 65, /* deprecated */
+	PERF_RECORD_HEADER_TRACING_DATA		= 66,
+	PERF_RECORD_HEADER_BUILD_ID		= 67,
+	PERF_RECORD_FINISHED_ROUND		= 68,
+	PERF_RECORD_ID_INDEX			= 69,
+	PERF_RECORD_AUXTRACE_INFO		= 70,
+	PERF_RECORD_AUXTRACE			= 71,
+	PERF_RECORD_AUXTRACE_ERROR		= 72,
+	PERF_RECORD_THREAD_MAP			= 73,
+	PERF_RECORD_CPU_MAP			= 74,
+	PERF_RECORD_STAT_CONFIG			= 75,
+	PERF_RECORD_STAT			= 76,
+	PERF_RECORD_STAT_ROUND			= 77,
+	PERF_RECORD_EVENT_UPDATE		= 78,
+	PERF_RECORD_TIME_CONV			= 79,
+	PERF_RECORD_HEADER_FEATURE		= 80,
+	PERF_RECORD_COMPRESSED			= 81,
+	PERF_RECORD_FINISHED_INIT		= 82,
+	PERF_RECORD_HEADER_MAX
+};
+
+union perf_event {
+	struct perf_event_header		header;
+	struct perf_record_mmap			mmap;
+	struct perf_record_mmap2		mmap2;
+	struct perf_record_comm			comm;
+	struct perf_record_namespaces		namespaces;
+	struct perf_record_cgroup		cgroup;
+	struct perf_record_fork			fork;
+	struct perf_record_lost			lost;
+	struct perf_record_lost_samples		lost_samples;
+	struct perf_record_read			read;
+	struct perf_record_throttle		throttle;
+	struct perf_record_sample		sample;
+	struct perf_record_bpf_event		bpf;
+	struct perf_record_ksymbol		ksymbol;
+	struct perf_record_text_poke_event	text_poke;
+	struct perf_record_header_attr		attr;
+	struct perf_record_event_update		event_update;
+	struct perf_record_header_event_type	event_type;
+	struct perf_record_header_tracing_data	tracing_data;
+	struct perf_record_header_build_id	build_id;
+	struct perf_record_id_index		id_index;
+	struct perf_record_auxtrace_info	auxtrace_info;
+	struct perf_record_auxtrace		auxtrace;
+	struct perf_record_auxtrace_error	auxtrace_error;
+	struct perf_record_aux			aux;
+	struct perf_record_itrace_start		itrace_start;
+	struct perf_record_aux_output_hw_id	aux_output_hw_id;
+	struct perf_record_switch		context_switch;
+	struct perf_record_thread_map		thread_map;
+	struct perf_record_cpu_map		cpu_map;
+	struct perf_record_stat_config		stat_config;
+	struct perf_record_stat			stat;
+	struct perf_record_stat_round		stat_round;
+	struct perf_record_time_conv		time_conv;
+	struct perf_record_header_feature	feat;
+	struct perf_record_compressed		pack;
+};
+
+#endif /* __LIBPERF_EVENT_H */
diff --git a/tools/lib/perf/include/perf/evlist.h b/tools/lib/perf/include/perf/evlist.h
new file mode 100644
index 000000000..e894b7707
--- /dev/null
+++ b/tools/lib/perf/include/perf/evlist.h
@@ -0,0 +1,51 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_EVLIST_H
+#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__add(struct perf_evlist *evlist,
+				  struct perf_evsel *evsel);
+LIBPERF_API void perf_evlist__remove(struct perf_evlist *evlist,
+				     struct perf_evsel *evsel);
+LIBPERF_API struct perf_evlist *perf_evlist__new(void);
+LIBPERF_API void perf_evlist__delete(struct perf_evlist *evlist);
+LIBPERF_API struct perf_evsel* perf_evlist__next(struct perf_evlist *evlist,
+						 struct perf_evsel *evsel);
+LIBPERF_API int perf_evlist__open(struct perf_evlist *evlist);
+LIBPERF_API void perf_evlist__close(struct perf_evlist *evlist);
+LIBPERF_API void perf_evlist__enable(struct perf_evlist *evlist);
+LIBPERF_API void perf_evlist__disable(struct perf_evlist *evlist);
+
+#define perf_evlist__for_each_evsel(evlist, pos)	\
+	for ((pos) = perf_evlist__next((evlist), NULL);	\
+	     (pos) != NULL;				\
+	     (pos) = perf_evlist__next((evlist), (pos)))
+
+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))
+
+LIBPERF_API void perf_evlist__set_leader(struct perf_evlist *evlist);
+LIBPERF_API int perf_evlist__nr_groups(struct perf_evlist *evlist);
+#endif /* __LIBPERF_EVLIST_H */
diff --git a/tools/lib/perf/include/perf/evsel.h b/tools/lib/perf/include/perf/evsel.h
new file mode 100644
index 000000000..6f9220407
--- /dev/null
+++ b/tools/lib/perf/include/perf/evsel.h
@@ -0,0 +1,50 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_EVSEL_H
+#define __LIBPERF_EVSEL_H
+
+#include <stdint.h>
+#include <perf/core.h>
+#include <stdbool.h>
+#include <linux/types.h>
+
+struct perf_evsel;
+struct perf_event_attr;
+struct perf_cpu_map;
+struct perf_thread_map;
+
+struct perf_counts_values {
+	union {
+		struct {
+			uint64_t val;
+			uint64_t ena;
+			uint64_t run;
+			uint64_t id;
+			uint64_t lost;
+		};
+		uint64_t values[5];
+	};
+};
+
+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,
+				 struct perf_thread_map *threads);
+LIBPERF_API void perf_evsel__close(struct perf_evsel *evsel);
+LIBPERF_API void perf_evsel__close_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+LIBPERF_API int perf_evsel__mmap(struct perf_evsel *evsel, int pages);
+LIBPERF_API void perf_evsel__munmap(struct perf_evsel *evsel);
+LIBPERF_API void *perf_evsel__mmap_base(struct perf_evsel *evsel, int cpu_map_idx, int thread);
+LIBPERF_API int perf_evsel__read(struct perf_evsel *evsel, int cpu_map_idx, int thread,
+				 struct perf_counts_values *count);
+LIBPERF_API int perf_evsel__enable(struct perf_evsel *evsel);
+LIBPERF_API int perf_evsel__enable_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+LIBPERF_API int perf_evsel__enable_thread(struct perf_evsel *evsel, int thread);
+LIBPERF_API int perf_evsel__disable(struct perf_evsel *evsel);
+LIBPERF_API int perf_evsel__disable_cpu(struct perf_evsel *evsel, int cpu_map_idx);
+LIBPERF_API struct perf_cpu_map *perf_evsel__cpus(struct perf_evsel *evsel);
+LIBPERF_API struct perf_thread_map *perf_evsel__threads(struct perf_evsel *evsel);
+LIBPERF_API struct perf_event_attr *perf_evsel__attr(struct perf_evsel *evsel);
+LIBPERF_API void perf_counts_values__scale(struct perf_counts_values *count,
+					   bool scale, __s8 *pscaled);
+
+#endif /* __LIBPERF_EVSEL_H */
diff --git a/tools/lib/perf/include/perf/mmap.h b/tools/lib/perf/include/perf/mmap.h
new file mode 100644
index 000000000..9508ad90d
--- /dev/null
+++ b/tools/lib/perf/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/lib/perf/include/perf/threadmap.h b/tools/lib/perf/include/perf/threadmap.h
new file mode 100644
index 000000000..8b40e7777
--- /dev/null
+++ b/tools/lib/perf/include/perf/threadmap.h
@@ -0,0 +1,21 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef __LIBPERF_THREADMAP_H
+#define __LIBPERF_THREADMAP_H
+
+#include <perf/core.h>
+#include <sys/types.h>
+
+struct perf_thread_map;
+
+LIBPERF_API struct perf_thread_map *perf_thread_map__new_dummy(void);
+LIBPERF_API struct perf_thread_map *perf_thread_map__new_array(int nr_threads, pid_t *array);
+
+LIBPERF_API void perf_thread_map__set_pid(struct perf_thread_map *map, int idx, pid_t pid);
+LIBPERF_API char *perf_thread_map__comm(struct perf_thread_map *map, int idx);
+LIBPERF_API int perf_thread_map__nr(struct perf_thread_map *threads);
+LIBPERF_API pid_t perf_thread_map__pid(struct perf_thread_map *map, int idx);
+
+LIBPERF_API struct perf_thread_map *perf_thread_map__get(struct perf_thread_map *map);
+LIBPERF_API void perf_thread_map__put(struct perf_thread_map *map);
+
+#endif /* __LIBPERF_THREADMAP_H */
diff --git a/tools/lib/perf/internal.h b/tools/lib/perf/internal.h
new file mode 100644
index 000000000..2c27e158d
--- /dev/null
+++ b/tools/lib/perf/internal.h
@@ -0,0 +1,23 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#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)));
+
+#define __pr(level, fmt, ...)   \
+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/lib/perf/lib.c b/tools/lib/perf/lib.c
new file mode 100644
index 000000000..696fb0ea6
--- /dev/null
+++ b/tools/lib/perf/lib.c
@@ -0,0 +1,68 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <unistd.h>
+#include <stdbool.h>
+#include <errno.h>
+#include <linux/kernel.h>
+#include <internal/lib.h>
+
+unsigned int page_size;
+
+static ssize_t ion(bool is_read, int fd, void *buf, size_t n)
+{
+	void *buf_start = buf;
+	size_t left = n;
+
+	while (left) {
+		/* buf must be treated as const if !is_read. */
+		ssize_t ret = is_read ? read(fd, buf, left) :
+					write(fd, buf, left);
+
+		if (ret < 0 && errno == EINTR)
+			continue;
+		if (ret <= 0)
+			return ret;
+
+		left -= ret;
+		buf  += ret;
+	}
+
+	BUG_ON((size_t)(buf - buf_start) != n);
+	return n;
+}
+
+/*
+ * Read exactly 'n' bytes or return an error.
+ */
+ssize_t readn(int fd, void *buf, size_t n)
+{
+	return ion(true, fd, buf, n);
+}
+
+ssize_t preadn(int fd, void *buf, size_t n, off_t offs)
+{
+	size_t left = n;
+
+	while (left) {
+		ssize_t ret = pread(fd, buf, left, offs);
+
+		if (ret < 0 && errno == EINTR)
+			continue;
+		if (ret <= 0)
+			return ret;
+
+		left -= ret;
+		buf  += ret;
+		offs += ret;
+	}
+
+	return n;
+}
+
+/*
+ * Write exactly 'n' bytes or return an error.
+ */
+ssize_t writen(int fd, const void *buf, size_t n)
+{
+	/* ion does not modify buf. */
+	return ion(false, fd, (void *)buf, n);
+}
diff --git a/tools/lib/perf/libperf.map b/tools/lib/perf/libperf.map
new file mode 100644
index 000000000..190b56ae9
--- /dev/null
+++ b/tools/lib/perf/libperf.map
@@ -0,0 +1,59 @@
+LIBPERF_0.0.1 {
+	global:
+		libperf_init;
+		perf_cpu_map__dummy_new;
+		perf_cpu_map__default_new;
+		perf_cpu_map__get;
+		perf_cpu_map__put;
+		perf_cpu_map__new;
+		perf_cpu_map__read;
+		perf_cpu_map__nr;
+		perf_cpu_map__cpu;
+		perf_cpu_map__empty;
+		perf_cpu_map__max;
+		perf_cpu_map__has;
+		perf_thread_map__new_array;
+		perf_thread_map__new_dummy;
+		perf_thread_map__set_pid;
+		perf_thread_map__comm;
+		perf_thread_map__nr;
+		perf_thread_map__pid;
+		perf_thread_map__get;
+		perf_thread_map__put;
+		perf_evsel__new;
+		perf_evsel__delete;
+		perf_evsel__enable;
+		perf_evsel__disable;
+		perf_evsel__open;
+		perf_evsel__close;
+		perf_evsel__mmap;
+		perf_evsel__munmap;
+		perf_evsel__mmap_base;
+		perf_evsel__read;
+		perf_evsel__cpus;
+		perf_evsel__threads;
+		perf_evsel__attr;
+		perf_evlist__new;
+		perf_evlist__delete;
+		perf_evlist__open;
+		perf_evlist__close;
+		perf_evlist__enable;
+		perf_evlist__disable;
+		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_evlist__set_leader;
+		perf_mmap__consume;
+		perf_mmap__read_init;
+		perf_mmap__read_done;
+		perf_mmap__read_event;
+		perf_counts_values__scale;
+	local:
+		*;
+};
diff --git a/tools/lib/perf/libperf.pc.template b/tools/lib/perf/libperf.pc.template
new file mode 100644
index 000000000..117e4a237
--- /dev/null
+++ b/tools/lib/perf/libperf.pc.template
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: (LGPL-2.1 OR BSD-2-Clause)
+
+prefix=@PREFIX@
+libdir=@LIBDIR@
+includedir=${prefix}/include
+
+Name: libperf
+Description: perf library
+Version: @VERSION@
+Libs: -L${libdir} -lperf
+Cflags: -I${includedir}
diff --git a/tools/lib/perf/mmap.c b/tools/lib/perf/mmap.c
new file mode 100644
index 000000000..2184814b3
--- /dev/null
+++ b/tools/lib/perf/mmap.c
@@ -0,0 +1,525 @@
+// 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 <perf/evsel.h>
+#include <internal/mmap.h>
+#include <internal/lib.h>
+#include <linux/kernel.h>
+#include <linux/math64.h>
+#include <linux/stringify.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, struct perf_cpu 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 from it.
+		 */
+		if (overwrite_rb_find_range(data, md->mask, &md->start, &md->end))
+			return -EINVAL;
+	}
+
+	return 0;
+}
+
+int perf_mmap__read_init(struct perf_mmap *map)
+{
+	/*
+	 * Check if event was unmapped due to a POLLHUP/POLLERR.
+	 */
+	if (!refcount_read(&map->refcnt))
+		return -ENOENT;
+
+	return __perf_mmap__read_init(map);
+}
+
+/*
+ * Mandatory for overwrite mode
+ * The direction of overwrite mode is backward.
+ * The last perf_mmap__read() will set tail to map->core.prev.
+ * Need to correct the map->core.prev to head which is the end of next read.
+ */
+void perf_mmap__read_done(struct perf_mmap *map)
+{
+	/*
+	 * Check if event was unmapped due to a POLLHUP/POLLERR.
+	 */
+	if (!refcount_read(&map->refcnt))
+		return;
+
+	map->prev = perf_mmap__read_head(map);
+}
+
+/* When check_messup is true, 'end' must points to a good entry */
+static union perf_event *perf_mmap__read(struct perf_mmap *map,
+					 u64 *startp, u64 end)
+{
+	unsigned char *data = map->base + page_size;
+	union perf_event *event = NULL;
+	int diff = end - *startp;
+
+	if (diff >= (int)sizeof(event->header)) {
+		size_t size;
+
+		event = (union perf_event *)&data[*startp & map->mask];
+		size = event->header.size;
+
+		if (size < sizeof(event->header) || diff < (int)size)
+			return NULL;
+
+		/*
+		 * Event straddles the mmap boundary -- header should always
+		 * be inside due to u64 alignment of output.
+		 */
+		if ((*startp & map->mask) + size != ((*startp + size) & map->mask)) {
+			unsigned int offset = *startp;
+			unsigned int len = min(sizeof(*event), size), cpy;
+			void *dst = map->event_copy;
+
+			do {
+				cpy = min(map->mask + 1 - (offset & map->mask), len);
+				memcpy(dst, &data[offset & map->mask], cpy);
+				offset += cpy;
+				dst += cpy;
+				len -= cpy;
+			} while (len);
+
+			event = (union perf_event *)map->event_copy;
+		}
+
+		*startp += size;
+	}
+
+	return event;
+}
+
+/*
+ * Read event from ring buffer one by one.
+ * Return one event for each call.
+ *
+ * Usage:
+ * perf_mmap__read_init()
+ * while(event = perf_mmap__read_event()) {
+ *	//process the event
+ *	perf_mmap__consume()
+ * }
+ * perf_mmap__read_done()
+ */
+union perf_event *perf_mmap__read_event(struct perf_mmap *map)
+{
+	union perf_event *event;
+
+	/*
+	 * Check if event was unmapped due to a POLLHUP/POLLERR.
+	 */
+	if (!refcount_read(&map->refcnt))
+		return NULL;
+
+	/* non-overwirte doesn't pause the ringbuffer */
+	if (!map->overwrite)
+		map->end = perf_mmap__read_head(map);
+
+	event = perf_mmap__read(map, &map->start, map->end);
+
+	if (!map->overwrite)
+		map->prev = map->start;
+
+	return event;
+}
+
+#if defined(__i386__) || defined(__x86_64__)
+static u64 read_perf_counter(unsigned int counter)
+{
+	unsigned int low, high;
+
+	asm volatile("rdpmc" : "=a" (low), "=d" (high) : "c" (counter));
+
+	return low | ((u64)high) << 32;
+}
+
+static u64 read_timestamp(void)
+{
+	unsigned int low, high;
+
+	asm volatile("rdtsc" : "=a" (low), "=d" (high));
+
+	return low | ((u64)high) << 32;
+}
+#elif defined(__aarch64__)
+#define read_sysreg(r) ({						\
+	u64 __val;							\
+	asm volatile("mrs %0, " __stringify(r) : "=r" (__val));		\
+	__val;								\
+})
+
+static u64 read_pmccntr(void)
+{
+	return read_sysreg(pmccntr_el0);
+}
+
+#define PMEVCNTR_READ(idx)					\
+	static u64 read_pmevcntr_##idx(void) {			\
+		return read_sysreg(pmevcntr##idx##_el0);	\
+	}
+
+PMEVCNTR_READ(0);
+PMEVCNTR_READ(1);
+PMEVCNTR_READ(2);
+PMEVCNTR_READ(3);
+PMEVCNTR_READ(4);
+PMEVCNTR_READ(5);
+PMEVCNTR_READ(6);
+PMEVCNTR_READ(7);
+PMEVCNTR_READ(8);
+PMEVCNTR_READ(9);
+PMEVCNTR_READ(10);
+PMEVCNTR_READ(11);
+PMEVCNTR_READ(12);
+PMEVCNTR_READ(13);
+PMEVCNTR_READ(14);
+PMEVCNTR_READ(15);
+PMEVCNTR_READ(16);
+PMEVCNTR_READ(17);
+PMEVCNTR_READ(18);
+PMEVCNTR_READ(19);
+PMEVCNTR_READ(20);
+PMEVCNTR_READ(21);
+PMEVCNTR_READ(22);
+PMEVCNTR_READ(23);
+PMEVCNTR_READ(24);
+PMEVCNTR_READ(25);
+PMEVCNTR_READ(26);
+PMEVCNTR_READ(27);
+PMEVCNTR_READ(28);
+PMEVCNTR_READ(29);
+PMEVCNTR_READ(30);
+
+/*
+ * Read a value direct from PMEVCNTR<idx>
+ */
+static u64 read_perf_counter(unsigned int counter)
+{
+	static u64 (* const read_f[])(void) = {
+		read_pmevcntr_0,
+		read_pmevcntr_1,
+		read_pmevcntr_2,
+		read_pmevcntr_3,
+		read_pmevcntr_4,
+		read_pmevcntr_5,
+		read_pmevcntr_6,
+		read_pmevcntr_7,
+		read_pmevcntr_8,
+		read_pmevcntr_9,
+		read_pmevcntr_10,
+		read_pmevcntr_11,
+		read_pmevcntr_13,
+		read_pmevcntr_12,
+		read_pmevcntr_14,
+		read_pmevcntr_15,
+		read_pmevcntr_16,
+		read_pmevcntr_17,
+		read_pmevcntr_18,
+		read_pmevcntr_19,
+		read_pmevcntr_20,
+		read_pmevcntr_21,
+		read_pmevcntr_22,
+		read_pmevcntr_23,
+		read_pmevcntr_24,
+		read_pmevcntr_25,
+		read_pmevcntr_26,
+		read_pmevcntr_27,
+		read_pmevcntr_28,
+		read_pmevcntr_29,
+		read_pmevcntr_30,
+		read_pmccntr
+	};
+
+	if (counter < ARRAY_SIZE(read_f))
+		return (read_f[counter])();
+
+	return 0;
+}
+
+static u64 read_timestamp(void) { return read_sysreg(cntvct_el0); }
+
+/* __riscv_xlen contains the witdh of the native base integer, here 64-bit */
+#elif defined(__riscv) && __riscv_xlen == 64
+
+/* TODO: implement rv32 support */
+
+#define CSR_CYCLE	0xc00
+#define CSR_TIME	0xc01
+
+#define csr_read(csr)						\
+({								\
+	register unsigned long __v;				\
+		__asm__ __volatile__ ("csrr %0, %1"		\
+		 : "=r" (__v)					\
+		 : "i" (csr) : );				\
+		 __v;						\
+})
+
+static unsigned long csr_read_num(int csr_num)
+{
+#define switchcase_csr_read(__csr_num, __val)           {\
+	case __csr_num:                                 \
+		__val = csr_read(__csr_num);            \
+		break; }
+#define switchcase_csr_read_2(__csr_num, __val)         {\
+	switchcase_csr_read(__csr_num + 0, __val)        \
+	switchcase_csr_read(__csr_num + 1, __val)}
+#define switchcase_csr_read_4(__csr_num, __val)         {\
+	switchcase_csr_read_2(__csr_num + 0, __val)      \
+	switchcase_csr_read_2(__csr_num + 2, __val)}
+#define switchcase_csr_read_8(__csr_num, __val)         {\
+	switchcase_csr_read_4(__csr_num + 0, __val)      \
+	switchcase_csr_read_4(__csr_num + 4, __val)}
+#define switchcase_csr_read_16(__csr_num, __val)        {\
+	switchcase_csr_read_8(__csr_num + 0, __val)      \
+	switchcase_csr_read_8(__csr_num + 8, __val)}
+#define switchcase_csr_read_32(__csr_num, __val)        {\
+	switchcase_csr_read_16(__csr_num + 0, __val)     \
+	switchcase_csr_read_16(__csr_num + 16, __val)}
+
+	unsigned long ret = 0;
+
+	switch (csr_num) {
+	switchcase_csr_read_32(CSR_CYCLE, ret)
+	default:
+		break;
+	}
+
+	return ret;
+#undef switchcase_csr_read_32
+#undef switchcase_csr_read_16
+#undef switchcase_csr_read_8
+#undef switchcase_csr_read_4
+#undef switchcase_csr_read_2
+#undef switchcase_csr_read
+}
+
+static u64 read_perf_counter(unsigned int counter)
+{
+	return csr_read_num(CSR_CYCLE + counter);
+}
+
+static u64 read_timestamp(void)
+{
+	return csr_read_num(CSR_TIME);
+}
+
+#else
+static u64 read_perf_counter(unsigned int counter __maybe_unused) { return 0; }
+static u64 read_timestamp(void) { return 0; }
+#endif
+
+int perf_mmap__read_self(struct perf_mmap *map, struct perf_counts_values *count)
+{
+	struct perf_event_mmap_page *pc = map->base;
+	u32 seq, idx, time_mult = 0, time_shift = 0;
+	u64 cnt, cyc = 0, time_offset = 0, time_cycles = 0, time_mask = ~0ULL;
+
+	if (!pc || !pc->cap_user_rdpmc)
+		return -1;
+
+	do {
+		seq = READ_ONCE(pc->lock);
+		barrier();
+
+		count->ena = READ_ONCE(pc->time_enabled);
+		count->run = READ_ONCE(pc->time_running);
+
+		if (pc->cap_user_time && count->ena != count->run) {
+			cyc = read_timestamp();
+			time_mult = READ_ONCE(pc->time_mult);
+			time_shift = READ_ONCE(pc->time_shift);
+			time_offset = READ_ONCE(pc->time_offset);
+
+			if (pc->cap_user_time_short) {
+				time_cycles = READ_ONCE(pc->time_cycles);
+				time_mask = READ_ONCE(pc->time_mask);
+			}
+		}
+
+		idx = READ_ONCE(pc->index);
+		cnt = READ_ONCE(pc->offset);
+		if (pc->cap_user_rdpmc && idx) {
+			s64 evcnt = read_perf_counter(idx - 1);
+			u16 width = READ_ONCE(pc->pmc_width);
+
+			evcnt <<= 64 - width;
+			evcnt >>= 64 - width;
+			cnt += evcnt;
+		} else
+			return -1;
+
+		barrier();
+	} while (READ_ONCE(pc->lock) != seq);
+
+	if (count->ena != count->run) {
+		u64 delta;
+
+		/* Adjust for cap_usr_time_short, a nop if not */
+		cyc = time_cycles + ((cyc - time_cycles) & time_mask);
+
+		delta = time_offset + mul_u64_u32_shr(cyc, time_mult, time_shift);
+
+		count->ena += delta;
+		if (idx)
+			count->run += delta;
+	}
+
+	count->val = cnt;
+
+	return 0;
+}
diff --git a/tools/lib/perf/tests/Build b/tools/lib/perf/tests/Build
new file mode 100644
index 000000000..56e81378d
--- /dev/null
+++ b/tools/lib/perf/tests/Build
@@ -0,0 +1,5 @@
+tests-y += main.o
+tests-y += test-evsel.o
+tests-y += test-evlist.o
+tests-y += test-cpumap.o
+tests-y += test-threadmap.o
diff --git a/tools/lib/perf/tests/main.c b/tools/lib/perf/tests/main.c
new file mode 100644
index 000000000..56423fd4d
--- /dev/null
+++ b/tools/lib/perf/tests/main.c
@@ -0,0 +1,15 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <internal/tests.h>
+#include "tests.h"
+
+int tests_failed;
+int tests_verbose;
+
+int main(int argc, char **argv)
+{
+	__T("test cpumap", !test_cpumap(argc, argv));
+	__T("test threadmap", !test_threadmap(argc, argv));
+	__T("test evlist", !test_evlist(argc, argv));
+	__T("test evsel", !test_evsel(argc, argv));
+	return 0;
+}
diff --git a/tools/lib/perf/tests/test-cpumap.c b/tools/lib/perf/tests/test-cpumap.c
new file mode 100644
index 000000000..87b0510a5
--- /dev/null
+++ b/tools/lib/perf/tests/test-cpumap.c
@@ -0,0 +1,43 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <stdarg.h>
+#include <stdio.h>
+#include <perf/cpumap.h>
+#include <internal/tests.h>
+#include "tests.h"
+
+static int libperf_print(enum libperf_print_level level,
+			 const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+int test_cpumap(int argc, char **argv)
+{
+	struct perf_cpu_map *cpus;
+	struct perf_cpu cpu;
+	int idx;
+
+	__T_START;
+
+	libperf_init(libperf_print);
+
+	cpus = perf_cpu_map__dummy_new();
+	if (!cpus)
+		return -1;
+
+	perf_cpu_map__get(cpus);
+	perf_cpu_map__put(cpus);
+	perf_cpu_map__put(cpus);
+
+	cpus = perf_cpu_map__default_new();
+	if (!cpus)
+		return -1;
+
+	perf_cpu_map__for_each_cpu(cpu, idx, cpus)
+		__T("wrong cpu number", cpu.cpu != -1);
+
+	perf_cpu_map__put(cpus);
+
+	__T_END;
+	return tests_failed == 0 ? 0 : -1;
+}
diff --git a/tools/lib/perf/tests/test-evlist.c b/tools/lib/perf/tests/test-evlist.c
new file mode 100644
index 000000000..ed616fc19
--- /dev/null
+++ b/tools/lib/perf/tests/test-evlist.c
@@ -0,0 +1,589 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE // needed for sched.h to get sched_[gs]etaffinity and CPU_(ZERO,SET)
+#include <inttypes.h>
+#include <sched.h>
+#include <stdio.h>
+#include <stdarg.h>
+#include <unistd.h>
+#include <stdlib.h>
+#include <linux/perf_event.h>
+#include <linux/limits.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+#include <sys/prctl.h>
+#include <perf/cpumap.h>
+#include <perf/threadmap.h>
+#include <perf/evlist.h>
+#include <perf/evsel.h>
+#include <perf/mmap.h>
+#include <perf/event.h>
+#include <internal/tests.h>
+#include <api/fs/fs.h>
+#include "tests.h"
+#include <internal/evsel.h>
+
+#define EVENT_NUM 15
+#define WAIT_COUNT 100000000UL
+
+static int libperf_print(enum libperf_print_level level,
+			 const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+static int test_stat_cpu(void)
+{
+	struct perf_cpu_map *cpus;
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel, *leader;
+	struct perf_event_attr attr1 = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_CPU_CLOCK,
+	};
+	struct perf_event_attr attr2 = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_TASK_CLOCK,
+	};
+	int err, idx;
+
+	cpus = perf_cpu_map__new(NULL);
+	__T("failed to create cpus", cpus);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	evsel = leader = perf_evsel__new(&attr1);
+	__T("failed to create evsel1", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	evsel = perf_evsel__new(&attr2);
+	__T("failed to create evsel2", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_leader(evlist);
+	__T("failed to set leader", leader->leader == leader);
+	__T("failed to set leader", evsel->leader  == leader);
+
+	perf_evlist__set_maps(evlist, cpus, NULL);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		cpus = perf_evsel__cpus(evsel);
+
+		for (idx = 0; idx < perf_cpu_map__nr(cpus); idx++) {
+			struct perf_counts_values counts = { .val = 0 };
+
+			perf_evsel__read(evsel, idx, 0, &counts);
+			__T("failed to read value for evsel", counts.val != 0);
+		}
+	}
+
+	perf_evlist__close(evlist);
+	perf_evlist__delete(evlist);
+
+	perf_cpu_map__put(cpus);
+	return 0;
+}
+
+static int test_stat_thread(void)
+{
+	struct perf_counts_values counts = { .val = 0 };
+	struct perf_thread_map *threads;
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel, *leader;
+	struct perf_event_attr attr1 = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_CPU_CLOCK,
+	};
+	struct perf_event_attr attr2 = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_TASK_CLOCK,
+	};
+	int err;
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	evsel = leader = perf_evsel__new(&attr1);
+	__T("failed to create evsel1", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	evsel = perf_evsel__new(&attr2);
+	__T("failed to create evsel2", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_leader(evlist);
+	__T("failed to set leader", leader->leader == leader);
+	__T("failed to set leader", evsel->leader  == leader);
+
+	perf_evlist__set_maps(evlist, NULL, threads);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		perf_evsel__read(evsel, 0, 0, &counts);
+		__T("failed to read value for evsel", counts.val != 0);
+	}
+
+	perf_evlist__close(evlist);
+	perf_evlist__delete(evlist);
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+static int test_stat_thread_enable(void)
+{
+	struct perf_counts_values counts = { .val = 0 };
+	struct perf_thread_map *threads;
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel, *leader;
+	struct perf_event_attr attr1 = {
+		.type	  = PERF_TYPE_SOFTWARE,
+		.config	  = PERF_COUNT_SW_CPU_CLOCK,
+		.disabled = 1,
+	};
+	struct perf_event_attr attr2 = {
+		.type	  = PERF_TYPE_SOFTWARE,
+		.config	  = PERF_COUNT_SW_TASK_CLOCK,
+		.disabled = 1,
+	};
+	int err;
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	evsel = leader = perf_evsel__new(&attr1);
+	__T("failed to create evsel1", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	evsel = perf_evsel__new(&attr2);
+	__T("failed to create evsel2", evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_leader(evlist);
+	__T("failed to set leader", leader->leader == leader);
+	__T("failed to set leader", evsel->leader  == leader);
+
+	perf_evlist__set_maps(evlist, NULL, threads);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		perf_evsel__read(evsel, 0, 0, &counts);
+		__T("failed to read value for evsel", counts.val == 0);
+	}
+
+	perf_evlist__enable(evlist);
+
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		perf_evsel__read(evsel, 0, 0, &counts);
+		__T("failed to read value for evsel", counts.val != 0);
+	}
+
+	perf_evlist__disable(evlist);
+
+	perf_evlist__close(evlist);
+	perf_evlist__delete(evlist);
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+static int test_mmap_thread(void)
+{
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel;
+	struct perf_mmap *map;
+	struct perf_cpu_map *cpus;
+	struct perf_thread_map *threads;
+	struct perf_event_attr attr = {
+		.type             = PERF_TYPE_TRACEPOINT,
+		.sample_period    = 1,
+		.wakeup_watermark = 1,
+		.disabled         = 1,
+	};
+	char path[PATH_MAX];
+	int id, err, pid, go_pipe[2];
+	union perf_event *event;
+	int count = 0;
+
+	snprintf(path, PATH_MAX, "%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id",
+		 sysfs__mountpoint());
+
+	if (filename__read_int(path, &id)) {
+		tests_failed++;
+		fprintf(stderr, "error: failed to get tracepoint id: %s\n", path);
+		return -1;
+	}
+
+	attr.config = id;
+
+	err = pipe(go_pipe);
+	__T("failed to create pipe", err == 0);
+
+	fflush(NULL);
+
+	pid = fork();
+	if (!pid) {
+		int i;
+		char bf;
+
+		read(go_pipe[0], &bf, 1);
+
+		/* Generate 100 prctl calls. */
+		for (i = 0; i < 100; i++)
+			prctl(0, 0, 0, 0, 0);
+
+		exit(0);
+	}
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	cpus = perf_cpu_map__dummy_new();
+	__T("failed to create cpus", cpus);
+
+	perf_thread_map__set_pid(threads, 0, pid);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel1", evsel);
+	__T("failed to set leader", evsel->leader == evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_maps(evlist, cpus, threads);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	err = perf_evlist__mmap(evlist, 4);
+	__T("failed to mmap evlist", err == 0);
+
+	perf_evlist__enable(evlist);
+
+	/* kick the child and wait for it to finish */
+	write(go_pipe[1], "A", 1);
+	waitpid(pid, NULL, 0);
+
+	/*
+	 * There's no need to call perf_evlist__disable,
+	 * monitored process is dead now.
+	 */
+
+	perf_evlist__for_each_mmap(evlist, map, false) {
+		if (perf_mmap__read_init(map) < 0)
+			continue;
+
+		while ((event = perf_mmap__read_event(map)) != NULL) {
+			count++;
+			perf_mmap__consume(map);
+		}
+
+		perf_mmap__read_done(map);
+	}
+
+	/* calls perf_evlist__munmap/perf_evlist__close */
+	perf_evlist__delete(evlist);
+
+	perf_thread_map__put(threads);
+	perf_cpu_map__put(cpus);
+
+	/*
+	 * The generated prctl calls should match the
+	 * number of events in the buffer.
+	 */
+	__T("failed count", count == 100);
+
+	return 0;
+}
+
+static int test_mmap_cpus(void)
+{
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel;
+	struct perf_mmap *map;
+	struct perf_cpu_map *cpus;
+	struct perf_event_attr attr = {
+		.type             = PERF_TYPE_TRACEPOINT,
+		.sample_period    = 1,
+		.wakeup_watermark = 1,
+		.disabled         = 1,
+	};
+	cpu_set_t saved_mask;
+	char path[PATH_MAX];
+	int id, err, tmp;
+	struct perf_cpu cpu;
+	union perf_event *event;
+	int count = 0;
+
+	snprintf(path, PATH_MAX, "%s/kernel/debug/tracing/events/syscalls/sys_enter_prctl/id",
+		 sysfs__mountpoint());
+
+	if (filename__read_int(path, &id)) {
+		fprintf(stderr, "error: failed to get tracepoint id: %s\n", path);
+		return -1;
+	}
+
+	attr.config = id;
+
+	cpus = perf_cpu_map__new(NULL);
+	__T("failed to create cpus", cpus);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel1", evsel);
+	__T("failed to set leader", evsel->leader == evsel);
+
+	perf_evlist__add(evlist, evsel);
+
+	perf_evlist__set_maps(evlist, cpus, NULL);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	err = perf_evlist__mmap(evlist, 4);
+	__T("failed to mmap evlist", err == 0);
+
+	perf_evlist__enable(evlist);
+
+	err = sched_getaffinity(0, sizeof(saved_mask), &saved_mask);
+	__T("sched_getaffinity failed", err == 0);
+
+	perf_cpu_map__for_each_cpu(cpu, tmp, cpus) {
+		cpu_set_t mask;
+
+		CPU_ZERO(&mask);
+		CPU_SET(cpu.cpu, &mask);
+
+		err = sched_setaffinity(0, sizeof(mask), &mask);
+		__T("sched_setaffinity failed", err == 0);
+
+		prctl(0, 0, 0, 0, 0);
+	}
+
+	err = sched_setaffinity(0, sizeof(saved_mask), &saved_mask);
+	__T("sched_setaffinity failed", err == 0);
+
+	perf_evlist__disable(evlist);
+
+	perf_evlist__for_each_mmap(evlist, map, false) {
+		if (perf_mmap__read_init(map) < 0)
+			continue;
+
+		while ((event = perf_mmap__read_event(map)) != NULL) {
+			count++;
+			perf_mmap__consume(map);
+		}
+
+		perf_mmap__read_done(map);
+	}
+
+	/* calls perf_evlist__munmap/perf_evlist__close */
+	perf_evlist__delete(evlist);
+
+	/*
+	 * The generated prctl events should match the
+	 * number of cpus or be bigger (we are system-wide).
+	 */
+	__T("failed count", count >= perf_cpu_map__nr(cpus));
+
+	perf_cpu_map__put(cpus);
+
+	return 0;
+}
+
+static double display_error(long long average,
+			    long long high,
+			    long long low,
+			    long long expected)
+{
+	double error;
+
+	error = (((double)average - expected) / expected) * 100.0;
+
+	__T_VERBOSE("   Expected: %lld\n", expected);
+	__T_VERBOSE("   High: %lld   Low:  %lld   Average:  %lld\n",
+		    high, low, average);
+
+	__T_VERBOSE("   Average Error = %.2f%%\n", error);
+
+	return error;
+}
+
+static int test_stat_multiplexing(void)
+{
+	struct perf_counts_values expected_counts = { .val = 0 };
+	struct perf_counts_values counts[EVENT_NUM] = {{ .val = 0 },};
+	struct perf_thread_map *threads;
+	struct perf_evlist *evlist;
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr = {
+		.type	     = PERF_TYPE_HARDWARE,
+		.config	     = PERF_COUNT_HW_INSTRUCTIONS,
+		.read_format = PERF_FORMAT_TOTAL_TIME_ENABLED |
+			       PERF_FORMAT_TOTAL_TIME_RUNNING,
+		.disabled    = 1,
+	};
+	int err, i, nonzero = 0;
+	unsigned long count;
+	long long max = 0, min = 0, avg = 0;
+	double error = 0.0;
+	s8 scaled = 0;
+
+	/* read for non-multiplexing event count */
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel", evsel);
+
+	err = perf_evsel__open(evsel, NULL, threads);
+	__T("failed to open evsel", err == 0);
+
+	err = perf_evsel__enable(evsel);
+	__T("failed to enable evsel", err == 0);
+
+	/* wait loop */
+	count = WAIT_COUNT;
+	while (count--)
+		;
+
+	perf_evsel__read(evsel, 0, 0, &expected_counts);
+	__T("failed to read value for evsel", expected_counts.val != 0);
+	__T("failed to read non-multiplexing event count",
+	    expected_counts.ena == expected_counts.run);
+
+	err = perf_evsel__disable(evsel);
+	__T("failed to enable evsel", err == 0);
+
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+
+	perf_thread_map__put(threads);
+
+	/* read for multiplexing event count */
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evlist = perf_evlist__new();
+	__T("failed to create evlist", evlist);
+
+	for (i = 0; i < EVENT_NUM; i++) {
+		evsel = perf_evsel__new(&attr);
+		__T("failed to create evsel", evsel);
+
+		perf_evlist__add(evlist, evsel);
+	}
+	perf_evlist__set_maps(evlist, NULL, threads);
+
+	err = perf_evlist__open(evlist);
+	__T("failed to open evlist", err == 0);
+
+	perf_evlist__enable(evlist);
+
+	/* wait loop */
+	count = WAIT_COUNT;
+	while (count--)
+		;
+
+	i = 0;
+	perf_evlist__for_each_evsel(evlist, evsel) {
+		perf_evsel__read(evsel, 0, 0, &counts[i]);
+		__T("failed to read value for evsel", counts[i].val != 0);
+		i++;
+	}
+
+	perf_evlist__disable(evlist);
+
+	min = counts[0].val;
+	for (i = 0; i < EVENT_NUM; i++) {
+		__T_VERBOSE("Event %2d -- Raw count = %" PRIu64 ", run = %" PRIu64 ", enable = %" PRIu64 "\n",
+			    i, counts[i].val, counts[i].run, counts[i].ena);
+
+		perf_counts_values__scale(&counts[i], true, &scaled);
+		if (scaled == 1) {
+			__T_VERBOSE("\t Scaled count = %" PRIu64 " (%.2lf%%, %" PRIu64 "/%" PRIu64 ")\n",
+				    counts[i].val,
+				    (double)counts[i].run / (double)counts[i].ena * 100.0,
+				    counts[i].run, counts[i].ena);
+		} else if (scaled == -1) {
+			__T_VERBOSE("\t Not Running\n");
+		} else {
+			__T_VERBOSE("\t Not Scaling\n");
+		}
+
+		if (counts[i].val > max)
+			max = counts[i].val;
+
+		if (counts[i].val < min)
+			min = counts[i].val;
+
+		avg += counts[i].val;
+
+		if (counts[i].val != 0)
+			nonzero++;
+	}
+
+	if (nonzero != 0)
+		avg = avg / nonzero;
+	else
+		avg = 0;
+
+	error = display_error(avg, max, min, expected_counts.val);
+
+	__T("Error out of range!", ((error <= 1.0) && (error >= -1.0)));
+
+	perf_evlist__close(evlist);
+	perf_evlist__delete(evlist);
+
+	perf_thread_map__put(threads);
+
+	return 0;
+}
+
+int test_evlist(int argc, char **argv)
+{
+	__T_START;
+
+	libperf_init(libperf_print);
+
+	test_stat_cpu();
+	test_stat_thread();
+	test_stat_thread_enable();
+	test_mmap_thread();
+	test_mmap_cpus();
+	test_stat_multiplexing();
+
+	__T_END;
+	return tests_failed == 0 ? 0 : -1;
+}
diff --git a/tools/lib/perf/tests/test-evsel.c b/tools/lib/perf/tests/test-evsel.c
new file mode 100644
index 000000000..a11fc51bf
--- /dev/null
+++ b/tools/lib/perf/tests/test-evsel.c
@@ -0,0 +1,367 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <stdarg.h>
+#include <stdio.h>
+#include <string.h>
+#include <linux/perf_event.h>
+#include <linux/kernel.h>
+#include <perf/cpumap.h>
+#include <perf/threadmap.h>
+#include <perf/evsel.h>
+#include <internal/evsel.h>
+#include <internal/tests.h>
+#include "tests.h"
+
+static int libperf_print(enum libperf_print_level level,
+			 const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+static int test_stat_cpu(void)
+{
+	struct perf_cpu_map *cpus;
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_CPU_CLOCK,
+	};
+	int err, idx;
+
+	cpus = perf_cpu_map__new(NULL);
+	__T("failed to create cpus", cpus);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel", evsel);
+
+	err = perf_evsel__open(evsel, cpus, NULL);
+	__T("failed to open evsel", err == 0);
+
+	for (idx = 0; idx < perf_cpu_map__nr(cpus); idx++) {
+		struct perf_counts_values counts = { .val = 0 };
+
+		perf_evsel__read(evsel, idx, 0, &counts);
+		__T("failed to read value for evsel", counts.val != 0);
+	}
+
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+
+	perf_cpu_map__put(cpus);
+	return 0;
+}
+
+static int test_stat_thread(void)
+{
+	struct perf_counts_values counts = { .val = 0 };
+	struct perf_thread_map *threads;
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_TASK_CLOCK,
+	};
+	int err;
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel", evsel);
+
+	err = perf_evsel__open(evsel, NULL, threads);
+	__T("failed to open evsel", err == 0);
+
+	perf_evsel__read(evsel, 0, 0, &counts);
+	__T("failed to read value for evsel", counts.val != 0);
+
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+static int test_stat_thread_enable(void)
+{
+	struct perf_counts_values counts = { .val = 0 };
+	struct perf_thread_map *threads;
+	struct perf_evsel *evsel;
+	struct perf_event_attr attr = {
+		.type	  = PERF_TYPE_SOFTWARE,
+		.config	  = PERF_COUNT_SW_TASK_CLOCK,
+		.disabled = 1,
+	};
+	int err;
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel", evsel);
+
+	err = perf_evsel__open(evsel, NULL, threads);
+	__T("failed to open evsel", err == 0);
+
+	perf_evsel__read(evsel, 0, 0, &counts);
+	__T("failed to read value for evsel", counts.val == 0);
+
+	err = perf_evsel__enable(evsel);
+	__T("failed to enable evsel", err == 0);
+
+	perf_evsel__read(evsel, 0, 0, &counts);
+	__T("failed to read value for evsel", counts.val != 0);
+
+	err = perf_evsel__disable(evsel);
+	__T("failed to enable evsel", err == 0);
+
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+static int test_stat_user_read(int event)
+{
+	struct perf_counts_values counts = { .val = 0 };
+	struct perf_thread_map *threads;
+	struct perf_evsel *evsel;
+	struct perf_event_mmap_page *pc;
+	struct perf_event_attr attr = {
+		.type	= PERF_TYPE_HARDWARE,
+		.config	= event,
+#ifdef __aarch64__
+		.config1 = 0x2,		/* Request user access */
+#endif
+	};
+	int err, i;
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	evsel = perf_evsel__new(&attr);
+	__T("failed to create evsel", evsel);
+
+	err = perf_evsel__open(evsel, NULL, threads);
+	__T("failed to open evsel", err == 0);
+
+	err = perf_evsel__mmap(evsel, 0);
+	__T("failed to mmap evsel", err == 0);
+
+	pc = perf_evsel__mmap_base(evsel, 0, 0);
+	__T("failed to get mmapped address", pc);
+
+#if defined(__i386__) || defined(__x86_64__) || defined(__aarch64__)
+	__T("userspace counter access not supported", pc->cap_user_rdpmc);
+	__T("userspace counter access not enabled", pc->index);
+	__T("userspace counter width not set", pc->pmc_width >= 32);
+#endif
+
+	perf_evsel__read(evsel, 0, 0, &counts);
+	__T("failed to read value for evsel", counts.val != 0);
+
+	for (i = 0; i < 5; i++) {
+		volatile int count = 0x10000 << i;
+		__u64 start, end, last = 0;
+
+		__T_VERBOSE("\tloop = %u, ", count);
+
+		perf_evsel__read(evsel, 0, 0, &counts);
+		start = counts.val;
+
+		while (count--) ;
+
+		perf_evsel__read(evsel, 0, 0, &counts);
+		end = counts.val;
+
+		__T("invalid counter data", (end - start) > last);
+		last = end - start;
+		__T_VERBOSE("count = %llu\n", end - start);
+	}
+
+	perf_evsel__munmap(evsel);
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+static int test_stat_read_format_single(struct perf_event_attr *attr, struct perf_thread_map *threads)
+{
+	struct perf_evsel *evsel;
+	struct perf_counts_values counts;
+	volatile int count = 0x100000;
+	int err;
+
+	evsel = perf_evsel__new(attr);
+	__T("failed to create evsel", evsel);
+
+	/* skip old kernels that don't support the format */
+	err = perf_evsel__open(evsel, NULL, threads);
+	if (err < 0)
+		return 0;
+
+	while (count--) ;
+
+	memset(&counts, -1, sizeof(counts));
+	perf_evsel__read(evsel, 0, 0, &counts);
+
+	__T("failed to read value", counts.val);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		__T("failed to read TOTAL_TIME_ENABLED", counts.ena);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		__T("failed to read TOTAL_TIME_RUNNING", counts.run);
+	if (attr->read_format & PERF_FORMAT_ID)
+		__T("failed to read ID", counts.id);
+	if (attr->read_format & PERF_FORMAT_LOST)
+		__T("failed to read LOST", counts.lost == 0);
+
+	perf_evsel__close(evsel);
+	perf_evsel__delete(evsel);
+	return 0;
+}
+
+static int test_stat_read_format_group(struct perf_event_attr *attr, struct perf_thread_map *threads)
+{
+	struct perf_evsel *leader, *member;
+	struct perf_counts_values counts;
+	volatile int count = 0x100000;
+	int err;
+
+	attr->read_format |= PERF_FORMAT_GROUP;
+	leader = perf_evsel__new(attr);
+	__T("failed to create leader", leader);
+
+	attr->read_format &= ~PERF_FORMAT_GROUP;
+	member = perf_evsel__new(attr);
+	__T("failed to create member", member);
+
+	member->leader = leader;
+	leader->nr_members = 2;
+
+	/* skip old kernels that don't support the format */
+	err = perf_evsel__open(leader, NULL, threads);
+	if (err < 0)
+		return 0;
+	err = perf_evsel__open(member, NULL, threads);
+	if (err < 0)
+		return 0;
+
+	while (count--) ;
+
+	memset(&counts, -1, sizeof(counts));
+	perf_evsel__read(leader, 0, 0, &counts);
+
+	__T("failed to read leader value", counts.val);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		__T("failed to read leader TOTAL_TIME_ENABLED", counts.ena);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		__T("failed to read leader TOTAL_TIME_RUNNING", counts.run);
+	if (attr->read_format & PERF_FORMAT_ID)
+		__T("failed to read leader ID", counts.id);
+	if (attr->read_format & PERF_FORMAT_LOST)
+		__T("failed to read leader LOST", counts.lost == 0);
+
+	memset(&counts, -1, sizeof(counts));
+	perf_evsel__read(member, 0, 0, &counts);
+
+	__T("failed to read member value", counts.val);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_ENABLED)
+		__T("failed to read member TOTAL_TIME_ENABLED", counts.ena);
+	if (attr->read_format & PERF_FORMAT_TOTAL_TIME_RUNNING)
+		__T("failed to read member TOTAL_TIME_RUNNING", counts.run);
+	if (attr->read_format & PERF_FORMAT_ID)
+		__T("failed to read member ID", counts.id);
+	if (attr->read_format & PERF_FORMAT_LOST)
+		__T("failed to read member LOST", counts.lost == 0);
+
+	perf_evsel__close(member);
+	perf_evsel__close(leader);
+	perf_evsel__delete(member);
+	perf_evsel__delete(leader);
+	return 0;
+}
+
+static int test_stat_read_format(void)
+{
+	struct perf_thread_map *threads;
+	struct perf_event_attr attr = {
+		.type	= PERF_TYPE_SOFTWARE,
+		.config	= PERF_COUNT_SW_TASK_CLOCK,
+	};
+	int err, i;
+
+#define FMT(_fmt)  PERF_FORMAT_ ## _fmt
+#define FMT_TIME  (FMT(TOTAL_TIME_ENABLED) | FMT(TOTAL_TIME_RUNNING))
+
+	uint64_t test_formats [] = {
+		0,
+		FMT_TIME,
+		FMT(ID),
+		FMT(LOST),
+		FMT_TIME | FMT(ID),
+		FMT_TIME | FMT(LOST),
+		FMT_TIME | FMT(ID) | FMT(LOST),
+		FMT(ID) | FMT(LOST),
+	};
+
+#undef FMT
+#undef FMT_TIME
+
+	threads = perf_thread_map__new_dummy();
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+
+	for (i = 0; i < (int)ARRAY_SIZE(test_formats); i++) {
+		attr.read_format = test_formats[i];
+		__T_VERBOSE("testing single read with read_format: %lx\n",
+			    (unsigned long)test_formats[i]);
+
+		err = test_stat_read_format_single(&attr, threads);
+		__T("failed to read single format", err == 0);
+	}
+
+	perf_thread_map__put(threads);
+
+	threads = perf_thread_map__new_array(2, NULL);
+	__T("failed to create threads", threads);
+
+	perf_thread_map__set_pid(threads, 0, 0);
+	perf_thread_map__set_pid(threads, 1, 0);
+
+	for (i = 0; i < (int)ARRAY_SIZE(test_formats); i++) {
+		attr.read_format = test_formats[i];
+		__T_VERBOSE("testing group read with read_format: %lx\n",
+			    (unsigned long)test_formats[i]);
+
+		err = test_stat_read_format_group(&attr, threads);
+		__T("failed to read group format", err == 0);
+	}
+
+	perf_thread_map__put(threads);
+	return 0;
+}
+
+int test_evsel(int argc, char **argv)
+{
+	__T_START;
+
+	libperf_init(libperf_print);
+
+	test_stat_cpu();
+	test_stat_thread();
+	test_stat_thread_enable();
+	test_stat_user_read(PERF_COUNT_HW_INSTRUCTIONS);
+	test_stat_user_read(PERF_COUNT_HW_CPU_CYCLES);
+	test_stat_read_format();
+
+	__T_END;
+	return tests_failed == 0 ? 0 : -1;
+}
diff --git a/tools/lib/perf/tests/test-threadmap.c b/tools/lib/perf/tests/test-threadmap.c
new file mode 100644
index 000000000..f728ad700
--- /dev/null
+++ b/tools/lib/perf/tests/test-threadmap.c
@@ -0,0 +1,73 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <stdarg.h>
+#include <stdio.h>
+#include <perf/threadmap.h>
+#include <internal/tests.h>
+#include "tests.h"
+
+static int libperf_print(enum libperf_print_level level,
+			 const char *fmt, va_list ap)
+{
+	return vfprintf(stderr, fmt, ap);
+}
+
+static int test_threadmap_array(int nr, pid_t *array)
+{
+	struct perf_thread_map *threads;
+	int i;
+
+	threads = perf_thread_map__new_array(nr, array);
+	__T("Failed to allocate new thread map", threads);
+
+	__T("Unexpected number of threads", perf_thread_map__nr(threads) == nr);
+
+	for (i = 0; i < nr; i++) {
+		__T("Unexpected initial value of thread",
+		    perf_thread_map__pid(threads, i) == (array ? array[i] : -1));
+	}
+
+	for (i = 1; i < nr; i++)
+		perf_thread_map__set_pid(threads, i, i * 100);
+
+	__T("Unexpected value of thread 0",
+	    perf_thread_map__pid(threads, 0) == (array ? array[0] : -1));
+
+	for (i = 1; i < nr; i++) {
+		__T("Unexpected thread value",
+		    perf_thread_map__pid(threads, i) == i * 100);
+	}
+
+	perf_thread_map__put(threads);
+
+	return 0;
+}
+
+#define THREADS_NR	10
+int test_threadmap(int argc, char **argv)
+{
+	struct perf_thread_map *threads;
+	pid_t thr_array[THREADS_NR];
+	int i;
+
+	__T_START;
+
+	libperf_init(libperf_print);
+
+	threads = perf_thread_map__new_dummy();
+	if (!threads)
+		return -1;
+
+	perf_thread_map__get(threads);
+	perf_thread_map__put(threads);
+	perf_thread_map__put(threads);
+
+	test_threadmap_array(THREADS_NR, NULL);
+
+	for (i = 0; i < THREADS_NR; i++)
+		thr_array[i] = i + 100;
+
+	test_threadmap_array(THREADS_NR, thr_array);
+
+	__T_END;
+	return tests_failed == 0 ? 0 : -1;
+}
diff --git a/tools/lib/perf/tests/tests.h b/tools/lib/perf/tests/tests.h
new file mode 100644
index 000000000..604838f21
--- /dev/null
+++ b/tools/lib/perf/tests/tests.h
@@ -0,0 +1,10 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef TESTS_H
+#define TESTS_H
+
+int test_cpumap(int argc, char **argv);
+int test_threadmap(int argc, char **argv);
+int test_evlist(int argc, char **argv);
+int test_evsel(int argc, char **argv);
+
+#endif /* TESTS_H */
diff --git a/tools/lib/perf/threadmap.c b/tools/lib/perf/threadmap.c
new file mode 100644
index 000000000..07968f3ea
--- /dev/null
+++ b/tools/lib/perf/threadmap.c
@@ -0,0 +1,101 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <perf/threadmap.h>
+#include <stdlib.h>
+#include <linux/refcount.h>
+#include <internal/threadmap.h>
+#include <string.h>
+#include <asm/bug.h>
+#include <stdio.h>
+
+static void perf_thread_map__reset(struct perf_thread_map *map, int start, int nr)
+{
+	size_t size = (nr - start) * sizeof(map->map[0]);
+
+	memset(&map->map[start], 0, size);
+	map->err_thread = -1;
+}
+
+struct perf_thread_map *perf_thread_map__realloc(struct perf_thread_map *map, int nr)
+{
+	size_t size = sizeof(*map) + sizeof(map->map[0]) * nr;
+	int start = map ? map->nr : 0;
+
+	map = realloc(map, size);
+	/*
+	 * We only realloc to add more items, let's reset new items.
+	 */
+	if (map)
+		perf_thread_map__reset(map, start, nr);
+
+	return map;
+}
+
+#define thread_map__alloc(__nr) perf_thread_map__realloc(NULL, __nr)
+
+void perf_thread_map__set_pid(struct perf_thread_map *map, int idx, pid_t pid)
+{
+	map->map[idx].pid = pid;
+}
+
+char *perf_thread_map__comm(struct perf_thread_map *map, int idx)
+{
+	return map->map[idx].comm;
+}
+
+struct perf_thread_map *perf_thread_map__new_array(int nr_threads, pid_t *array)
+{
+	struct perf_thread_map *threads = thread_map__alloc(nr_threads);
+	int i;
+
+	if (!threads)
+		return NULL;
+
+	for (i = 0; i < nr_threads; i++)
+		perf_thread_map__set_pid(threads, i, array ? array[i] : -1);
+
+	threads->nr = nr_threads;
+	refcount_set(&threads->refcnt, 1);
+
+	return threads;
+}
+
+struct perf_thread_map *perf_thread_map__new_dummy(void)
+{
+	return perf_thread_map__new_array(1, NULL);
+}
+
+static void perf_thread_map__delete(struct perf_thread_map *threads)
+{
+	if (threads) {
+		int i;
+
+		WARN_ONCE(refcount_read(&threads->refcnt) != 0,
+			  "thread map refcnt unbalanced\n");
+		for (i = 0; i < threads->nr; i++)
+			free(perf_thread_map__comm(threads, i));
+		free(threads);
+	}
+}
+
+struct perf_thread_map *perf_thread_map__get(struct perf_thread_map *map)
+{
+	if (map)
+		refcount_inc(&map->refcnt);
+	return map;
+}
+
+void perf_thread_map__put(struct perf_thread_map *map)
+{
+	if (map && refcount_dec_and_test(&map->refcnt))
+		perf_thread_map__delete(map);
+}
+
+int perf_thread_map__nr(struct perf_thread_map *threads)
+{
+	return threads ? threads->nr : 1;
+}
+
+pid_t perf_thread_map__pid(struct perf_thread_map *map, int idx)
+{
+	return map->map[idx].pid;
+}
diff --git a/tools/lib/perf/xyarray.c b/tools/lib/perf/xyarray.c
new file mode 100644
index 000000000..dcd901d15
--- /dev/null
+++ b/tools/lib/perf/xyarray.c
@@ -0,0 +1,33 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <internal/xyarray.h>
+#include <linux/zalloc.h>
+#include <stdlib.h>
+#include <string.h>
+
+struct xyarray *xyarray__new(int xlen, int ylen, size_t entry_size)
+{
+	size_t row_size = ylen * entry_size;
+	struct xyarray *xy = zalloc(sizeof(*xy) + xlen * row_size);
+
+	if (xy != NULL) {
+		xy->entry_size = entry_size;
+		xy->row_size   = row_size;
+		xy->entries    = xlen * ylen;
+		xy->max_x      = xlen;
+		xy->max_y      = ylen;
+	}
+
+	return xy;
+}
+
+void xyarray__reset(struct xyarray *xy)
+{
+	size_t n = xy->entries * xy->entry_size;
+
+	memset(xy->contents, 0, n);
+}
+
+void xyarray__delete(struct xyarray *xy)
+{
+	free(xy);
+}