Merge tag 'perf-core-for-mingo-4.11-20170306' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core

Pull perf/core improvements and fixes from Arnaldo Carvalho de Melo:

New features:

- Allow sorting by symbol_size in 'perf report' and 'perf top' (Charles Baylis)

  E.g.:

  # perf report -s symbol_size,symbol

  Samples: 9K of event 'cycles:k', Event count (approx.): 2870461623
  Overhead  Symbol size  Symbol
    14.55%          326  [k] flush_tlb_mm_range
     7.20%         1045  [k] filemap_map_pages
     5.82%          124  [k] vma_interval_tree_insert
     5.18%         2430  [k] unmap_page_range
     2.57%          571  [k] vma_interval_tree_remove
     1.94%          494  [k] page_add_file_rmap
     1.82%          740  [k] page_remove_rmap
     1.66%         1017  [k] release_pages
     1.57%         1636  [k] update_blocked_averages
     1.57%           76  [k] unlock_page

- Add support for -p/--pid, -a/--all-cpus and -C/--cpu in 'perf ftrace' (Namhyung Kim)

Change in behaviour:

- Make system wide (-a) the default option if no target was specified and one
  of following conditions is met:

  - No workload specified (current behaviour)

  - A workload is specified but all requested events are system wide ones,
    like uncore ones. (Jiri Olsa)

Fixes:

- Add missing initialization to the instruction decoder used in the
  intel PT/BTS code, which was causing lots of failures in 'perf test',
  looking for a value when there was none (Adrian Hunter)

Infrastructure changes:

- Add arch code needed to adopt the kernel's refcount_t to aid in
  catching bugs when using atomic_t as a reference counter, basically
  cmpxchg related functions (Arnaldo Carvalho de Melo)

- Convert the code using atomic_t as reference counts to refcount_t
  (Elena Rashetova)

- Add feature test for sched_getcpu() to more easily check for its
  presence in the many libc implementations and accross different
  versions of such C libraries (Arnaldo Carvalho de Melo)

- Issue a HW watchdog disable hint in 'perf stat' for when some of the
  requested events can't get counted because a PMU counter is taken by that
  watchdog (Borislav Petkov).

- Add mapping for Intel's KnightsMill PMU events (Karol Wachowski)

Documentation changes:

- Clarify the term 'convergence' in:

   perf bench numa numa-mem -h --show_convergence (Jiri Olsa)

Kernel code changes:

- Ensure probe location is at function entry in kretprobes (Naveen N. Rao)

- Allow return probes with offsets and absolute addresses (Naveen N. Rao)

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Ingo Molnar <mingo@kernel.org>

71 files changed, 743 insertions, 166 deletions

diff --git a/Documentation/trace/kprobetrace.txt b/Documentation/trace/kprobetrace.txt
index e4991fb1eedc..41ef9d8efe95 100644
--- a/Documentation/trace/kprobetrace.txt
+++ b/Documentation/trace/kprobetrace.txt
@@ -12,7 +12,7 @@ kprobes can probe (this means, all functions body except for __kprobes
 functions). Unlike the Tracepoint based event, this can be added and removed
 dynamically, on the fly.
 
-To enable this feature, build your kernel with CONFIG_KPROBE_EVENT=y.
+To enable this feature, build your kernel with CONFIG_KPROBE_EVENTS=y.
 
 Similar to the events tracer, this doesn't need to be activated via
 current_tracer. Instead of that, add probe points via
diff --git a/Documentation/trace/uprobetracer.txt b/Documentation/trace/uprobetracer.txt
index fa7b680ee8a0..bf526a7c5559 100644
--- a/Documentation/trace/uprobetracer.txt
+++ b/Documentation/trace/uprobetracer.txt
@@ -7,7 +7,7 @@
 Overview
 --------
 Uprobe based trace events are similar to kprobe based trace events.
-To enable this feature, build your kernel with CONFIG_UPROBE_EVENT=y.
+To enable this feature, build your kernel with CONFIG_UPROBE_EVENTS=y.
 
 Similar to the kprobe-event tracer, this doesn't need to be activated via
 current_tracer. Instead of that, add probe points via
diff --git a/arch/s390/configs/default_defconfig b/arch/s390/configs/default_defconfig
index 143b1e00b818..4b176fe83da4 100644
--- a/arch/s390/configs/default_defconfig
+++ b/arch/s390/configs/default_defconfig
@@ -609,7 +609,7 @@ CONFIG_SCHED_TRACER=y
 CONFIG_FTRACE_SYSCALLS=y
 CONFIG_STACK_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
-CONFIG_UPROBE_EVENT=y
+CONFIG_UPROBE_EVENTS=y
 CONFIG_FUNCTION_PROFILER=y
 CONFIG_HIST_TRIGGERS=y
 CONFIG_TRACE_ENUM_MAP_FILE=y
diff --git a/arch/s390/configs/gcov_defconfig b/arch/s390/configs/gcov_defconfig
index f05d2d6e1087..0de46cc397f6 100644
--- a/arch/s390/configs/gcov_defconfig
+++ b/arch/s390/configs/gcov_defconfig
@@ -560,7 +560,7 @@ CONFIG_SCHED_TRACER=y
 CONFIG_FTRACE_SYSCALLS=y
 CONFIG_STACK_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
-CONFIG_UPROBE_EVENT=y
+CONFIG_UPROBE_EVENTS=y
 CONFIG_FUNCTION_PROFILER=y
 CONFIG_HIST_TRIGGERS=y
 CONFIG_TRACE_ENUM_MAP_FILE=y
diff --git a/arch/s390/configs/performance_defconfig b/arch/s390/configs/performance_defconfig
index 2358bf33c5ef..e167557b434c 100644
--- a/arch/s390/configs/performance_defconfig
+++ b/arch/s390/configs/performance_defconfig
@@ -558,7 +558,7 @@ CONFIG_SCHED_TRACER=y
 CONFIG_FTRACE_SYSCALLS=y
 CONFIG_STACK_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
-CONFIG_UPROBE_EVENT=y
+CONFIG_UPROBE_EVENTS=y
 CONFIG_FUNCTION_PROFILER=y
 CONFIG_HIST_TRIGGERS=y
 CONFIG_TRACE_ENUM_MAP_FILE=y
diff --git a/arch/s390/defconfig b/arch/s390/defconfig
index 68bfd09f1b02..97189dbaf34b 100644
--- a/arch/s390/defconfig
+++ b/arch/s390/defconfig
@@ -179,7 +179,7 @@ CONFIG_FTRACE_SYSCALLS=y
 CONFIG_TRACER_SNAPSHOT_PER_CPU_SWAP=y
 CONFIG_STACK_TRACER=y
 CONFIG_BLK_DEV_IO_TRACE=y
-CONFIG_UPROBE_EVENT=y
+CONFIG_UPROBE_EVENTS=y
 CONFIG_FUNCTION_PROFILER=y
 CONFIG_TRACE_ENUM_MAP_FILE=y
 CONFIG_KPROBES_SANITY_TEST=y
diff --git a/arch/x86/kernel/kprobes/common.h b/arch/x86/kernel/kprobes/common.h
index c6ee63f927ab..d688826e5736 100644
--- a/arch/x86/kernel/kprobes/common.h
+++ b/arch/x86/kernel/kprobes/common.h
@@ -67,7 +67,7 @@
 #endif
 
 /* Ensure if the instruction can be boostable */
-extern int can_boost(kprobe_opcode_t *instruction);
+extern int can_boost(kprobe_opcode_t *instruction, void *addr);
 /* Recover instruction if given address is probed */
 extern unsigned long recover_probed_instruction(kprobe_opcode_t *buf,
 					 unsigned long addr);
diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
index 6384eb754a58..993fa4fe4f68 100644
--- a/arch/x86/kernel/kprobes/core.c
+++ b/arch/x86/kernel/kprobes/core.c
@@ -167,12 +167,12 @@ NOKPROBE_SYMBOL(skip_prefixes);
  * Returns non-zero if opcode is boostable.
  * RIP relative instructions are adjusted at copying time in 64 bits mode
  */
-int can_boost(kprobe_opcode_t *opcodes)
+int can_boost(kprobe_opcode_t *opcodes, void *addr)
 {
 	kprobe_opcode_t opcode;
 	kprobe_opcode_t *orig_opcodes = opcodes;
 
-	if (search_exception_tables((unsigned long)opcodes))
+	if (search_exception_tables((unsigned long)addr))
 		return 0;	/* Page fault may occur on this address. */
 
 retry:
@@ -417,7 +417,7 @@ static int arch_copy_kprobe(struct kprobe *p)
 	 * __copy_instruction can modify the displacement of the instruction,
 	 * but it doesn't affect boostable check.
 	 */
-	if (can_boost(p->ainsn.insn))
+	if (can_boost(p->ainsn.insn, p->addr))
 		p->ainsn.boostable = 0;
 	else
 		p->ainsn.boostable = -1;
diff --git a/arch/x86/kernel/kprobes/opt.c b/arch/x86/kernel/kprobes/opt.c
index 3d1bee9d6a72..3e7c6e5a08ff 100644
--- a/arch/x86/kernel/kprobes/opt.c
+++ b/arch/x86/kernel/kprobes/opt.c
@@ -178,7 +178,7 @@ static int copy_optimized_instructions(u8 *dest, u8 *src)
 
 	while (len < RELATIVEJUMP_SIZE) {
 		ret = __copy_instruction(dest + len, src + len);
-		if (!ret || !can_boost(dest + len))
+		if (!ret || !can_boost(dest + len, src + len))
 			return -EINVAL;
 		len += ret;
 	}
diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
index c328e4f7dcad..177bdf6c6aeb 100644
--- a/include/linux/kprobes.h
+++ b/include/linux/kprobes.h
@@ -267,6 +267,7 @@ extern int arch_init_kprobes(void);
 extern void show_registers(struct pt_regs *regs);
 extern void kprobes_inc_nmissed_count(struct kprobe *p);
 extern bool arch_within_kprobe_blacklist(unsigned long addr);
+extern bool arch_function_offset_within_entry(unsigned long offset);
 
 extern bool within_kprobe_blacklist(unsigned long addr);
 
diff --git a/kernel/kprobes.c b/kernel/kprobes.c
index 699c5bc51a92..448759d4a263 100644
--- a/kernel/kprobes.c
+++ b/kernel/kprobes.c
@@ -1875,12 +1875,25 @@ static int pre_handler_kretprobe(struct kprobe *p, struct pt_regs *regs)
 }
 NOKPROBE_SYMBOL(pre_handler_kretprobe);
 
+bool __weak arch_function_offset_within_entry(unsigned long offset)
+{
+	return !offset;
+}
+
 int register_kretprobe(struct kretprobe *rp)
 {
 	int ret = 0;
 	struct kretprobe_instance *inst;
 	int i;
 	void *addr;
+	unsigned long offset;
+
+	addr = kprobe_addr(&rp->kp);
+	if (!kallsyms_lookup_size_offset((unsigned long)addr, NULL, &offset))
+		return -EINVAL;
+
+	if (!arch_function_offset_within_entry(offset))
+		return -EINVAL;
 
 	if (kretprobe_blacklist_size) {
 		addr = kprobe_addr(&rp->kp);
diff --git a/kernel/trace/Kconfig b/kernel/trace/Kconfig
index d5038005eb5d..d4a06e714645 100644
--- a/kernel/trace/Kconfig
+++ b/kernel/trace/Kconfig
@@ -429,7 +429,7 @@ config BLK_DEV_IO_TRACE
 
 	  If unsure, say N.
 
-config KPROBE_EVENT
+config KPROBE_EVENTS
 	depends on KPROBES
 	depends on HAVE_REGS_AND_STACK_ACCESS_API
 	bool "Enable kprobes-based dynamic events"
@@ -447,7 +447,7 @@ config KPROBE_EVENT
 	  This option is also required by perf-probe subcommand of perf tools.
 	  If you want to use perf tools, this option is strongly recommended.
 
-config UPROBE_EVENT
+config UPROBE_EVENTS
 	bool "Enable uprobes-based dynamic events"
 	depends on ARCH_SUPPORTS_UPROBES
 	depends on MMU
@@ -466,7 +466,7 @@ config UPROBE_EVENT
 
 config BPF_EVENTS
 	depends on BPF_SYSCALL
-	depends on (KPROBE_EVENT || UPROBE_EVENT) && PERF_EVENTS
+	depends on (KPROBE_EVENTS || UPROBE_EVENTS) && PERF_EVENTS
 	bool
 	default y
 	help
diff --git a/kernel/trace/Makefile b/kernel/trace/Makefile
index e57980845549..90f2701d92a7 100644
--- a/kernel/trace/Makefile
+++ b/kernel/trace/Makefile
@@ -57,7 +57,7 @@ obj-$(CONFIG_EVENT_TRACING) += trace_events_filter.o
 obj-$(CONFIG_EVENT_TRACING) += trace_events_trigger.o
 obj-$(CONFIG_HIST_TRIGGERS) += trace_events_hist.o
 obj-$(CONFIG_BPF_EVENTS) += bpf_trace.o
-obj-$(CONFIG_KPROBE_EVENT) += trace_kprobe.o
+obj-$(CONFIG_KPROBE_EVENTS) += trace_kprobe.o
 obj-$(CONFIG_TRACEPOINTS) += power-traces.o
 ifeq ($(CONFIG_PM),y)
 obj-$(CONFIG_TRACEPOINTS) += rpm-traces.o
@@ -66,7 +66,7 @@ ifeq ($(CONFIG_TRACING),y)
 obj-$(CONFIG_KGDB_KDB) += trace_kdb.o
 endif
 obj-$(CONFIG_PROBE_EVENTS) += trace_probe.o
-obj-$(CONFIG_UPROBE_EVENT) += trace_uprobe.o
+obj-$(CONFIG_UPROBE_EVENTS) += trace_uprobe.o
 
 obj-$(CONFIG_TRACEPOINT_BENCHMARK) += trace_benchmark.o
 
diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c
index 707445ceb7ef..0ed834d6beb0 100644
--- a/kernel/trace/trace.c
+++ b/kernel/trace/trace.c
@@ -4341,22 +4341,23 @@ static const char readme_msg[] =
 	"\t\t\t  traces\n"
 #endif
 #endif /* CONFIG_STACK_TRACER */
-#ifdef CONFIG_KPROBE_EVENT
+#ifdef CONFIG_KPROBE_EVENTS
 	"  kprobe_events\t\t- Add/remove/show the kernel dynamic events\n"
 	"\t\t\t  Write into this file to define/undefine new trace events.\n"
 #endif
-#ifdef CONFIG_UPROBE_EVENT
+#ifdef CONFIG_UPROBE_EVENTS
 	"  uprobe_events\t\t- Add/remove/show the userspace dynamic events\n"
 	"\t\t\t  Write into this file to define/undefine new trace events.\n"
 #endif
-#if defined(CONFIG_KPROBE_EVENT) || defined(CONFIG_UPROBE_EVENT)
+#if defined(CONFIG_KPROBE_EVENTS) || defined(CONFIG_UPROBE_EVENTS)
 	"\t  accepts: event-definitions (one definition per line)\n"
 	"\t   Format: p|r[:[<group>/]<event>] <place> [<args>]\n"
 	"\t           -:[<group>/]<event>\n"
-#ifdef CONFIG_KPROBE_EVENT
+#ifdef CONFIG_KPROBE_EVENTS
 	"\t    place: [<module>:]<symbol>[+<offset>]|<memaddr>\n"
+  "place (kretprobe): [<module>:]<symbol>[+<offset>]|<memaddr>\n"
 #endif
-#ifdef CONFIG_UPROBE_EVENT
+#ifdef CONFIG_UPROBE_EVENTS
 	"\t    place: <path>:<offset>\n"
 #endif
 	"\t     args: <name>=fetcharg[:type]\n"
diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c
index 5f688cc724f0..12fb540da0e5 100644
--- a/kernel/trace/trace_kprobe.c
+++ b/kernel/trace/trace_kprobe.c
@@ -681,10 +681,6 @@ static int create_trace_kprobe(int argc, char **argv)
 		return -EINVAL;
 	}
 	if (isdigit(argv[1][0])) {
-		if (is_return) {
-			pr_info("Return probe point must be a symbol.\n");
-			return -EINVAL;
-		}
 		/* an address specified */
 		ret = kstrtoul(&argv[1][0], 0, (unsigned long *)&addr);
 		if (ret) {
@@ -700,8 +696,9 @@ static int create_trace_kprobe(int argc, char **argv)
 			pr_info("Failed to parse symbol.\n");
 			return ret;
 		}
-		if (offset && is_return) {
-			pr_info("Return probe must be used without offset.\n");
+		if (offset && is_return &&
+		    !arch_function_offset_within_entry(offset)) {
+			pr_info("Given offset is not valid for return probe.\n");
 			return -EINVAL;
 		}
 	}
diff --git a/kernel/trace/trace_probe.h b/kernel/trace/trace_probe.h
index 0c0ae54d44c6..903273c93e61 100644
--- a/kernel/trace/trace_probe.h
+++ b/kernel/trace/trace_probe.h
@@ -248,7 +248,7 @@ ASSIGN_FETCH_FUNC(file_offset, ftype),			\
 #define FETCH_TYPE_STRING	0
 #define FETCH_TYPE_STRSIZE	1
 
-#ifdef CONFIG_KPROBE_EVENT
+#ifdef CONFIG_KPROBE_EVENTS
 struct symbol_cache;
 unsigned long update_symbol_cache(struct symbol_cache *sc);
 void free_symbol_cache(struct symbol_cache *sc);
@@ -278,7 +278,7 @@ alloc_symbol_cache(const char *sym, long offset)
 {
 	return NULL;
 }
-#endif /* CONFIG_KPROBE_EVENT */
+#endif /* CONFIG_KPROBE_EVENTS */
 
 struct probe_arg {
 	struct fetch_param	fetch;
diff --git a/tools/arch/x86/include/asm/atomic.h b/tools/arch/x86/include/asm/atomic.h
index 059e33e94260..328eeceec709 100644
--- a/tools/arch/x86/include/asm/atomic.h
+++ b/tools/arch/x86/include/asm/atomic.h
@@ -7,6 +7,8 @@
 
 #define LOCK_PREFIX "\n\tlock; "
 
+#include <asm/cmpxchg.h>
+
 /*
  * Atomic operations that C can't guarantee us.  Useful for
  * resource counting etc..
@@ -62,4 +64,9 @@ static inline int atomic_dec_and_test(atomic_t *v)
 	GEN_UNARY_RMWcc(LOCK_PREFIX "decl", v->counter, "%0", "e");
 }
 
+static __always_inline int atomic_cmpxchg(atomic_t *v, int old, int new)
+{
+	return cmpxchg(&v->counter, old, new);
+}
+
 #endif /* _TOOLS_LINUX_ASM_X86_ATOMIC_H */
diff --git a/tools/arch/x86/include/asm/cmpxchg.h b/tools/arch/x86/include/asm/cmpxchg.h
new file mode 100644
index 000000000000..f5253260f3cc
--- /dev/null
+++ b/tools/arch/x86/include/asm/cmpxchg.h
@@ -0,0 +1,89 @@
+#ifndef TOOLS_ASM_X86_CMPXCHG_H
+#define TOOLS_ASM_X86_CMPXCHG_H
+
+#include <linux/compiler.h>
+
+/*
+ * Non-existant functions to indicate usage errors at link time
+ * (or compile-time if the compiler implements __compiletime_error().
+ */
+extern void __cmpxchg_wrong_size(void)
+	__compiletime_error("Bad argument size for cmpxchg");
+
+/*
+ * Constants for operation sizes. On 32-bit, the 64-bit size it set to
+ * -1 because sizeof will never return -1, thereby making those switch
+ * case statements guaranteeed dead code which the compiler will
+ * eliminate, and allowing the "missing symbol in the default case" to
+ * indicate a usage error.
+ */
+#define __X86_CASE_B	1
+#define __X86_CASE_W	2
+#define __X86_CASE_L	4
+#ifdef __x86_64__
+#define __X86_CASE_Q	8
+#else
+#define	__X86_CASE_Q	-1		/* sizeof will never return -1 */
+#endif
+
+/*
+ * Atomic compare and exchange.  Compare OLD with MEM, if identical,
+ * store NEW in MEM.  Return the initial value in MEM.  Success is
+ * indicated by comparing RETURN with OLD.
+ */
+#define __raw_cmpxchg(ptr, old, new, size, lock)			\
+({									\
+	__typeof__(*(ptr)) __ret;					\
+	__typeof__(*(ptr)) __old = (old);				\
+	__typeof__(*(ptr)) __new = (new);				\
+	switch (size) {							\
+	case __X86_CASE_B:						\
+	{								\
+		volatile u8 *__ptr = (volatile u8 *)(ptr);		\
+		asm volatile(lock "cmpxchgb %2,%1"			\
+			     : "=a" (__ret), "+m" (*__ptr)		\
+			     : "q" (__new), "0" (__old)			\
+			     : "memory");				\
+		break;							\
+	}								\
+	case __X86_CASE_W:						\
+	{								\
+		volatile u16 *__ptr = (volatile u16 *)(ptr);		\
+		asm volatile(lock "cmpxchgw %2,%1"			\
+			     : "=a" (__ret), "+m" (*__ptr)		\
+			     : "r" (__new), "0" (__old)			\
+			     : "memory");				\
+		break;							\
+	}								\
+	case __X86_CASE_L:						\
+	{								\
+		volatile u32 *__ptr = (volatile u32 *)(ptr);		\
+		asm volatile(lock "cmpxchgl %2,%1"			\
+			     : "=a" (__ret), "+m" (*__ptr)		\
+			     : "r" (__new), "0" (__old)			\
+			     : "memory");				\
+		break;							\
+	}								\
+	case __X86_CASE_Q:						\
+	{								\
+		volatile u64 *__ptr = (volatile u64 *)(ptr);		\
+		asm volatile(lock "cmpxchgq %2,%1"			\
+			     : "=a" (__ret), "+m" (*__ptr)		\
+			     : "r" (__new), "0" (__old)			\
+			     : "memory");				\
+		break;							\
+	}								\
+	default:							\
+		__cmpxchg_wrong_size();					\
+	}								\
+	__ret;								\
+})
+
+#define __cmpxchg(ptr, old, new, size)					\
+	__raw_cmpxchg((ptr), (old), (new), (size), LOCK_PREFIX)
+
+#define cmpxchg(ptr, old, new)						\
+	__cmpxchg(ptr, old, new, sizeof(*(ptr)))
+
+
+#endif	/* TOOLS_ASM_X86_CMPXCHG_H */
diff --git a/tools/build/Makefile.feature b/tools/build/Makefile.feature
index e3fb5ecbdcb6..523911f316ce 100644
--- a/tools/build/Makefile.feature
+++ b/tools/build/Makefile.feature
@@ -63,6 +63,7 @@ FEATURE_TESTS_BASIC :=                  \
         lzma                            \
         get_cpuid                       \
         bpf                             \
+        sched_getcpu			\
         sdt
 
 # FEATURE_TESTS_BASIC + FEATURE_TESTS_EXTRA is the complete list
diff --git a/tools/build/feature/Makefile b/tools/build/feature/Makefile
index b564a2eea039..09c9626ea666 100644
--- a/tools/build/feature/Makefile
+++ b/tools/build/feature/Makefile
@@ -48,12 +48,13 @@ FILES=                                          \
          test-get_cpuid.bin                     \
          test-sdt.bin                           \
          test-cxx.bin                           \
-         test-jvmti.bin
+         test-jvmti.bin				\
+         test-sched_getcpu.bin
 
 FILES := $(addprefix $(OUTPUT),$(FILES))
 
-CC := $(CROSS_COMPILE)gcc -MD
-CXX := $(CROSS_COMPILE)g++ -MD
+CC ?= $(CROSS_COMPILE)gcc -MD
+CXX ?= $(CROSS_COMPILE)g++ -MD
 PKG_CONFIG := $(CROSS_COMPILE)pkg-config
 LLVM_CONFIG ?= llvm-config
 
@@ -91,6 +92,9 @@ $(OUTPUT)test-libelf.bin:
 $(OUTPUT)test-glibc.bin:
 	$(BUILD)
 
+$(OUTPUT)test-sched_getcpu.bin:
+	$(BUILD)
+
 DWARFLIBS := -ldw
 ifeq ($(findstring -static,${LDFLAGS}),-static)
 DWARFLIBS += -lelf -lebl -lz -llzma -lbz2
diff --git a/tools/build/feature/test-all.c b/tools/build/feature/test-all.c
index 699e43627397..cc6c7c01f4ca 100644
--- a/tools/build/feature/test-all.c
+++ b/tools/build/feature/test-all.c
@@ -117,6 +117,10 @@
 # include "test-pthread-attr-setaffinity-np.c"
 #undef main
 
+#define main main_test_sched_getcpu
+# include "test-sched_getcpu.c"
+#undef main
+
 # if 0
 /*
  * Disable libbabeltrace check for test-all, because the requested
@@ -182,6 +186,7 @@ int main(int argc, char *argv[])
 	main_test_get_cpuid();
 	main_test_bpf();
 	main_test_libcrypto();
+	main_test_sched_getcpu();
 	main_test_sdt();
 
 	return 0;
diff --git a/tools/build/feature/test-sched_getcpu.c b/tools/build/feature/test-sched_getcpu.c
new file mode 100644
index 000000000000..c4a148dd7104
--- /dev/null
+++ b/tools/build/feature/test-sched_getcpu.c
@@ -0,0 +1,7 @@
+#define _GNU_SOURCE
+#include <sched.h>
+
+int main(void)
+{
+	return sched_getcpu();
+}
diff --git a/tools/include/asm-generic/atomic-gcc.h b/tools/include/asm-generic/atomic-gcc.h
index 2ba78c9f5701..5e9738f97bf3 100644
--- a/tools/include/asm-generic/atomic-gcc.h
+++ b/tools/include/asm-generic/atomic-gcc.h
@@ -60,4 +60,12 @@ static inline int atomic_dec_and_test(atomic_t *v)
 	return __sync_sub_and_fetch(&v->counter, 1) == 0;
 }
 
+#define cmpxchg(ptr, oldval, newval) \
+	__sync_val_compare_and_swap(ptr, oldval, newval)
+
+static inline int atomic_cmpxchg(atomic_t *v, int oldval, int newval)
+{
+	return cmpxchg(&(v)->counter, oldval, newval);
+}
+
 #endif /* __TOOLS_ASM_GENERIC_ATOMIC_H */
diff --git a/tools/include/linux/atomic.h b/tools/include/linux/atomic.h
index 4e3d3d18ebab..9f21fc2b092b 100644
--- a/tools/include/linux/atomic.h
+++ b/tools/include/linux/atomic.h
@@ -3,4 +3,10 @@
 
 #include <asm/atomic.h>
 
+/* atomic_cmpxchg_relaxed */
+#ifndef atomic_cmpxchg_relaxed
+#define  atomic_cmpxchg_relaxed		atomic_cmpxchg
+#define  atomic_cmpxchg_release         atomic_cmpxchg
+#endif /* atomic_cmpxchg_relaxed */
+
 #endif /* __TOOLS_LINUX_ATOMIC_H */
diff --git a/tools/include/linux/compiler-gcc.h b/tools/include/linux/compiler-gcc.h
index 48af2f10a42d..616935f1ff56 100644
--- a/tools/include/linux/compiler-gcc.h
+++ b/tools/include/linux/compiler-gcc.h
@@ -12,3 +12,7 @@
 #if GCC_VERSION >= 70000 && !defined(__CHECKER__)
 # define __fallthrough __attribute__ ((fallthrough))
 #endif
+
+#if GCC_VERSION >= 40300
+# define __compiletime_error(message) __attribute__((error(message)))
+#endif /* GCC_VERSION >= 40300 */
diff --git a/tools/include/linux/compiler.h b/tools/include/linux/compiler.h
index 8de163b17c0d..c9e65e8faacd 100644
--- a/tools/include/linux/compiler.h
+++ b/tools/include/linux/compiler.h
@@ -5,6 +5,10 @@
 #include <linux/compiler-gcc.h>
 #endif
 
+#ifndef __compiletime_error
+# define __compiletime_error(message)
+#endif
+
 /* Optimization barrier */
 /* The "volatile" is due to gcc bugs */
 #define barrier() __asm__ __volatile__("": : :"memory")
diff --git a/tools/include/linux/kernel.h b/tools/include/linux/kernel.h
index 28607db02bd3..adb4d0147755 100644
--- a/tools/include/linux/kernel.h
+++ b/tools/include/linux/kernel.h
@@ -5,6 +5,10 @@
 #include <stddef.h>
 #include <assert.h>
 
+#ifndef UINT_MAX
+#define UINT_MAX	(~0U)
+#endif
+
 #define DIV_ROUND_UP(n,d) (((n) + (d) - 1) / (d))
 
 #define PERF_ALIGN(x, a)	__PERF_ALIGN_MASK(x, (typeof(x))(a)-1)
diff --git a/tools/include/linux/refcount.h b/tools/include/linux/refcount.h
new file mode 100644
index 000000000000..a0177c1f55b1
--- /dev/null
+++ b/tools/include/linux/refcount.h
@@ -0,0 +1,151 @@
+#ifndef _TOOLS_LINUX_REFCOUNT_H
+#define _TOOLS_LINUX_REFCOUNT_H
+
+/*
+ * Variant of atomic_t specialized for reference counts.
+ *
+ * The interface matches the atomic_t interface (to aid in porting) but only
+ * provides the few functions one should use for reference counting.
+ *
+ * It differs in that the counter saturates at UINT_MAX and will not move once
+ * there. This avoids wrapping the counter and causing 'spurious'
+ * use-after-free issues.
+ *
+ * Memory ordering rules are slightly relaxed wrt regular atomic_t functions
+ * and provide only what is strictly required for refcounts.
+ *
+ * The increments are fully relaxed; these will not provide ordering. The
+ * rationale is that whatever is used to obtain the object we're increasing the
+ * reference count on will provide the ordering. For locked data structures,
+ * its the lock acquire, for RCU/lockless data structures its the dependent
+ * load.
+ *
+ * Do note that inc_not_zero() provides a control dependency which will order
+ * future stores against the inc, this ensures we'll never modify the object
+ * if we did not in fact acquire a reference.
+ *
+ * The decrements will provide release order, such that all the prior loads and
+ * stores will be issued before, it also provides a control dependency, which
+ * will order us against the subsequent free().
+ *
+ * The control dependency is against the load of the cmpxchg (ll/sc) that
+ * succeeded. This means the stores aren't fully ordered, but this is fine
+ * because the 1->0 transition indicates no concurrency.
+ *
+ * Note that the allocator is responsible for ordering things between free()
+ * and alloc().
+ *
+ */
+
+#include <linux/atomic.h>
+#include <linux/kernel.h>
+
+#ifdef NDEBUG
+#define REFCOUNT_WARN(cond, str) (void)(cond)
+#define __refcount_check
+#else
+#define REFCOUNT_WARN(cond, str) BUG_ON(cond)
+#define __refcount_check	__must_check
+#endif
+
+typedef struct refcount_struct {
+	atomic_t refs;
+} refcount_t;
+
+#define REFCOUNT_INIT(n)	{ .refs = ATOMIC_INIT(n), }
+
+static inline void refcount_set(refcount_t *r, unsigned int n)
+{
+	atomic_set(&r->refs, n);
+}
+
+static inline unsigned int refcount_read(const refcount_t *r)
+{
+	return atomic_read(&r->refs);
+}
+
+/*
+ * Similar to atomic_inc_not_zero(), will saturate at UINT_MAX and WARN.
+ *
+ * Provides no memory ordering, it is assumed the caller has guaranteed the
+ * object memory to be stable (RCU, etc.). It does provide a control dependency
+ * and thereby orders future stores. See the comment on top.
+ */
+static inline __refcount_check
+bool refcount_inc_not_zero(refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		new = val + 1;
+
+		if (!val)
+			return false;
+
+		if (unlikely(!new))
+			return true;
+
+		old = atomic_cmpxchg_relaxed(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	REFCOUNT_WARN(new == UINT_MAX, "refcount_t: saturated; leaking memory.\n");
+
+	return true;
+}
+
+/*
+ * Similar to atomic_inc(), will saturate at UINT_MAX and WARN.
+ *
+ * Provides no memory ordering, it is assumed the caller already has a
+ * reference on the object, will WARN when this is not so.
+ */
+static inline void refcount_inc(refcount_t *r)
+{
+	REFCOUNT_WARN(!refcount_inc_not_zero(r), "refcount_t: increment on 0; use-after-free.\n");
+}
+
+/*
+ * Similar to atomic_dec_and_test(), it will WARN on underflow and fail to
+ * decrement when saturated at UINT_MAX.
+ *
+ * Provides release memory ordering, such that prior loads and stores are done
+ * before, and provides a control dependency such that free() must come after.
+ * See the comment on top.
+ */
+static inline __refcount_check
+bool refcount_sub_and_test(unsigned int i, refcount_t *r)
+{
+	unsigned int old, new, val = atomic_read(&r->refs);
+
+	for (;;) {
+		if (unlikely(val == UINT_MAX))
+			return false;
+
+		new = val - i;
+		if (new > val) {
+			REFCOUNT_WARN(new > val, "refcount_t: underflow; use-after-free.\n");
+			return false;
+		}
+
+		old = atomic_cmpxchg_release(&r->refs, val, new);
+		if (old == val)
+			break;
+
+		val = old;
+	}
+
+	return !new;
+}
+
+static inline __refcount_check
+bool refcount_dec_and_test(refcount_t *r)
+{
+	return refcount_sub_and_test(1, r);
+}
+
+
+#endif /* _ATOMIC_LINUX_REFCOUNT_H */
diff --git a/tools/perf/Documentation/perf-ftrace.txt b/tools/perf/Documentation/perf-ftrace.txt
index 2d96de6132a9..6e6a8b22c859 100644
--- a/tools/perf/Documentation/perf-ftrace.txt
+++ b/tools/perf/Documentation/perf-ftrace.txt
@@ -30,6 +30,24 @@ OPTIONS
 --verbose=::
         Verbosity level.
 
+-p::
+--pid=::
+	Trace on existing process id (comma separated list).
+
+-a::
+--all-cpus::
+	Force system-wide collection.  Scripts run without a <command>
+	normally use -a by default, while scripts run with a <command>
+	normally don't - this option allows the latter to be run in
+	system-wide mode.
+
+-C::
+--cpu=::
+	Only trace for the list of CPUs provided.  Multiple CPUs can
+	be provided as a comma separated list with no space like: 0,1.
+	Ranges of CPUs are specified with -: 0-2.
+	Default is to trace on all online CPUs.
+
 
 SEE ALSO
 --------
diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index c04cc0647c16..33f91906f5dc 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -80,6 +80,7 @@ OPTIONS
 	- pid: command and tid of the task
 	- dso: name of library or module executed at the time of sample
 	- symbol: name of function executed at the time of sample
+	- symbol_size: size of function executed at the time of sample
 	- parent: name of function matched to the parent regex filter. Unmatched
 	entries are displayed as "[other]".
 	- cpu: cpu number the task ran at the time of sample
diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST
index 8672f835ae4e..28648c09dcd6 100644
--- a/tools/perf/MANIFEST
+++ b/tools/perf/MANIFEST
@@ -12,6 +12,7 @@ tools/arch/sparc/include/asm/barrier_32.h
 tools/arch/sparc/include/asm/barrier_64.h
 tools/arch/tile/include/asm/barrier.h
 tools/arch/x86/include/asm/barrier.h
+tools/arch/x86/include/asm/cmpxchg.h
 tools/arch/x86/include/asm/cpufeatures.h
 tools/arch/x86/include/asm/disabled-features.h
 tools/arch/x86/include/asm/required-features.h
@@ -78,6 +79,7 @@ tools/include/uapi/linux/perf_event.h
 tools/include/linux/poison.h
 tools/include/linux/rbtree.h
 tools/include/linux/rbtree_augmented.h
+tools/include/linux/refcount.h
 tools/include/linux/string.h
 tools/include/linux/stringify.h
 tools/include/linux/types.h
diff --git a/tools/perf/Makefile.config b/tools/perf/Makefile.config
index 27c9fbca7bd9..2b656de99495 100644
--- a/tools/perf/Makefile.config
+++ b/tools/perf/Makefile.config
@@ -317,6 +317,10 @@ ifdef NO_DWARF
   NO_LIBDW_DWARF_UNWIND := 1
 endif
 
+ifeq ($(feature-sched_getcpu), 1)
+  CFLAGS += -DHAVE_SCHED_GETCPU_SUPPORT
+endif
+
 ifndef NO_LIBELF
   CFLAGS += -DHAVE_LIBELF_SUPPORT
   EXTLIBS += -lelf
diff --git a/tools/perf/bench/futex-hash.c b/tools/perf/bench/futex-hash.c
index da04b8c5568a..2499e1b0c6fb 100644
--- a/tools/perf/bench/futex-hash.c
+++ b/tools/perf/bench/futex-hash.c
@@ -9,6 +9,7 @@
  */
 
 /* For the CLR_() macros */
+#include <string.h>
 #include <pthread.h>
 
 #include <errno.h>
diff --git a/tools/perf/bench/futex-lock-pi.c b/tools/perf/bench/futex-lock-pi.c
index 91877777ec6e..a20814d94af1 100644
--- a/tools/perf/bench/futex-lock-pi.c
+++ b/tools/perf/bench/futex-lock-pi.c
@@ -3,6 +3,7 @@
  */
 
 /* For the CLR_() macros */
+#include <string.h>
 #include <pthread.h>
 
 #include <signal.h>
diff --git a/tools/perf/bench/futex-requeue.c b/tools/perf/bench/futex-requeue.c
index 2b9705a8734c..9fad1e4fcd3e 100644
--- a/tools/perf/bench/futex-requeue.c
+++ b/tools/perf/bench/futex-requeue.c
@@ -9,6 +9,7 @@
  */
 
 /* For the CLR_() macros */
+#include <string.h>
 #include <pthread.h>
 
 #include <signal.h>
diff --git a/tools/perf/bench/futex-wake-parallel.c b/tools/perf/bench/futex-wake-parallel.c
index 2c8fa67ad537..40f5fcf1d120 100644
--- a/tools/perf/bench/futex-wake-parallel.c
+++ b/tools/perf/bench/futex-wake-parallel.c
@@ -8,6 +8,7 @@
  */
 
 /* For the CLR_() macros */
+#include <string.h>
 #include <pthread.h>
 
 #include <signal.h>
diff --git a/tools/perf/bench/futex-wake.c b/tools/perf/bench/futex-wake.c
index e246b1b8388a..789490281ae3 100644
--- a/tools/perf/bench/futex-wake.c
+++ b/tools/perf/bench/futex-wake.c
@@ -9,6 +9,7 @@
  */
 
 /* For the CLR_() macros */
+#include <string.h>
 #include <pthread.h>
 
 #include <signal.h>
diff --git a/tools/perf/bench/futex.h b/tools/perf/bench/futex.h
index b2e06d1190d0..e44fd3239530 100644
--- a/tools/perf/bench/futex.h
+++ b/tools/perf/bench/futex.h
@@ -88,13 +88,11 @@ futex_cmp_requeue(u_int32_t *uaddr, u_int32_t val, u_int32_t *uaddr2, int nr_wak
 
 #ifndef HAVE_PTHREAD_ATTR_SETAFFINITY_NP
 #include <pthread.h>
-static inline int pthread_attr_setaffinity_np(pthread_attr_t *attr,
-					      size_t cpusetsize,
-					      cpu_set_t *cpuset)
+#include <linux/compiler.h>
+static inline int pthread_attr_setaffinity_np(pthread_attr_t *attr __maybe_unused,
+					      size_t cpusetsize __maybe_unused,
+					      cpu_set_t *cpuset __maybe_unused)
 {
-	attr = attr;
-	cpusetsize = cpusetsize;
-	cpuset = cpuset;
 	return 0;
 }
 #endif
diff --git a/tools/perf/bench/numa.c b/tools/perf/bench/numa.c
index 3083fc36282b..6bd0581de298 100644
--- a/tools/perf/bench/numa.c
+++ b/tools/perf/bench/numa.c
@@ -187,7 +187,8 @@ static const struct option options[] = {
 	OPT_INCR   ('d', "show_details"	, &p0.show_details,	"Show details"),
 	OPT_INCR   ('a', "all"		, &p0.run_all,		"Run all tests in the suite"),
 	OPT_INTEGER('H', "thp"		, &p0.thp,		"MADV_NOHUGEPAGE < 0 < MADV_HUGEPAGE"),
-	OPT_BOOLEAN('c', "show_convergence", &p0.show_convergence, "show convergence details"),
+	OPT_BOOLEAN('c', "show_convergence", &p0.show_convergence, "show convergence details, "
+		    "convergence is reached when each process (all its threads) is running on a single NUMA node."),
 	OPT_BOOLEAN('m', "measure_convergence",	&p0.measure_convergence, "measure convergence latency"),
 	OPT_BOOLEAN('q', "quiet"	, &p0.show_quiet,	"quiet mode"),
 	OPT_BOOLEAN('S', "serialize-startup", &p0.serialize_startup,"serialize thread startup"),
diff --git a/tools/perf/builtin-ftrace.c b/tools/perf/builtin-ftrace.c
index c3e643666c72..6087295f8827 100644
--- a/tools/perf/builtin-ftrace.c
+++ b/tools/perf/builtin-ftrace.c
@@ -11,11 +11,13 @@
 
 #include <unistd.h>
 #include <signal.h>
+#include <fcntl.h>
 
 #include "debug.h"
 #include <subcmd/parse-options.h>
 #include "evlist.h"
 #include "target.h"
+#include "cpumap.h"
 #include "thread_map.h"
 #include "util/config.h"
 
@@ -50,11 +52,12 @@ static void ftrace__workload_exec_failed_signal(int signo __maybe_unused,
 	done = true;
 }
 
-static int write_tracing_file(const char *name, const char *val)
+static int __write_tracing_file(const char *name, const char *val, bool append)
 {
 	char *file;
 	int fd, ret = -1;
 	ssize_t size = strlen(val);
+	int flags = O_WRONLY;
 
 	file = get_tracing_file(name);
 	if (!file) {
@@ -62,7 +65,12 @@ static int write_tracing_file(const char *name, const char *val)
 		return -1;
 	}
 
-	fd = open(file, O_WRONLY);
+	if (append)
+		flags |= O_APPEND;
+	else
+		flags |= O_TRUNC;
+
+	fd = open(file, flags);
 	if (fd < 0) {
 		pr_debug("cannot open tracing file: %s\n", name);
 		goto out;
@@ -79,6 +87,18 @@ out:
 	return ret;
 }
 
+static int write_tracing_file(const char *name, const char *val)
+{
+	return __write_tracing_file(name, val, false);
+}
+
+static int append_tracing_file(const char *name, const char *val)
+{
+	return __write_tracing_file(name, val, true);
+}
+
+static int reset_tracing_cpu(void);
+
 static int reset_tracing_files(struct perf_ftrace *ftrace __maybe_unused)
 {
 	if (write_tracing_file("tracing_on", "0") < 0)
@@ -90,14 +110,78 @@ static int reset_tracing_files(struct perf_ftrace *ftrace __maybe_unused)
 	if (write_tracing_file("set_ftrace_pid", " ") < 0)
 		return -1;
 
+	if (reset_tracing_cpu() < 0)
+		return -1;
+
+	return 0;
+}
+
+static int set_tracing_pid(struct perf_ftrace *ftrace)
+{
+	int i;
+	char buf[16];
+
+	if (target__has_cpu(&ftrace->target))
+		return 0;
+
+	for (i = 0; i < thread_map__nr(ftrace->evlist->threads); i++) {
+		scnprintf(buf, sizeof(buf), "%d",
+			  ftrace->evlist->threads->map[i]);
+		if (append_tracing_file("set_ftrace_pid", buf) < 0)
+			return -1;
+	}
 	return 0;
 }
 
+static int set_tracing_cpumask(struct cpu_map *cpumap)
+{
+	char *cpumask;
+	size_t mask_size;
+	int ret;
+	int last_cpu;
+
+	last_cpu = cpu_map__cpu(cpumap, cpumap->nr - 1);
+	mask_size = (last_cpu + 3) / 4 + 1;
+	mask_size += last_cpu / 32; /* ',' is needed for every 32th cpus */
+
+	cpumask = malloc(mask_size);
+	if (cpumask == NULL) {
+		pr_debug("failed to allocate cpu mask\n");
+		return -1;
+	}
+
+	cpu_map__snprint_mask(cpumap, cpumask, mask_size);
+
+	ret = write_tracing_file("tracing_cpumask", cpumask);
+
+	free(cpumask);
+	return ret;
+}
+
+static int set_tracing_cpu(struct perf_ftrace *ftrace)
+{
+	struct cpu_map *cpumap = ftrace->evlist->cpus;
+
+	if (!target__has_cpu(&ftrace->target))
+		return 0;
+
+	return set_tracing_cpumask(cpumap);
+}
+
+static int reset_tracing_cpu(void)
+{
+	struct cpu_map *cpumap = cpu_map__new(NULL);
+	int ret;
+
+	ret = set_tracing_cpumask(cpumap);
+	cpu_map__put(cpumap);
+	return ret;
+}
+
 static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 {
 	char *trace_file;
 	int trace_fd;
-	char *trace_pid;
 	char buf[4096];
 	struct pollfd pollfd = {
 		.events = POLLIN,
@@ -108,42 +192,43 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 		return -1;
 	}
 
-	if (argc < 1)
-		return -1;
-
 	signal(SIGINT, sig_handler);
 	signal(SIGUSR1, sig_handler);
 	signal(SIGCHLD, sig_handler);
+	signal(SIGPIPE, sig_handler);
 
-	reset_tracing_files(ftrace);
+	if (reset_tracing_files(ftrace) < 0)
+		goto out;
 
 	/* reset ftrace buffer */
 	if (write_tracing_file("trace", "0") < 0)
 		goto out;
 
-	if (perf_evlist__prepare_workload(ftrace->evlist, &ftrace->target,
-					  argv, false, ftrace__workload_exec_failed_signal) < 0)
+	if (argc && perf_evlist__prepare_workload(ftrace->evlist,
+				&ftrace->target, argv, false,
+				ftrace__workload_exec_failed_signal) < 0) {
 		goto out;
+	}
 
-	if (write_tracing_file("current_tracer", ftrace->tracer) < 0) {
-		pr_err("failed to set current_tracer to %s\n", ftrace->tracer);
-		goto out;
+	if (set_tracing_pid(ftrace) < 0) {
+		pr_err("failed to set ftrace pid\n");
+		goto out_reset;
 	}
 
-	if (asprintf(&trace_pid, "%d", thread_map__pid(ftrace->evlist->threads, 0)) < 0) {
-		pr_err("failed to allocate pid string\n");
-		goto out;
+	if (set_tracing_cpu(ftrace) < 0) {
+		pr_err("failed to set tracing cpumask\n");
+		goto out_reset;
 	}
 
-	if (write_tracing_file("set_ftrace_pid", trace_pid) < 0) {
-		pr_err("failed to set pid: %s\n", trace_pid);
-		goto out_free_pid;
+	if (write_tracing_file("current_tracer", ftrace->tracer) < 0) {
+		pr_err("failed to set current_tracer to %s\n", ftrace->tracer);
+		goto out_reset;
 	}
 
 	trace_file = get_tracing_file("trace_pipe");
 	if (!trace_file) {
 		pr_err("failed to open trace_pipe\n");
-		goto out_free_pid;
+		goto out_reset;
 	}
 
 	trace_fd = open(trace_file, O_RDONLY);
@@ -152,7 +237,7 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 
 	if (trace_fd < 0) {
 		pr_err("failed to open trace_pipe\n");
-		goto out_free_pid;
+		goto out_reset;
 	}
 
 	fcntl(trace_fd, F_SETFL, O_NONBLOCK);
@@ -163,6 +248,8 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 		goto out_close_fd;
 	}
 
+	setup_pager();
+
 	perf_evlist__start_workload(ftrace->evlist);
 
 	while (!done) {
@@ -191,11 +278,9 @@ static int __cmd_ftrace(struct perf_ftrace *ftrace, int argc, const char **argv)
 
 out_close_fd:
 	close(trace_fd);
-out_free_pid:
-	free(trace_pid);
-out:
+out_reset:
 	reset_tracing_files(ftrace);
-
+out:
 	return done ? 0 : -1;
 }
 
@@ -227,15 +312,21 @@ int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
 		.target = { .uid = UINT_MAX, },
 	};
 	const char * const ftrace_usage[] = {
-		"perf ftrace [<options>] <command>",
+		"perf ftrace [<options>] [<command>]",
 		"perf ftrace [<options>] -- <command> [<options>]",
 		NULL
 	};
 	const struct option ftrace_options[] = {
 	OPT_STRING('t', "tracer", &ftrace.tracer, "tracer",
 		   "tracer to use: function_graph(default) or function"),
+	OPT_STRING('p', "pid", &ftrace.target.pid, "pid",
+		   "trace on existing process id"),
 	OPT_INCR('v', "verbose", &verbose,
 		 "be more verbose"),
+	OPT_BOOLEAN('a', "all-cpus", &ftrace.target.system_wide,
+		    "system-wide collection from all CPUs"),
+	OPT_STRING('C', "cpu", &ftrace.target.cpu_list, "cpu",
+		    "list of cpus to monitor"),
 	OPT_END()
 	};
 
@@ -245,9 +336,18 @@ int cmd_ftrace(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	argc = parse_options(argc, argv, ftrace_options, ftrace_usage,
 			    PARSE_OPT_STOP_AT_NON_OPTION);
-	if (!argc)
+	if (!argc && target__none(&ftrace.target))
 		usage_with_options(ftrace_usage, ftrace_options);
 
+	ret = target__validate(&ftrace.target);
+	if (ret) {
+		char errbuf[512];
+
+		target__strerror(&ftrace.target, ret, errbuf, 512);
+		pr_err("%s\n", errbuf);
+		return -EINVAL;
+	}
+
 	ftrace.evlist = perf_evlist__new();
 	if (ftrace.evlist == NULL)
 		return -ENOMEM;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 13b54999ad79..f53f449d864d 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -146,6 +146,7 @@ static aggr_get_id_t		aggr_get_id;
 static bool			append_file;
 static const char		*output_name;
 static int			output_fd;
+static int			print_free_counters_hint;
 
 struct perf_stat {
 	bool			 record;
@@ -1109,6 +1110,9 @@ static void printout(int id, int nr, struct perf_evsel *counter, double uval,
 			counter->supported ? CNTR_NOT_COUNTED : CNTR_NOT_SUPPORTED,
 			csv_sep);
 
+		if (counter->supported)
+			print_free_counters_hint = 1;
+
 		fprintf(stat_config.output, "%-*s%s",
 			csv_output ? 0 : unit_width,
 			counter->unit, csv_sep);
@@ -1477,6 +1481,13 @@ static void print_footer(void)
 				avg_stats(&walltime_nsecs_stats));
 	}
 	fprintf(output, "\n\n");
+
+	if (print_free_counters_hint)
+		fprintf(output,
+"Some events weren't counted. Try disabling the NMI watchdog:\n"
+"	echo 0 > /proc/sys/kernel/nmi_watchdog\n"
+"	perf stat ...\n"
+"	echo 1 > /proc/sys/kernel/nmi_watchdog\n");
 }
 
 static void print_counters(struct timespec *ts, int argc, const char **argv)
@@ -2339,6 +2350,35 @@ static int __cmd_report(int argc, const char **argv)
 	return 0;
 }
 
+static void setup_system_wide(int forks)
+{
+	/*
+	 * Make system wide (-a) the default target if
+	 * no target was specified and one of following
+	 * conditions is met:
+	 *
+	 *   - there's no workload specified
+	 *   - there is workload specified but all requested
+	 *     events are system wide events
+	 */
+	if (!target__none(&target))
+		return;
+
+	if (!forks)
+		target.system_wide = true;
+	else {
+		struct perf_evsel *counter;
+
+		evlist__for_each_entry(evsel_list, counter) {
+			if (!counter->system_wide)
+				return;
+		}
+
+		if (evsel_list->nr_entries)
+			target.system_wide = true;
+	}
+}
+
 int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	const char * const stat_usage[] = {
@@ -2445,9 +2485,7 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 	} else if (big_num_opt == 0) /* User passed --no-big-num */
 		big_num = false;
 
-	/* Make system wide (-a) the default target. */
-	if (!argc && target__none(&target))
-		target.system_wide = true;
+	setup_system_wide(argc);
 
 	if (run_count < 0) {
 		pr_err("Run count must be a positive number\n");
diff --git a/tools/perf/pmu-events/arch/x86/mapfile.csv b/tools/perf/pmu-events/arch/x86/mapfile.csv
index 12181bb1da2a..d1a12e584c1b 100644
--- a/tools/perf/pmu-events/arch/x86/mapfile.csv
+++ b/tools/perf/pmu-events/arch/x86/mapfile.csv
@@ -17,6 +17,7 @@ GenuineIntel-6-3A,v18,ivybridge,core
 GenuineIntel-6-3E,v19,ivytown,core
 GenuineIntel-6-2D,v20,jaketown,core
 GenuineIntel-6-57,v9,knightslanding,core
+GenuineIntel-6-85,v9,knightslanding,core
 GenuineIntel-6-1E,v2,nehalemep,core
 GenuineIntel-6-1F,v2,nehalemep,core
 GenuineIntel-6-1A,v2,nehalemep,core
diff --git a/tools/perf/tests/cpumap.c b/tools/perf/tests/cpumap.c
index f168a85992d0..4478773cdb97 100644
--- a/tools/perf/tests/cpumap.c
+++ b/tools/perf/tests/cpumap.c
@@ -66,7 +66,7 @@ static int process_event_cpus(struct perf_tool *tool __maybe_unused,
 	TEST_ASSERT_VAL("wrong nr",  map->nr == 2);
 	TEST_ASSERT_VAL("wrong cpu", map->map[0] == 1);
 	TEST_ASSERT_VAL("wrong cpu", map->map[1] == 256);
-	TEST_ASSERT_VAL("wrong refcnt", atomic_read(&map->refcnt) == 1);
+	TEST_ASSERT_VAL("wrong refcnt", refcount_read(&map->refcnt) == 1);
 	cpu_map__put(map);
 	return 0;
 }
diff --git a/tools/perf/tests/thread-map.c b/tools/perf/tests/thread-map.c
index f2d2e542d0ee..a63d6945807b 100644
--- a/tools/perf/tests/thread-map.c
+++ b/tools/perf/tests/thread-map.c
@@ -29,7 +29,7 @@ int test__thread_map(int subtest __maybe_unused)
 			thread_map__comm(map, 0) &&
 			!strcmp(thread_map__comm(map, 0), NAME));
 	TEST_ASSERT_VAL("wrong refcnt",
-			atomic_read(&map->refcnt) == 1);
+			refcount_read(&map->refcnt) == 1);
 	thread_map__put(map);
 
 	/* test dummy pid */
@@ -44,7 +44,7 @@ int test__thread_map(int subtest __maybe_unused)
 			thread_map__comm(map, 0) &&
 			!strcmp(thread_map__comm(map, 0), "dummy"));
 	TEST_ASSERT_VAL("wrong refcnt",
-			atomic_read(&map->refcnt) == 1);
+			refcount_read(&map->refcnt) == 1);
 	thread_map__put(map);
 	return 0;
 }
@@ -71,7 +71,7 @@ static int process_event(struct perf_tool *tool __maybe_unused,
 			thread_map__comm(threads, 0) &&
 			!strcmp(thread_map__comm(threads, 0), NAME));
 	TEST_ASSERT_VAL("wrong refcnt",
-			atomic_read(&threads->refcnt) == 1);
+			refcount_read(&threads->refcnt) == 1);
 	thread_map__put(threads);
 	return 0;
 }
diff --git a/tools/perf/tests/thread-mg-share.c b/tools/perf/tests/thread-mg-share.c
index 188b63140fc8..76686dd6f5ec 100644
--- a/tools/perf/tests/thread-mg-share.c
+++ b/tools/perf/tests/thread-mg-share.c
@@ -43,7 +43,7 @@ int test__thread_mg_share(int subtest __maybe_unused)
 			leader && t1 && t2 && t3 && other);
 
 	mg = leader->mg;
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&mg->refcnt), 4);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&mg->refcnt), 4);
 
 	/* test the map groups pointer is shared */
 	TEST_ASSERT_VAL("map groups don't match", mg == t1->mg);
@@ -71,25 +71,25 @@ int test__thread_mg_share(int subtest __maybe_unused)
 	machine__remove_thread(machine, other_leader);
 
 	other_mg = other->mg;
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&other_mg->refcnt), 2);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&other_mg->refcnt), 2);
 
 	TEST_ASSERT_VAL("map groups don't match", other_mg == other_leader->mg);
 
 	/* release thread group */
 	thread__put(leader);
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&mg->refcnt), 3);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&mg->refcnt), 3);
 
 	thread__put(t1);
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&mg->refcnt), 2);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&mg->refcnt), 2);
 
 	thread__put(t2);
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&mg->refcnt), 1);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&mg->refcnt), 1);
 
 	thread__put(t3);
 
 	/* release other group  */
 	thread__put(other_leader);
-	TEST_ASSERT_EQUAL("wrong refcnt", atomic_read(&other_mg->refcnt), 1);
+	TEST_ASSERT_EQUAL("wrong refcnt", refcount_read(&other_mg->refcnt), 1);
 
 	thread__put(other);
 
diff --git a/tools/perf/util/cgroup.c b/tools/perf/util/cgroup.c
index eafbf11442b2..86399eda3684 100644
--- a/tools/perf/util/cgroup.c
+++ b/tools/perf/util/cgroup.c
@@ -127,19 +127,19 @@ static int add_cgroup(struct perf_evlist *evlist, char *str)
 			goto found;
 		n++;
 	}
-	if (atomic_read(&cgrp->refcnt) == 0)
+	if (refcount_read(&cgrp->refcnt) == 0)
 		free(cgrp);
 
 	return -1;
 found:
-	atomic_inc(&cgrp->refcnt);
+	refcount_inc(&cgrp->refcnt);
 	counter->cgrp = cgrp;
 	return 0;
 }
 
 void close_cgroup(struct cgroup_sel *cgrp)
 {
-	if (cgrp && atomic_dec_and_test(&cgrp->refcnt)) {
+	if (cgrp && refcount_dec_and_test(&cgrp->refcnt)) {
 		close(cgrp->fd);
 		zfree(&cgrp->name);
 		free(cgrp);
diff --git a/tools/perf/util/cgroup.h b/tools/perf/util/cgroup.h
index 31f8dcdbd7ef..d91966b97cbd 100644
--- a/tools/perf/util/cgroup.h
+++ b/tools/perf/util/cgroup.h
@@ -1,14 +1,14 @@
 #ifndef __CGROUP_H__
 #define __CGROUP_H__
 
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 
 struct option;
 
 struct cgroup_sel {
 	char *name;
 	int fd;
-	atomic_t refcnt;
+	refcount_t refcnt;
 };
 
 
diff --git a/tools/perf/util/cloexec.h b/tools/perf/util/cloexec.h
index d0d465953d36..94a5a7d829d5 100644
--- a/tools/perf/util/cloexec.h
+++ b/tools/perf/util/cloexec.h
@@ -3,10 +3,4 @@
 
 unsigned long perf_event_open_cloexec_flag(void);
 
-#ifdef __GLIBC_PREREQ
-#if !__GLIBC_PREREQ(2, 6) && !defined(__UCLIBC__)
-int sched_getcpu(void) __THROW;
-#endif
-#endif
-
 #endif /* __PERF_CLOEXEC_H */
diff --git a/tools/perf/util/comm.c b/tools/perf/util/comm.c
index 21b7ff382c3f..32837b6f7879 100644
--- a/tools/perf/util/comm.c
+++ b/tools/perf/util/comm.c
@@ -2,12 +2,12 @@
 #include "util.h"
 #include <stdlib.h>
 #include <stdio.h>
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 
 struct comm_str {
 	char *str;
 	struct rb_node rb_node;
-	atomic_t refcnt;
+	refcount_t refcnt;
 };
 
 /* Should perhaps be moved to struct machine */
@@ -16,13 +16,13 @@ static struct rb_root comm_str_root;
 static struct comm_str *comm_str__get(struct comm_str *cs)
 {
 	if (cs)
-		atomic_inc(&cs->refcnt);
+		refcount_inc(&cs->refcnt);
 	return cs;
 }
 
 static void comm_str__put(struct comm_str *cs)
 {
-	if (cs && atomic_dec_and_test(&cs->refcnt)) {
+	if (cs && refcount_dec_and_test(&cs->refcnt)) {
 		rb_erase(&cs->rb_node, &comm_str_root);
 		zfree(&cs->str);
 		free(cs);
@@ -43,7 +43,7 @@ static struct comm_str *comm_str__alloc(const char *str)
 		return NULL;
 	}
 
-	atomic_set(&cs->refcnt, 0);
+	refcount_set(&cs->refcnt, 1);
 
 	return cs;
 }
@@ -61,7 +61,7 @@ static struct comm_str *comm_str__findnew(const char *str, struct rb_root *root)
 
 		cmp = strcmp(str, iter->str);
 		if (!cmp)
-			return iter;
+			return comm_str__get(iter);
 
 		if (cmp < 0)
 			p = &(*p)->rb_left;
@@ -95,8 +95,6 @@ struct comm *comm__new(const char *str, u64 timestamp, bool exec)
 		return NULL;
 	}
 
-	comm_str__get(comm->comm_str);
-
 	return comm;
 }
 
@@ -108,7 +106,6 @@ int comm__override(struct comm *comm, const char *str, u64 timestamp, bool exec)
 	if (!new)
 		return -ENOMEM;
 
-	comm_str__get(new);
 	comm_str__put(old);
 	comm->comm_str = new;
 	comm->start = timestamp;
diff --git a/tools/perf/util/cpumap.c b/tools/perf/util/cpumap.c
index 8c7504939113..061018b42393 100644
--- a/tools/perf/util/cpumap.c
+++ b/tools/perf/util/cpumap.c
@@ -29,7 +29,7 @@ static struct cpu_map *cpu_map__default_new(void)
 			cpus->map[i] = i;
 
 		cpus->nr = nr_cpus;
-		atomic_set(&cpus->refcnt, 1);
+		refcount_set(&cpus->refcnt, 1);
 	}
 
 	return cpus;
@@ -43,7 +43,7 @@ static struct cpu_map *cpu_map__trim_new(int nr_cpus, int *tmp_cpus)
 	if (cpus != NULL) {
 		cpus->nr = nr_cpus;
 		memcpy(cpus->map, tmp_cpus, payload_size);
-		atomic_set(&cpus->refcnt, 1);
+		refcount_set(&cpus->refcnt, 1);
 	}
 
 	return cpus;
@@ -252,7 +252,7 @@ struct cpu_map *cpu_map__dummy_new(void)
 	if (cpus != NULL) {
 		cpus->nr = 1;
 		cpus->map[0] = -1;
-		atomic_set(&cpus->refcnt, 1);
+		refcount_set(&cpus->refcnt, 1);
 	}
 
 	return cpus;
@@ -269,7 +269,7 @@ struct cpu_map *cpu_map__empty_new(int nr)
 		for (i = 0; i < nr; i++)
 			cpus->map[i] = -1;
 
-		atomic_set(&cpus->refcnt, 1);
+		refcount_set(&cpus->refcnt, 1);
 	}
 
 	return cpus;
@@ -278,7 +278,7 @@ struct cpu_map *cpu_map__empty_new(int nr)
 static void cpu_map__delete(struct cpu_map *map)
 {
 	if (map) {
-		WARN_ONCE(atomic_read(&map->refcnt) != 0,
+		WARN_ONCE(refcount_read(&map->refcnt) != 0,
 			  "cpu_map refcnt unbalanced\n");
 		free(map);
 	}
@@ -287,13 +287,13 @@ static void cpu_map__delete(struct cpu_map *map)
 struct cpu_map *cpu_map__get(struct cpu_map *map)
 {
 	if (map)
-		atomic_inc(&map->refcnt);
+		refcount_inc(&map->refcnt);
 	return map;
 }
 
 void cpu_map__put(struct cpu_map *map)
 {
-	if (map && atomic_dec_and_test(&map->refcnt))
+	if (map && refcount_dec_and_test(&map->refcnt))
 		cpu_map__delete(map);
 }
 
@@ -357,7 +357,7 @@ int cpu_map__build_map(struct cpu_map *cpus, struct cpu_map **res,
 	/* ensure we process id in increasing order */
 	qsort(c->map, c->nr, sizeof(int), cmp_ids);
 
-	atomic_set(&c->refcnt, 1);
+	refcount_set(&c->refcnt, 1);
 	*res = c;
 	return 0;
 }
@@ -673,3 +673,49 @@ size_t cpu_map__snprint(struct cpu_map *map, char *buf, size_t size)
 	pr_debug("cpumask list: %s\n", buf);
 	return ret;
 }
+
+static char hex_char(unsigned char val)
+{
+	if (val < 10)
+		return val + '0';
+	if (val < 16)
+		return val - 10 + 'a';
+	return '?';
+}
+
+size_t cpu_map__snprint_mask(struct cpu_map *map, char *buf, size_t size)
+{
+	int i, cpu;
+	char *ptr = buf;
+	unsigned char *bitmap;
+	int last_cpu = cpu_map__cpu(map, map->nr - 1);
+
+	bitmap = zalloc((last_cpu + 7) / 8);
+	if (bitmap == NULL) {
+		buf[0] = '\0';
+		return 0;
+	}
+
+	for (i = 0; i < map->nr; i++) {
+		cpu = cpu_map__cpu(map, i);
+		bitmap[cpu / 8] |= 1 << (cpu % 8);
+	}
+
+	for (cpu = last_cpu / 4 * 4; cpu >= 0; cpu -= 4) {
+		unsigned char bits = bitmap[cpu / 8];
+
+		if (cpu % 8)
+			bits >>= 4;
+		else
+			bits &= 0xf;
+
+		*ptr++ = hex_char(bits);
+		if ((cpu % 32) == 0 && cpu > 0)
+			*ptr++ = ',';
+	}
+	*ptr = '\0';
+	free(bitmap);
+
+	buf[size - 1] = '\0';
+	return ptr - buf;
+}
diff --git a/tools/perf/util/cpumap.h b/tools/perf/util/cpumap.h
index 1a0549af8f5c..6b8bff87481d 100644
--- a/tools/perf/util/cpumap.h
+++ b/tools/perf/util/cpumap.h
@@ -3,13 +3,13 @@
 
 #include <stdio.h>
 #include <stdbool.h>
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 
 #include "perf.h"
 #include "util/debug.h"
 
 struct cpu_map {
-	atomic_t refcnt;
+	refcount_t refcnt;
 	int nr;
 	int map[];
 };
@@ -20,6 +20,7 @@ struct cpu_map *cpu_map__dummy_new(void);
 struct cpu_map *cpu_map__new_data(struct cpu_map_data *data);
 struct cpu_map *cpu_map__read(FILE *file);
 size_t cpu_map__snprint(struct cpu_map *map, char *buf, size_t size);
+size_t cpu_map__snprint_mask(struct cpu_map *map, char *buf, size_t size);
 size_t cpu_map__fprintf(struct cpu_map *map, FILE *fp);
 int cpu_map__get_socket_id(int cpu);
 int cpu_map__get_socket(struct cpu_map *map, int idx, void *data);
diff --git a/tools/perf/util/dso.c b/tools/perf/util/dso.c
index d38b62a700ca..42db00d78573 100644
--- a/tools/perf/util/dso.c
+++ b/tools/perf/util/dso.c
@@ -1109,7 +1109,7 @@ struct dso *dso__new(const char *name)
 		INIT_LIST_HEAD(&dso->node);
 		INIT_LIST_HEAD(&dso->data.open_entry);
 		pthread_mutex_init(&dso->lock, NULL);
-		atomic_set(&dso->refcnt, 1);
+		refcount_set(&dso->refcnt, 1);
 	}
 
 	return dso;
@@ -1147,13 +1147,13 @@ void dso__delete(struct dso *dso)
 struct dso *dso__get(struct dso *dso)
 {
 	if (dso)
-		atomic_inc(&dso->refcnt);
+		refcount_inc(&dso->refcnt);
 	return dso;
 }
 
 void dso__put(struct dso *dso)
 {
-	if (dso && atomic_dec_and_test(&dso->refcnt))
+	if (dso && refcount_dec_and_test(&dso->refcnt))
 		dso__delete(dso);
 }
 
diff --git a/tools/perf/util/dso.h b/tools/perf/util/dso.h
index ecc4bbd3f82e..12350b171727 100644
--- a/tools/perf/util/dso.h
+++ b/tools/perf/util/dso.h
@@ -1,7 +1,7 @@
 #ifndef __PERF_DSO
 #define __PERF_DSO
 
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 #include <linux/types.h>
 #include <linux/rbtree.h>
 #include <sys/types.h>
@@ -187,7 +187,7 @@ struct dso {
 		void	 *priv;
 		u64	 db_id;
 	};
-	atomic_t	 refcnt;
+	refcount_t	 refcnt;
 	char		 name[0];
 };
 
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index b601f2814a30..50420cd35446 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -777,7 +777,7 @@ union perf_event *perf_mmap__read_forward(struct perf_mmap *md, bool check_messu
 	/*
 	 * Check if event was unmapped due to a POLLHUP/POLLERR.
 	 */
-	if (!atomic_read(&md->refcnt))
+	if (!refcount_read(&md->refcnt))
 		return NULL;
 
 	head = perf_mmap__read_head(md);
@@ -794,7 +794,7 @@ perf_mmap__read_backward(struct perf_mmap *md)
 	/*
 	 * Check if event was unmapped due to a POLLHUP/POLLERR.
 	 */
-	if (!atomic_read(&md->refcnt))
+	if (!refcount_read(&md->refcnt))
 		return NULL;
 
 	head = perf_mmap__read_head(md);
@@ -856,7 +856,7 @@ void perf_mmap__read_catchup(struct perf_mmap *md)
 {
 	u64 head;
 
-	if (!atomic_read(&md->refcnt))
+	if (!refcount_read(&md->refcnt))
 		return;
 
 	head = perf_mmap__read_head(md);
@@ -875,14 +875,14 @@ static bool perf_mmap__empty(struct perf_mmap *md)
 
 static void perf_mmap__get(struct perf_mmap *map)
 {
-	atomic_inc(&map->refcnt);
+	refcount_inc(&map->refcnt);
 }
 
 static void perf_mmap__put(struct perf_mmap *md)
 {
-	BUG_ON(md->base && atomic_read(&md->refcnt) == 0);
+	BUG_ON(md->base && refcount_read(&md->refcnt) == 0);
 
-	if (atomic_dec_and_test(&md->refcnt))
+	if (refcount_dec_and_test(&md->refcnt))
 		perf_mmap__munmap(md);
 }
 
@@ -894,7 +894,7 @@ void perf_mmap__consume(struct perf_mmap *md, bool overwrite)
 		perf_mmap__write_tail(md, old);
 	}
 
-	if (atomic_read(&md->refcnt) == 1 && perf_mmap__empty(md))
+	if (refcount_read(&md->refcnt) == 1 && perf_mmap__empty(md))
 		perf_mmap__put(md);
 }
 
@@ -937,7 +937,7 @@ static void perf_mmap__munmap(struct perf_mmap *map)
 		munmap(map->base, perf_mmap__mmap_len(map));
 		map->base = NULL;
 		map->fd = -1;
-		atomic_set(&map->refcnt, 0);
+		refcount_set(&map->refcnt, 0);
 	}
 	auxtrace_mmap__munmap(&map->auxtrace_mmap);
 }
@@ -974,8 +974,19 @@ static struct perf_mmap *perf_evlist__alloc_mmap(struct perf_evlist *evlist)
 	if (!map)
 		return NULL;
 
-	for (i = 0; i < evlist->nr_mmaps; i++)
+	for (i = 0; i < evlist->nr_mmaps; i++) {
 		map[i].fd = -1;
+		/*
+		 * When the perf_mmap() call is made we grab one refcount, plus
+		 * one extra to let perf_evlist__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.
+		 */
+		refcount_set(&map[i].refcnt, 0);
+	}
 	return map;
 }
 
@@ -1001,7 +1012,7 @@ static int perf_mmap__mmap(struct perf_mmap *map,
 	 * evlist layer can't just drop it when filtering events in
 	 * perf_evlist__filter_pollfd().
 	 */
-	atomic_set(&map->refcnt, 2);
+	refcount_set(&map->refcnt, 2);
 	map->prev = 0;
 	map->mask = mp->mask;
 	map->base = mmap(NULL, perf_mmap__mmap_len(map), mp->prot,
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 389b9ccdf8c7..39942995f537 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -1,7 +1,7 @@
 #ifndef __PERF_EVLIST_H
 #define __PERF_EVLIST_H 1
 
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 #include <linux/list.h>
 #include <api/fd/array.h>
 #include <stdio.h>
@@ -29,7 +29,7 @@ struct perf_mmap {
 	void		 *base;
 	int		 mask;
 	int		 fd;
-	atomic_t	 refcnt;
+	refcount_t	 refcnt;
 	u64		 prev;
 	struct auxtrace_mmap auxtrace_mmap;
 	char		 event_copy[PERF_SAMPLE_MAX_SIZE] __attribute__((aligned(8)));
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index 28c216e3d5b7..2e839bf40bdd 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -57,6 +57,7 @@ enum hist_column {
 	HISTC_SRCLINE_FROM,
 	HISTC_SRCLINE_TO,
 	HISTC_TRACE,
+	HISTC_SYM_SIZE,
 	HISTC_NR_COLS, /* Last entry */
 };
 
diff --git a/tools/perf/util/intel-pt-decoder/intel-pt-insn-decoder.c b/tools/perf/util/intel-pt-decoder/intel-pt-insn-decoder.c
index 7913363bde5c..55b6250350d7 100644
--- a/tools/perf/util/intel-pt-decoder/intel-pt-insn-decoder.c
+++ b/tools/perf/util/intel-pt-decoder/intel-pt-insn-decoder.c
@@ -39,6 +39,8 @@ static void intel_pt_insn_decoder(struct insn *insn,
 	enum intel_pt_insn_branch branch = INTEL_PT_BR_NO_BRANCH;
 	int ext;
 
+	intel_pt_insn->rel = 0;
+
 	if (insn_is_avx(insn)) {
 		intel_pt_insn->op = INTEL_PT_OP_OTHER;
 		intel_pt_insn->branch = INTEL_PT_BR_NO_BRANCH;
diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index 71c9720d4973..b9974fe41bc1 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -1439,7 +1439,7 @@ static void __machine__remove_thread(struct machine *machine, struct thread *th,
 	if (machine->last_match == th)
 		machine->last_match = NULL;
 
-	BUG_ON(atomic_read(&th->refcnt) == 0);
+	BUG_ON(refcount_read(&th->refcnt) == 0);
 	if (lock)
 		pthread_rwlock_wrlock(&machine->threads_lock);
 	rb_erase_init(&th->rb_node, &machine->threads);
diff --git a/tools/perf/util/map.c b/tools/perf/util/map.c
index 0a943e7b1ea7..1d9ebcf9e38e 100644
--- a/tools/perf/util/map.c
+++ b/tools/perf/util/map.c
@@ -141,7 +141,7 @@ void map__init(struct map *map, enum map_type type,
 	RB_CLEAR_NODE(&map->rb_node);
 	map->groups   = NULL;
 	map->erange_warned = false;
-	atomic_set(&map->refcnt, 1);
+	refcount_set(&map->refcnt, 1);
 }
 
 struct map *map__new(struct machine *machine, u64 start, u64 len,
@@ -255,7 +255,7 @@ void map__delete(struct map *map)
 
 void map__put(struct map *map)
 {
-	if (map && atomic_dec_and_test(&map->refcnt))
+	if (map && refcount_dec_and_test(&map->refcnt))
 		map__delete(map);
 }
 
@@ -354,7 +354,7 @@ struct map *map__clone(struct map *from)
 	struct map *map = memdup(from, sizeof(*map));
 
 	if (map != NULL) {
-		atomic_set(&map->refcnt, 1);
+		refcount_set(&map->refcnt, 1);
 		RB_CLEAR_NODE(&map->rb_node);
 		dso__get(map->dso);
 		map->groups = NULL;
@@ -485,7 +485,7 @@ void map_groups__init(struct map_groups *mg, struct machine *machine)
 		maps__init(&mg->maps[i]);
 	}
 	mg->machine = machine;
-	atomic_set(&mg->refcnt, 1);
+	refcount_set(&mg->refcnt, 1);
 }
 
 static void __maps__purge(struct maps *maps)
@@ -547,7 +547,7 @@ void map_groups__delete(struct map_groups *mg)
 
 void map_groups__put(struct map_groups *mg)
 {
-	if (mg && atomic_dec_and_test(&mg->refcnt))
+	if (mg && refcount_dec_and_test(&mg->refcnt))
 		map_groups__delete(mg);
 }
 
diff --git a/tools/perf/util/map.h b/tools/perf/util/map.h
index abdacf800c98..c8a5a644c0a9 100644
--- a/tools/perf/util/map.h
+++ b/tools/perf/util/map.h
@@ -1,7 +1,7 @@
 #ifndef __PERF_MAP_H
 #define __PERF_MAP_H
 
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 #include <linux/compiler.h>
 #include <linux/list.h>
 #include <linux/rbtree.h>
@@ -51,7 +51,7 @@ struct map {
 
 	struct dso		*dso;
 	struct map_groups	*groups;
-	atomic_t		refcnt;
+	refcount_t		refcnt;
 };
 
 struct kmap {
@@ -67,7 +67,7 @@ struct maps {
 struct map_groups {
 	struct maps	 maps[MAP__NR_TYPES];
 	struct machine	 *machine;
-	atomic_t	 refcnt;
+	refcount_t	 refcnt;
 };
 
 struct map_groups *map_groups__new(struct machine *machine);
@@ -77,7 +77,7 @@ bool map_groups__empty(struct map_groups *mg);
 static inline struct map_groups *map_groups__get(struct map_groups *mg)
 {
 	if (mg)
-		atomic_inc(&mg->refcnt);
+		refcount_inc(&mg->refcnt);
 	return mg;
 }
 
@@ -150,7 +150,7 @@ struct map *map__clone(struct map *map);
 static inline struct map *map__get(struct map *map)
 {
 	if (map)
-		atomic_inc(&map->refcnt);
+		refcount_inc(&map->refcnt);
 	return map;
 }
 
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index 67a8aebc67ab..54355d3caf09 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -316,8 +316,9 @@ __add_event(struct list_head *list, int *idx,
 		return NULL;
 
 	(*idx)++;
-	evsel->cpus     = cpu_map__get(cpus);
-	evsel->own_cpus = cpu_map__get(cpus);
+	evsel->cpus        = cpu_map__get(cpus);
+	evsel->own_cpus    = cpu_map__get(cpus);
+	evsel->system_wide = !!cpus;
 
 	if (name)
 		evsel->name = strdup(name);
diff --git a/tools/perf/util/probe-file.c b/tools/perf/util/probe-file.c
index 436b64731f65..1a62daceb028 100644
--- a/tools/perf/util/probe-file.c
+++ b/tools/perf/util/probe-file.c
@@ -70,7 +70,7 @@ static void print_both_open_warning(int kerr, int uerr)
 	}
 }
 
-static int open_probe_events(const char *trace_file, bool readwrite)
+int open_trace_file(const char *trace_file, bool readwrite)
 {
 	char buf[PATH_MAX];
 	int ret;
@@ -92,12 +92,12 @@ static int open_probe_events(const char *trace_file, bool readwrite)
 
 static int open_kprobe_events(bool readwrite)
 {
-	return open_probe_events("kprobe_events", readwrite);
+	return open_trace_file("kprobe_events", readwrite);
 }
 
 static int open_uprobe_events(bool readwrite)
 {
-	return open_probe_events("uprobe_events", readwrite);
+	return open_trace_file("uprobe_events", readwrite);
 }
 
 int probe_file__open(int flag)
@@ -899,6 +899,7 @@ bool probe_type_is_available(enum probe_type type)
 	size_t len = 0;
 	bool target_line = false;
 	bool ret = probe_type_table[type].avail;
+	int fd;
 
 	if (type >= PROBE_TYPE_END)
 		return false;
@@ -906,14 +907,16 @@ bool probe_type_is_available(enum probe_type type)
 	if (ret || probe_type_table[type].checked)
 		return ret;
 
-	if (asprintf(&buf, "%s/README", tracing_path) < 0)
+	fd = open_trace_file("README", false);
+	if (fd < 0)
 		return ret;
 
-	fp = fopen(buf, "r");
-	if (!fp)
-		goto end;
+	fp = fdopen(fd, "r");
+	if (!fp) {
+		close(fd);
+		return ret;
+	}
 
-	zfree(&buf);
 	while (getline(&buf, &len, fp) > 0 && !ret) {
 		if (!target_line) {
 			target_line = !!strstr(buf, " type: ");
@@ -928,7 +931,6 @@ bool probe_type_is_available(enum probe_type type)
 	probe_type_table[type].avail = ret;
 
 	fclose(fp);
-end:
 	free(buf);
 
 	return ret;
diff --git a/tools/perf/util/probe-file.h b/tools/perf/util/probe-file.h
index eba44c3e9dca..a17a82eff8a0 100644
--- a/tools/perf/util/probe-file.h
+++ b/tools/perf/util/probe-file.h
@@ -35,6 +35,7 @@ enum probe_type {
 
 /* probe-file.c depends on libelf */
 #ifdef HAVE_LIBELF_SUPPORT
+int open_trace_file(const char *trace_file, bool readwrite);
 int probe_file__open(int flag);
 int probe_file__open_both(int *kfd, int *ufd, int flag);
 struct strlist *probe_file__get_namelist(int fd);
diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index 0ff622288d24..f8f16c0e20b6 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -1396,6 +1396,46 @@ struct sort_entry sort_transaction = {
 	.se_width_idx	= HISTC_TRANSACTION,
 };
 
+/* --sort symbol_size */
+
+static int64_t _sort__sym_size_cmp(struct symbol *sym_l, struct symbol *sym_r)
+{
+	int64_t size_l = sym_l != NULL ? symbol__size(sym_l) : 0;
+	int64_t size_r = sym_r != NULL ? symbol__size(sym_r) : 0;
+
+	return size_l < size_r ? -1 :
+		size_l == size_r ? 0 : 1;
+}
+
+static int64_t
+sort__sym_size_cmp(struct hist_entry *left, struct hist_entry *right)
+{
+	return _sort__sym_size_cmp(right->ms.sym, left->ms.sym);
+}
+
+static int _hist_entry__sym_size_snprintf(struct symbol *sym, char *bf,
+					  size_t bf_size, unsigned int width)
+{
+	if (sym)
+		return repsep_snprintf(bf, bf_size, "%*d", width, symbol__size(sym));
+
+	return repsep_snprintf(bf, bf_size, "%*s", width, "unknown");
+}
+
+static int hist_entry__sym_size_snprintf(struct hist_entry *he, char *bf,
+					 size_t size, unsigned int width)
+{
+	return _hist_entry__sym_size_snprintf(he->ms.sym, bf, size, width);
+}
+
+struct sort_entry sort_sym_size = {
+	.se_header	= "Symbol size",
+	.se_cmp		= sort__sym_size_cmp,
+	.se_snprintf	= hist_entry__sym_size_snprintf,
+	.se_width_idx	= HISTC_SYM_SIZE,
+};
+
+
 struct sort_dimension {
 	const char		*name;
 	struct sort_entry	*entry;
@@ -1418,6 +1458,7 @@ static struct sort_dimension common_sort_dimensions[] = {
 	DIM(SORT_GLOBAL_WEIGHT, "weight", sort_global_weight),
 	DIM(SORT_TRANSACTION, "transaction", sort_transaction),
 	DIM(SORT_TRACE, "trace", sort_trace),
+	DIM(SORT_SYM_SIZE, "symbol_size", sort_sym_size),
 };
 
 #undef DIM
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index 796c847e2f00..f583325a3743 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -211,6 +211,7 @@ enum sort_type {
 	SORT_GLOBAL_WEIGHT,
 	SORT_TRANSACTION,
 	SORT_TRACE,
+	SORT_SYM_SIZE,
 
 	/* branch stack specific sort keys */
 	__SORT_BRANCH_STACK,
diff --git a/tools/perf/util/thread.c b/tools/perf/util/thread.c
index f5af87f66663..74e79d26b421 100644
--- a/tools/perf/util/thread.c
+++ b/tools/perf/util/thread.c
@@ -53,7 +53,7 @@ struct thread *thread__new(pid_t pid, pid_t tid)
 			goto err_thread;
 
 		list_add(&comm->list, &thread->comm_list);
-		atomic_set(&thread->refcnt, 1);
+		refcount_set(&thread->refcnt, 1);
 		RB_CLEAR_NODE(&thread->rb_node);
 	}
 
@@ -88,13 +88,13 @@ void thread__delete(struct thread *thread)
 struct thread *thread__get(struct thread *thread)
 {
 	if (thread)
-		atomic_inc(&thread->refcnt);
+		refcount_inc(&thread->refcnt);
 	return thread;
 }
 
 void thread__put(struct thread *thread)
 {
-	if (thread && atomic_dec_and_test(&thread->refcnt)) {
+	if (thread && refcount_dec_and_test(&thread->refcnt)) {
 		/*
 		 * Remove it from the dead_threads list, as last reference
 		 * is gone.
diff --git a/tools/perf/util/thread.h b/tools/perf/util/thread.h
index 99263cb6e6b6..e57188546465 100644
--- a/tools/perf/util/thread.h
+++ b/tools/perf/util/thread.h
@@ -1,7 +1,7 @@
 #ifndef __PERF_THREAD_H
 #define __PERF_THREAD_H
 
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 #include <linux/rbtree.h>
 #include <linux/list.h>
 #include <unistd.h>
@@ -23,7 +23,7 @@ struct thread {
 	pid_t			tid;
 	pid_t			ppid;
 	int			cpu;
-	atomic_t		refcnt;
+	refcount_t		refcnt;
 	char			shortname[3];
 	bool			comm_set;
 	int			comm_len;
diff --git a/tools/perf/util/thread_map.c b/tools/perf/util/thread_map.c
index 7c3fcc538a70..9026408ea55b 100644
--- a/tools/perf/util/thread_map.c
+++ b/tools/perf/util/thread_map.c
@@ -66,7 +66,7 @@ struct thread_map *thread_map__new_by_pid(pid_t pid)
 		for (i = 0; i < items; i++)
 			thread_map__set_pid(threads, i, atoi(namelist[i]->d_name));
 		threads->nr = items;
-		atomic_set(&threads->refcnt, 1);
+		refcount_set(&threads->refcnt, 1);
 	}
 
 	for (i=0; i<items; i++)
@@ -83,7 +83,7 @@ struct thread_map *thread_map__new_by_tid(pid_t tid)
 	if (threads != NULL) {
 		thread_map__set_pid(threads, 0, tid);
 		threads->nr = 1;
-		atomic_set(&threads->refcnt, 1);
+		refcount_set(&threads->refcnt, 1);
 	}
 
 	return threads;
@@ -105,7 +105,7 @@ struct thread_map *thread_map__new_by_uid(uid_t uid)
 		goto out_free_threads;
 
 	threads->nr = 0;
-	atomic_set(&threads->refcnt, 1);
+	refcount_set(&threads->refcnt, 1);
 
 	while ((dirent = readdir(proc)) != NULL) {
 		char *end;
@@ -235,7 +235,7 @@ static struct thread_map *thread_map__new_by_pid_str(const char *pid_str)
 out:
 	strlist__delete(slist);
 	if (threads)
-		atomic_set(&threads->refcnt, 1);
+		refcount_set(&threads->refcnt, 1);
 	return threads;
 
 out_free_namelist:
@@ -255,7 +255,7 @@ struct thread_map *thread_map__new_dummy(void)
 	if (threads != NULL) {
 		thread_map__set_pid(threads, 0, -1);
 		threads->nr = 1;
-		atomic_set(&threads->refcnt, 1);
+		refcount_set(&threads->refcnt, 1);
 	}
 	return threads;
 }
@@ -300,7 +300,7 @@ struct thread_map *thread_map__new_by_tid_str(const char *tid_str)
 	}
 out:
 	if (threads)
-		atomic_set(&threads->refcnt, 1);
+		refcount_set(&threads->refcnt, 1);
 	return threads;
 
 out_free_threads:
@@ -326,7 +326,7 @@ static void thread_map__delete(struct thread_map *threads)
 	if (threads) {
 		int i;
 
-		WARN_ONCE(atomic_read(&threads->refcnt) != 0,
+		WARN_ONCE(refcount_read(&threads->refcnt) != 0,
 			  "thread map refcnt unbalanced\n");
 		for (i = 0; i < threads->nr; i++)
 			free(thread_map__comm(threads, i));
@@ -337,13 +337,13 @@ static void thread_map__delete(struct thread_map *threads)
 struct thread_map *thread_map__get(struct thread_map *map)
 {
 	if (map)
-		atomic_inc(&map->refcnt);
+		refcount_inc(&map->refcnt);
 	return map;
 }
 
 void thread_map__put(struct thread_map *map)
 {
-	if (map && atomic_dec_and_test(&map->refcnt))
+	if (map && refcount_dec_and_test(&map->refcnt))
 		thread_map__delete(map);
 }
 
@@ -423,7 +423,7 @@ static void thread_map__copy_event(struct thread_map *threads,
 		threads->map[i].comm = strndup(event->entries[i].comm, 16);
 	}
 
-	atomic_set(&threads->refcnt, 1);
+	refcount_set(&threads->refcnt, 1);
 }
 
 struct thread_map *thread_map__new_event(struct thread_map_event *event)
diff --git a/tools/perf/util/thread_map.h b/tools/perf/util/thread_map.h
index ea0ef08c6303..bd34d7a0b9fa 100644
--- a/tools/perf/util/thread_map.h
+++ b/tools/perf/util/thread_map.h
@@ -3,7 +3,7 @@
 
 #include <sys/types.h>
 #include <stdio.h>
-#include <linux/atomic.h>
+#include <linux/refcount.h>
 
 struct thread_map_data {
 	pid_t    pid;
@@ -11,7 +11,7 @@ struct thread_map_data {
 };
 
 struct thread_map {
-	atomic_t refcnt;
+	refcount_t refcnt;
 	int nr;
 	struct thread_map_data map[];
 };
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index c74708da8571..b2cfa47990dc 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -355,8 +355,8 @@ void print_binary(unsigned char *data, size_t len,
 		  size_t bytes_per_line, print_binary_t printer,
 		  void *extra);
 
-#if !defined(__GLIBC__) && !defined(__ANDROID__)
-extern int sched_getcpu(void);
+#ifndef HAVE_SCHED_GETCPU_SUPPORT
+int sched_getcpu(void);
 #endif
 
 int is_printable_array(char *p, unsigned int len);
diff --git a/tools/scripts/Makefile.include b/tools/scripts/Makefile.include
index 621578aa12d6..fc74db62fef4 100644
--- a/tools/scripts/Makefile.include
+++ b/tools/scripts/Makefile.include
@@ -43,6 +43,15 @@ ifneq ($(CC), clang)
 EXTRA_WARNINGS += -Wstrict-aliasing=3
 endif
 
+# Hack to avoid type-punned warnings on old systems such as RHEL5:
+# We should be changing CFLAGS and checking gcc version, but this
+# will do for now and keep the above -Wstrict-aliasing=3 in place
+# in newer systems.
+# Needed for the __raw_cmpxchg in tools/arch/x86/include/asm/cmpxchg.h
+ifneq ($(filter 3.%,$(MAKE_VERSION)),)  # make-3
+EXTRA_WARNINGS += -fno-strict-aliasing
+endif
+
 ifneq ($(findstring $(MAKEFLAGS), w),w)
 PRINT_DIR = --no-print-directory
 else