Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux…

…/kernel/git/acme/linux into perf/core

perf/core improvements and fixes:

 * Improve 'perf bench' docs, by Namhyung Kim

 * Fix build when O= is not used, from David Ahern

 * Fix cross compilation build, from Namhyung Kim

 * Fix pipe mode when callchains are used, from David Ahern

 * Follow .gnu_debuglink section to find separate symbols, from Pierre-Loup A. Griffais

 * Fix 'perf test' raw events entries, from Jiri Olsa

 * Use the events description in the perf.data file, not the sysfs ones.

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

tools/perf/Documentation/perf-bench.txt

@@ -12,7 +12,7 @@ SYNOPSIS
 
 DESCRIPTION
 -----------
-This 'perf bench' command is general framework for benchmark suites.
+This 'perf bench' command is a general framework for benchmark suites.
 
 COMMON OPTIONS
 --------------
@@ -45,14 +45,20 @@ SUBSYSTEM
 'sched'::
 	Scheduler and IPC mechanisms.
 
+'mem'::
+	Memory access performance.
+
+'all'::
+	All benchmark subsystems.
+
 SUITES FOR 'sched'
 ~~~~~~~~~~~~~~~~~~
 *messaging*::
 Suite for evaluating performance of scheduler and IPC mechanisms.
 Based on hackbench by Rusty Russell.
 
-Options of *pipe*
-^^^^^^^^^^^^^^^^^
+Options of *messaging*
+^^^^^^^^^^^^^^^^^^^^^^
 -p::
 --pipe::
 Use pipe() instead of socketpair()
@@ -115,6 +121,72 @@ Example of *pipe*
                 59004 ops/sec
 ---------------------
 
+SUITES FOR 'mem'
+~~~~~~~~~~~~~~~~
+*memcpy*::
+Suite for evaluating performance of simple memory copy in various ways.
+
+Options of *memcpy*
+^^^^^^^^^^^^^^^^^^^
+-l::
+--length::
+Specify length of memory to copy (default: 1MB).
+Available units are B, KB, MB, GB and TB (case insensitive).
+
+-r::
+--routine::
+Specify routine to copy (default: default).
+Available routines are depend on the architecture.
+On x86-64, x86-64-unrolled, x86-64-movsq and x86-64-movsb are supported.
+
+-i::
+--iterations::
+Repeat memcpy invocation this number of times.
+
+-c::
+--clock::
+Use perf's cpu-cycles event instead of gettimeofday syscall.
+
+-o::
+--only-prefault::
+Show only the result with page faults before memcpy.
+
+-n::
+--no-prefault::
+Show only the result without page faults before memcpy.
+
+*memset*::
+Suite for evaluating performance of simple memory set in various ways.
+
+Options of *memset*
+^^^^^^^^^^^^^^^^^^^
+-l::
+--length::
+Specify length of memory to set (default: 1MB).
+Available units are B, KB, MB, GB and TB (case insensitive).
+
+-r::
+--routine::
+Specify routine to set (default: default).
+Available routines are depend on the architecture.
+On x86-64, x86-64-unrolled, x86-64-stosq and x86-64-stosb are supported.
+
+-i::
+--iterations::
+Repeat memset invocation this number of times.
+
+-c::
+--clock::
+Use perf's cpu-cycles event instead of gettimeofday syscall.
+
+-o::
+--only-prefault::
+Show only the result with page faults before memset.
+
+-n::
+--no-prefault::
+Show only the result without page faults before memset.
+
 SEE ALSO
 --------
 linkperf:perf[1]

tools/perf/Makefile

@@ -155,7 +155,7 @@ endif
 
 ### --- END CONFIGURATION SECTION ---
 
-BASIC_CFLAGS = -Iutil/include -Iarch/$(ARCH)/include -I$(OUTPUT)/util -I$(TRACE_EVENT_DIR) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE
+BASIC_CFLAGS = -Iutil/include -Iarch/$(ARCH)/include -I$(OUTPUT)util -I$(TRACE_EVENT_DIR) -D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -D_GNU_SOURCE
 BASIC_LDFLAGS =
 
 # Guard against environment variables

tools/perf/bench/mem-memcpy.c

@@ -32,13 +32,13 @@ static bool		no_prefault;
 static const struct option options[] = {
 	OPT_STRING('l', "length", &length_str, "1MB",
 		    "Specify length of memory to copy. "
-		    "available unit: B, MB, GB (upper and lower)"),
+		    "Available units: B, KB, MB, GB and TB (upper and lower)"),
 	OPT_STRING('r', "routine", &routine, "default",
 		    "Specify routine to copy"),
 	OPT_INTEGER('i', "iterations", &iterations,
 		    "repeat memcpy() invocation this number of times"),
 	OPT_BOOLEAN('c', "clock", &use_clock,
-		    "Use CPU clock for measuring"),
+		    "Use cycles event instead of gettimeofday() for measuring"),
 	OPT_BOOLEAN('o', "only-prefault", &only_prefault,
 		    "Show only the result with page faults before memcpy()"),
 	OPT_BOOLEAN('n', "no-prefault", &no_prefault,

tools/perf/bench/mem-memset.c

@@ -31,14 +31,14 @@ static bool		no_prefault;
 
 static const struct option options[] = {
 	OPT_STRING('l', "length", &length_str, "1MB",
-		    "Specify length of memory to copy. "
-		    "available unit: B, MB, GB (upper and lower)"),
+		    "Specify length of memory to set. "
+		    "Available units: B, KB, MB, GB and TB (upper and lower)"),
 	OPT_STRING('r', "routine", &routine, "default",
-		    "Specify routine to copy"),
+		    "Specify routine to set"),
 	OPT_INTEGER('i', "iterations", &iterations,
 		    "repeat memset() invocation this number of times"),
 	OPT_BOOLEAN('c', "clock", &use_clock,
-		    "Use CPU clock for measuring"),
+		    "Use cycles event instead of gettimeofday() for measuring"),
 	OPT_BOOLEAN('o', "only-prefault", &only_prefault,
 		    "Show only the result with page faults before memset()"),
 	OPT_BOOLEAN('n', "no-prefault", &no_prefault,

tools/perf/builtin-bench.c

@@ -33,7 +33,7 @@ struct bench_suite {
 };
 						\
 /* sentinel: easy for help */
-#define suite_all { "all", "test all suite (pseudo suite)", NULL }
+#define suite_all { "all", "Test all benchmark suites", NULL }
 
 static struct bench_suite sched_suites[] = {
 	{ "messaging",
@@ -75,7 +75,7 @@ static struct bench_subsys subsystems[] = {
 	  "memory access performance",
 	  mem_suites },
 	{ "all",		/* sentinel: easy for help */
-	  "test all subsystem (pseudo subsystem)",
+	  "all benchmark subsystem",
 	  NULL },
 	{ NULL,
 	  NULL,

tools/perf/builtin-kmem.c

@@ -57,6 +57,11 @@ static unsigned long nr_allocs, nr_cross_allocs;
 
 #define PATH_SYS_NODE	"/sys/devices/system/node"
 
+struct perf_kmem {
+	struct perf_tool    tool;
+	struct perf_session *session;
+};
+
 static void init_cpunode_map(void)
 {
 	FILE *fp;
@@ -278,14 +283,16 @@ static void process_free_event(void *data,
 	s_alloc->alloc_cpu = -1;
 }
 
-static void process_raw_event(union perf_event *raw_event __used, void *data,
+static void process_raw_event(struct perf_tool *tool,
+			      union perf_event *raw_event __used, void *data,
 			      int cpu, u64 timestamp, struct thread *thread)
 {
+	struct perf_kmem *kmem = container_of(tool, struct perf_kmem, tool);
 	struct event_format *event;
 	int type;
 
-	type = trace_parse_common_type(data);
-	event = trace_find_event(type);
+	type = trace_parse_common_type(kmem->session->pevent, data);
+	event = pevent_find_event(kmem->session->pevent, type);
 
 	if (!strcmp(event->name, "kmalloc") ||
 	    !strcmp(event->name, "kmem_cache_alloc")) {
@@ -306,7 +313,7 @@ static void process_raw_event(union perf_event *raw_event __used, void *data,
 	}
 }
 
-static int process_sample_event(struct perf_tool *tool __used,
+static int process_sample_event(struct perf_tool *tool,
 				union perf_event *event,
 				struct perf_sample *sample,
 				struct perf_evsel *evsel __used,
@@ -322,16 +329,18 @@ static int process_sample_event(struct perf_tool *tool __used,
 
 	dump_printf(" ... thread: %s:%d\n", thread->comm, thread->pid);
 
-	process_raw_event(event, sample->raw_data, sample->cpu,
+	process_raw_event(tool, event, sample->raw_data, sample->cpu,
 			  sample->time, thread);
 
 	return 0;
 }
 
-static struct perf_tool perf_kmem = {
-	.sample			= process_sample_event,
-	.comm			= perf_event__process_comm,
-	.ordered_samples	= true,
+static struct perf_kmem perf_kmem = {
+	.tool = {
+		.sample			= process_sample_event,
+		.comm			= perf_event__process_comm,
+		.ordered_samples	= true,
+	},
 };
 
 static double fragmentation(unsigned long n_req, unsigned long n_alloc)
@@ -486,19 +495,23 @@ static void sort_result(void)
 static int __cmd_kmem(void)
 {
 	int err = -EINVAL;
-	struct perf_session *session = perf_session__new(input_name, O_RDONLY,
-							 0, false, &perf_kmem);
+	struct perf_session *session;
+
+	session = perf_session__new(input_name, O_RDONLY, 0, false,
+				    &perf_kmem.tool);
 	if (session == NULL)
 		return -ENOMEM;
 
+	perf_kmem.session = session;
+
 	if (perf_session__create_kernel_maps(session) < 0)
 		goto out_delete;
 
 	if (!perf_session__has_traces(session, "kmem record"))
 		goto out_delete;
 
 	setup_pager();
-	err = perf_session__process_events(session, &perf_kmem);
+	err = perf_session__process_events(session, &perf_kmem.tool);
 	if (err != 0)
 		goto out_delete;
 	sort_result();

tools/perf/builtin-lock.c

@@ -724,8 +724,8 @@ process_raw_event(void *data, int cpu, u64 timestamp, struct thread *thread)
 	struct event_format *event;
 	int type;
 
-	type = trace_parse_common_type(data);
-	event = trace_find_event(type);
+	type = trace_parse_common_type(session->pevent, data);
+	event = pevent_find_event(session->pevent, type);
 
 	if (!strcmp(event->name, "lock_acquire"))
 		process_lock_acquire_event(data, event, cpu, timestamp, thread);

tools/perf/builtin-report.c

@@ -245,11 +245,12 @@ static int process_read_event(struct perf_tool *tool,
 	return 0;
 }
 
+/* For pipe mode, sample_type is not currently set */
 static int perf_report__setup_sample_type(struct perf_report *rep)
 {
 	struct perf_session *self = rep->session;
 
-	if (!(self->sample_type & PERF_SAMPLE_CALLCHAIN)) {
+	if (!self->fd_pipe && !(self->sample_type & PERF_SAMPLE_CALLCHAIN)) {
 		if (sort__has_parent) {
 			ui__error("Selected --sort parent, but no "
 				    "callchain data. Did you call "
@@ -272,7 +273,8 @@ static int perf_report__setup_sample_type(struct perf_report *rep)
 	}
 
 	if (sort__branch_mode == 1) {
-		if (!(self->sample_type & PERF_SAMPLE_BRANCH_STACK)) {
+		if (!self->fd_pipe &&
+		    !(self->sample_type & PERF_SAMPLE_BRANCH_STACK)) {
 			ui__error("Selected -b but no branch data. "
 				  "Did you call perf record without -b?\n");
 			return -1;

tools/perf/builtin-sched.c

@@ -43,6 +43,11 @@ static u64			sleep_measurement_overhead;
 
 static unsigned long		nr_tasks;
 
+struct perf_sched {
+	struct perf_tool    tool;
+	struct perf_session *session;
+};
+
 struct sched_atom;
 
 struct task_desc {
@@ -1597,6 +1602,8 @@ static int perf_sched__process_tracepoint_sample(struct perf_tool *tool,
 						 struct perf_evsel *evsel,
 						 struct machine *machine)
 {
+	struct perf_sched *sched = container_of(tool, struct perf_sched, tool);
+	struct pevent *pevent = sched->session->pevent;
 	struct thread *thread = machine__findnew_thread(machine, sample->pid);
 
 	if (thread == NULL) {
@@ -1612,20 +1619,23 @@ static int perf_sched__process_tracepoint_sample(struct perf_tool *tool,
 		tracepoint_handler f = evsel->handler.func;
 
 		if (evsel->handler.data == NULL)
-			evsel->handler.data = trace_find_event(evsel->attr.config);
+			evsel->handler.data = pevent_find_event(pevent,
+							  evsel->attr.config);
 
 		f(tool, evsel->handler.data, sample, machine, thread);
 	}
 
 	return 0;
 }
 
-static struct perf_tool perf_sched = {
-	.sample			= perf_sched__process_tracepoint_sample,
-	.comm			= perf_event__process_comm,
-	.lost			= perf_event__process_lost,
-	.fork			= perf_event__process_task,
-	.ordered_samples	= true,
+static struct perf_sched perf_sched = {
+	.tool = {
+		.sample		 = perf_sched__process_tracepoint_sample,
+		.comm		 = perf_event__process_comm,
+		.lost		 = perf_event__process_lost,
+		.fork		 = perf_event__process_task,
+		.ordered_samples = true,
+	},
 };
 
 static void read_events(bool destroy, struct perf_session **psession)
@@ -1640,16 +1650,20 @@ static void read_events(bool destroy, struct perf_session **psession)
 		{ "sched:sched_process_exit", process_sched_exit_event, },
 		{ "sched:sched_migrate_task", process_sched_migrate_task_event, },
 	};
-	struct perf_session *session = perf_session__new(input_name, O_RDONLY,
-							 0, false, &perf_sched);
+	struct perf_session *session;
+
+	session = perf_session__new(input_name, O_RDONLY, 0, false,
+				    &perf_sched.tool);
 	if (session == NULL)
 		die("No Memory");
 
-	err = perf_evlist__set_tracepoints_handlers_array(session->evlist, handlers);
+	perf_sched.session = session;
+
+	err = perf_session__set_tracepoints_handlers(session, handlers);
 	assert(err == 0);
 
 	if (perf_session__has_traces(session, "record -R")) {
-		err = perf_session__process_events(session, &perf_sched);
+		err = perf_session__process_events(session, &perf_sched.tool);
 		if (err)
 			die("Failed to process events, error %d", err);