[GIT PULL 00/38] perf/core improvements and fixes

[GIT PULL 00/38] perf/core improvements and fixes

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>

Hi Ingo,

	There are more patchkits outstanding that I have to revisit/retest, but
lets get this first batch in to keep pull reqs at a reasonable size.

	Please consider pulling,

- Arnaldo

The following changes since commit 1341f3e4c0276aae3de6d902c9202265d89fe438:

  Merge tag 'perf-core-for-mingo' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux into perf/core (2014-01-12 17:39:47 +0100)

are available in the git repository at:


  git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux tags/perf-core-for-mingo

for you to fetch changes up to 26f7f9877234e6b9ed87eff4ca450631bafe0182:

  perf tools: Remove unused test-volatile-register-var.c (2014-01-13 17:20:05 -0300)

----------------------------------------------------------------
New features:

. perf record: Add --initial-delay option (Andi Kleen)

. Column colouring improvements in 'diff' (Ramkumar Ramachandra)

Fixes:

. Don't show counter information when workload fails (Arnaldo Carvalho de Melo)

. Fixup leak on error path in parse events test. (Arnaldo Carvalho de Melo)

. Fix --delay option in 'stat' man page (Andi Kleen)

. Use the DWARF unwind info only if loaded (Jean Pihet):

Developer stuff:

. Improve forked workload error reporting by sending the errno in the signal
  data queueing integer field, using sigqueue and by doing the signal setup in
  the evlist methods, removing open coded equivalents in various tools. (Arnaldo Carvalho de Melo)

. Do more auto exit cleanup shores in the 'evlist' destructor, so that the tools
  don't have to all do that sequence. (Arnaldo Carvalho de Melo)

. Pack 'struct perf_session_env' and 'struct trace' (Arnaldo Carvalho de Melo)

. Include tools/lib/api/ in MANIFEST, fixing detached tarballs (Arnaldo Carvalho de Melo)

. Add test for building detached source tarballs (Arnaldo Carvalho de Melo)

. Shut up libtracevent plugins make message (Jiri Olsa)

. Fix installation tests path setup (Jiri Olsa)

. Fix id_hdr_size initialization (Jiri Olsa)

. Move some header files from tools/perf/ to tools/include/ to make them available to
  other tools/ dwelling codebases (Namhyung Kim)

. Fix 'probe' build when DWARF support libraries not present (Arnaldo Carvalho de Melo)

Refactorings:

. Move logic to warn about kptr_restrict'ed kernels to separate
  function in 'report' (Arnaldo Carvalho de Melo)

. Move hist browser selection code to separate function (Arnaldo Carvalho de Melo)

. Move histogram entries collapsing to separate function (Arnaldo Carvalho de Melo)

. Introduce evlist__for_each() & friends (Arnaldo Carvalho de Melo)

. Automate setup of FEATURE_CHECK_(C|LD)FLAGS-all variables (Jiri Olsa)

. Move arch setup into seprate Makefile (Jiri Olsa)

Trivial stuff:

. Remove misplaced __maybe_unused in 'stat' (Arnaldo Carvalho de Melo)

. Remove old evsel_list usage in 'record' (Arnaldo Carvalho de Melo)

. Comment typo fix (Cody P Schafer)

. Remove unused test-volatile-register-var.c (Yann Droneaud)

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

----------------------------------------------------------------
Andi Kleen (2):
      perf stat: Fix --delay option in man page
      perf record: Add --initial-delay option

Arnaldo Carvalho de Melo (18):
      perf stat: Don't show counter information when workload fails
      perf evlist: Send the errno in the signal when workload fails
      perf evlist: Move the SIGUSR1 error reporting logic to prepare_workload
      perf record: Remove old evsel_list usage
      perf evlist: Move destruction of maps to evlist destructor
      perf evlist: Close fds on destructor
      perf evlist: Auto unmap on destructor
      perf tests: Fixup leak on error path in parse events test
      perf stat: Remove misplaced __maybe_unused
      perf header: Pack 'struct perf_session_env'
      perf trace: Pack 'struct trace'
      perf report: Move logic to warn about kptr_restrict'ed kernels to separate function
      perf report: Move hist browser selection code to separate function
      perf report: Move histogram entries collapsing to separate function
      perf evlist: Introduce evlist__for_each() & friends
      perf tools: Include tools/lib/api/ in MANIFEST
      perf tools: Add test for building detached source tarballs
      perf probe: Fix build when DWARF support libraries not present

Cody P Schafer (1):
      tools perf: Comment typo fix

Jean Pihet (1):
      perf tools: Use the DWARF unwind info only if loaded

Jiri Olsa (7):
      perf tools: Move arch setup into seprate Makefile
      perf tests: Fix installation tests path setup
      tools lib traceevent: Replace tabs with spaces for all non-commands statements
      tools lib traceevent: Shut up plugins make message
      perf tools: Automate setup of FEATURE_CHECK_(C|LD)FLAGS-all variables
      perf machine: Fix id_hdr_size initialization
      perf tools: Make perf_event__synthesize_mmap_events global

Namhyung Kim (4):
      tools include: Move perf's linux/compiler.h to a generic place
      tools include: Define likely/unlikely in linux/compiler.h
      tools include: Move perf's bug.h to a generic place
      tools include: Include <linux/compiler.h> from asm/bug.h

Ramkumar Ramachandra (4):
      perf tools: Generalize percent_color_snprintf()
      perf diff: Color the Delta column
      perf diff: Color the Ratio column
      perf diff: Color the Weighted Diff column

Yann Droneaud (1):
      perf tools: Remove unused test-volatile-register-var.c

 tools/{perf/util => }/include/asm/bug.h            |   9 +-
 tools/{perf/util => }/include/linux/compiler.h     |  12 +-
 tools/lib/traceevent/Makefile                      |  59 ++++---
 tools/perf/Documentation/perf-record.txt           |   4 +
 tools/perf/Documentation/perf-stat.txt             |   2 +-
 tools/perf/MANIFEST                                |   4 +-
 tools/perf/Makefile.perf                           |   4 +-
 tools/perf/builtin-annotate.c                      |   2 +-
 tools/perf/builtin-diff.c                          |  98 ++++++++++-
 tools/perf/builtin-evlist.c                        |   2 +-
 tools/perf/builtin-inject.c                        |   2 +-
 tools/perf/builtin-kvm.c                           |   6 +-
 tools/perf/builtin-record.c                        |  82 +++++----
 tools/perf/builtin-report.c                        | 192 ++++++++++++---------
 tools/perf/builtin-script.c                        |   5 +-
 tools/perf/builtin-stat.c                          |  66 ++++---
 tools/perf/builtin-top.c                           |  14 +-
 tools/perf/builtin-trace.c                         |  30 ++--
 tools/perf/config/Makefile                         |  67 ++++---
 tools/perf/config/Makefile.arch                    |  22 +++
 .../feature-checks/test-volatile-register-var.c    |   6 -
 tools/perf/perf.h                                  |   1 +
 tools/perf/tests/code-reading.c                    |   7 +-
 tools/perf/tests/evsel-roundtrip-name.c            |   2 +-
 tools/perf/tests/hists_link.c                      |   4 +-
 tools/perf/tests/keep-tracking.c                   |   7 +-
 tools/perf/tests/make                              |  40 +++--
 tools/perf/tests/mmap-basic.c                      |  25 ++-
 tools/perf/tests/open-syscall-tp-fields.c          |  16 +-
 tools/perf/tests/parse-events.c                    |  10 +-
 tools/perf/tests/perf-record.c                     |  21 +--
 tools/perf/tests/perf-targz-src-pkg                |  21 +++
 tools/perf/tests/perf-time-to-tsc.c                |   6 -
 tools/perf/tests/sw-clock.c                        |  18 +-
 tools/perf/tests/task-exit.c                       |  33 ++--
 tools/perf/ui/browsers/hists.c                     |   5 +-
 tools/perf/ui/gtk/hists.c                          |   2 +-
 tools/perf/util/cgroup.c                           |   4 +-
 tools/perf/util/color.c                            |  15 +-
 tools/perf/util/color.h                            |   1 +
 tools/perf/util/event.c                            |  12 +-
 tools/perf/util/event.h                            |   7 +
 tools/perf/util/evlist.c                           |  78 +++++----
 tools/perf/util/evlist.h                           |  69 +++++++-
 tools/perf/util/evsel.c                            |   3 +-
 tools/perf/util/header.c                           |  19 +-
 tools/perf/util/header.h                           |  10 +-
 tools/perf/util/machine.c                          |   1 +
 tools/perf/util/parse-events.c                     |   5 +-
 tools/perf/util/pmu.c                              |   2 +-
 tools/perf/util/probe-event.c                      |   5 +-
 tools/perf/util/python.c                           |   3 +-
 tools/perf/util/record.c                           |   6 +-
 tools/perf/util/session.c                          |   6 +-
 tools/perf/util/unwind.c                           |   8 +-
 55 files changed, 724 insertions(+), 436 deletions(-)
 rename tools/{perf/util => }/include/asm/bug.h (81%)
 rename tools/{perf/util => }/include/linux/compiler.h (64%)
 create mode 100644 tools/perf/config/Makefile.arch
 delete mode 100644 tools/perf/config/feature-checks/test-volatile-register-var.c
 create mode 100755 tools/perf/tests/perf-targz-src-pkg

[PATCH 01/38] perf stat: Don't show counter information when workload fails

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

When starting a workload 'stat' wasn't using prepare_workload evlist
method's signal based exec() error reporting mechanism.

Use it so that the we don't report 'not counted' counters.

Before:

  [acme@zoo linux]$ perf stat dfadsfa
  dfadsfa: No such file or directory

   Performance counter stats for 'dfadsfa':

       <not counted>      task-clock
       <not counted>      context-switches
       <not counted>      cpu-migrations
       <not counted>      page-faults
       <not counted>      cycles
       <not counted>      stalled-cycles-frontend
     <not supported>      stalled-cycles-backend
       <not counted>      instructions
       <not counted>      branches
       <not counted>      branch-misses

         0.001831462 seconds time elapsed

  [acme@zoo linux]$

After:

  [acme@zoo linux]$ perf stat dfadsfa
  dfadsfa: No such file or directory
  [acme@zoo linux]$

Reported-by: David Ahern <dsahern@gmail.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-5yui3bv7e3hitxucnjsn6z8q@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 106a5e5b7842..1c76c7a66f78 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -509,6 +509,18 @@ static void handle_initial_delay(void)
 	}
 }
 
+static volatile bool workload_exec_failed;
+
+/*
+ * perf_evlist__prepare_workload will send a SIGUSR1
+ * if the fork fails, since we asked by setting its
+ * want_signal to true.
+ */
+static void workload_exec_failed_signal(int signo __maybe_unused)
+{
+	workload_exec_failed = true;
+}
+
 static int __run_perf_stat(int argc, const char **argv)
 {
 	char msg[512];
@@ -529,7 +541,7 @@ static int __run_perf_stat(int argc, const char **argv)
 
 	if (forks) {
 		if (perf_evlist__prepare_workload(evsel_list, &target, argv,
-						  false, false) < 0) {
+						  false, true) < 0) {
 			perror("failed to prepare workload");
 			return -1;
 		}
@@ -584,6 +596,14 @@ static int __run_perf_stat(int argc, const char **argv)
 	clock_gettime(CLOCK_MONOTONIC, &ref_time);
 
 	if (forks) {
+		/*
+		 * perf_evlist__prepare_workload will, after we call
+		 * perf_evlist__start_Workload, send a SIGUSR1 if the exec call
+		 * fails, that we will catch in workload_signal to flip
+		 * workload_exec_failed.
+ 		 */
+		signal(SIGUSR1, workload_exec_failed_signal);
+
 		perf_evlist__start_workload(evsel_list);
 		handle_initial_delay();
 
@@ -594,6 +614,10 @@ static int __run_perf_stat(int argc, const char **argv)
 			}
 		}
 		wait(&status);
+
+		if (workload_exec_failed)
+			return -1;
+
 		if (WIFSIGNALED(status))
 			psignal(WTERMSIG(status), argv[0]);
 	} else {
-- 
1.8.1.4

[PATCH 02/38] perf evlist: Send the errno in the signal when workload fails

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

When a tool uses perf_evlist__start_workload and the supplied workload
fails (e.g.: its binary wasn't found), perror was being used to print
the error reason.

This is undesirable, as the caller may be a GUI, when it wants to have
total control of the error reporting process.

So move to using sigaction(SA_SIGINFO) + siginfo_t->sa_value->sival_int
to communicate to the caller the errno and let it print it using the UI
of its choosing.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-epgcv7kjq8ll2udqfken92pz@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c | 39 +++++++++++++++++++++++++++++++++++++--
 tools/perf/builtin-stat.c   | 21 +++++++++++++++------
 tools/perf/util/evlist.c    | 11 ++++++++---
 3 files changed, 60 insertions(+), 11 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 6ec0cbc2a5d5..f987d385c6f0 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -341,6 +341,22 @@ static void record__init_features(struct record *rec)
 		perf_header__clear_feat(&session->header, HEADER_BRANCH_STACK);
 }
 
+static volatile int workload_exec_errno;
+
+/*
+ * perf_evlist__prepare_workload will send a SIGUSR1
+ * if the fork fails, since we asked by setting its
+ * want_signal to true.
+ */
+static void workload_exec_failed_signal(int signo, siginfo_t *info,
+					void *ucontext __maybe_unused)
+{
+	workload_exec_errno = info->si_value.sival_int;
+	done = 1;
+	signr = signo;
+	child_finished = 1;
+}
+
 static int __cmd_record(struct record *rec, int argc, const char **argv)
 {
 	int err;
@@ -359,7 +375,6 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	on_exit(record__sig_exit, rec);
 	signal(SIGCHLD, sig_handler);
 	signal(SIGINT, sig_handler);
-	signal(SIGUSR1, sig_handler);
 	signal(SIGTERM, sig_handler);
 
 	session = perf_session__new(file, false, NULL);
@@ -492,8 +507,20 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	/*
 	 * Let the child rip
 	 */
-	if (forks)
+	if (forks) {
+		struct sigaction act = {
+			.sa_flags     = SA_SIGINFO,
+			.sa_sigaction = workload_exec_failed_signal,
+		};
+		/*
+		 * perf_evlist__prepare_workload will, after we call
+		 * perf_evlist__start_Workload, send a SIGUSR1 if the exec call
+		 * fails, that we will catch in workload_signal to flip
+		 * workload_exec_errno.
+ 		 */
+		sigaction(SIGUSR1, &act, NULL);
 		perf_evlist__start_workload(evsel_list);
+	}
 
 	for (;;) {
 		int hits = rec->samples;
@@ -521,6 +548,14 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		}
 	}
 
+	if (forks && workload_exec_errno) {
+		char msg[512];
+		const char *emsg = strerror_r(workload_exec_errno, msg, sizeof(msg));
+		pr_err("Workload failed: %s\n", emsg);
+		err = -1;
+		goto out_delete_session;
+	}
+
 	if (quiet || signr == SIGUSR1)
 		return 0;
 
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 1c76c7a66f78..9d0d52d55ee6 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -58,6 +58,7 @@
 #include "util/thread.h"
 #include "util/thread_map.h"
 
+#include <signal.h>
 #include <stdlib.h>
 #include <sys/prctl.h>
 #include <locale.h>
@@ -509,16 +510,17 @@ static void handle_initial_delay(void)
 	}
 }
 
-static volatile bool workload_exec_failed;
+static volatile int workload_exec_errno;
 
 /*
  * perf_evlist__prepare_workload will send a SIGUSR1
  * if the fork fails, since we asked by setting its
  * want_signal to true.
  */
-static void workload_exec_failed_signal(int signo __maybe_unused)
+static void workload_exec_failed_signal(int signo __maybe_unused, siginfo_t *info,
+					void *ucontext __maybe_unused)
 {
-	workload_exec_failed = true;
+	workload_exec_errno = info->si_value.sival_int;
 }
 
 static int __run_perf_stat(int argc, const char **argv)
@@ -596,13 +598,17 @@ static int __run_perf_stat(int argc, const char **argv)
 	clock_gettime(CLOCK_MONOTONIC, &ref_time);
 
 	if (forks) {
+		struct sigaction act = {
+			.sa_flags     = SA_SIGINFO,
+			.sa_sigaction = workload_exec_failed_signal,
+		};
 		/*
 		 * perf_evlist__prepare_workload will, after we call
 		 * perf_evlist__start_Workload, send a SIGUSR1 if the exec call
 		 * fails, that we will catch in workload_signal to flip
-		 * workload_exec_failed.
+		 * workload_exec_errno.
  		 */
-		signal(SIGUSR1, workload_exec_failed_signal);
+		sigaction(SIGUSR1, &act, NULL);
 
 		perf_evlist__start_workload(evsel_list);
 		handle_initial_delay();
@@ -615,8 +621,11 @@ static int __run_perf_stat(int argc, const char **argv)
 		}
 		wait(&status);
 
-		if (workload_exec_failed)
+		if (workload_exec_errno) {
+			const char *emsg = strerror_r(workload_exec_errno, msg, sizeof(msg));
+			pr_err("Workload failed: %s\n", emsg);
 			return -1;
+		}
 
 		if (WIFSIGNALED(status))
 			psignal(WTERMSIG(status), argv[0]);
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index b08a7ecdcea1..4a30c87d24ec 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -1073,9 +1073,14 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist, struct target *tar
 
 		execvp(argv[0], (char **)argv);
 
-		perror(argv[0]);
-		if (want_signal)
-			kill(getppid(), SIGUSR1);
+		if (want_signal) {
+			union sigval val;
+
+			val.sival_int = errno;
+			if (sigqueue(getppid(), SIGUSR1, val))
+				perror(argv[0]);
+		} else
+			perror(argv[0]);
 		exit(-1);
 	}
 
-- 
1.8.1.4

[PATCH 03/38] perf evlist: Move the SIGUSR1 error reporting logic to prepare_workload

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

So that we have the boilerplate in the preparation method, instead of
open coded in tools wanting the reporting when the exec fails.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-purbdzcphdveskh7wwmnm4t7@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c    | 16 ++--------------
 tools/perf/builtin-stat.c      | 17 ++---------------
 tools/perf/builtin-trace.c     |  2 +-
 tools/perf/tests/perf-record.c |  3 +--
 tools/perf/tests/task-exit.c   | 19 ++++++++++++++-----
 tools/perf/util/evlist.c       | 12 ++++++++++--
 tools/perf/util/evlist.h       |  3 ++-
 7 files changed, 32 insertions(+), 40 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index f987d385c6f0..ea7c3060e8e7 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -390,7 +390,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	if (forks) {
 		err = perf_evlist__prepare_workload(evsel_list, &opts->target,
 						    argv, file->is_pipe,
-						    true);
+						    workload_exec_failed_signal);
 		if (err < 0) {
 			pr_err("Couldn't run the workload!\n");
 			goto out_delete_session;
@@ -507,20 +507,8 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	/*
 	 * Let the child rip
 	 */
-	if (forks) {
-		struct sigaction act = {
-			.sa_flags     = SA_SIGINFO,
-			.sa_sigaction = workload_exec_failed_signal,
-		};
-		/*
-		 * perf_evlist__prepare_workload will, after we call
-		 * perf_evlist__start_Workload, send a SIGUSR1 if the exec call
-		 * fails, that we will catch in workload_signal to flip
-		 * workload_exec_errno.
- 		 */
-		sigaction(SIGUSR1, &act, NULL);
+	if (forks)
 		perf_evlist__start_workload(evsel_list);
-	}
 
 	for (;;) {
 		int hits = rec->samples;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 9d0d52d55ee6..f8456cad656d 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -58,7 +58,6 @@
 #include "util/thread.h"
 #include "util/thread_map.h"
 
-#include <signal.h>
 #include <stdlib.h>
 #include <sys/prctl.h>
 #include <locale.h>
@@ -542,8 +541,8 @@ static int __run_perf_stat(int argc, const char **argv)
 	}
 
 	if (forks) {
-		if (perf_evlist__prepare_workload(evsel_list, &target, argv,
-						  false, true) < 0) {
+		if (perf_evlist__prepare_workload(evsel_list, &target, argv, false,
+						  workload_exec_failed_signal) < 0) {
 			perror("failed to prepare workload");
 			return -1;
 		}
@@ -598,18 +597,6 @@ static int __run_perf_stat(int argc, const char **argv)
 	clock_gettime(CLOCK_MONOTONIC, &ref_time);
 
 	if (forks) {
-		struct sigaction act = {
-			.sa_flags     = SA_SIGINFO,
-			.sa_sigaction = workload_exec_failed_signal,
-		};
-		/*
-		 * perf_evlist__prepare_workload will, after we call
-		 * perf_evlist__start_Workload, send a SIGUSR1 if the exec call
-		 * fails, that we will catch in workload_signal to flip
-		 * workload_exec_errno.
- 		 */
-		sigaction(SIGUSR1, &act, NULL);
-
 		perf_evlist__start_workload(evsel_list);
 		handle_initial_delay();
 
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index c5b4bc51175c..5498eacf8fc6 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1895,7 +1895,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 
 	if (forks) {
 		err = perf_evlist__prepare_workload(evlist, &trace->opts.target,
-						    argv, false, false);
+						    argv, false, NULL);
 		if (err < 0) {
 			fprintf(trace->output, "Couldn't run the workload!\n");
 			goto out_delete_maps;
diff --git a/tools/perf/tests/perf-record.c b/tools/perf/tests/perf-record.c
index eeba562920e9..fa0ed35afb9a 100644
--- a/tools/perf/tests/perf-record.c
+++ b/tools/perf/tests/perf-record.c
@@ -83,8 +83,7 @@ int test__PERF_RECORD(void)
 	 * so that we have time to open the evlist (calling sys_perf_event_open
 	 * on all the fds) and then mmap them.
 	 */
-	err = perf_evlist__prepare_workload(evlist, &opts.target, argv,
-					    false, false);
+	err = perf_evlist__prepare_workload(evlist, &opts.target, argv, false, NULL);
 	if (err < 0) {
 		pr_debug("Couldn't run the workload!\n");
 		goto out_delete_maps;
diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c
index d09ab579119e..44e339d4e297 100644
--- a/tools/perf/tests/task-exit.c
+++ b/tools/perf/tests/task-exit.c
@@ -9,12 +9,21 @@
 static int exited;
 static int nr_exit;
 
-static void sig_handler(int sig)
+static void sig_handler(int sig __maybe_unused)
 {
 	exited = 1;
+}
 
-	if (sig == SIGUSR1)
-		nr_exit = -1;
+/*
+ * perf_evlist__prepare_workload will send a SIGUSR1 if the fork fails, since
+ * we asked by setting its exec_error to this handler.
+ */
+static void workload_exec_failed_signal(int signo __maybe_unused,
+					siginfo_t *info __maybe_unused,
+					void *ucontext __maybe_unused)
+{
+	exited	= 1;
+	nr_exit = -1;
 }
 
 /*
@@ -35,7 +44,6 @@ int test__task_exit(void)
 	const char *argv[] = { "true", NULL };
 
 	signal(SIGCHLD, sig_handler);
-	signal(SIGUSR1, sig_handler);
 
 	evlist = perf_evlist__new_default();
 	if (evlist == NULL) {
@@ -57,7 +65,8 @@ int test__task_exit(void)
 		goto out_delete_maps;
 	}
 
-	err = perf_evlist__prepare_workload(evlist, &target, argv, false, true);
+	err = perf_evlist__prepare_workload(evlist, &target, argv, false,
+					    workload_exec_failed_signal);
 	if (err < 0) {
 		pr_debug("Couldn't run the workload!\n");
 		goto out_delete_maps;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 4a30c87d24ec..96b3ef547db4 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -1029,7 +1029,7 @@ out_err:
 
 int perf_evlist__prepare_workload(struct perf_evlist *evlist, struct target *target,
 				  const char *argv[], bool pipe_output,
-				  bool want_signal)
+				  void (*exec_error)(int signo, siginfo_t *info, void *ucontext))
 {
 	int child_ready_pipe[2], go_pipe[2];
 	char bf;
@@ -1073,7 +1073,7 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist, struct target *tar
 
 		execvp(argv[0], (char **)argv);
 
-		if (want_signal) {
+		if (exec_error) {
 			union sigval val;
 
 			val.sival_int = errno;
@@ -1084,6 +1084,14 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist, struct target *tar
 		exit(-1);
 	}
 
+	if (exec_error) {
+		struct sigaction act = {
+			.sa_flags     = SA_SIGINFO,
+			.sa_sigaction = exec_error,
+		};
+		sigaction(SIGUSR1, &act, NULL);
+	}
+
 	if (target__none(target))
 		evlist->threads->map[0] = evlist->workload.pid;
 
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 2fe51958ed85..18d1222c0762 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -103,7 +103,8 @@ int record_opts__config(struct record_opts *opts);
 int perf_evlist__prepare_workload(struct perf_evlist *evlist,
 				  struct target *target,
 				  const char *argv[], bool pipe_output,
-				  bool want_signal);
+				  void (*exec_error)(int signo, siginfo_t *info,
+						     void *ucontext));
 int perf_evlist__start_workload(struct perf_evlist *evlist);
 
 int perf_evlist__parse_mmap_pages(const struct option *opt,
-- 
1.8.1.4

[PATCH 04/38] perf record: Remove old evsel_list usage

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

To be consistent with other places, use just 'evlist' for the evsel list
variable, and since we have it in 'struct record', use it directly from
there.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-396bnfvmlxrsj3o2tk47b8t1@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c | 43 +++++++++++++++++++------------------------
 1 file changed, 19 insertions(+), 24 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index ea7c3060e8e7..6dcb8aae908c 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -324,7 +324,6 @@ out:
 
 static void record__init_features(struct record *rec)
 {
-	struct perf_evlist *evsel_list = rec->evlist;
 	struct perf_session *session = rec->session;
 	int feat;
 
@@ -334,7 +333,7 @@ static void record__init_features(struct record *rec)
 	if (rec->no_buildid)
 		perf_header__clear_feat(&session->header, HEADER_BUILD_ID);
 
-	if (!have_tracepoints(&evsel_list->entries))
+	if (!have_tracepoints(&rec->evlist->entries))
 		perf_header__clear_feat(&session->header, HEADER_TRACING_DATA);
 
 	if (!rec->opts.branch_stack)
@@ -365,7 +364,6 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	struct machine *machine;
 	struct perf_tool *tool = &rec->tool;
 	struct record_opts *opts = &rec->opts;
-	struct perf_evlist *evsel_list = rec->evlist;
 	struct perf_data_file *file = &rec->file;
 	struct perf_session *session;
 	bool disabled = false;
@@ -388,7 +386,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	record__init_features(rec);
 
 	if (forks) {
-		err = perf_evlist__prepare_workload(evsel_list, &opts->target,
+		err = perf_evlist__prepare_workload(rec->evlist, &opts->target,
 						    argv, file->is_pipe,
 						    workload_exec_failed_signal);
 		if (err < 0) {
@@ -402,7 +400,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		goto out_delete_session;
 	}
 
-	if (!evsel_list->nr_groups)
+	if (!rec->evlist->nr_groups)
 		perf_header__clear_feat(&session->header, HEADER_GROUP_DESC);
 
 	/*
@@ -415,7 +413,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		if (err < 0)
 			goto out_delete_session;
 	} else {
-		err = perf_session__write_header(session, evsel_list,
+		err = perf_session__write_header(session, rec->evlist,
 						 file->fd, false);
 		if (err < 0)
 			goto out_delete_session;
@@ -439,7 +437,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 			goto out_delete_session;
 		}
 
-		if (have_tracepoints(&evsel_list->entries)) {
+		if (have_tracepoints(&rec->evlist->entries)) {
 			/*
 			 * FIXME err <= 0 here actually means that
 			 * there were no tracepoints so its not really
@@ -448,7 +446,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 			 * return this more properly and also
 			 * propagate errors that now are calling die()
 			 */
-			err = perf_event__synthesize_tracing_data(tool, file->fd, evsel_list,
+			err = perf_event__synthesize_tracing_data(tool, file->fd, rec->evlist,
 								  process_synthesized_event);
 			if (err <= 0) {
 				pr_err("Couldn't record tracing data.\n");
@@ -480,7 +478,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 					 perf_event__synthesize_guest_os, tool);
 	}
 
-	err = __machine__synthesize_threads(machine, tool, &opts->target, evsel_list->threads,
+	err = __machine__synthesize_threads(machine, tool, &opts->target, rec->evlist->threads,
 					    process_synthesized_event, opts->sample_address);
 	if (err != 0)
 		goto out_delete_session;
@@ -502,13 +500,13 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	 * so don't spoil it by prematurely enabling them.
 	 */
 	if (!target__none(&opts->target))
-		perf_evlist__enable(evsel_list);
+		perf_evlist__enable(rec->evlist);
 
 	/*
 	 * Let the child rip
 	 */
 	if (forks)
-		perf_evlist__start_workload(evsel_list);
+		perf_evlist__start_workload(rec->evlist);
 
 	for (;;) {
 		int hits = rec->samples;
@@ -521,7 +519,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		if (hits == rec->samples) {
 			if (done)
 				break;
-			err = poll(evsel_list->pollfd, evsel_list->nr_fds, -1);
+			err = poll(rec->evlist->pollfd, rec->evlist->nr_fds, -1);
 			waking++;
 		}
 
@@ -531,7 +529,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 		 * disable events in this case.
 		 */
 		if (done && !disabled && !target__none(&opts->target)) {
-			perf_evlist__disable(evsel_list);
+			perf_evlist__disable(rec->evlist);
 			disabled = true;
 		}
 	}
@@ -901,16 +899,13 @@ const struct option record_options[] = {
 int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 {
 	int err = -ENOMEM;
-	struct perf_evlist *evsel_list;
 	struct record *rec = &record;
 	char errbuf[BUFSIZ];
 
-	evsel_list = perf_evlist__new();
-	if (evsel_list == NULL)
+	rec->evlist = perf_evlist__new();
+	if (rec->evlist == NULL)
 		return -ENOMEM;
 
-	rec->evlist = evsel_list;
-
 	argc = parse_options(argc, argv, record_options, record_usage,
 			    PARSE_OPT_STOP_AT_NON_OPTION);
 	if (!argc && target__none(&rec->opts.target))
@@ -937,8 +932,8 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 	if (rec->no_buildid_cache || rec->no_buildid)
 		disable_buildid_cache();
 
-	if (evsel_list->nr_entries == 0 &&
-	    perf_evlist__add_default(evsel_list) < 0) {
+	if (rec->evlist->nr_entries == 0 &&
+	    perf_evlist__add_default(rec->evlist) < 0) {
 		pr_err("Not enough memory for event selector list\n");
 		goto out_symbol_exit;
 	}
@@ -964,7 +959,7 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 	}
 
 	err = -ENOMEM;
-	if (perf_evlist__create_maps(evsel_list, &rec->opts.target) < 0)
+	if (perf_evlist__create_maps(rec->evlist, &rec->opts.target) < 0)
 		usage_with_options(record_usage, record_options);
 
 	if (record_opts__config(&rec->opts)) {
@@ -974,10 +969,10 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	err = __cmd_record(&record, argc, argv);
 
-	perf_evlist__munmap(evsel_list);
-	perf_evlist__close(evsel_list);
+	perf_evlist__munmap(rec->evlist);
+	perf_evlist__close(rec->evlist);
 out_free_fd:
-	perf_evlist__delete_maps(evsel_list);
+	perf_evlist__delete_maps(rec->evlist);
 out_symbol_exit:
 	symbol__exit();
 	return err;
-- 
1.8.1.4

[PATCH 05/38] perf evlist: Move destruction of maps to evlist destructor

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Instead of requiring tools to do an extra destructor call just before
calling perf_evlist__delete.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-0jd2ptzyikxb5wp7inzz2ah2@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-kvm.c                  |  4 +---
 tools/perf/builtin-record.c               |  4 +---
 tools/perf/builtin-stat.c                 |  8 +++-----
 tools/perf/builtin-top.c                  |  6 ++----
 tools/perf/builtin-trace.c                |  6 ++----
 tools/perf/tests/code-reading.c           |  5 ++---
 tools/perf/tests/keep-tracking.c          |  5 ++---
 tools/perf/tests/mmap-basic.c             |  2 ++
 tools/perf/tests/open-syscall-tp-fields.c |  4 +---
 tools/perf/tests/perf-record.c            | 10 ++++------
 tools/perf/tests/perf-time-to-tsc.c       |  4 ----
 tools/perf/tests/sw-clock.c               | 10 ++++------
 tools/perf/tests/task-exit.c              |  9 ++++-----
 tools/perf/util/evlist.c                  | 12 ++++--------
 tools/perf/util/evlist.h                  |  1 -
 15 files changed, 32 insertions(+), 58 deletions(-)

diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index a6ec1052c291..858b11bf6711 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -1556,10 +1556,8 @@ out:
 	if (kvm->session)
 		perf_session__delete(kvm->session);
 	kvm->session = NULL;
-	if (kvm->evlist) {
-		perf_evlist__delete_maps(kvm->evlist);
+	if (kvm->evlist)
 		perf_evlist__delete(kvm->evlist);
-	}
 
 	return err;
 }
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 6dcb8aae908c..5149b41e63d1 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -964,15 +964,13 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	if (record_opts__config(&rec->opts)) {
 		err = -EINVAL;
-		goto out_free_fd;
+		goto out_symbol_exit;
 	}
 
 	err = __cmd_record(&record, argc, argv);
 
 	perf_evlist__munmap(rec->evlist);
 	perf_evlist__close(rec->evlist);
-out_free_fd:
-	perf_evlist__delete_maps(rec->evlist);
 out_symbol_exit:
 	symbol__exit();
 	return err;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index f8456cad656d..6ca076660de5 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -1782,14 +1782,14 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 	if (interval && interval < 100) {
 		pr_err("print interval must be >= 100ms\n");
 		parse_options_usage(stat_usage, options, "I", 1);
-		goto out_free_maps;
+		goto out;
 	}
 
 	if (perf_evlist__alloc_stats(evsel_list, interval))
-		goto out_free_maps;
+		goto out;
 
 	if (perf_stat_init_aggr_mode())
-		goto out_free_maps;
+		goto out;
 
 	/*
 	 * We dont want to block the signals - that would cause
@@ -1821,8 +1821,6 @@ int cmd_stat(int argc, const char **argv, const char *prefix __maybe_unused)
 		print_stat(argc, argv);
 
 	perf_evlist__free_stats(evsel_list);
-out_free_maps:
-	perf_evlist__delete_maps(evsel_list);
 out:
 	perf_evlist__delete(evsel_list);
 	return status;
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 172e91a9ce62..e0fd0aa57f06 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1171,7 +1171,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 	if (!top.evlist->nr_entries &&
 	    perf_evlist__add_default(top.evlist) < 0) {
 		ui__error("Not enough memory for event selector list\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	symbol_conf.nr_events = top.evlist->nr_entries;
@@ -1181,7 +1181,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	if (record_opts__config(opts)) {
 		status = -EINVAL;
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	top.sym_evsel = perf_evlist__first(top.evlist);
@@ -1206,8 +1206,6 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	status = __cmd_top(&top);
 
-out_delete_maps:
-	perf_evlist__delete_maps(top.evlist);
 out_delete_evlist:
 	perf_evlist__delete(top.evlist);
 
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 5498eacf8fc6..f4ddd14b8a9b 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1885,7 +1885,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 	err = trace__symbols_init(trace, evlist);
 	if (err < 0) {
 		fprintf(trace->output, "Problems initializing symbol libraries!\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	perf_evlist__config(evlist, &trace->opts);
@@ -1898,7 +1898,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 						    argv, false, NULL);
 		if (err < 0) {
 			fprintf(trace->output, "Couldn't run the workload!\n");
-			goto out_delete_maps;
+			goto out_delete_evlist;
 		}
 	}
 
@@ -1996,8 +1996,6 @@ out_disable:
 	perf_evlist__munmap(evlist);
 out_close_evlist:
 	perf_evlist__close(evlist);
-out_delete_maps:
-	perf_evlist__delete_maps(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/code-reading.c b/tools/perf/tests/code-reading.c
index 4248d1e96848..ddbc7756620c 100644
--- a/tools/perf/tests/code-reading.c
+++ b/tools/perf/tests/code-reading.c
@@ -543,11 +543,10 @@ out_err:
 		perf_evlist__munmap(evlist);
 		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
-	}
-	if (cpus)
+	} else {
 		cpu_map__delete(cpus);
-	if (threads)
 		thread_map__delete(threads);
+	}
 	machines__destroy_kernel_maps(&machines);
 	machine__delete_threads(machine);
 	machines__exit(&machines);
diff --git a/tools/perf/tests/keep-tracking.c b/tools/perf/tests/keep-tracking.c
index 27eb75142b88..f9bc1fcd17cd 100644
--- a/tools/perf/tests/keep-tracking.c
+++ b/tools/perf/tests/keep-tracking.c
@@ -145,11 +145,10 @@ out_err:
 		perf_evlist__munmap(evlist);
 		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
-	}
-	if (cpus)
+	} else {
 		cpu_map__delete(cpus);
-	if (threads)
 		thread_map__delete(threads);
+	}
 
 	return err;
 }
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index d64ab79c6d35..bbb334d2a864 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -143,6 +143,8 @@ out_close_fd:
 		perf_evsel__close_fd(evsels[i], 1, threads->nr);
 out_free_evlist:
 	perf_evlist__delete(evlist);
+	cpus	= NULL;
+	threads = NULL;
 out_free_cpus:
 	cpu_map__delete(cpus);
 out_free_threads:
diff --git a/tools/perf/tests/open-syscall-tp-fields.c b/tools/perf/tests/open-syscall-tp-fields.c
index 774620a5aecb..595b577172a6 100644
--- a/tools/perf/tests/open-syscall-tp-fields.c
+++ b/tools/perf/tests/open-syscall-tp-fields.c
@@ -48,7 +48,7 @@ int test__syscall_open_tp_fields(void)
 	err = perf_evlist__open(evlist);
 	if (err < 0) {
 		pr_debug("perf_evlist__open: %s\n", strerror(errno));
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	err = perf_evlist__mmap(evlist, UINT_MAX, false);
@@ -114,8 +114,6 @@ out_munmap:
 	perf_evlist__munmap(evlist);
 out_close_evlist:
 	perf_evlist__close(evlist);
-out_delete_maps:
-	perf_evlist__delete_maps(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-record.c b/tools/perf/tests/perf-record.c
index fa0ed35afb9a..266da9d41300 100644
--- a/tools/perf/tests/perf-record.c
+++ b/tools/perf/tests/perf-record.c
@@ -86,7 +86,7 @@ int test__PERF_RECORD(void)
 	err = perf_evlist__prepare_workload(evlist, &opts.target, argv, false, NULL);
 	if (err < 0) {
 		pr_debug("Couldn't run the workload!\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	/*
@@ -101,7 +101,7 @@ int test__PERF_RECORD(void)
 	err = sched__get_first_possible_cpu(evlist->workload.pid, &cpu_mask);
 	if (err < 0) {
 		pr_debug("sched__get_first_possible_cpu: %s\n", strerror(errno));
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	cpu = err;
@@ -111,7 +111,7 @@ int test__PERF_RECORD(void)
 	 */
 	if (sched_setaffinity(evlist->workload.pid, cpu_mask_size, &cpu_mask) < 0) {
 		pr_debug("sched_setaffinity: %s\n", strerror(errno));
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	/*
@@ -121,7 +121,7 @@ int test__PERF_RECORD(void)
 	err = perf_evlist__open(evlist);
 	if (err < 0) {
 		pr_debug("perf_evlist__open: %s\n", strerror(errno));
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	/*
@@ -306,8 +306,6 @@ out_err:
 	perf_evlist__munmap(evlist);
 out_close_evlist:
 	perf_evlist__close(evlist);
-out_delete_maps:
-	perf_evlist__delete_maps(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-time-to-tsc.c b/tools/perf/tests/perf-time-to-tsc.c
index c6398b90e897..97d08fff1f8d 100644
--- a/tools/perf/tests/perf-time-to-tsc.c
+++ b/tools/perf/tests/perf-time-to-tsc.c
@@ -170,10 +170,6 @@ out_err:
 		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
 	}
-	if (cpus)
-		cpu_map__delete(cpus);
-	if (threads)
-		thread_map__delete(threads);
 
 	return err;
 }
diff --git a/tools/perf/tests/sw-clock.c b/tools/perf/tests/sw-clock.c
index 6664a7cd828c..266d38127ba6 100644
--- a/tools/perf/tests/sw-clock.c
+++ b/tools/perf/tests/sw-clock.c
@@ -45,7 +45,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
 	evsel = perf_evsel__new(&attr);
 	if (evsel == NULL) {
 		pr_debug("perf_evsel__new\n");
-		goto out_free_evlist;
+		goto out_delete_evlist;
 	}
 	perf_evlist__add(evlist, evsel);
 
@@ -54,7 +54,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
 	if (!evlist->cpus || !evlist->threads) {
 		err = -ENOMEM;
 		pr_debug("Not enough memory to create thread/cpu maps\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	if (perf_evlist__open(evlist)) {
@@ -63,7 +63,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
 		err = -errno;
 		pr_debug("Couldn't open evlist: %s\nHint: check %s, using %" PRIu64 " in this test.\n",
 			 strerror(errno), knob, (u64)attr.sample_freq);
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	err = perf_evlist__mmap(evlist, 128, true);
@@ -109,9 +109,7 @@ out_unmap_evlist:
 	perf_evlist__munmap(evlist);
 out_close_evlist:
 	perf_evlist__close(evlist);
-out_delete_maps:
-	perf_evlist__delete_maps(evlist);
-out_free_evlist:
+out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
 }
diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c
index 44e339d4e297..fdeb2aa46a50 100644
--- a/tools/perf/tests/task-exit.c
+++ b/tools/perf/tests/task-exit.c
@@ -62,14 +62,14 @@ int test__task_exit(void)
 	if (!evlist->cpus || !evlist->threads) {
 		err = -ENOMEM;
 		pr_debug("Not enough memory to create thread/cpu maps\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	err = perf_evlist__prepare_workload(evlist, &target, argv, false,
 					    workload_exec_failed_signal);
 	if (err < 0) {
 		pr_debug("Couldn't run the workload!\n");
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	evsel = perf_evlist__first(evlist);
@@ -83,7 +83,7 @@ int test__task_exit(void)
 	err = perf_evlist__open(evlist);
 	if (err < 0) {
 		pr_debug("Couldn't open the evlist: %s\n", strerror(-err));
-		goto out_delete_maps;
+		goto out_delete_evlist;
 	}
 
 	if (perf_evlist__mmap(evlist, 128, true) < 0) {
@@ -115,8 +115,7 @@ retry:
 	perf_evlist__munmap(evlist);
 out_close_evlist:
 	perf_evlist__close(evlist);
-out_delete_maps:
-	perf_evlist__delete_maps(evlist);
+out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
 }
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 96b3ef547db4..143eaf05178f 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -107,6 +107,10 @@ void perf_evlist__exit(struct perf_evlist *evlist)
 
 void perf_evlist__delete(struct perf_evlist *evlist)
 {
+	cpu_map__delete(evlist->cpus);
+	thread_map__delete(evlist->threads);
+	evlist->cpus = NULL;
+	evlist->threads = NULL;
 	perf_evlist__purge(evlist);
 	perf_evlist__exit(evlist);
 	free(evlist);
@@ -833,14 +837,6 @@ out_delete_threads:
 	return -1;
 }
 
-void perf_evlist__delete_maps(struct perf_evlist *evlist)
-{
-	cpu_map__delete(evlist->cpus);
-	thread_map__delete(evlist->threads);
-	evlist->cpus	= NULL;
-	evlist->threads = NULL;
-}
-
 int perf_evlist__apply_filters(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 18d1222c0762..518e521c801d 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -135,7 +135,6 @@ static inline void perf_evlist__set_maps(struct perf_evlist *evlist,
 }
 
 int perf_evlist__create_maps(struct perf_evlist *evlist, struct target *target);
-void perf_evlist__delete_maps(struct perf_evlist *evlist);
 int perf_evlist__apply_filters(struct perf_evlist *evlist);
 
 void __perf_evlist__set_leader(struct list_head *list);
-- 
1.8.1.4

[PATCH 06/38] perf evlist: Close fds on destructor

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Since it is safe to call perf_evlist__close() multiple times, autoclose
it and remove the calls to the close from existing tools, reducing the
tooling boilerplate.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-2kq9v7p1rude1tqxa0aue2tk@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c               |  1 -
 tools/perf/builtin-trace.c                |  4 +---
 tools/perf/tests/code-reading.c           |  1 -
 tools/perf/tests/keep-tracking.c          |  1 -
 tools/perf/tests/mmap-basic.c             | 11 ++++-------
 tools/perf/tests/open-syscall-tp-fields.c |  4 +---
 tools/perf/tests/perf-record.c            |  4 +---
 tools/perf/tests/perf-time-to-tsc.c       |  1 -
 tools/perf/tests/sw-clock.c               |  4 +---
 tools/perf/tests/task-exit.c              |  4 +---
 tools/perf/util/evlist.c                  |  1 +
 11 files changed, 10 insertions(+), 26 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 5149b41e63d1..b7f5e438276b 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -970,7 +970,6 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 	err = __cmd_record(&record, argc, argv);
 
 	perf_evlist__munmap(rec->evlist);
-	perf_evlist__close(rec->evlist);
 out_symbol_exit:
 	symbol__exit();
 	return err;
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index f4ddd14b8a9b..aa8a5f46c361 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1909,7 +1909,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv)
 	err = perf_evlist__mmap(evlist, trace->opts.mmap_pages, false);
 	if (err < 0) {
 		fprintf(trace->output, "Couldn't mmap the events: %s\n", strerror(errno));
-		goto out_close_evlist;
+		goto out_delete_evlist;
 	}
 
 	perf_evlist__enable(evlist);
@@ -1994,8 +1994,6 @@ out_disable:
 	}
 
 	perf_evlist__munmap(evlist);
-out_close_evlist:
-	perf_evlist__close(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/code-reading.c b/tools/perf/tests/code-reading.c
index ddbc7756620c..2c0ce723749f 100644
--- a/tools/perf/tests/code-reading.c
+++ b/tools/perf/tests/code-reading.c
@@ -541,7 +541,6 @@ static int do_test_code_reading(bool try_kcore)
 out_err:
 	if (evlist) {
 		perf_evlist__munmap(evlist);
-		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
 	} else {
 		cpu_map__delete(cpus);
diff --git a/tools/perf/tests/keep-tracking.c b/tools/perf/tests/keep-tracking.c
index f9bc1fcd17cd..dd1c677510bd 100644
--- a/tools/perf/tests/keep-tracking.c
+++ b/tools/perf/tests/keep-tracking.c
@@ -143,7 +143,6 @@ out_err:
 	if (evlist) {
 		perf_evlist__disable(evlist);
 		perf_evlist__munmap(evlist);
-		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
 	} else {
 		cpu_map__delete(cpus);
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index bbb334d2a864..111dd4a917ad 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -68,7 +68,7 @@ int test__basic_mmap(void)
 		evsels[i] = perf_evsel__newtp("syscalls", name);
 		if (evsels[i] == NULL) {
 			pr_debug("perf_evsel__new\n");
-			goto out_free_evlist;
+			goto out_delete_evlist;
 		}
 
 		evsels[i]->attr.wakeup_events = 1;
@@ -80,7 +80,7 @@ int test__basic_mmap(void)
 			pr_debug("failed to open counter: %s, "
 				 "tweak /proc/sys/kernel/perf_event_paranoid?\n",
 				 strerror(errno));
-			goto out_close_fd;
+			goto out_delete_evlist;
 		}
 
 		nr_events[i] = 0;
@@ -90,7 +90,7 @@ int test__basic_mmap(void)
 	if (perf_evlist__mmap(evlist, 128, true) < 0) {
 		pr_debug("failed to mmap events: %d (%s)\n", errno,
 			 strerror(errno));
-		goto out_close_fd;
+		goto out_delete_evlist;
 	}
 
 	for (i = 0; i < nsyscalls; ++i)
@@ -138,10 +138,7 @@ int test__basic_mmap(void)
 
 out_munmap:
 	perf_evlist__munmap(evlist);
-out_close_fd:
-	for (i = 0; i < nsyscalls; ++i)
-		perf_evsel__close_fd(evsels[i], 1, threads->nr);
-out_free_evlist:
+out_delete_evlist:
 	perf_evlist__delete(evlist);
 	cpus	= NULL;
 	threads = NULL;
diff --git a/tools/perf/tests/open-syscall-tp-fields.c b/tools/perf/tests/open-syscall-tp-fields.c
index 595b577172a6..0a00638a9ebb 100644
--- a/tools/perf/tests/open-syscall-tp-fields.c
+++ b/tools/perf/tests/open-syscall-tp-fields.c
@@ -54,7 +54,7 @@ int test__syscall_open_tp_fields(void)
 	err = perf_evlist__mmap(evlist, UINT_MAX, false);
 	if (err < 0) {
 		pr_debug("perf_evlist__mmap: %s\n", strerror(errno));
-		goto out_close_evlist;
+		goto out_delete_evlist;
 	}
 
 	perf_evlist__enable(evlist);
@@ -112,8 +112,6 @@ out_ok:
 	err = 0;
 out_munmap:
 	perf_evlist__munmap(evlist);
-out_close_evlist:
-	perf_evlist__close(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-record.c b/tools/perf/tests/perf-record.c
index 266da9d41300..682978e66efe 100644
--- a/tools/perf/tests/perf-record.c
+++ b/tools/perf/tests/perf-record.c
@@ -132,7 +132,7 @@ int test__PERF_RECORD(void)
 	err = perf_evlist__mmap(evlist, opts.mmap_pages, false);
 	if (err < 0) {
 		pr_debug("perf_evlist__mmap: %s\n", strerror(errno));
-		goto out_close_evlist;
+		goto out_delete_evlist;
 	}
 
 	/*
@@ -304,8 +304,6 @@ found_exit:
 	}
 out_err:
 	perf_evlist__munmap(evlist);
-out_close_evlist:
-	perf_evlist__close(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-time-to-tsc.c b/tools/perf/tests/perf-time-to-tsc.c
index 97d08fff1f8d..3d50f2d2a4d5 100644
--- a/tools/perf/tests/perf-time-to-tsc.c
+++ b/tools/perf/tests/perf-time-to-tsc.c
@@ -167,7 +167,6 @@ out_err:
 	if (evlist) {
 		perf_evlist__disable(evlist);
 		perf_evlist__munmap(evlist);
-		perf_evlist__close(evlist);
 		perf_evlist__delete(evlist);
 	}
 
diff --git a/tools/perf/tests/sw-clock.c b/tools/perf/tests/sw-clock.c
index 266d38127ba6..b366f0edca79 100644
--- a/tools/perf/tests/sw-clock.c
+++ b/tools/perf/tests/sw-clock.c
@@ -70,7 +70,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
 	if (err < 0) {
 		pr_debug("failed to mmap event: %d (%s)\n", errno,
 			 strerror(errno));
-		goto out_close_evlist;
+		goto out_delete_evlist;
 	}
 
 	perf_evlist__enable(evlist);
@@ -107,8 +107,6 @@ next_event:
 
 out_unmap_evlist:
 	perf_evlist__munmap(evlist);
-out_close_evlist:
-	perf_evlist__close(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c
index fdeb2aa46a50..5511a0a929cd 100644
--- a/tools/perf/tests/task-exit.c
+++ b/tools/perf/tests/task-exit.c
@@ -89,7 +89,7 @@ int test__task_exit(void)
 	if (perf_evlist__mmap(evlist, 128, true) < 0) {
 		pr_debug("failed to mmap events: %d (%s)\n", errno,
 			 strerror(errno));
-		goto out_close_evlist;
+		goto out_delete_evlist;
 	}
 
 	perf_evlist__start_workload(evlist);
@@ -113,8 +113,6 @@ retry:
 	}
 
 	perf_evlist__munmap(evlist);
-out_close_evlist:
-	perf_evlist__close(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 143eaf05178f..a083bdc4680e 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -107,6 +107,7 @@ void perf_evlist__exit(struct perf_evlist *evlist)
 
 void perf_evlist__delete(struct perf_evlist *evlist)
 {
+	perf_evlist__close(evlist);
 	cpu_map__delete(evlist->cpus);
 	thread_map__delete(evlist->threads);
 	evlist->cpus = NULL;
-- 
1.8.1.4

[PATCH 07/38] perf evlist: Auto unmap on destructor

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Removing further boilerplate after making sure perf_evlist__munmap can
be called multiple times for the same evlist.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-o0luenuld4abupm4nmrgzm6f@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c               |  2 --
 tools/perf/builtin-trace.c                |  1 -
 tools/perf/tests/code-reading.c           |  1 -
 tools/perf/tests/keep-tracking.c          |  1 -
 tools/perf/tests/mmap-basic.c             | 10 ++++------
 tools/perf/tests/open-syscall-tp-fields.c |  8 +++-----
 tools/perf/tests/perf-record.c            |  4 +---
 tools/perf/tests/perf-time-to-tsc.c       |  1 -
 tools/perf/tests/sw-clock.c               |  4 +---
 tools/perf/tests/task-exit.c              |  1 -
 tools/perf/util/evlist.c                  |  4 ++++
 11 files changed, 13 insertions(+), 24 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index b7f5e438276b..cb00b53c200f 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -968,8 +968,6 @@ int cmd_record(int argc, const char **argv, const char *prefix __maybe_unused)
 	}
 
 	err = __cmd_record(&record, argc, argv);
-
-	perf_evlist__munmap(rec->evlist);
 out_symbol_exit:
 	symbol__exit();
 	return err;
diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index aa8a5f46c361..399b4b91b456 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1993,7 +1993,6 @@ out_disable:
 		}
 	}
 
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/code-reading.c b/tools/perf/tests/code-reading.c
index 2c0ce723749f..653a8fe2db95 100644
--- a/tools/perf/tests/code-reading.c
+++ b/tools/perf/tests/code-reading.c
@@ -540,7 +540,6 @@ static int do_test_code_reading(bool try_kcore)
 		err = TEST_CODE_READING_OK;
 out_err:
 	if (evlist) {
-		perf_evlist__munmap(evlist);
 		perf_evlist__delete(evlist);
 	} else {
 		cpu_map__delete(cpus);
diff --git a/tools/perf/tests/keep-tracking.c b/tools/perf/tests/keep-tracking.c
index dd1c677510bd..497957f269d8 100644
--- a/tools/perf/tests/keep-tracking.c
+++ b/tools/perf/tests/keep-tracking.c
@@ -142,7 +142,6 @@ int test__keep_tracking(void)
 out_err:
 	if (evlist) {
 		perf_evlist__disable(evlist);
-		perf_evlist__munmap(evlist);
 		perf_evlist__delete(evlist);
 	} else {
 		cpu_map__delete(cpus);
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index 111dd4a917ad..aef25f0ff9cc 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -105,13 +105,13 @@ int test__basic_mmap(void)
 		if (event->header.type != PERF_RECORD_SAMPLE) {
 			pr_debug("unexpected %s event\n",
 				 perf_event__name(event->header.type));
-			goto out_munmap;
+			goto out_delete_evlist;
 		}
 
 		err = perf_evlist__parse_sample(evlist, event, &sample);
 		if (err) {
 			pr_err("Can't parse sample, err = %d\n", err);
-			goto out_munmap;
+			goto out_delete_evlist;
 		}
 
 		err = -1;
@@ -119,7 +119,7 @@ int test__basic_mmap(void)
 		if (evsel == NULL) {
 			pr_debug("event with id %" PRIu64
 				 " doesn't map to an evsel\n", sample.id);
-			goto out_munmap;
+			goto out_delete_evlist;
 		}
 		nr_events[evsel->idx]++;
 		perf_evlist__mmap_consume(evlist, 0);
@@ -132,12 +132,10 @@ int test__basic_mmap(void)
 				 expected_nr_events[evsel->idx],
 				 perf_evsel__name(evsel), nr_events[evsel->idx]);
 			err = -1;
-			goto out_munmap;
+			goto out_delete_evlist;
 		}
 	}
 
-out_munmap:
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 	cpus	= NULL;
diff --git a/tools/perf/tests/open-syscall-tp-fields.c b/tools/perf/tests/open-syscall-tp-fields.c
index 0a00638a9ebb..5a016f66f5d2 100644
--- a/tools/perf/tests/open-syscall-tp-fields.c
+++ b/tools/perf/tests/open-syscall-tp-fields.c
@@ -85,7 +85,7 @@ int test__syscall_open_tp_fields(void)
 				err = perf_evsel__parse_sample(evsel, event, &sample);
 				if (err) {
 					pr_err("Can't parse sample, err = %d\n", err);
-					goto out_munmap;
+					goto out_delete_evlist;
 				}
 
 				tp_flags = perf_evsel__intval(evsel, &sample, "flags");
@@ -93,7 +93,7 @@ int test__syscall_open_tp_fields(void)
 				if (flags != tp_flags) {
 					pr_debug("%s: Expected flags=%#x, got %#x\n",
 						 __func__, flags, tp_flags);
-					goto out_munmap;
+					goto out_delete_evlist;
 				}
 
 				goto out_ok;
@@ -105,13 +105,11 @@ int test__syscall_open_tp_fields(void)
 
 		if (++nr_polls > 5) {
 			pr_debug("%s: no events!\n", __func__);
-			goto out_munmap;
+			goto out_delete_evlist;
 		}
 	}
 out_ok:
 	err = 0;
-out_munmap:
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-record.c b/tools/perf/tests/perf-record.c
index 682978e66efe..39cc7c3c0d0c 100644
--- a/tools/perf/tests/perf-record.c
+++ b/tools/perf/tests/perf-record.c
@@ -165,7 +165,7 @@ int test__PERF_RECORD(void)
 					if (verbose)
 						perf_event__fprintf(event, stderr);
 					pr_debug("Couldn't parse sample\n");
-					goto out_err;
+					goto out_delete_evlist;
 				}
 
 				if (verbose) {
@@ -302,8 +302,6 @@ found_exit:
 		pr_debug("PERF_RECORD_MMAP for %s missing!\n", "[vdso]");
 		++errs;
 	}
-out_err:
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 out:
diff --git a/tools/perf/tests/perf-time-to-tsc.c b/tools/perf/tests/perf-time-to-tsc.c
index 3d50f2d2a4d5..47146d388dbf 100644
--- a/tools/perf/tests/perf-time-to-tsc.c
+++ b/tools/perf/tests/perf-time-to-tsc.c
@@ -166,7 +166,6 @@ next_event:
 out_err:
 	if (evlist) {
 		perf_evlist__disable(evlist);
-		perf_evlist__munmap(evlist);
 		perf_evlist__delete(evlist);
 	}
 
diff --git a/tools/perf/tests/sw-clock.c b/tools/perf/tests/sw-clock.c
index b366f0edca79..983d6b8562a8 100644
--- a/tools/perf/tests/sw-clock.c
+++ b/tools/perf/tests/sw-clock.c
@@ -90,7 +90,7 @@ static int __test__sw_clock_freq(enum perf_sw_ids clock_id)
 		err = perf_evlist__parse_sample(evlist, event, &sample);
 		if (err < 0) {
 			pr_debug("Error during parse sample\n");
-			goto out_unmap_evlist;
+			goto out_delete_evlist;
 		}
 
 		total_periods += sample.period;
@@ -105,8 +105,6 @@ next_event:
 		err = -1;
 	}
 
-out_unmap_evlist:
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
diff --git a/tools/perf/tests/task-exit.c b/tools/perf/tests/task-exit.c
index 5511a0a929cd..5ff3db318f12 100644
--- a/tools/perf/tests/task-exit.c
+++ b/tools/perf/tests/task-exit.c
@@ -112,7 +112,6 @@ retry:
 		err = -1;
 	}
 
-	perf_evlist__munmap(evlist);
 out_delete_evlist:
 	perf_evlist__delete(evlist);
 	return err;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index a083bdc4680e..0810f5ca77d7 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -107,6 +107,7 @@ void perf_evlist__exit(struct perf_evlist *evlist)
 
 void perf_evlist__delete(struct perf_evlist *evlist)
 {
+	perf_evlist__munmap(evlist);
 	perf_evlist__close(evlist);
 	cpu_map__delete(evlist->cpus);
 	thread_map__delete(evlist->threads);
@@ -587,6 +588,9 @@ void perf_evlist__munmap(struct perf_evlist *evlist)
 {
 	int i;
 
+	if (evlist->mmap == NULL)
+		return;
+
 	for (i = 0; i < evlist->nr_mmaps; i++)
 		__perf_evlist__munmap(evlist, i);
 
-- 
1.8.1.4

[PATCH 08/38] perf tests: Fixup leak on error path in parse events test

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

We need to call the evlist destructor when failing to parse events.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ilslu69s7v7bpvdgqtrlp8f5@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/tests/parse-events.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c
index e4ce8aed29d3..41dee5e67f63 100644
--- a/tools/perf/tests/parse-events.c
+++ b/tools/perf/tests/parse-events.c
@@ -1385,10 +1385,10 @@ static int test_event(struct evlist_test *e)
 	if (ret) {
 		pr_debug("failed to parse event '%s', err %d\n",
 			 e->name, ret);
-		return ret;
+	} else {
+		ret = e->check(evlist);
 	}
-
-	ret = e->check(evlist);
+	
 	perf_evlist__delete(evlist);
 
 	return ret;
-- 
1.8.1.4

[PATCH 09/38] perf stat: Remove misplaced __maybe_unused

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

That 'argc' argument _is_ being used.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-t2gsxc15zulkorieg8zq996o@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 6ca076660de5..b27b2645eb3c 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -645,7 +645,7 @@ static int __run_perf_stat(int argc, const char **argv)
 	return WEXITSTATUS(status);
 }
 
-static int run_perf_stat(int argc __maybe_unused, const char **argv)
+static int run_perf_stat(int argc, const char **argv)
 {
 	int ret;
 
-- 
1.8.1.4

[PATCH 10/38] perf tools: Move arch setup into seprate Makefile

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

I need to use arch related setup in the tests/make, so moving arch setup
into Makefile.arch.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1388759553-12974-1-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/config/Makefile      | 52 ++++++++++++++---------------------------
 tools/perf/config/Makefile.arch | 22 +++++++++++++++++
 2 files changed, 40 insertions(+), 34 deletions(-)
 create mode 100644 tools/perf/config/Makefile.arch

diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index 14faeeb0d752..f2bc659d20af 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -1,28 +1,26 @@
-uname_M := $(shell uname -m 2>/dev/null || echo not)
 
-ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/ -e s/sun4u/sparc64/ \
-                                  -e s/arm.*/arm/ -e s/sa110/arm/ \
-                                  -e s/s390x/s390/ -e s/parisc64/parisc/ \
-                                  -e s/ppc.*/powerpc/ -e s/mips.*/mips/ \
-                                  -e s/sh[234].*/sh/ -e s/aarch64.*/arm64/ )
-NO_PERF_REGS := 1
-CFLAGS := $(EXTRA_CFLAGS) $(EXTRA_WARNINGS)
+ifeq ($(src-perf),)
+src-perf := $(srctree)/tools/perf
+endif
 
-# Additional ARCH settings for x86
-ifeq ($(ARCH),i386)
-  override ARCH := x86
-  NO_PERF_REGS := 0
-  LIBUNWIND_LIBS = -lunwind -lunwind-x86
+ifeq ($(obj-perf),)
+obj-perf := $(OUTPUT)
 endif
 
-ifeq ($(ARCH),x86_64)
-  override ARCH := x86
-  IS_X86_64 := 0
-  ifeq (, $(findstring m32,$(CFLAGS)))
-    IS_X86_64 := $(shell echo __x86_64__ | ${CC} -E -x c - | tail -n 1)
-  endif
+ifneq ($(obj-perf),)
+obj-perf := $(abspath $(obj-perf))/
+endif
+
+LIB_INCLUDE := $(srctree)/tools/lib/
+CFLAGS := $(EXTRA_CFLAGS) $(EXTRA_WARNINGS)
+
+include $(src-perf)/config/Makefile.arch
+
+NO_PERF_REGS := 1
+
+# Additional ARCH settings for x86
+ifeq ($(ARCH),x86)
   ifeq (${IS_X86_64}, 1)
-    RAW_ARCH := x86_64
     CFLAGS += -DHAVE_ARCH_X86_64_SUPPORT
     ARCH_INCLUDE = ../../arch/x86/lib/memcpy_64.S ../../arch/x86/lib/memset_64.S
     LIBUNWIND_LIBS = -lunwind -lunwind-x86_64
@@ -64,20 +62,6 @@ ifeq ($(NO_PERF_REGS),0)
   CFLAGS += -DHAVE_PERF_REGS_SUPPORT
 endif
 
-ifeq ($(src-perf),)
-src-perf := $(srctree)/tools/perf
-endif
-
-ifeq ($(obj-perf),)
-obj-perf := $(OUTPUT)
-endif
-
-ifneq ($(obj-perf),)
-obj-perf := $(abspath $(obj-perf))/
-endif
-
-LIB_INCLUDE := $(srctree)/tools/lib/
-
 # include ARCH specific config
 -include $(src-perf)/arch/$(ARCH)/Makefile
 
diff --git a/tools/perf/config/Makefile.arch b/tools/perf/config/Makefile.arch
new file mode 100644
index 000000000000..fef8ae922800
--- /dev/null
+++ b/tools/perf/config/Makefile.arch
@@ -0,0 +1,22 @@
+
+uname_M := $(shell uname -m 2>/dev/null || echo not)
+
+ARCH ?= $(shell echo $(uname_M) | sed -e s/i.86/i386/ -e s/sun4u/sparc64/ \
+                                  -e s/arm.*/arm/ -e s/sa110/arm/ \
+                                  -e s/s390x/s390/ -e s/parisc64/parisc/ \
+                                  -e s/ppc.*/powerpc/ -e s/mips.*/mips/ \
+                                  -e s/sh[234].*/sh/ -e s/aarch64.*/arm64/ )
+
+# Additional ARCH settings for x86
+ifeq ($(ARCH),i386)
+  override ARCH := x86
+endif
+
+ifeq ($(ARCH),x86_64)
+  override ARCH := x86
+  IS_X86_64 := 0
+  ifeq (, $(findstring m32,$(CFLAGS)))
+    IS_X86_64 := $(shell echo __x86_64__ | ${CC} -E -x c - | tail -n 1)
+    RAW_ARCH := x86_64
+  endif
+endif
-- 
1.8.1.4

[PATCH 11/38] perf tests: Fix installation tests path setup

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Currently installation tests work only over x86_64, adding arch check to
make it work over i386 as well.

NOTE looks like x86 is the only arch running tests, we need some
IS_(32/64) flag to make this generic.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1388759553-12974-2-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/tests/make | 30 ++++++++++++++++++++----------
 1 file changed, 20 insertions(+), 10 deletions(-)

diff --git a/tools/perf/tests/make b/tools/perf/tests/make
index f641c35f2321..e341088de8f3 100644
--- a/tools/perf/tests/make
+++ b/tools/perf/tests/make
@@ -1,6 +1,16 @@
 PERF := .
 MK   := Makefile
 
+include config/Makefile.arch
+
+# FIXME looks like x86 is the only arch running tests ;-)
+# we need some IS_(32/64) flag to make this generic
+ifeq ($(IS_X86_64),1)
+lib = lib64
+else
+lib = lib
+endif
+
 has = $(shell which $1 2>/dev/null)
 
 # standard single make variable specified
@@ -118,16 +128,16 @@ installed_files_bin := bin/perf
 installed_files_bin += etc/bash_completion.d/perf
 installed_files_bin += libexec/perf-core/perf-archive
 
-installed_files_plugins := lib64/traceevent/plugins/plugin_cfg80211.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_scsi.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_xen.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_function.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_sched_switch.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_mac80211.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_kvm.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_kmem.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_hrtimer.so
-installed_files_plugins += lib64/traceevent/plugins/plugin_jbd2.so
+installed_files_plugins := $(lib)/traceevent/plugins/plugin_cfg80211.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_scsi.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_xen.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_function.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_sched_switch.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_mac80211.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_kvm.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_kmem.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_hrtimer.so
+installed_files_plugins += $(lib)/traceevent/plugins/plugin_jbd2.so
 
 installed_files_all := $(installed_files_bin)
 installed_files_all += $(installed_files_plugins)
-- 
1.8.1.4

[PATCH 12/38] tools lib traceevent: Replace tabs with spaces for all non-commands statements

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

The tabbed indentation in non-commands statements could be sometimes
considered as follow up for the rule command in the Makefile.

This error is hard to find, so as a precaution replacing tabs with
spaces for all non-commands statements.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://marc.info/?t=136484403900003&r=1&w=2
Link: http://lkml.kernel.org/r/20140102095304.GA1196@krava.brq.redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/lib/traceevent/Makefile | 54 +++++++++++++++++++++----------------------
 1 file changed, 27 insertions(+), 27 deletions(-)

diff --git a/tools/lib/traceevent/Makefile b/tools/lib/traceevent/Makefile
index ca4ab78425d1..76fe0aeb874e 100644
--- a/tools/lib/traceevent/Makefile
+++ b/tools/lib/traceevent/Makefile
@@ -86,8 +86,8 @@ ifeq ($(BUILD_SRC),)
 ifneq ($(OUTPUT),)
 
 define build_output
-	$(if $(VERBOSE:1=),@)+$(MAKE) -C $(OUTPUT) \
-	BUILD_SRC=$(CURDIR)/ -f $(CURDIR)/Makefile $1
+  $(if $(VERBOSE:1=),@)+$(MAKE) -C $(OUTPUT) \
+  BUILD_SRC=$(CURDIR)/ -f $(CURDIR)/Makefile $1
 endef
 
 all: sub-make
@@ -221,23 +221,23 @@ $(PLUGINS): %.so: %.o
 	$(QUIET_LINK)$(CC) $(CFLAGS) -shared -nostartfiles -o $@ $<
 
 define make_version.h
-	(echo '/* This file is automatically generated. Do not modify. */';		\
-	echo \#define VERSION_CODE $(shell						\
-	expr $(VERSION) \* 256 + $(PATCHLEVEL));					\
-	echo '#define EXTRAVERSION ' $(EXTRAVERSION);					\
-	echo '#define VERSION_STRING "'$(VERSION).$(PATCHLEVEL).$(EXTRAVERSION)'"';	\
-	echo '#define FILE_VERSION '$(FILE_VERSION);					\
-	) > $1
+  (echo '/* This file is automatically generated. Do not modify. */';		\
+   echo \#define VERSION_CODE $(shell						\
+   expr $(VERSION) \* 256 + $(PATCHLEVEL));					\
+   echo '#define EXTRAVERSION ' $(EXTRAVERSION);				\
+   echo '#define VERSION_STRING "'$(VERSION).$(PATCHLEVEL).$(EXTRAVERSION)'"';	\
+   echo '#define FILE_VERSION '$(FILE_VERSION);					\
+  ) > $1
 endef
 
 define update_version.h
-	($(call make_version.h, $@.tmp);		\
-	if [ -r $@ ] && cmp -s $@ $@.tmp; then		\
-		rm -f $@.tmp;				\
-	else						\
-		echo '  UPDATE                 $@';	\
-		mv -f $@.tmp $@;			\
-	fi);
+  ($(call make_version.h, $@.tmp);		\
+    if [ -r $@ ] && cmp -s $@ $@.tmp; then	\
+      rm -f $@.tmp;				\
+    else					\
+      echo '  UPDATE                 $@';	\
+      mv -f $@.tmp $@;				\
+    fi);
 endef
 
 ep_version.h: force
@@ -246,13 +246,13 @@ ep_version.h: force
 VERSION_FILES = ep_version.h
 
 define update_dir
-	(echo $1 > $@.tmp;	\
-	if [ -r $@ ] && cmp -s $@ $@.tmp; then		\
-		rm -f $@.tmp;				\
-	else						\
-		echo '  UPDATE                 $@';	\
-		mv -f $@.tmp $@;			\
-	fi);
+  (echo $1 > $@.tmp;				\
+   if [ -r $@ ] && cmp -s $@ $@.tmp; then	\
+     rm -f $@.tmp;				\
+   else						\
+     echo '  UPDATE                 $@';	\
+     mv -f $@.tmp $@;				\
+   fi);
 endef
 
 ## make deps
@@ -262,10 +262,10 @@ all_deps := $(all_objs:%.o=.%.d)
 
 # let .d file also depends on the source and header files
 define check_deps
-		@set -e; $(RM) $@; \
-		$(CC) -MM $(CFLAGS) $< > $@.$$$$; \
-		sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@.$$$$ > $@; \
-		$(RM) $@.$$$$
+  @set -e; $(RM) $@; \
+  $(CC) -MM $(CFLAGS) $< > $@.$$$$; \
+  sed 's,\($*\)\.o[ :]*,\1.o $@ : ,g' < $@.$$$$ > $@; \
+  $(RM) $@.$$$$
 endef
 
 $(all_deps): .%.d: $(src)/%.c
-- 
1.8.1.4

[PATCH 13/38] tools lib traceevent: Shut up plugins make message

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Getting rid of following build output:

  $ make O=/tmp/build/perf -C tools/perf/ install-bin
  ...
  make[3]: Nothing to be done for `plugins'.
  make[2]: Nothing to be done for `plugins'.
  ...

which triggers when traceevent library needs to be rebuilt, but we have
plugins built already.

Adding extra 'plugins' target with nop which is visible and triggers in
both Makefile parts (for detached output directory (O=...) the
traceevent Makefile spawns sub make for the build itself).

Reported-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1388595050-23005-2-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/lib/traceevent/Makefile | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/tools/lib/traceevent/Makefile b/tools/lib/traceevent/Makefile
index 76fe0aeb874e..f778d48ac609 100644
--- a/tools/lib/traceevent/Makefile
+++ b/tools/lib/traceevent/Makefile
@@ -329,9 +329,12 @@ clean:
 
 endif # skip-makefile
 
-PHONY += force
+PHONY += force plugins
 force:
 
+plugins:
+	@echo > /dev/null
+
 # Declare the contents of the .PHONY variable as phony.  We keep that
 # information in a variable so we can use it in if_changed and friends.
 .PHONY: $(PHONY)
-- 
1.8.1.4

[PATCH 14/38] perf header: Pack 'struct perf_session_env'

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Initial struct:

[acme@ssdandy linux]$ pahole -C perf_session_env ~/bin/perf
struct perf_session_env {
	char *                     hostname;             /*     0     8 */
	char *                     os_release;           /*     8     8 */
	char *                     version;              /*    16     8 */
	char *                     arch;                 /*    24     8 */
	int                        nr_cpus_online;       /*    32     4 */
	int                        nr_cpus_avail;        /*    36     4 */
	char *                     cpu_desc;             /*    40     8 */
	char *                     cpuid;                /*    48     8 */
	long long unsigned int     total_mem;            /*    56     8 */
	/* --- cacheline 1 boundary (64 bytes) --- */
	int                        nr_cmdline;           /*    64     4 */

	/* XXX 4 bytes hole, try to pack */

	char *                     cmdline;              /*    72     8 */
	int                        nr_sibling_cores;     /*    80     4 */

	/* XXX 4 bytes hole, try to pack */

	char *                     sibling_cores;        /*    88     8 */
	int                        nr_sibling_threads;   /*    96     4 */

	/* XXX 4 bytes hole, try to pack */

	char *                     sibling_threads;      /*   104     8 */
	int                        nr_numa_nodes;        /*   112     4 */

	/* XXX 4 bytes hole, try to pack */

	char *                     numa_nodes;           /*   120     8 */
	/* --- cacheline 2 boundary (128 bytes) --- */
	int                        nr_pmu_mappings;      /*   128     4 */

	/* XXX 4 bytes hole, try to pack */

	char *                     pmu_mappings;         /*   136     8 */
	int                        nr_groups;            /*   144     4 */

	/* size: 152, cachelines: 3, members: 20 */
	/* sum members: 128, holes: 5, sum holes: 20 */
	/* padding: 4 */
	/* last cacheline: 24 bytes */
};
[acme@ssdandy linux]$

[acme@ssdandy linux]$ pahole -C perf_session_env --reorganize --show_reorg_steps ~/bin/perf | grep ^/ | grep -v Final
/* Moving 'nr_sibling_cores' from after 'cmdline' to after 'nr_cmdline' */
/* Moving 'nr_numa_nodes' from after 'sibling_threads' to after 'nr_sibling_threads' */
/* Moving 'nr_groups' from after 'pmu_mappings' to after 'nr_pmu_mappings' */
[acme@ssdandy linux]$

Final struct stats:

[acme@ssdandy linux]$ pahole -C perf_session_env --reorganize --show_reorg_steps ~/bin/perf | tail -4
	/* --- cacheline 2 boundary (128 bytes) --- */

	/* size: 128, cachelines: 2, members: 20 */
};   /* saved 24 bytes and 1 cacheline! */
[acme@ssdandy linux]$

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-3d9tshamloinzxcqeb7mtd1n@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/header.h | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/tools/perf/util/header.h b/tools/perf/util/header.h
index 307c9aed972e..a2d047bdf4ef 100644
--- a/tools/perf/util/header.h
+++ b/tools/perf/util/header.h
@@ -77,16 +77,16 @@ struct perf_session_env {
 	unsigned long long	total_mem;
 
 	int			nr_cmdline;
-	char			*cmdline;
 	int			nr_sibling_cores;
-	char			*sibling_cores;
 	int			nr_sibling_threads;
-	char			*sibling_threads;
 	int			nr_numa_nodes;
-	char			*numa_nodes;
 	int			nr_pmu_mappings;
-	char			*pmu_mappings;
 	int			nr_groups;
+	char			*cmdline;
+	char			*sibling_cores;
+	char			*sibling_threads;
+	char			*numa_nodes;
+	char			*pmu_mappings;
 };
 
 struct perf_header {
-- 
1.8.1.4

[PATCH 15/38] perf trace: Pack 'struct trace'

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Initial struct stats:

	/* size: 368, cachelines: 6, members: 24 */
	/* sum members: 353, holes: 3, sum holes: 15 */
	/* last cacheline: 48 bytes */

After reorg:

[acme@ssdandy linux]$ pahole -C trace ~/bin/trace | tail -4
	/* size: 360, cachelines: 6, members: 24 */
	/* padding: 7 */
	/* last cacheline: 40 bytes */
};
[acme@ssdandy linux]$

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-6jimc80yu89qkx6zb8465s6t@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-trace.c | 17 +++++++++--------
 1 file changed, 9 insertions(+), 8 deletions(-)

diff --git a/tools/perf/builtin-trace.c b/tools/perf/builtin-trace.c
index 399b4b91b456..4bd44aba343e 100644
--- a/tools/perf/builtin-trace.c
+++ b/tools/perf/builtin-trace.c
@@ -1160,26 +1160,27 @@ struct trace {
 	struct record_opts	opts;
 	struct machine		*host;
 	u64			base_time;
-	bool			full_time;
 	FILE			*output;
 	unsigned long		nr_events;
 	struct strlist		*ev_qualifier;
-	bool			not_ev_qualifier;
-	bool			live;
 	const char 		*last_vfs_getname;
 	struct intlist		*tid_list;
 	struct intlist		*pid_list;
+	double			duration_filter;
+	double			runtime_ms;
+	struct {
+		u64		vfs_getname,
+				proc_getname;
+	} stats;
+	bool			not_ev_qualifier;
+	bool			live;
+	bool			full_time;
 	bool			sched;
 	bool			multiple_threads;
 	bool			summary;
 	bool			summary_only;
 	bool			show_comm;
 	bool			show_tool_stats;
-	double			duration_filter;
-	double			runtime_ms;
-	struct {
-		u64		vfs_getname, proc_getname;
-	} stats;
 };
 
 static int trace__set_fd_pathname(struct thread *thread, int fd, const char *pathname)
-- 
1.8.1.4

[PATCH 16/38] perf tools: Automate setup of FEATURE_CHECK_(C|LD)FLAGS-all variables

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Instead of explicitly adding same value into
FEATURE_CHECK_(C|LD)FLAGS-all variables we can do that automatically.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1389098853-14466-2-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/config/Makefile | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index f2bc659d20af..5d15b43b03c6 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -53,9 +53,6 @@ else
   FEATURE_CHECK_LDFLAGS-libunwind = $(LIBUNWIND_LDFLAGS)
   FEATURE_CHECK_CFLAGS-libunwind-debug-frame = $(LIBUNWIND_CFLAGS)
   FEATURE_CHECK_LDFLAGS-libunwind-debug-frame = $(LIBUNWIND_LDFLAGS)
-  #  and the flags for the test-all case
-  FEATURE_CHECK_CFLAGS-all += $(LIBUNWIND_CFLAGS)
-  FEATURE_CHECK_LDFLAGS-all += $(LIBUNWIND_LDFLAGS)
 endif
 
 ifeq ($(NO_PERF_REGS),0)
@@ -152,6 +149,17 @@ CORE_FEATURE_TESTS =			\
 	stackprotector-all		\
 	timerfd
 
+# Set FEATURE_CHECK_(C|LD)FLAGS-all for all CORE_FEATURE_TESTS features.
+# If in the future we need per-feature checks/flags for features not
+# mentioned in this list we need to refactor this ;-).
+set_test_all_flags = $(eval $(set_test_all_flags_code))
+define set_test_all_flags_code
+  FEATURE_CHECK_CFLAGS-all  += $(FEATURE_CHECK_CFLAGS-$(1))
+  FEATURE_CHECK_LDFLAGS-all += $(FEATURE_CHECK_LDFLAGS-$(1))
+endef
+
+$(foreach feat,$(CORE_FEATURE_TESTS),$(call set_test_all_flags,$(feat)))
+
 #
 # So here we detect whether test-all was rebuilt, to be able
 # to skip the print-out of the long features list if the file
-- 
1.8.1.4

[PATCH 17/38] perf machine: Fix id_hdr_size initialization

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

The id_hdr_size field was not properly initialized, set it to zero, as
the machine struct may have come from some non zeroing allocation
routine or from the stack without any field being initialized.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1389098853-14466-3-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/machine.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/perf/util/machine.c b/tools/perf/util/machine.c
index a98538dc465a..0130279aac51 100644
--- a/tools/perf/util/machine.c
+++ b/tools/perf/util/machine.c
@@ -27,6 +27,7 @@ int machine__init(struct machine *machine, const char *root_dir, pid_t pid)
 	machine->pid = pid;
 
 	machine->symbol_filter = NULL;
+	machine->id_hdr_size = 0;
 
 	machine->root_dir = strdup(root_dir);
 	if (machine->root_dir == NULL)
-- 
1.8.1.4

[PATCH 18/38] perf tools: Make perf_event__synthesize_mmap_events global

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Making perf_event__synthesize_mmap_events global, it will be used in
following patch from test code.

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jean Pihet <jean.pihet@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1389098853-14466-4-git-send-email-jolsa@redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/event.c | 12 ++++++------
 tools/perf/util/event.h |  7 +++++++
 2 files changed, 13 insertions(+), 6 deletions(-)

diff --git a/tools/perf/util/event.c b/tools/perf/util/event.c
index 45a76c69a9ed..1fc1c2f04772 100644
--- a/tools/perf/util/event.c
+++ b/tools/perf/util/event.c
@@ -175,12 +175,12 @@ out:
 	return tgid;
 }
 
-static int perf_event__synthesize_mmap_events(struct perf_tool *tool,
-					      union perf_event *event,
-					      pid_t pid, pid_t tgid,
-					      perf_event__handler_t process,
-					      struct machine *machine,
-					      bool mmap_data)
+int perf_event__synthesize_mmap_events(struct perf_tool *tool,
+				       union perf_event *event,
+				       pid_t pid, pid_t tgid,
+				       perf_event__handler_t process,
+				       struct machine *machine,
+				       bool mmap_data)
 {
 	char filename[PATH_MAX];
 	FILE *fp;
diff --git a/tools/perf/util/event.h b/tools/perf/util/event.h
index 30fec9901e44..faf6e219be21 100644
--- a/tools/perf/util/event.h
+++ b/tools/perf/util/event.h
@@ -266,6 +266,13 @@ int perf_event__synthesize_sample(union perf_event *event, u64 type,
 				  const struct perf_sample *sample,
 				  bool swapped);
 
+int perf_event__synthesize_mmap_events(struct perf_tool *tool,
+				       union perf_event *event,
+				       pid_t pid, pid_t tgid,
+				       perf_event__handler_t process,
+				       struct machine *machine,
+				       bool mmap_data);
+
 size_t perf_event__fprintf_comm(union perf_event *event, FILE *fp);
 size_t perf_event__fprintf_mmap(union perf_event *event, FILE *fp);
 size_t perf_event__fprintf_mmap2(union perf_event *event, FILE *fp);
-- 
1.8.1.4

[PATCH 19/38] perf stat: Fix --delay option in man page

[Arnaldo Carvalho de Melo]

From: Andi Kleen <ak@linux.intel.com>

The --delay option was documented as --initial-delay in the manpage. Fix this.

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1389132847-31982-1-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-stat.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/Documentation/perf-stat.txt b/tools/perf/Documentation/perf-stat.txt
index 80c7da6732f2..29ee857c09c6 100644
--- a/tools/perf/Documentation/perf-stat.txt
+++ b/tools/perf/Documentation/perf-stat.txt
@@ -133,7 +133,7 @@ use --per-core in addition to -a. (system-wide).  The output includes the
 core number and the number of online logical processors on that physical processor.
 
 -D msecs::
---initial-delay msecs::
+--delay msecs::
 After starting the program, wait msecs before measuring. This is useful to
 filter out the startup phase of the program, which is often very different.
 
-- 
1.8.1.4

[PATCH 20/38] tools perf: Comment typo fix

[Arnaldo Carvalho de Melo]

From: Cody P Schafer <cody@linux.vnet.ibm.com>

s/temr/term/

Signed-off-by: Cody P Schafer <cody@linux.vnet.ibm.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Link: http://lkml.kernel.org/r/1389199434-21761-1-git-send-email-cody@linux.vnet.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/pmu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/util/pmu.c b/tools/perf/util/pmu.c
index 0934d645ebdc..d9cab4d27192 100644
--- a/tools/perf/util/pmu.c
+++ b/tools/perf/util/pmu.c
@@ -505,7 +505,7 @@ static __u64 pmu_format_value(unsigned long *format, __u64 value)
 
 /*
  * Setup one of config[12] attr members based on the
- * user input data - temr parameter.
+ * user input data - term parameter.
  */
 static int pmu_config_term(struct list_head *formats,
 			   struct perf_event_attr *attr,
-- 
1.8.1.4

[PATCH 21/38] perf report: Move logic to warn about kptr_restrict'ed kernels to separate function

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Its too big, better have a separate function for it so that the main
logic gets shorter/clearer.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-ahh6vfzyh8fsygjwrsbroeu0@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-report.c | 56 ++++++++++++++++++++++++---------------------
 1 file changed, 30 insertions(+), 26 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index bf8dd2e893e4..f2ff8609e98e 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -412,14 +412,41 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,
 	return 0;
 }
 
+static void report__warn_kptr_restrict(const struct report *rep)
+{
+	struct map *kernel_map = rep->session->machines.host.vmlinux_maps[MAP__FUNCTION];
+	struct kmap *kernel_kmap = map__kmap(kernel_map);
+
+	if (kernel_map == NULL ||
+	    (kernel_map->dso->hit &&
+	     (kernel_kmap->ref_reloc_sym == NULL ||
+	      kernel_kmap->ref_reloc_sym->addr == 0))) {
+		const char *desc =
+		    "As no suitable kallsyms nor vmlinux was found, kernel samples\n"
+		    "can't be resolved.";
+
+		if (kernel_map) {
+			const struct dso *kdso = kernel_map->dso;
+			if (!RB_EMPTY_ROOT(&kdso->symbols[MAP__FUNCTION])) {
+				desc = "If some relocation was applied (e.g. "
+				       "kexec) symbols may be misresolved.";
+			}
+		}
+
+		ui__warning(
+"Kernel address maps (/proc/{kallsyms,modules}) were restricted.\n\n"
+"Check /proc/sys/kernel/kptr_restrict before running 'perf record'.\n\n%s\n\n"
+"Samples in kernel modules can't be resolved as well.\n\n",
+		desc);
+	}
+}
+
 static int __cmd_report(struct report *rep)
 {
 	int ret = -EINVAL;
 	u64 nr_samples;
 	struct perf_session *session = rep->session;
 	struct perf_evsel *pos;
-	struct map *kernel_map;
-	struct kmap *kernel_kmap;
 	const char *help = "For a higher level overview, try: perf report --sort comm,dso";
 	struct ui_progress prog;
 	struct perf_data_file *file = session->file;
@@ -444,30 +471,7 @@ static int __cmd_report(struct report *rep)
 	if (ret)
 		return ret;
 
-	kernel_map = session->machines.host.vmlinux_maps[MAP__FUNCTION];
-	kernel_kmap = map__kmap(kernel_map);
-	if (kernel_map == NULL ||
-	    (kernel_map->dso->hit &&
-	     (kernel_kmap->ref_reloc_sym == NULL ||
-	      kernel_kmap->ref_reloc_sym->addr == 0))) {
-		const char *desc =
-		    "As no suitable kallsyms nor vmlinux was found, kernel samples\n"
-		    "can't be resolved.";
-
-		if (kernel_map) {
-			const struct dso *kdso = kernel_map->dso;
-			if (!RB_EMPTY_ROOT(&kdso->symbols[MAP__FUNCTION])) {
-				desc = "If some relocation was applied (e.g. "
-				       "kexec) symbols may be misresolved.";
-			}
-		}
-
-		ui__warning(
-"Kernel address maps (/proc/{kallsyms,modules}) were restricted.\n\n"
-"Check /proc/sys/kernel/kptr_restrict before running 'perf record'.\n\n%s\n\n"
-"Samples in kernel modules can't be resolved as well.\n\n",
-		desc);
-	}
+	report__warn_kptr_restrict(rep);
 
 	if (use_browser == 0) {
 		if (verbose > 3)
-- 
1.8.1.4

[PATCH 22/38] perf report: Move hist browser selection code to separate function

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

To unclutter the main function.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-agvxwpazlucy6h5sejuttw9t@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-report.c | 79 ++++++++++++++++++++++++++-------------------
 1 file changed, 46 insertions(+), 33 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index f2ff8609e98e..03941ad8fc46 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -441,13 +441,57 @@ static void report__warn_kptr_restrict(const struct report *rep)
 	}
 }
 
+static int report__gtk_browse_hists(struct report *rep, const char *help)
+{
+	int (*hist_browser)(struct perf_evlist *evlist, const char *help,
+			    struct hist_browser_timer *timer, float min_pcnt);
+
+	hist_browser = dlsym(perf_gtk_handle, "perf_evlist__gtk_browse_hists");
+
+	if (hist_browser == NULL) {
+		ui__error("GTK browser not found!\n");
+		return -1;
+	}
+
+	return hist_browser(rep->session->evlist, help, NULL, rep->min_percent);
+}
+
+static int report__browse_hists(struct report *rep)
+{
+	int ret;
+	struct perf_session *session = rep->session;
+	struct perf_evlist *evlist = session->evlist;
+	const char *help = "For a higher level overview, try: perf report --sort comm,dso";
+
+	switch (use_browser) {
+	case 1:
+		ret = perf_evlist__tui_browse_hists(evlist, help, NULL,
+						    rep->min_percent,
+						    &session->header.env);
+		/*
+		 * Usually "ret" is the last pressed key, and we only
+		 * care if the key notifies us to switch data file.
+		 */
+		if (ret != K_SWITCH_INPUT_DATA)
+			ret = 0;
+		break;
+	case 2:
+		ret = report__gtk_browse_hists(rep, help);
+		break;
+	default:
+		ret = perf_evlist__tty_browse_hists(evlist, rep, help);
+		break;
+	}
+
+	return ret;
+}
+
 static int __cmd_report(struct report *rep)
 {
 	int ret = -EINVAL;
 	u64 nr_samples;
 	struct perf_session *session = rep->session;
 	struct perf_evsel *pos;
-	const char *help = "For a higher level overview, try: perf report --sort comm,dso";
 	struct ui_progress prog;
 	struct perf_data_file *file = session->file;
 
@@ -524,38 +568,7 @@ static int __cmd_report(struct report *rep)
 	list_for_each_entry(pos, &session->evlist->entries, node)
 		hists__output_resort(&pos->hists);
 
-	if (use_browser > 0) {
-		if (use_browser == 1) {
-			ret = perf_evlist__tui_browse_hists(session->evlist,
-							help, NULL,
-							rep->min_percent,
-							&session->header.env);
-			/*
-			 * Usually "ret" is the last pressed key, and we only
-			 * care if the key notifies us to switch data file.
-			 */
-			if (ret != K_SWITCH_INPUT_DATA)
-				ret = 0;
-
-		} else if (use_browser == 2) {
-			int (*hist_browser)(struct perf_evlist *,
-					    const char *,
-					    struct hist_browser_timer *,
-					    float min_pcnt);
-
-			hist_browser = dlsym(perf_gtk_handle,
-					     "perf_evlist__gtk_browse_hists");
-			if (hist_browser == NULL) {
-				ui__error("GTK browser not found!\n");
-				return ret;
-			}
-			hist_browser(session->evlist, help, NULL,
-				     rep->min_percent);
-		}
-	} else
-		perf_evlist__tty_browse_hists(session->evlist, rep, help);
-
-	return ret;
+	return report__browse_hists(rep);
 }
 
 static int
-- 
1.8.1.4

[PATCH 23/38] perf report: Move histogram entries collapsing to separate function

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Further uncluttering the main 'report' function by group related code in
separate function.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-b594zsbwke8khir13kudwqmj@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-report.c | 73 ++++++++++++++++++++++++++++-----------------
 1 file changed, 45 insertions(+), 28 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 03941ad8fc46..cff9465847f2 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -486,13 +486,55 @@ static int report__browse_hists(struct report *rep)
 	return ret;
 }
 
+static u64 report__collapse_hists(struct report *rep)
+{
+	struct ui_progress prog;
+	struct perf_evsel *pos;
+	u64 nr_samples = 0;
+	/*
+ 	 * Count number of histogram entries to use when showing progress,
+ 	 * reusing nr_samples variable.
+ 	 */
+	list_for_each_entry(pos, &rep->session->evlist->entries, node)
+		nr_samples += pos->hists.nr_entries;
+
+	ui_progress__init(&prog, nr_samples, "Merging related events...");
+	/*
+	 * Count total number of samples, will be used to check if this
+ 	 * session had any.
+ 	 */
+	nr_samples = 0;
+
+	list_for_each_entry(pos, &rep->session->evlist->entries, node) {
+		struct hists *hists = &pos->hists;
+
+		if (pos->idx == 0)
+			hists->symbol_filter_str = rep->symbol_filter_str;
+
+		hists__collapse_resort(hists, &prog);
+		nr_samples += hists->stats.nr_events[PERF_RECORD_SAMPLE];
+
+		/* Non-group events are considered as leader */
+		if (symbol_conf.event_group &&
+		    !perf_evsel__is_group_leader(pos)) {
+			struct hists *leader_hists = &pos->leader->hists;
+
+			hists__match(leader_hists, hists);
+			hists__link(leader_hists, hists);
+		}
+	}
+
+	ui_progress__finish();
+
+	return nr_samples;
+}
+
 static int __cmd_report(struct report *rep)
 {
-	int ret = -EINVAL;
+	int ret;
 	u64 nr_samples;
 	struct perf_session *session = rep->session;
 	struct perf_evsel *pos;
-	struct ui_progress prog;
 	struct perf_data_file *file = session->file;
 
 	signal(SIGINT, sig_handler);
@@ -530,32 +572,7 @@ static int __cmd_report(struct report *rep)
 		}
 	}
 
-	nr_samples = 0;
-	list_for_each_entry(pos, &session->evlist->entries, node)
-		nr_samples += pos->hists.nr_entries;
-
-	ui_progress__init(&prog, nr_samples, "Merging related events...");
-
-	nr_samples = 0;
-	list_for_each_entry(pos, &session->evlist->entries, node) {
-		struct hists *hists = &pos->hists;
-
-		if (pos->idx == 0)
-			hists->symbol_filter_str = rep->symbol_filter_str;
-
-		hists__collapse_resort(hists, &prog);
-		nr_samples += hists->stats.nr_events[PERF_RECORD_SAMPLE];
-
-		/* Non-group events are considered as leader */
-		if (symbol_conf.event_group &&
-		    !perf_evsel__is_group_leader(pos)) {
-			struct hists *leader_hists = &pos->leader->hists;
-
-			hists__match(leader_hists, hists);
-			hists__link(leader_hists, hists);
-		}
-	}
-	ui_progress__finish();
+	nr_samples = report__collapse_hists(rep);
 
 	if (session_done())
 		return 0;
-- 
1.8.1.4

[PATCH 24/38] perf evlist: Introduce evlist__for_each() & friends

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

For the common evsel list traversal, so that it becomes more compact.

Use the opportunity to start ditching the 'perf_' from 'perf_evlist__',
as discussed, as the whole conversion touches a lot of places, lets do
it piecemeal when we have the chance due to other work, like in this
case.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-qnkx7dzm2h6m6uptkfk03ni6@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-annotate.c           |  2 +-
 tools/perf/builtin-diff.c               |  7 ++--
 tools/perf/builtin-evlist.c             |  2 +-
 tools/perf/builtin-inject.c             |  2 +-
 tools/perf/builtin-kvm.c                |  2 +-
 tools/perf/builtin-record.c             |  2 +-
 tools/perf/builtin-report.c             |  8 ++--
 tools/perf/builtin-script.c             |  5 +--
 tools/perf/builtin-stat.c               | 32 ++++++++--------
 tools/perf/builtin-top.c                |  8 ++--
 tools/perf/tests/evsel-roundtrip-name.c |  2 +-
 tools/perf/tests/hists_link.c           |  4 +-
 tools/perf/tests/mmap-basic.c           |  2 +-
 tools/perf/tests/parse-events.c         |  4 +-
 tools/perf/ui/browsers/hists.c          |  5 ++-
 tools/perf/ui/gtk/hists.c               |  2 +-
 tools/perf/util/cgroup.c                |  4 +-
 tools/perf/util/evlist.c                | 40 ++++++++++----------
 tools/perf/util/evlist.h                | 65 +++++++++++++++++++++++++++++++++
 tools/perf/util/header.c                | 19 +++++-----
 tools/perf/util/parse-events.c          |  5 +--
 tools/perf/util/python.c                |  3 +-
 tools/perf/util/record.c                |  6 +--
 tools/perf/util/session.c               |  6 +--
 24 files changed, 151 insertions(+), 86 deletions(-)

diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index ab65057a0317..0da603b79b61 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -232,7 +232,7 @@ static int __cmd_annotate(struct perf_annotate *ann)
 		perf_session__fprintf_dsos(session, stdout);
 
 	total_nr_samples = 0;
-	list_for_each_entry(pos, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, pos) {
 		struct hists *hists = &pos->hists;
 		u32 nr_samples = hists->stats.nr_events[PERF_RECORD_SAMPLE];
 
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index e6a0844bc2f0..987cac3b4bba 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -356,9 +356,10 @@ static struct perf_evsel *evsel_match(struct perf_evsel *evsel,
 {
 	struct perf_evsel *e;
 
-	list_for_each_entry(e, &evlist->entries, node)
+	evlist__for_each(evlist, e) {
 		if (perf_evsel__match2(evsel, e))
 			return e;
+	}
 
 	return NULL;
 }
@@ -367,7 +368,7 @@ static void perf_evlist__collapse_resort(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		struct hists *hists = &evsel->hists;
 
 		hists__collapse_resort(hists, NULL);
@@ -614,7 +615,7 @@ static void data_process(void)
 	struct perf_evsel *evsel_base;
 	bool first = true;
 
-	list_for_each_entry(evsel_base, &evlist_base->entries, node) {
+	evlist__for_each(evlist_base, evsel_base) {
 		struct data__file *d;
 		int i;
 
diff --git a/tools/perf/builtin-evlist.c b/tools/perf/builtin-evlist.c
index 20b0f12763b0..c99e0de7e54a 100644
--- a/tools/perf/builtin-evlist.c
+++ b/tools/perf/builtin-evlist.c
@@ -29,7 +29,7 @@ static int __cmd_evlist(const char *file_name, struct perf_attr_details *details
 	if (session == NULL)
 		return -ENOMEM;
 
-	list_for_each_entry(pos, &session->evlist->entries, node)
+	evlist__for_each(session->evlist, pos)
 		perf_evsel__fprintf(pos, details, stdout);
 
 	perf_session__delete(session);
diff --git a/tools/perf/builtin-inject.c b/tools/perf/builtin-inject.c
index c9f6d74e1fd7..b3466018bbd7 100644
--- a/tools/perf/builtin-inject.c
+++ b/tools/perf/builtin-inject.c
@@ -369,7 +369,7 @@ static int __cmd_inject(struct perf_inject *inject)
 
 		inject->tool.ordered_samples = true;
 
-		list_for_each_entry(evsel, &session->evlist->entries, node) {
+		evlist__for_each(session->evlist, evsel) {
 			const char *name = perf_evsel__name(evsel);
 
 			if (!strcmp(name, "sched:sched_switch")) {
diff --git a/tools/perf/builtin-kvm.c b/tools/perf/builtin-kvm.c
index 858b11bf6711..a7350519c63f 100644
--- a/tools/perf/builtin-kvm.c
+++ b/tools/perf/builtin-kvm.c
@@ -1174,7 +1174,7 @@ static int kvm_live_open_events(struct perf_kvm_stat *kvm)
 	 * Note: exclude_{guest,host} do not apply here.
 	 *       This command processes KVM tracepoints from host only
 	 */
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		struct perf_event_attr *attr = &pos->attr;
 
 		/* make sure these *are* set */
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index cb00b53c200f..88600158400e 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -183,7 +183,7 @@ static int record__open(struct record *rec)
 
 	perf_evlist__config(evlist, opts);
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 try_again:
 		if (perf_evsel__open(pos, evlist->cpus, evlist->threads) < 0) {
 			if (perf_evsel__fallback(pos, errno, msg, sizeof(msg))) {
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index cff9465847f2..46864dd7eb83 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -384,7 +384,7 @@ static int perf_evlist__tty_browse_hists(struct perf_evlist *evlist,
 {
 	struct perf_evsel *pos;
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		struct hists *hists = &pos->hists;
 		const char *evname = perf_evsel__name(pos);
 
@@ -495,7 +495,7 @@ static u64 report__collapse_hists(struct report *rep)
  	 * Count number of histogram entries to use when showing progress,
  	 * reusing nr_samples variable.
  	 */
-	list_for_each_entry(pos, &rep->session->evlist->entries, node)
+	evlist__for_each(rep->session->evlist, pos)
 		nr_samples += pos->hists.nr_entries;
 
 	ui_progress__init(&prog, nr_samples, "Merging related events...");
@@ -505,7 +505,7 @@ static u64 report__collapse_hists(struct report *rep)
  	 */
 	nr_samples = 0;
 
-	list_for_each_entry(pos, &rep->session->evlist->entries, node) {
+	evlist__for_each(rep->session->evlist, pos) {
 		struct hists *hists = &pos->hists;
 
 		if (pos->idx == 0)
@@ -582,7 +582,7 @@ static int __cmd_report(struct report *rep)
 		return 0;
 	}
 
-	list_for_each_entry(pos, &session->evlist->entries, node)
+	evlist__for_each(session->evlist, pos)
 		hists__output_resort(&pos->hists);
 
 	return report__browse_hists(rep);
diff --git a/tools/perf/builtin-script.c b/tools/perf/builtin-script.c
index 6040000bdfa6..9e9c91f5b7fa 100644
--- a/tools/perf/builtin-script.c
+++ b/tools/perf/builtin-script.c
@@ -603,7 +603,7 @@ static int process_attr(struct perf_tool *tool, union perf_event *event,
 	if (evsel->attr.type >= PERF_TYPE_MAX)
 		return 0;
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		if (pos->attr.type == evsel->attr.type && pos != evsel)
 			return 0;
 	}
@@ -1309,8 +1309,7 @@ static int check_ev_match(char *dir_name, char *scriptname,
 			snprintf(evname, len + 1, "%s", p);
 
 			match = 0;
-			list_for_each_entry(pos,
-					&session->evlist->entries, node) {
+			evlist__for_each(session->evlist, pos) {
 				if (!strcmp(perf_evsel__name(pos), evname)) {
 					match = 1;
 					break;
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index b27b2645eb3c..8b0e1c9234d9 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -214,7 +214,7 @@ static void perf_evlist__free_stats(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		perf_evsel__free_stat_priv(evsel);
 		perf_evsel__free_counts(evsel);
 		perf_evsel__free_prev_raw_counts(evsel);
@@ -225,7 +225,7 @@ static int perf_evlist__alloc_stats(struct perf_evlist *evlist, bool alloc_raw)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (perf_evsel__alloc_stat_priv(evsel) < 0 ||
 		    perf_evsel__alloc_counts(evsel, perf_evsel__nr_cpus(evsel)) < 0 ||
 		    (alloc_raw && perf_evsel__alloc_prev_raw_counts(evsel) < 0))
@@ -259,7 +259,7 @@ static void perf_stat__reset_stats(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		perf_evsel__reset_stat_priv(evsel);
 		perf_evsel__reset_counts(evsel, perf_evsel__nr_cpus(evsel));
 	}
@@ -326,13 +326,13 @@ static struct perf_evsel *nth_evsel(int n)
 
 	/* Assumes this only called when evsel_list does not change anymore. */
 	if (!array) {
-		list_for_each_entry(ev, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, ev)
 			array_len++;
 		array = malloc(array_len * sizeof(void *));
 		if (!array)
 			exit(ENOMEM);
 		j = 0;
-		list_for_each_entry(ev, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, ev)
 			array[j++] = ev;
 	}
 	if (n < array_len)
@@ -440,13 +440,13 @@ static void print_interval(void)
 	char prefix[64];
 
 	if (aggr_mode == AGGR_GLOBAL) {
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+		evlist__for_each(evsel_list, counter) {
 			ps = counter->priv;
 			memset(ps->res_stats, 0, sizeof(ps->res_stats));
 			read_counter_aggr(counter);
 		}
 	} else	{
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+		evlist__for_each(evsel_list, counter) {
 			ps = counter->priv;
 			memset(ps->res_stats, 0, sizeof(ps->res_stats));
 			read_counter(counter);
@@ -483,12 +483,12 @@ static void print_interval(void)
 		print_aggr(prefix);
 		break;
 	case AGGR_NONE:
-		list_for_each_entry(counter, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, counter)
 			print_counter(counter, prefix);
 		break;
 	case AGGR_GLOBAL:
 	default:
-		list_for_each_entry(counter, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, counter)
 			print_counter_aggr(counter, prefix);
 	}
 
@@ -504,7 +504,7 @@ static void handle_initial_delay(void)
 			nthreads = thread_map__nr(evsel_list->threads);
 
 		usleep(initial_delay * 1000);
-		list_for_each_entry(counter, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, counter)
 			perf_evsel__enable(counter, ncpus, nthreads);
 	}
 }
@@ -552,7 +552,7 @@ static int __run_perf_stat(int argc, const char **argv)
 	if (group)
 		perf_evlist__set_leader(evsel_list);
 
-	list_for_each_entry(counter, &evsel_list->entries, node) {
+	evlist__for_each(evsel_list, counter) {
 		if (create_perf_stat_counter(counter) < 0) {
 			/*
 			 * PPC returns ENXIO for HW counters until 2.6.37
@@ -630,13 +630,13 @@ static int __run_perf_stat(int argc, const char **argv)
 	update_stats(&walltime_nsecs_stats, t1 - t0);
 
 	if (aggr_mode == AGGR_GLOBAL) {
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+		evlist__for_each(evsel_list, counter) {
 			read_counter_aggr(counter);
 			perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter),
 					     thread_map__nr(evsel_list->threads));
 		}
 	} else {
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+		evlist__for_each(evsel_list, counter) {
 			read_counter(counter);
 			perf_evsel__close_fd(counter, perf_evsel__nr_cpus(counter), 1);
 		}
@@ -1117,7 +1117,7 @@ static void print_aggr(char *prefix)
 
 	for (s = 0; s < aggr_map->nr; s++) {
 		id = aggr_map->map[s];
-		list_for_each_entry(counter, &evsel_list->entries, node) {
+		evlist__for_each(evsel_list, counter) {
 			val = ena = run = 0;
 			nr = 0;
 			for (cpu = 0; cpu < perf_evsel__nr_cpus(counter); cpu++) {
@@ -1328,11 +1328,11 @@ static void print_stat(int argc, const char **argv)
 		print_aggr(NULL);
 		break;
 	case AGGR_GLOBAL:
-		list_for_each_entry(counter, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, counter)
 			print_counter_aggr(counter, NULL);
 		break;
 	case AGGR_NONE:
-		list_for_each_entry(counter, &evsel_list->entries, node)
+		evlist__for_each(evsel_list, counter)
 			print_counter(counter, NULL);
 		break;
 	default:
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index e0fd0aa57f06..569dd87690ef 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -482,7 +482,7 @@ static bool perf_top__handle_keypress(struct perf_top *top, int c)
 
 				fprintf(stderr, "\nAvailable events:");
 
-				list_for_each_entry(top->sym_evsel, &top->evlist->entries, node)
+				evlist__for_each(top->evlist, top->sym_evsel)
 					fprintf(stderr, "\n\t%d %s", top->sym_evsel->idx, perf_evsel__name(top->sym_evsel));
 
 				prompt_integer(&counter, "Enter details event counter");
@@ -493,7 +493,7 @@ static bool perf_top__handle_keypress(struct perf_top *top, int c)
 					sleep(1);
 					break;
 				}
-				list_for_each_entry(top->sym_evsel, &top->evlist->entries, node)
+				evlist__for_each(top->evlist, top->sym_evsel)
 					if (top->sym_evsel->idx == counter)
 						break;
 			} else
@@ -575,7 +575,7 @@ static void *display_thread_tui(void *arg)
 	 * Zooming in/out UIDs. For now juse use whatever the user passed
 	 * via --uid.
 	 */
-	list_for_each_entry(pos, &top->evlist->entries, node)
+	evlist__for_each(top->evlist, pos)
 		pos->hists.uid_filter_str = top->record_opts.target.uid_str;
 
 	perf_evlist__tui_browse_hists(top->evlist, help, &hbt, top->min_percent,
@@ -858,7 +858,7 @@ static int perf_top__start_counters(struct perf_top *top)
 
 	perf_evlist__config(evlist, opts);
 
-	list_for_each_entry(counter, &evlist->entries, node) {
+	evlist__for_each(evlist, counter) {
 try_again:
 		if (perf_evsel__open(counter, top->evlist->cpus,
 				     top->evlist->threads) < 0) {
diff --git a/tools/perf/tests/evsel-roundtrip-name.c b/tools/perf/tests/evsel-roundtrip-name.c
index 0197bda9c461..465cdbc345cf 100644
--- a/tools/perf/tests/evsel-roundtrip-name.c
+++ b/tools/perf/tests/evsel-roundtrip-name.c
@@ -79,7 +79,7 @@ static int __perf_evsel__name_array_test(const char *names[], int nr_names)
 	}
 
 	err = 0;
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (strcmp(perf_evsel__name(evsel), names[evsel->idx])) {
 			--err;
 			pr_debug("%s != %s\n", perf_evsel__name(evsel), names[evsel->idx]);
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index 173bf42cc03e..2b6519e0e36f 100644
--- a/tools/perf/tests/hists_link.c
+++ b/tools/perf/tests/hists_link.c
@@ -208,7 +208,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 	 * However the second evsel also has a collapsed entry for
 	 * "bash [libc] malloc" so total 9 entries will be in the tree.
 	 */
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		for (k = 0; k < ARRAY_SIZE(fake_common_samples); k++) {
 			const union perf_event event = {
 				.header = {
@@ -466,7 +466,7 @@ int test__hists_link(void)
 	if (err < 0)
 		goto out;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		hists__collapse_resort(&evsel->hists, NULL);
 
 		if (verbose > 2)
diff --git a/tools/perf/tests/mmap-basic.c b/tools/perf/tests/mmap-basic.c
index aef25f0ff9cc..142263492f6f 100644
--- a/tools/perf/tests/mmap-basic.c
+++ b/tools/perf/tests/mmap-basic.c
@@ -126,7 +126,7 @@ int test__basic_mmap(void)
 	}
 
 	err = 0;
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (nr_events[evsel->idx] != expected_nr_events[evsel->idx]) {
 			pr_debug("expected %d %s events, got %d\n",
 				 expected_nr_events[evsel->idx],
diff --git a/tools/perf/tests/parse-events.c b/tools/perf/tests/parse-events.c
index 41dee5e67f63..4db0ae617d70 100644
--- a/tools/perf/tests/parse-events.c
+++ b/tools/perf/tests/parse-events.c
@@ -30,7 +30,7 @@ static int test__checkevent_tracepoint_multi(struct perf_evlist *evlist)
 	TEST_ASSERT_VAL("wrong number of entries", evlist->nr_entries > 1);
 	TEST_ASSERT_VAL("wrong number of groups", 0 == evlist->nr_groups);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		TEST_ASSERT_VAL("wrong type",
 			PERF_TYPE_TRACEPOINT == evsel->attr.type);
 		TEST_ASSERT_VAL("wrong sample_type",
@@ -201,7 +201,7 @@ test__checkevent_tracepoint_multi_modifier(struct perf_evlist *evlist)
 
 	TEST_ASSERT_VAL("wrong number of entries", evlist->nr_entries > 1);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		TEST_ASSERT_VAL("wrong exclude_user",
 				!evsel->attr.exclude_user);
 		TEST_ASSERT_VAL("wrong exclude_kernel",
diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
index a7045ea6d1d5..b720b92eba6e 100644
--- a/tools/perf/ui/browsers/hists.c
+++ b/tools/perf/ui/browsers/hists.c
@@ -1938,7 +1938,7 @@ static int __perf_evlist__tui_browse_hists(struct perf_evlist *evlist,
 
 	ui_helpline__push("Press ESC to exit");
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		const char *ev_name = perf_evsel__name(pos);
 		size_t line_len = strlen(ev_name) + 7;
 
@@ -1970,9 +1970,10 @@ single_entry:
 		struct perf_evsel *pos;
 
 		nr_entries = 0;
-		list_for_each_entry(pos, &evlist->entries, node)
+		evlist__for_each(evlist, pos) {
 			if (perf_evsel__is_group_leader(pos))
 				nr_entries++;
+		}
 
 		if (nr_entries == 1)
 			goto single_entry;
diff --git a/tools/perf/ui/gtk/hists.c b/tools/perf/ui/gtk/hists.c
index 2ca66cc1160f..5b95c44f3435 100644
--- a/tools/perf/ui/gtk/hists.c
+++ b/tools/perf/ui/gtk/hists.c
@@ -375,7 +375,7 @@ int perf_evlist__gtk_browse_hists(struct perf_evlist *evlist,
 
 	gtk_container_add(GTK_CONTAINER(window), vbox);
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		struct hists *hists = &pos->hists;
 		const char *evname = perf_evsel__name(pos);
 		GtkWidget *scrolled_window;
diff --git a/tools/perf/util/cgroup.c b/tools/perf/util/cgroup.c
index 0922aa4218c2..88f7be399432 100644
--- a/tools/perf/util/cgroup.c
+++ b/tools/perf/util/cgroup.c
@@ -81,7 +81,7 @@ static int add_cgroup(struct perf_evlist *evlist, char *str)
 	/*
 	 * check if cgrp is already defined, if so we reuse it
 	 */
-	list_for_each_entry(counter, &evlist->entries, node) {
+	evlist__for_each(evlist, counter) {
 		cgrp = counter->cgrp;
 		if (!cgrp)
 			continue;
@@ -110,7 +110,7 @@ static int add_cgroup(struct perf_evlist *evlist, char *str)
 	 * if add cgroup N, then need to find event N
 	 */
 	n = 0;
-	list_for_each_entry(counter, &evlist->entries, node) {
+	evlist__for_each(evlist, counter) {
 		if (n == nr_cgroups)
 			goto found;
 		n++;
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 0810f5ca77d7..40bd2c04df8a 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -81,7 +81,7 @@ static void perf_evlist__update_id_pos(struct perf_evlist *evlist)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node)
+	evlist__for_each(evlist, evsel)
 		perf_evsel__calc_id_pos(evsel);
 
 	perf_evlist__set_id_pos(evlist);
@@ -91,7 +91,7 @@ static void perf_evlist__purge(struct perf_evlist *evlist)
 {
 	struct perf_evsel *pos, *n;
 
-	list_for_each_entry_safe(pos, n, &evlist->entries, node) {
+	evlist__for_each_safe(evlist, n, pos) {
 		list_del_init(&pos->node);
 		perf_evsel__delete(pos);
 	}
@@ -148,7 +148,7 @@ void __perf_evlist__set_leader(struct list_head *list)
 
 	leader->nr_members = evsel->idx - leader->idx + 1;
 
-	list_for_each_entry(evsel, list, node) {
+	__evlist__for_each(list, evsel) {
 		evsel->leader = leader;
 	}
 }
@@ -207,7 +207,7 @@ static int perf_evlist__add_attrs(struct perf_evlist *evlist,
 	return 0;
 
 out_delete_partial_list:
-	list_for_each_entry_safe(evsel, n, &head, node)
+	__evlist__for_each_safe(&head, n, evsel)
 		perf_evsel__delete(evsel);
 	return -1;
 }
@@ -228,7 +228,7 @@ perf_evlist__find_tracepoint_by_id(struct perf_evlist *evlist, int id)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (evsel->attr.type   == PERF_TYPE_TRACEPOINT &&
 		    (int)evsel->attr.config == id)
 			return evsel;
@@ -243,7 +243,7 @@ perf_evlist__find_tracepoint_by_name(struct perf_evlist *evlist,
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if ((evsel->attr.type == PERF_TYPE_TRACEPOINT) &&
 		    (strcmp(evsel->name, name) == 0))
 			return evsel;
@@ -273,7 +273,7 @@ void perf_evlist__disable(struct perf_evlist *evlist)
 	int nr_threads = thread_map__nr(evlist->threads);
 
 	for (cpu = 0; cpu < nr_cpus; cpu++) {
-		list_for_each_entry(pos, &evlist->entries, node) {
+		evlist__for_each(evlist, pos) {
 			if (!perf_evsel__is_group_leader(pos) || !pos->fd)
 				continue;
 			for (thread = 0; thread < nr_threads; thread++)
@@ -291,7 +291,7 @@ void perf_evlist__enable(struct perf_evlist *evlist)
 	int nr_threads = thread_map__nr(evlist->threads);
 
 	for (cpu = 0; cpu < nr_cpus; cpu++) {
-		list_for_each_entry(pos, &evlist->entries, node) {
+		evlist__for_each(evlist, pos) {
 			if (!perf_evsel__is_group_leader(pos) || !pos->fd)
 				continue;
 			for (thread = 0; thread < nr_threads; thread++)
@@ -630,7 +630,7 @@ static int perf_evlist__mmap_per_evsel(struct perf_evlist *evlist, int idx,
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		int fd = FD(evsel, cpu, thread);
 
 		if (*output == -1) {
@@ -806,7 +806,7 @@ int perf_evlist__mmap(struct perf_evlist *evlist, unsigned int pages,
 	pr_debug("mmap size %zuB\n", evlist->mmap_len);
 	mask = evlist->mmap_len - page_size - 1;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if ((evsel->attr.read_format & PERF_FORMAT_ID) &&
 		    evsel->sample_id == NULL &&
 		    perf_evsel__alloc_id(evsel, cpu_map__nr(cpus), threads->nr) < 0)
@@ -849,7 +849,7 @@ int perf_evlist__apply_filters(struct perf_evlist *evlist)
 	const int ncpus = cpu_map__nr(evlist->cpus),
 		  nthreads = thread_map__nr(evlist->threads);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (evsel->filter == NULL)
 			continue;
 
@@ -868,7 +868,7 @@ int perf_evlist__set_filter(struct perf_evlist *evlist, const char *filter)
 	const int ncpus = cpu_map__nr(evlist->cpus),
 		  nthreads = thread_map__nr(evlist->threads);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		err = perf_evsel__set_filter(evsel, ncpus, nthreads, filter);
 		if (err)
 			break;
@@ -887,7 +887,7 @@ bool perf_evlist__valid_sample_type(struct perf_evlist *evlist)
 	if (evlist->id_pos < 0 || evlist->is_pos < 0)
 		return false;
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		if (pos->id_pos != evlist->id_pos ||
 		    pos->is_pos != evlist->is_pos)
 			return false;
@@ -903,7 +903,7 @@ u64 __perf_evlist__combined_sample_type(struct perf_evlist *evlist)
 	if (evlist->combined_sample_type)
 		return evlist->combined_sample_type;
 
-	list_for_each_entry(evsel, &evlist->entries, node)
+	evlist__for_each(evlist, evsel)
 		evlist->combined_sample_type |= evsel->attr.sample_type;
 
 	return evlist->combined_sample_type;
@@ -921,7 +921,7 @@ bool perf_evlist__valid_read_format(struct perf_evlist *evlist)
 	u64 read_format = first->attr.read_format;
 	u64 sample_type = first->attr.sample_type;
 
-	list_for_each_entry_continue(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		if (read_format != pos->attr.read_format)
 			return false;
 	}
@@ -978,7 +978,7 @@ bool perf_evlist__valid_sample_id_all(struct perf_evlist *evlist)
 {
 	struct perf_evsel *first = perf_evlist__first(evlist), *pos = first;
 
-	list_for_each_entry_continue(pos, &evlist->entries, node) {
+	evlist__for_each_continue(evlist, pos) {
 		if (first->attr.sample_id_all != pos->attr.sample_id_all)
 			return false;
 	}
@@ -1004,7 +1004,7 @@ void perf_evlist__close(struct perf_evlist *evlist)
 	int ncpus = cpu_map__nr(evlist->cpus);
 	int nthreads = thread_map__nr(evlist->threads);
 
-	list_for_each_entry_reverse(evsel, &evlist->entries, node)
+	evlist__for_each_reverse(evlist, evsel)
 		perf_evsel__close(evsel, ncpus, nthreads);
 }
 
@@ -1015,7 +1015,7 @@ int perf_evlist__open(struct perf_evlist *evlist)
 
 	perf_evlist__update_id_pos(evlist);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		err = perf_evsel__open(evsel, evlist->cpus, evlist->threads);
 		if (err < 0)
 			goto out_err;
@@ -1154,7 +1154,7 @@ size_t perf_evlist__fprintf(struct perf_evlist *evlist, FILE *fp)
 	struct perf_evsel *evsel;
 	size_t printed = 0;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		printed += fprintf(fp, "%s%s", evsel->idx ? ", " : "",
 				   perf_evsel__name(evsel));
 	}
@@ -1233,7 +1233,7 @@ void perf_evlist__to_front(struct perf_evlist *evlist,
 	if (move_evsel == perf_evlist__first(evlist))
 		return;
 
-	list_for_each_entry_safe(evsel, n, &evlist->entries, node) {
+	evlist__for_each_safe(evlist, n, evsel) {
 		if (evsel->leader == move_evsel->leader)
 			list_move_tail(&evsel->node, &move);
 	}
diff --git a/tools/perf/util/evlist.h b/tools/perf/util/evlist.h
index 518e521c801d..f5173cd63693 100644
--- a/tools/perf/util/evlist.h
+++ b/tools/perf/util/evlist.h
@@ -196,5 +196,70 @@ bool perf_evlist__can_select_event(struct perf_evlist *evlist, const char *str);
 void perf_evlist__to_front(struct perf_evlist *evlist,
 			   struct perf_evsel *move_evsel);
 
+/**
+ * __evlist__for_each - iterate thru all the evsels
+ * @list: list_head instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define __evlist__for_each(list, evsel) \
+        list_for_each_entry(evsel, list, node)
+
+/**
+ * evlist__for_each - iterate thru all the evsels
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define evlist__for_each(evlist, evsel) \
+	__evlist__for_each(&(evlist)->entries, evsel)
+
+/**
+ * __evlist__for_each_continue - continue iteration thru all the evsels
+ * @list: list_head instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define __evlist__for_each_continue(list, evsel) \
+        list_for_each_entry_continue(evsel, list, node)
+
+/**
+ * evlist__for_each_continue - continue iteration thru all the evsels
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define evlist__for_each_continue(evlist, evsel) \
+	__evlist__for_each_continue(&(evlist)->entries, evsel)
+
+/**
+ * __evlist__for_each_reverse - iterate thru all the evsels in reverse order
+ * @list: list_head instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define __evlist__for_each_reverse(list, evsel) \
+        list_for_each_entry_reverse(evsel, list, node)
+
+/**
+ * evlist__for_each_reverse - iterate thru all the evsels in reverse order
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ */
+#define evlist__for_each_reverse(evlist, evsel) \
+	__evlist__for_each_reverse(&(evlist)->entries, evsel)
+
+/**
+ * __evlist__for_each_safe - safely iterate thru all the evsels
+ * @list: list_head instance to iterate
+ * @tmp: struct evsel temp iterator
+ * @evsel: struct evsel iterator
+ */
+#define __evlist__for_each_safe(list, tmp, evsel) \
+        list_for_each_entry_safe(evsel, tmp, list, node)
+
+/**
+ * evlist__for_each_safe - safely iterate thru all the evsels
+ * @evlist: evlist instance to iterate
+ * @evsel: struct evsel iterator
+ * @tmp: struct evsel temp iterator
+ */
+#define evlist__for_each_safe(evlist, tmp, evsel) \
+	__evlist__for_each_safe(&(evlist)->entries, tmp, evsel)
 
 #endif /* __PERF_EVLIST_H */
diff --git a/tools/perf/util/header.c b/tools/perf/util/header.c
index a4a60b7887ee..bb3e0ede6183 100644
--- a/tools/perf/util/header.c
+++ b/tools/perf/util/header.c
@@ -643,8 +643,7 @@ static int write_event_desc(int fd, struct perf_header *h __maybe_unused,
 	if (ret < 0)
 		return ret;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
-
+	evlist__for_each(evlist, evsel) {
 		ret = do_write(fd, &evsel->attr, sz);
 		if (ret < 0)
 			return ret;
@@ -1092,7 +1091,7 @@ static int write_group_desc(int fd, struct perf_header *h __maybe_unused,
 	if (ret < 0)
 		return ret;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (perf_evsel__is_group_leader(evsel) &&
 		    evsel->nr_members > 1) {
 			const char *name = evsel->group_name ?: "{anon_group}";
@@ -1487,7 +1486,7 @@ static void print_group_desc(struct perf_header *ph, int fd __maybe_unused,
 
 	session = container_of(ph, struct perf_session, header);
 
-	list_for_each_entry(evsel, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, evsel) {
 		if (perf_evsel__is_group_leader(evsel) &&
 		    evsel->nr_members > 1) {
 			fprintf(fp, "# group: %s{%s", evsel->group_name ?: "",
@@ -1768,7 +1767,7 @@ perf_evlist__find_by_index(struct perf_evlist *evlist, int idx)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		if (evsel->idx == idx)
 			return evsel;
 	}
@@ -2071,7 +2070,7 @@ static int process_group_desc(struct perf_file_section *section __maybe_unused,
 	session->evlist->nr_groups = nr_groups;
 
 	i = nr = 0;
-	list_for_each_entry(evsel, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, evsel) {
 		if (evsel->idx == (int) desc[i].leader_idx) {
 			evsel->leader = evsel;
 			/* {anon_group} is a dummy name */
@@ -2298,7 +2297,7 @@ int perf_session__write_header(struct perf_session *session,
 
 	lseek(fd, sizeof(f_header), SEEK_SET);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(session->evlist, evsel) {
 		evsel->id_offset = lseek(fd, 0, SEEK_CUR);
 		err = do_write(fd, evsel->id, evsel->ids * sizeof(u64));
 		if (err < 0) {
@@ -2309,7 +2308,7 @@ int perf_session__write_header(struct perf_session *session,
 
 	attr_offset = lseek(fd, 0, SEEK_CUR);
 
-	list_for_each_entry(evsel, &evlist->entries, node) {
+	evlist__for_each(evlist, evsel) {
 		f_attr = (struct perf_file_attr){
 			.attr = evsel->attr,
 			.ids  = {
@@ -2742,7 +2741,7 @@ static int perf_evlist__prepare_tracepoint_events(struct perf_evlist *evlist,
 {
 	struct perf_evsel *pos;
 
-	list_for_each_entry(pos, &evlist->entries, node) {
+	evlist__for_each(evlist, pos) {
 		if (pos->attr.type == PERF_TYPE_TRACEPOINT &&
 		    perf_evsel__prepare_tracepoint_event(pos, pevent))
 			return -1;
@@ -2890,7 +2889,7 @@ int perf_event__synthesize_attrs(struct perf_tool *tool,
 	struct perf_evsel *evsel;
 	int err = 0;
 
-	list_for_each_entry(evsel, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, evsel) {
 		err = perf_event__synthesize_attr(tool, &evsel->attr, evsel->ids,
 						  evsel->id, process);
 		if (err) {
diff --git a/tools/perf/util/parse-events.c b/tools/perf/util/parse-events.c
index 0153435b8427..a7f1b6a91fdd 100644
--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -820,8 +820,7 @@ int parse_events__modifier_event(struct list_head *list, char *str, bool add)
 	if (!add && get_event_modifier(&mod, str, NULL))
 		return -EINVAL;
 
-	list_for_each_entry(evsel, list, node) {
-
+	__evlist__for_each(list, evsel) {
 		if (add && get_event_modifier(&mod, str, evsel))
 			return -EINVAL;
 
@@ -845,7 +844,7 @@ int parse_events_name(struct list_head *list, char *name)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, list, node) {
+	__evlist__for_each(list, evsel) {
 		if (!evsel->name)
 			evsel->name = strdup(name);
 	}
diff --git a/tools/perf/util/python.c b/tools/perf/util/python.c
index 4bf8ace7f511..122669c18ff4 100644
--- a/tools/perf/util/python.c
+++ b/tools/perf/util/python.c
@@ -908,9 +908,10 @@ static PyObject *pyrf_evlist__item(PyObject *obj, Py_ssize_t i)
 	if (i >= pevlist->evlist.nr_entries)
 		return NULL;
 
-	list_for_each_entry(pos, &pevlist->evlist.entries, node)
+	evlist__for_each(&pevlist->evlist, pos) {
 		if (i-- == 0)
 			break;
+	}
 
 	return Py_BuildValue("O", container_of(pos, struct pyrf_evsel, evsel));
 }
diff --git a/tools/perf/util/record.c b/tools/perf/util/record.c
index 104a47563d39..373762501dad 100644
--- a/tools/perf/util/record.c
+++ b/tools/perf/util/record.c
@@ -89,19 +89,19 @@ void perf_evlist__config(struct perf_evlist *evlist, struct record_opts *opts)
 	if (evlist->cpus->map[0] < 0)
 		opts->no_inherit = true;
 
-	list_for_each_entry(evsel, &evlist->entries, node)
+	evlist__for_each(evlist, evsel)
 		perf_evsel__config(evsel, opts);
 
 	if (evlist->nr_entries > 1) {
 		struct perf_evsel *first = perf_evlist__first(evlist);
 
-		list_for_each_entry(evsel, &evlist->entries, node) {
+		evlist__for_each(evlist, evsel) {
 			if (evsel->attr.sample_type == first->attr.sample_type)
 				continue;
 			use_sample_identifier = perf_can_sample_identifier();
 			break;
 		}
-		list_for_each_entry(evsel, &evlist->entries, node)
+		evlist__for_each(evlist, evsel)
 			perf_evsel__set_sample_id(evsel, use_sample_identifier);
 	}
 
diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 8ffe29c55d0f..7acc03e8f3b2 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1384,7 +1384,7 @@ bool perf_session__has_traces(struct perf_session *session, const char *msg)
 {
 	struct perf_evsel *evsel;
 
-	list_for_each_entry(evsel, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, evsel) {
 		if (evsel->attr.type == PERF_TYPE_TRACEPOINT)
 			return true;
 	}
@@ -1442,7 +1442,7 @@ size_t perf_session__fprintf_nr_events(struct perf_session *session, FILE *fp)
 
 	ret += events_stats__fprintf(&session->stats, fp);
 
-	list_for_each_entry(pos, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, pos) {
 		ret += fprintf(fp, "%s stats:\n", perf_evsel__name(pos));
 		ret += events_stats__fprintf(&pos->hists.stats, fp);
 	}
@@ -1464,7 +1464,7 @@ struct perf_evsel *perf_session__find_first_evtype(struct perf_session *session,
 {
 	struct perf_evsel *pos;
 
-	list_for_each_entry(pos, &session->evlist->entries, node) {
+	evlist__for_each(session->evlist, pos) {
 		if (pos->attr.type == type)
 			return pos;
 	}
-- 
1.8.1.4

[PATCH 25/38] tools include: Move perf's linux/compiler.h to a generic place

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

So that it can be shared with others like libtraceevent.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Borislav Petkov <bp@suse.de>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1389276059-8829-2-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/include/linux/compiler.h           | 30 ++++++++++++++++++++++++++++++
 tools/perf/Makefile.perf                 |  2 +-
 tools/perf/config/Makefile               |  1 +
 tools/perf/util/include/linux/compiler.h | 30 ------------------------------
 4 files changed, 32 insertions(+), 31 deletions(-)
 create mode 100644 tools/include/linux/compiler.h
 delete mode 100644 tools/perf/util/include/linux/compiler.h

diff --git a/tools/include/linux/compiler.h b/tools/include/linux/compiler.h
new file mode 100644
index 000000000000..0135ccf2a00c
--- /dev/null
+++ b/tools/include/linux/compiler.h
@@ -0,0 +1,30 @@
+#ifndef _TOOLS_LINUX_COMPILER_H_
+#define _TOOLS_LINUX_COMPILER_H_
+
+#ifndef __always_inline
+# define __always_inline	inline __attribute__((always_inline))
+#endif
+
+#define __user
+
+#ifndef __attribute_const__
+# define __attribute_const__
+#endif
+
+#ifndef __maybe_unused
+# define __maybe_unused		__attribute__((unused))
+#endif
+
+#ifndef __packed
+# define __packed		__attribute__((__packed__))
+#endif
+
+#ifndef __force
+# define __force
+#endif
+
+#ifndef __weak
+# define __weak			__attribute__((weak))
+#endif
+
+#endif /* _TOOLS_LINUX_COMPILER_H */
diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index 3638b0bd20dc..6be06767da7b 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -211,7 +211,7 @@ LIB_H += ../../include/linux/hash.h
 LIB_H += ../../include/linux/stringify.h
 LIB_H += util/include/linux/bitmap.h
 LIB_H += util/include/linux/bitops.h
-LIB_H += util/include/linux/compiler.h
+LIB_H += ../include/linux/compiler.h
 LIB_H += util/include/linux/const.h
 LIB_H += util/include/linux/ctype.h
 LIB_H += util/include/linux/kernel.h
diff --git a/tools/perf/config/Makefile b/tools/perf/config/Makefile
index 5d15b43b03c6..01dd43df0d04 100644
--- a/tools/perf/config/Makefile
+++ b/tools/perf/config/Makefile
@@ -232,6 +232,7 @@ endif
 
 CFLAGS += -I$(src-perf)/util/include
 CFLAGS += -I$(src-perf)/arch/$(ARCH)/include
+CFLAGS += -I$(srctree)/tools/include/
 CFLAGS += -I$(srctree)/arch/$(ARCH)/include/uapi
 CFLAGS += -I$(srctree)/arch/$(ARCH)/include
 CFLAGS += -I$(srctree)/include/uapi
diff --git a/tools/perf/util/include/linux/compiler.h b/tools/perf/util/include/linux/compiler.h
deleted file mode 100644
index b003ad7200b2..000000000000
--- a/tools/perf/util/include/linux/compiler.h
+++ /dev/null
@@ -1,30 +0,0 @@
-#ifndef _PERF_LINUX_COMPILER_H_
-#define _PERF_LINUX_COMPILER_H_
-
-#ifndef __always_inline
-# define __always_inline	inline __attribute__((always_inline))
-#endif
-
-#define __user
-
-#ifndef __attribute_const__
-# define __attribute_const__
-#endif
-
-#ifndef __maybe_unused
-# define __maybe_unused		__attribute__((unused))
-#endif
-
-#ifndef __packed
-# define __packed		__attribute__((__packed__))
-#endif
-
-#ifndef __force
-# define __force
-#endif
-
-#ifndef __weak
-# define __weak			__attribute__((weak))
-#endif
-
-#endif
-- 
1.8.1.4

[PATCH 26/38] tools include: Define likely/unlikely in linux/compiler.h

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1389276059-8829-3-git-send-email-namhyung@kernel.org
[ Added the new header to tools/perf/MANIFEST ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/include/linux/compiler.h | 8 ++++++++
 tools/perf/MANIFEST            | 1 +
 2 files changed, 9 insertions(+)

diff --git a/tools/include/linux/compiler.h b/tools/include/linux/compiler.h
index 0135ccf2a00c..fbc6665c6d53 100644
--- a/tools/include/linux/compiler.h
+++ b/tools/include/linux/compiler.h
@@ -27,4 +27,12 @@
 # define __weak			__attribute__((weak))
 #endif
 
+#ifndef likely
+# define likely(x)		__builtin_expect(!!(x), 1)
+#endif
+
+#ifndef unlikely
+# define unlikely(x)		__builtin_expect(!!(x), 0)
+#endif
+
 #endif /* _TOOLS_LINUX_COMPILER_H */
diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST
index 3170a7ff5782..285f28f7f821 100644
--- a/tools/perf/MANIFEST
+++ b/tools/perf/MANIFEST
@@ -4,6 +4,7 @@ tools/lib/traceevent
 tools/lib/lk
 tools/lib/symbol/kallsyms.c
 tools/lib/symbol/kallsyms.h
+tools/include/linux/compiler.h
 include/linux/const.h
 include/linux/perf_event.h
 include/linux/rbtree.h
-- 
1.8.1.4

[PATCH 27/38] tools include: Move perf's bug.h to a generic place

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

So that it can be shared with others like libtraceevent.

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1389276059-8829-4-git-send-email-namhyung@kernel.org
[ Added the new header to tools/perf/MANIFEST ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/include/asm/bug.h           | 23 +++++++++++++++++++++++
 tools/perf/MANIFEST               |  1 +
 tools/perf/Makefile.perf          |  2 +-
 tools/perf/util/include/asm/bug.h | 22 ----------------------
 4 files changed, 25 insertions(+), 23 deletions(-)
 create mode 100644 tools/include/asm/bug.h
 delete mode 100644 tools/perf/util/include/asm/bug.h

diff --git a/tools/include/asm/bug.h b/tools/include/asm/bug.h
new file mode 100644
index 000000000000..eca78df7b8f2
--- /dev/null
+++ b/tools/include/asm/bug.h
@@ -0,0 +1,23 @@
+#ifndef _TOOLS_ASM_BUG_H
+#define _TOOLS_ASM_BUG_H
+
+#define __WARN_printf(arg...)	do { fprintf(stderr, arg); } while (0)
+
+#define WARN(condition, format...) ({		\
+	int __ret_warn_on = !!(condition);	\
+	if (unlikely(__ret_warn_on))		\
+		__WARN_printf(format);		\
+	unlikely(__ret_warn_on);		\
+})
+
+#define WARN_ONCE(condition, format...)	({	\
+	static int __warned;			\
+	int __ret_warn_once = !!(condition);	\
+						\
+	if (unlikely(__ret_warn_once))		\
+		if (WARN(!__warned, format)) 	\
+			__warned = 1;		\
+	unlikely(__ret_warn_once);		\
+})
+
+#endif /* _TOOLS_ASM_BUG_H */
diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST
index 285f28f7f821..4e535354f162 100644
--- a/tools/perf/MANIFEST
+++ b/tools/perf/MANIFEST
@@ -4,6 +4,7 @@ tools/lib/traceevent
 tools/lib/lk
 tools/lib/symbol/kallsyms.c
 tools/lib/symbol/kallsyms.h
+tools/include/asm/bug.h
 tools/include/linux/compiler.h
 include/linux/const.h
 include/linux/perf_event.h
diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index 6be06767da7b..87d7726cee2d 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -226,7 +226,7 @@ LIB_H += util/include/linux/string.h
 LIB_H += util/include/linux/types.h
 LIB_H += util/include/linux/linkage.h
 LIB_H += util/include/asm/asm-offsets.h
-LIB_H += util/include/asm/bug.h
+LIB_H += ../include/asm/bug.h
 LIB_H += util/include/asm/byteorder.h
 LIB_H += util/include/asm/hweight.h
 LIB_H += util/include/asm/swab.h
diff --git a/tools/perf/util/include/asm/bug.h b/tools/perf/util/include/asm/bug.h
deleted file mode 100644
index 7fcc6810adc2..000000000000
--- a/tools/perf/util/include/asm/bug.h
+++ /dev/null
@@ -1,22 +0,0 @@
-#ifndef _PERF_ASM_GENERIC_BUG_H
-#define _PERF_ASM_GENERIC_BUG_H
-
-#define __WARN_printf(arg...)	do { fprintf(stderr, arg); } while (0)
-
-#define WARN(condition, format...) ({		\
-	int __ret_warn_on = !!(condition);	\
-	if (unlikely(__ret_warn_on))		\
-		__WARN_printf(format);		\
-	unlikely(__ret_warn_on);		\
-})
-
-#define WARN_ONCE(condition, format...)	({	\
-	static int __warned;			\
-	int __ret_warn_once = !!(condition);	\
-						\
-	if (unlikely(__ret_warn_once))		\
-		if (WARN(!__warned, format)) 	\
-			__warned = 1;		\
-	unlikely(__ret_warn_once);		\
-})
-#endif
-- 
1.8.1.4

[PATCH 28/38] perf tools: Include tools/lib/api/ in MANIFEST

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

When 553873e1df63 renamed tools/lib/lk to tools/lib/api we forgot to
do the switch in tools/perf/MANIFEST, breaking tarball building:

  [acme@ssdandy linux]$ make perf-targz-src-pkg
    TAR
  [acme@ssdandy linux]$ tar xf perf-3.13.0-rc4.tar.gz -C /tmp/tmp.OgdYyvp77p/
  [acme@ssdandy linux]$ make -C /tmp/tmp.OgdYyvp77p/perf-3.13.0-rc4/tools/perf
  make: Entering directory
  `/tmp/tmp.OgdYyvp77p/perf-3.13.0-rc4/tools/perf'
    BUILD:   Doing 'make -j8' parallel build
    FLEX     util/pmu-flex.c
    CC       util/evlist.o
    CC       util/evsel.o
  util/evsel.c:12:28: fatal error: api/fs/debugfs.h: No such file or directory compilation terminated.
  In file included from util/cache.h:5:0,
  <SNIP>

Fix it.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-1wwjs01rt3xbyhn6kjl2gfs9@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/MANIFEST | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/MANIFEST b/tools/perf/MANIFEST
index 4e535354f162..f41572d0dd76 100644
--- a/tools/perf/MANIFEST
+++ b/tools/perf/MANIFEST
@@ -1,7 +1,7 @@
 tools/perf
 tools/scripts
 tools/lib/traceevent
-tools/lib/lk
+tools/lib/api
 tools/lib/symbol/kallsyms.c
 tools/lib/symbol/kallsyms.h
 tools/include/asm/bug.h
-- 
1.8.1.4

[PATCH 29/38] perf tools: Add test for building detached source tarballs

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

Test one of the main kernel Makefile targets to generate a perf sources
tarball suitable for build outside the full kernel sources.

This is to test that the tools/perf/MANIFEST file lists all the files
needed to be in such tarball, which sometimes gets broken when we move
files around, like when we made some files that were in tools/perf/
available to other tools/ codebases by moving it to tools/include/, etc.

Now everytime we use 'make -C tools/perf -f tests/make' this test will
be performed, helping detect such problems earlier in the devel cycle.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-gyivwbbu2j7c4j4pwpmttg2p@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/tests/make               | 10 ++++++++--
 tools/perf/tests/perf-targz-src-pkg | 21 +++++++++++++++++++++
 2 files changed, 29 insertions(+), 2 deletions(-)
 create mode 100755 tools/perf/tests/perf-targz-src-pkg

diff --git a/tools/perf/tests/make b/tools/perf/tests/make
index e341088de8f3..00544b8b644b 100644
--- a/tools/perf/tests/make
+++ b/tools/perf/tests/make
@@ -216,10 +216,16 @@ $(run_O):
 	rm -rf $$TMP_O \
 	rm -rf $$TMP_DEST
 
-all: $(run) $(run_O)
+tarpkg:
+	@cmd="$(PERF)/tests/perf-targz-src-pkg $(PERF)"; \
+	echo "- $@: $$cmd" && echo $$cmd > $@ && \
+	( eval $$cmd ) >> $@ 2>&1
+	
+
+all: $(run) $(run_O) tarpkg
 	@echo OK
 
 out: $(run_O)
 	@echo OK
 
-.PHONY: all $(run) $(run_O) clean
+.PHONY: all $(run) $(run_O) tarpkg clean
diff --git a/tools/perf/tests/perf-targz-src-pkg b/tools/perf/tests/perf-targz-src-pkg
new file mode 100755
index 000000000000..238aa3927c71
--- /dev/null
+++ b/tools/perf/tests/perf-targz-src-pkg
@@ -0,0 +1,21 @@
+#!/bin/sh
+# Test one of the main kernel Makefile targets to generate a perf sources tarball
+# suitable for build outside the full kernel sources.
+#
+# This is to test that the tools/perf/MANIFEST file lists all the files needed to
+# be in such tarball, which sometimes gets broken when we move files around,
+# like when we made some files that were in tools/perf/ available to other tools/
+# codebases by moving it to tools/include/, etc.
+
+PERF=$1
+cd ${PERF}/../..
+make perf-targz-src-pkg > /dev/null
+TARBALL=$(ls -rt perf-*.tar.gz)
+TMP_DEST=$(mktemp -d)
+tar xf ${TARBALL} -C $TMP_DEST
+rm -f ${TARBALL}
+cd - > /dev/null
+make -C $TMP_DEST/perf*/tools/perf > /dev/null 2>&1
+RC=$?
+rm -rf ${TMP_DEST}
+exit $RC
-- 
1.8.1.4

[PATCH 30/38] perf tools: Use the DWARF unwind info only if loaded

[Arnaldo Carvalho de Melo]

From: Jean Pihet <jean.pihet@linaro.org>

Use the info only if it has been found in the .debug_frame section of
the ELF binary.

Signed-off-by: Jean Pihet <jean.pihet@linaro.org>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linaro-kernel@lists.linaro.org
Cc: patches@linaro.org
Link: http://lkml.kernel.org/r/1387212194-8028-1-git-send-email-jean.pihet@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/unwind.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/tools/perf/util/unwind.c b/tools/perf/util/unwind.c
index 0efd5393de85..416f22bf3693 100644
--- a/tools/perf/util/unwind.c
+++ b/tools/perf/util/unwind.c
@@ -340,10 +340,10 @@ find_proc_info(unw_addr_space_t as, unw_word_t ip, unw_proc_info_t *pi,
 	/* Check the .debug_frame section for unwinding info */
 	if (!read_unwind_spec_debug_frame(map->dso, ui->machine, &segbase)) {
 		memset(&di, 0, sizeof(di));
-		dwarf_find_debug_frame(0, &di, ip, 0, map->dso->name,
-				       map->start, map->end);
-		return dwarf_search_unwind_table(as, ip, &di, pi,
-						 need_unwind_info, arg);
+		if (dwarf_find_debug_frame(0, &di, ip, 0, map->dso->name,
+					   map->start, map->end))
+			return dwarf_search_unwind_table(as, ip, &di, pi,
+							 need_unwind_info, arg);
 	}
 #endif
 
-- 
1.8.1.4

[PATCH 31/38] perf record: Add --initial-delay option

[Arnaldo Carvalho de Melo]

From: Andi Kleen <ak@linux.intel.com>

perf stat has a --delay option to delay measuring the workload.

This is useful to skip measuring the startup phase of the program, which
is often very different from the main workload.

The same is useful for perf record when sampling.

--no-delay was already taken, so add a --initial-delay
to perf record too.
-D was already taken for record, so there is only a long option.

v2: Don't disable group members (Namhyung Kim)
v3: port to latest perf/core
    rename to --initial-delay to avoid conflict with --no-delay

Signed-off-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1389476307-2124-1-git-send-email-andi@firstfloor.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/Documentation/perf-record.txt | 4 ++++
 tools/perf/builtin-record.c              | 9 ++++++++-
 tools/perf/perf.h                        | 1 +
 tools/perf/util/evsel.c                  | 3 ++-
 4 files changed, 15 insertions(+), 2 deletions(-)

diff --git a/tools/perf/Documentation/perf-record.txt b/tools/perf/Documentation/perf-record.txt
index c407897f0435..82bffac036e1 100644
--- a/tools/perf/Documentation/perf-record.txt
+++ b/tools/perf/Documentation/perf-record.txt
@@ -209,6 +209,10 @@ overrides that and uses per-thread mmaps.  A side-effect of that is that
 inheritance is automatically disabled.  --per-thread is ignored with a warning
 if combined with -a or -C options.
 
+--initial-delay msecs::
+After starting the program, wait msecs before measuring. This is useful to
+filter out the startup phase of the program, which is often very different.
+
 SEE ALSO
 --------
 linkperf:perf-stat[1], linkperf:perf-list[1]
diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 88600158400e..07d4cf8d3fd3 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -499,7 +499,7 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	 * (apart from group members) have enable_on_exec=1 set,
 	 * so don't spoil it by prematurely enabling them.
 	 */
-	if (!target__none(&opts->target))
+	if (!target__none(&opts->target) && !opts->initial_delay)
 		perf_evlist__enable(rec->evlist);
 
 	/*
@@ -508,6 +508,11 @@ static int __cmd_record(struct record *rec, int argc, const char **argv)
 	if (forks)
 		perf_evlist__start_workload(rec->evlist);
 
+	if (opts->initial_delay) {
+		usleep(opts->initial_delay * 1000);
+		perf_evlist__enable(rec->evlist);
+	}
+
 	for (;;) {
 		int hits = rec->samples;
 
@@ -877,6 +882,8 @@ const struct option record_options[] = {
 	OPT_CALLBACK('G', "cgroup", &record.evlist, "name",
 		     "monitor event in cgroup name only",
 		     parse_cgroups),
+	OPT_UINTEGER(0, "initial-delay", &record.opts.initial_delay,
+		  "ms to wait before starting measurement after program start"),
 	OPT_STRING('u', "uid", &record.opts.target.uid_str, "user",
 		   "user to profile"),
 
diff --git a/tools/perf/perf.h b/tools/perf/perf.h
index b1cc84b01d5b..af1ce6e14a93 100644
--- a/tools/perf/perf.h
+++ b/tools/perf/perf.h
@@ -269,6 +269,7 @@ struct record_opts {
 	u64	     user_interval;
 	u16	     stack_dump_size;
 	bool	     sample_transaction;
+	unsigned     initial_delay;
 };
 
 #endif
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index ade8d9c1c431..cd4630abfa43 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -658,7 +658,8 @@ void perf_evsel__config(struct perf_evsel *evsel, struct record_opts *opts)
 	 * Setting enable_on_exec for independent events and
 	 * group leaders for traced executed by perf.
 	 */
-	if (target__none(&opts->target) && perf_evsel__is_group_leader(evsel))
+	if (target__none(&opts->target) && perf_evsel__is_group_leader(evsel) &&
+		!opts->initial_delay)
 		attr->enable_on_exec = 1;
 }
 
-- 
1.8.1.4

[PATCH 32/38] tools include: Include <linux/compiler.h> from asm/bug.h

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung@kernel.org>

Since it uses unlikely() macro inside WARN()

Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung.kim@lge.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Link: http://lkml.kernel.org/r/1389276059-8829-5-git-send-email-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/include/asm/bug.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/tools/include/asm/bug.h b/tools/include/asm/bug.h
index eca78df7b8f2..9e5f4846967f 100644
--- a/tools/include/asm/bug.h
+++ b/tools/include/asm/bug.h
@@ -1,6 +1,8 @@
 #ifndef _TOOLS_ASM_BUG_H
 #define _TOOLS_ASM_BUG_H
 
+#include <linux/compiler.h>
+
 #define __WARN_printf(arg...)	do { fprintf(stderr, arg); } while (0)
 
 #define WARN(condition, format...) ({		\
-- 
1.8.1.4

[PATCH 33/38] perf tools: Generalize percent_color_snprintf()

[Arnaldo Carvalho de Melo]

From: Ramkumar Ramachandra <artagnon@gmail.com>

Make percent_color_snprintf() handle negative values correctly.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1388388861-7931-2-git-send-email-artagnon@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/color.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/color.c b/tools/perf/util/color.c
index 66e44a5019d5..8cda46c43e74 100644
--- a/tools/perf/util/color.c
+++ b/tools/perf/util/color.c
@@ -1,6 +1,7 @@
 #include <linux/kernel.h>
 #include "cache.h"
 #include "color.h"
+#include <math.h>
 
 int perf_use_color_default = -1;
 
@@ -298,10 +299,10 @@ const char *get_percent_color(double percent)
 	 * entries in green - and keep the low overhead places
 	 * normal:
 	 */
-	if (percent >= MIN_RED)
+	if (fabs(percent) >= MIN_RED)
 		color = PERF_COLOR_RED;
 	else {
-		if (percent > MIN_GREEN)
+		if (fabs(percent) > MIN_GREEN)
 			color = PERF_COLOR_GREEN;
 	}
 	return color;
-- 
1.8.1.4

[PATCH 34/38] perf diff: Color the Delta column

[Arnaldo Carvalho de Melo]

From: Ramkumar Ramachandra <artagnon@gmail.com>

Color the numbers in the Delta column using percent_color_snprintf().

Generalize the coloring function so that we can accommodate all three
comparison methods in future patches: delta, ratio, and wdiff.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1388388861-7931-3-git-send-email-artagnon@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-diff.c | 49 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 48 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 987cac3b4bba..6c3f220115ac 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -770,6 +770,45 @@ static int hpp__entry_baseline(struct hist_entry *he, char *buf, size_t size)
 	return ret;
 }
 
+static int __hpp__color_compare(struct perf_hpp_fmt *fmt,
+				struct perf_hpp *hpp, struct hist_entry *he,
+				int comparison_method)
+{
+	struct diff_hpp_fmt *dfmt =
+		container_of(fmt, struct diff_hpp_fmt, fmt);
+	struct hist_entry *pair = get_pair_fmt(he, dfmt);
+	double diff;
+	char pfmt[20] = " ";
+
+	if (!pair)
+		goto dummy_print;
+
+	switch (comparison_method) {
+	case COMPUTE_DELTA:
+		if (pair->diff.computed)
+			diff = pair->diff.period_ratio_delta;
+		else
+			diff = compute_delta(he, pair);
+
+		if (fabs(diff) < 0.01)
+			goto dummy_print;
+		scnprintf(pfmt, 20, "%%%+d.2f%%%%", dfmt->header_width - 1);
+		return percent_color_snprintf(hpp->buf, hpp->size,
+					pfmt, diff);
+	default:
+		BUG_ON(1);
+	}
+dummy_print:
+	return scnprintf(hpp->buf, hpp->size, "%*s",
+			dfmt->header_width, pfmt);
+}
+
+static int hpp__color_delta(struct perf_hpp_fmt *fmt,
+			struct perf_hpp *hpp, struct hist_entry *he)
+{
+	return __hpp__color_compare(fmt, hpp, he, COMPUTE_DELTA);
+}
+
 static void
 hpp__entry_unpair(struct hist_entry *he, int idx, char *buf, size_t size)
 {
@@ -941,8 +980,16 @@ static void data__hpp_register(struct data__file *d, int idx)
 	fmt->entry  = hpp__entry_global;
 
 	/* TODO more colors */
-	if (idx == PERF_HPP_DIFF__BASELINE)
+	switch (idx) {
+	case PERF_HPP_DIFF__BASELINE:
 		fmt->color = hpp__color_baseline;
+		break;
+	case PERF_HPP_DIFF__DELTA:
+		fmt->color = hpp__color_delta;
+		break;
+	default:
+		break;
+	}
 
 	init_header(d, dfmt);
 	perf_hpp__column_register(fmt);
-- 
1.8.1.4

[PATCH 35/38] perf diff: Color the Ratio column

[Arnaldo Carvalho de Melo]

From: Ramkumar Ramachandra <artagnon@gmail.com>

In

  $ perf diff -c ratio

color the Ratio column using value_color_snprintf(), a new function that
operates exactly like percent_color_snprintf().

At first glance, it looks like percent_color_snprintf() can be turned
into a non-variadic function simplifying things; however, 53805ec (perf
tools: Remove cast of non-variadic function to variadic, 2013-10-31)
explains why it needs to be a variadic function.

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1388388861-7931-4-git-send-email-artagnon@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-diff.c | 20 ++++++++++++++++++++
 tools/perf/util/color.c   | 10 +++++++---
 tools/perf/util/color.h   |  1 +
 3 files changed, 28 insertions(+), 3 deletions(-)

diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 6c3f220115ac..73d8bffdd484 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -795,6 +795,17 @@ static int __hpp__color_compare(struct perf_hpp_fmt *fmt,
 		scnprintf(pfmt, 20, "%%%+d.2f%%%%", dfmt->header_width - 1);
 		return percent_color_snprintf(hpp->buf, hpp->size,
 					pfmt, diff);
+	case COMPUTE_RATIO:
+		if (he->dummy)
+			goto dummy_print;
+		if (pair->diff.computed)
+			diff = pair->diff.period_ratio;
+		else
+			diff = compute_ratio(he, pair);
+
+		scnprintf(pfmt, 20, "%%%d.6f", dfmt->header_width);
+		return value_color_snprintf(hpp->buf, hpp->size,
+					pfmt, diff);
 	default:
 		BUG_ON(1);
 	}
@@ -809,6 +820,12 @@ static int hpp__color_delta(struct perf_hpp_fmt *fmt,
 	return __hpp__color_compare(fmt, hpp, he, COMPUTE_DELTA);
 }
 
+static int hpp__color_ratio(struct perf_hpp_fmt *fmt,
+			struct perf_hpp *hpp, struct hist_entry *he)
+{
+	return __hpp__color_compare(fmt, hpp, he, COMPUTE_RATIO);
+}
+
 static void
 hpp__entry_unpair(struct hist_entry *he, int idx, char *buf, size_t size)
 {
@@ -987,6 +1004,9 @@ static void data__hpp_register(struct data__file *d, int idx)
 	case PERF_HPP_DIFF__DELTA:
 		fmt->color = hpp__color_delta;
 		break;
+	case PERF_HPP_DIFF__RATIO:
+		fmt->color = hpp__color_ratio;
+		break;
 	default:
 		break;
 	}
diff --git a/tools/perf/util/color.c b/tools/perf/util/color.c
index 8cda46c43e74..87b8672eb413 100644
--- a/tools/perf/util/color.c
+++ b/tools/perf/util/color.c
@@ -319,15 +319,19 @@ int percent_color_fprintf(FILE *fp, const char *fmt, double percent)
 	return r;
 }
 
+int value_color_snprintf(char *bf, size_t size, const char *fmt, double value)
+{
+	const char *color = get_percent_color(value);
+	return color_snprintf(bf, size, color, fmt, value);
+}
+
 int percent_color_snprintf(char *bf, size_t size, const char *fmt, ...)
 {
 	va_list args;
 	double percent;
-	const char *color;
 
 	va_start(args, fmt);
 	percent = va_arg(args, double);
 	va_end(args);
-	color = get_percent_color(percent);
-	return color_snprintf(bf, size, color, fmt, percent);
+	return value_color_snprintf(bf, size, fmt, percent);
 }
diff --git a/tools/perf/util/color.h b/tools/perf/util/color.h
index fced3840e99c..7ff30a62a132 100644
--- a/tools/perf/util/color.h
+++ b/tools/perf/util/color.h
@@ -39,6 +39,7 @@ int color_fprintf(FILE *fp, const char *color, const char *fmt, ...);
 int color_snprintf(char *bf, size_t size, const char *color, const char *fmt, ...);
 int color_fprintf_ln(FILE *fp, const char *color, const char *fmt, ...);
 int color_fwrite_lines(FILE *fp, const char *color, size_t count, const char *buf);
+int value_color_snprintf(char *bf, size_t size, const char *fmt, double value);
 int percent_color_snprintf(char *bf, size_t size, const char *fmt, ...);
 int percent_color_fprintf(FILE *fp, const char *fmt, double percent);
 const char *get_percent_color(double percent);
-- 
1.8.1.4

[PATCH 36/38] perf diff: Color the Weighted Diff column

[Arnaldo Carvalho de Melo]

From: Ramkumar Ramachandra <artagnon@gmail.com>

In

  $ perf diff -c wdiff:M,N

color the numbers in the Weighted Diff column using color_snprintf(),
picking the colors using get_percent_color().

Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/1388390555-10808-1-git-send-email-artagnon@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-diff.c | 22 ++++++++++++++++++++++
 1 file changed, 22 insertions(+)

diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 73d8bffdd484..a77e31246c00 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -778,6 +778,7 @@ static int __hpp__color_compare(struct perf_hpp_fmt *fmt,
 		container_of(fmt, struct diff_hpp_fmt, fmt);
 	struct hist_entry *pair = get_pair_fmt(he, dfmt);
 	double diff;
+	s64 wdiff;
 	char pfmt[20] = " ";
 
 	if (!pair)
@@ -806,6 +807,18 @@ static int __hpp__color_compare(struct perf_hpp_fmt *fmt,
 		scnprintf(pfmt, 20, "%%%d.6f", dfmt->header_width);
 		return value_color_snprintf(hpp->buf, hpp->size,
 					pfmt, diff);
+	case COMPUTE_WEIGHTED_DIFF:
+		if (he->dummy)
+			goto dummy_print;
+		if (pair->diff.computed)
+			wdiff = pair->diff.wdiff;
+		else
+			wdiff = compute_wdiff(he, pair);
+
+		scnprintf(pfmt, 20, "%%14ld", dfmt->header_width);
+		return color_snprintf(hpp->buf, hpp->size,
+				get_percent_color(wdiff),
+				pfmt, wdiff);
 	default:
 		BUG_ON(1);
 	}
@@ -826,6 +839,12 @@ static int hpp__color_ratio(struct perf_hpp_fmt *fmt,
 	return __hpp__color_compare(fmt, hpp, he, COMPUTE_RATIO);
 }
 
+static int hpp__color_wdiff(struct perf_hpp_fmt *fmt,
+			struct perf_hpp *hpp, struct hist_entry *he)
+{
+	return __hpp__color_compare(fmt, hpp, he, COMPUTE_WEIGHTED_DIFF);
+}
+
 static void
 hpp__entry_unpair(struct hist_entry *he, int idx, char *buf, size_t size)
 {
@@ -1007,6 +1026,9 @@ static void data__hpp_register(struct data__file *d, int idx)
 	case PERF_HPP_DIFF__RATIO:
 		fmt->color = hpp__color_ratio;
 		break;
+	case PERF_HPP_DIFF__WEIGHTED_DIFF:
+		fmt->color = hpp__color_wdiff;
+		break;
 	default:
 		break;
 	}
-- 
1.8.1.4

[PATCH 37/38] perf probe: Fix build when DWARF support libraries not present

[Arnaldo Carvalho de Melo]

From: Arnaldo Carvalho de Melo <acme@redhat.com>

On a freshly installed system, after libelf-dev is installed we get:

    CC       /tmp/build/perf/util/probe-event.o
  util/probe-event.c: In function ‘try_to_find_probe_trace_events’:
  util/probe-event.c:753:46: error: unused parameter ‘target’ [-Werror=unused-parameter]
       int max_tevs __maybe_unused, const char *target)
                                                ^
    CC       /tmp/build/perf/util/cgroup.o
  util/probe-event.c: At top level:
  util/probe-event.c:193:12: error: ‘get_text_start_address’ defined but not used [-Werror=unused-function]
   static int get_text_start_address(const char *exec, unsigned long *address)
            ^
  cc1: all warnings being treated as errors
  make[1]: *** [/tmp/build/perf/util/probe-event.o] Error 1
  make[1]: *** Waiting for unfinished jobs....
  make: *** [install] Error 2

Fix it by enclosing functions only used when those libraries are installed
under the suitable preprocessor define and using __maybe_unused to a function
that is only built when DWARF support is disabled.

Problem introduced in this changeset:

  commit fb7345bbf7fad9bf72ef63a19c707970b9685812
  Author: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
  Date:   Thu Dec 26 05:41:53 2013 +0000

      perf probe: Support basic dwarf-based operations on uprobe events

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/n/tip-73kc2fopt81517hrdgdra18o@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/probe-event.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/tools/perf/util/probe-event.c b/tools/perf/util/probe-event.c
index 86ed8580c3cb..a4ee6b4a840f 100644
--- a/tools/perf/util/probe-event.c
+++ b/tools/perf/util/probe-event.c
@@ -172,6 +172,7 @@ const char *kernel_get_module_path(const char *module)
 	return (dso) ? dso->long_name : NULL;
 }
 
+#ifdef HAVE_DWARF_SUPPORT
 /* Copied from unwind.c */
 static Elf_Scn *elf_section_by_name(Elf *elf, GElf_Ehdr *ep,
 				    GElf_Shdr *shp, const char *name)
@@ -217,6 +218,7 @@ out:
 	elf_end(elf);
 	return ret;
 }
+#endif
 
 static int init_user_exec(void)
 {
@@ -750,7 +752,8 @@ static int kprobe_convert_to_perf_probe(struct probe_trace_point *tp,
 
 static int try_to_find_probe_trace_events(struct perf_probe_event *pev,
 				struct probe_trace_event **tevs __maybe_unused,
-				int max_tevs __maybe_unused, const char *target)
+				int max_tevs __maybe_unused,
+				const char *target __maybe_unused)
 {
 	if (perf_probe_event_need_dwarf(pev)) {
 		pr_warning("Debuginfo-analysis is not supported.\n");
-- 
1.8.1.4

[PATCH 38/38] perf tools: Remove unused test-volatile-register-var.c

[Arnaldo Carvalho de Melo]

From: Yann Droneaud <ydroneaud@opteya.com>

Since commit 01287e2cb7ad, test-volatile-register-var.c is no more built
as part of the automatic feature check.

This patch remove the unneeded file.

Signed-off-by: Yann Droneaud <ydroneaud@opteya.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/339d86ad76741ed929defd18541f774b404003b4.1389461371.git.ydroneaud@opteya.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/config/feature-checks/test-volatile-register-var.c | 6 ------
 1 file changed, 6 deletions(-)
 delete mode 100644 tools/perf/config/feature-checks/test-volatile-register-var.c

diff --git a/tools/perf/config/feature-checks/test-volatile-register-var.c b/tools/perf/config/feature-checks/test-volatile-register-var.c
deleted file mode 100644
index c9f398d87868..000000000000
--- a/tools/perf/config/feature-checks/test-volatile-register-var.c
+++ /dev/null
@@ -1,6 +0,0 @@
-#include <stdio.h>
-
-int main(void)
-{
-	return puts("hi");
-}
-- 
1.8.1.4

Re: [GIT PULL 00/38] perf/core improvements and fixes

[Ingo Molnar]

* Arnaldo Carvalho de Melo <acme@infradead.org> wrote:

> Andi Kleen (2):
>       perf stat: Fix --delay option in man page
>       perf record: Add --initial-delay option

Hm, so we now have two options that do something similar, one named 
--delay, the other --initial-delay? That looks rather inconsistent.

Furthermore, it looks silly to rename things --initial-delay just to 
preserve a rarely used option, --no-delay. Instead the better option 
would be to rename --no-delay to --no-buffering, and use --delay and 
use it consistently within the tools.

Thanks,

	Ingo

Re: [GIT PULL 00/38] perf/core improvements and fixes

[Arnaldo Carvalho de Melo]

Em Tue, Jan 14, 2014 at 02:41:01PM +0100, Ingo Molnar escreveu:
> * Arnaldo Carvalho de Melo <acme@infradead.org> wrote:

> > Andi Kleen (2):
> >       perf stat: Fix --delay option in man page
> >       perf record: Add --initial-delay option

> Hm, so we now have two options that do something similar, one named
> --delay, the other --initial-delay? That looks rather inconsistent.

> Furthermore, it looks silly to rename things --initial-delay just to
> preserve a rarely used option, --no-delay. Instead the better option
> would be to rename --no-delay to --no-buffering, and use --delay and
> use it consistently within the tools.

Ok, agreed, I'll do that, adding a patch that renames the old option,
then reworking Andi's patches to use --delay consistently.

- Arnaldo

Re: [GIT PULL 00/38] perf/core improvements and fixes

[Ingo Molnar]

* Arnaldo Carvalho de Melo <acme@ghostprotocols.net> wrote:

> Em Tue, Jan 14, 2014 at 02:41:01PM +0100, Ingo Molnar escreveu:
> > * Arnaldo Carvalho de Melo <acme@infradead.org> wrote:
> 
> > > Andi Kleen (2):
> > >       perf stat: Fix --delay option in man page
> > >       perf record: Add --initial-delay option
> 
> > Hm, so we now have two options that do something similar, one named
> > --delay, the other --initial-delay? That looks rather inconsistent.
> 
> > Furthermore, it looks silly to rename things --initial-delay just to
> > preserve a rarely used option, --no-delay. Instead the better option
> > would be to rename --no-delay to --no-buffering, and use --delay and
> > use it consistently within the tools.
> 
> Ok, agreed, I'll do that, adding a patch that renames the old option,
> then reworking Andi's patches to use --delay consistently.

Ok, thanks - I'll pull the existing lineup meanwhile, it's not a 
showstopper.

Thanks,

	Ingo