[GIT PULL 00/16] perf/core fixes and improvements

[GIT PULL 00/16] perf/core fixes and improvements

[Arnaldo Carvalho de Melo]

Hi Ingo,

	This is on top of a 15 patches long series still outstanding on my
perf/core branch, please consider pulling,

- Arnaldo

The following changes since commit dc41b9b8f02dbe2228ae787d525dac43beebb7fa:

  perf ui: Change fallback policy of setup_browser() (2012-05-02 16:17:37 -0300)

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 5a5626b1b4bf8467891c9297ffda979db97ed5ec:

  perf hists browser: Use '/' for search/filter instead of 's' (2012-05-10 13:07:59 -0300)

----------------------------------------------------------------
Fixes and improvements for perf/core:

. perf_target: abstraction for --uid, --pid, --tid, --cpu, --all-cpus handling,
  eliminating code duplicated in the tools, having constraints that apply to
  all of them, from Namhyung Kim

. Fixes for handling fallback to cpu-clock on PPC, from David Ahern

. Fix for processing events with unknown size, from Jiri Olsa

. Compilation fix on 32-bit, from Jiri Olsa

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

----------------------------------------------------------------
Arnaldo Carvalho de Melo (3):
      perf top: Set target.system_wide
      perf top: Default to system wide using perf_target methods
      perf hists browser: Use '/' for search/filter instead of 's'

David Ahern (5):
      perf record: Fix fallback to cpu-clock on ppc
      perf stat: handle ENXIO error for perf_event_open
      perf top: Update event name when falling back to cpu-clock
      perf record: Reset event name when falling back to cpu-clock
      perf annotate: shorten helpline so it fits in visible space

Jiri Olsa (2):
      perf session: Fail on processing event with unknown size
      perf report: Fix format string for x86-32 compilation

Namhyung Kim (6):
      perf evlist: Fix creation of cpu map
      perf target: Introduce perf_target_errno
      perf target: Introduce perf_target__parse_uid()
      perf tools: Introduce perf_target__strerror()
      perf target: Consolidate target task/cpu checking
      perf stat: Use perf_evlist__create_maps

 tools/perf/builtin-record.c       |   36 ++++++++---
 tools/perf/builtin-report.c       |    2 +-
 tools/perf/builtin-stat.c         |   41 +++++++------
 tools/perf/builtin-top.c          |   29 +++++++--
 tools/perf/ui/browsers/annotate.c |    6 +-
 tools/perf/ui/browsers/hists.c    |    4 +-
 tools/perf/util/debug.c           |    1 +
 tools/perf/util/evlist.c          |    9 ++-
 tools/perf/util/evsel.c           |    8 +--
 tools/perf/util/session.c         |   30 +++-------
 tools/perf/util/target.c          |  119 +++++++++++++++++++++++++++++++++----
 tools/perf/util/target.h          |   48 ++++++++++++++-
 tools/perf/util/usage.c           |   31 ----------
 tools/perf/util/util.h            |    3 -
 14 files changed, 250 insertions(+), 117 deletions(-)

[PATCH 01/16] perf session: Fail on processing event with unknown size

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Currently if we cannot decide the size of the event, we guess next
event possition by:
  "... check alignment, and increment a single u64 in the hope
  to catch on again 'soon'"

This usually ends up with segfault or endless loop. It's better
to admit the failure right away, then pretend nothing happened.
It makes the life easier ;)

Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/20120416184251.GA11503@m.brq.redhat.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/session.c |   30 +++++++++---------------------
 1 file changed, 9 insertions(+), 21 deletions(-)

diff --git a/tools/perf/util/session.c b/tools/perf/util/session.c
index 9412e3b..f992ae3 100644
--- a/tools/perf/util/session.c
+++ b/tools/perf/util/session.c
@@ -1100,16 +1100,10 @@ more:
 	}
 
 	if ((skip = perf_session__process_event(self, &event, tool, head)) < 0) {
-		dump_printf("%#" PRIx64 " [%#x]: skipping unknown header type: %d\n",
-			    head, event.header.size, event.header.type);
-		/*
-		 * assume we lost track of the stream, check alignment, and
-		 * increment a single u64 in the hope to catch on again 'soon'.
-		 */
-		if (unlikely(head & 7))
-			head &= ~7ULL;
-
-		size = 8;
+		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
+		       head, event.header.size, event.header.type);
+		err = -EINVAL;
+		goto out_err;
 	}
 
 	head += size;
@@ -1218,17 +1212,11 @@ more:
 
 	if (size == 0 ||
 	    perf_session__process_event(session, event, tool, file_pos) < 0) {
-		dump_printf("%#" PRIx64 " [%#x]: skipping unknown header type: %d\n",
-			    file_offset + head, event->header.size,
-			    event->header.type);
-		/*
-		 * assume we lost track of the stream, check alignment, and
-		 * increment a single u64 in the hope to catch on again 'soon'.
-		 */
-		if (unlikely(head & 7))
-			head &= ~7ULL;
-
-		size = 8;
+		pr_err("%#" PRIx64 " [%#x]: failed to process type: %d\n",
+		       file_offset + head, event->header.size,
+		       event->header.type);
+		err = -EINVAL;
+		goto out_err;
 	}
 
 	head += size;
-- 
1.7.9.2.358.g22243

[PATCH 02/16] perf top: Set target.system_wide

[Arnaldo Carvalho de Melo]

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

Check if neither of --pid, --tid or --uid was specified and if so, set
system_wide appropriately.

Namhyung's patch would make using any of the above target specifiers
emit a warning in perf_target__validate, since it would see
target.system_wide set and one of the others as well.

So set system_wide after validation.

Suggested-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
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-6e4zrji1uw0rinfyoitl0wi4@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-top.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 2a0ec09..e40f86e 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1258,6 +1258,10 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
 	if (top.target.uid_str != NULL && top.target.uid == UINT_MAX - 1)
 		goto out_delete_evlist;
 
+	if (top.target.tid == 0 && top.target.pid == 0 &&
+	    top.target.uid_str == NULL)
+		top.target.system_wide = true;
+
 	if (perf_evlist__create_maps(top.evlist, &top.target) < 0)
 		usage_with_options(top_usage, options);
 
-- 
1.7.9.2.358.g22243

[PATCH 03/16] perf evlist: Fix creation of cpu map

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

Currently, 'perf record -- sleep 1' creates a cpu map for all online
cpus since it turns out calling cpu_map__new(NULL). Fix it.

Also it is guaranteed that cpu_list is NULL if PID/TID is given by
calling perf_target__validate(), so we can make the conditional bit
simpler.

This also fixes perf test 7 (Validate) failure on my 6 core machine:

  $ cat /sys/devices/system/cpu/online
  0-11
  $ ./perf test -v 7
   7: Validate PERF_RECORD_* events & perf_sample fields:
  --- start ---
  perf_evlist__mmap: Operation not permitted
  ---- end ----
  Validate PERF_RECORD_* events & perf_sample fields: FAILED!

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-3-git-send-email-namhyung.kim@lge.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/evlist.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 3032862..183b199 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -609,8 +609,9 @@ int perf_evlist__create_maps(struct perf_evlist *evlist,
 	if (evlist->threads == NULL)
 		return -1;
 
-	if (target->uid != UINT_MAX ||
-	    (target->cpu_list == NULL && target->tid))
+	if (target->uid != UINT_MAX || target->tid)
+		evlist->cpus = cpu_map__dummy_new();
+	else if (!target->system_wide && target->cpu_list == NULL)
 		evlist->cpus = cpu_map__dummy_new();
 	else
 		evlist->cpus = cpu_map__new(target->cpu_list);
-- 
1.7.9.2.358.g22243

[PATCH 04/16] perf target: Introduce perf_target_errno

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

The perf_target_errno enumerations are used to indicate specific error
cases on perf target operations. It'd help libperf being a more generic
library.

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Suggested-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-4-git-send-email-namhyung.kim@lge.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/util/target.c |   33 ++++++++++++++++++++++-----------
 tools/perf/util/target.h |   25 ++++++++++++++++++++++++-
 2 files changed, 46 insertions(+), 12 deletions(-)

diff --git a/tools/perf/util/target.c b/tools/perf/util/target.c
index 3fadf85..5c59dcf 100644
--- a/tools/perf/util/target.c
+++ b/tools/perf/util/target.c
@@ -10,36 +10,47 @@
 #include "debug.h"
 
 
-void perf_target__validate(struct perf_target *target)
+enum perf_target_errno perf_target__validate(struct perf_target *target)
 {
+	enum perf_target_errno ret = PERF_ERRNO_TARGET__SUCCESS;
+
 	if (target->pid)
 		target->tid = target->pid;
 
 	/* CPU and PID are mutually exclusive */
 	if (target->tid && target->cpu_list) {
-		ui__warning("WARNING: PID switch overriding CPU\n");
-		sleep(1);
 		target->cpu_list = NULL;
+		if (ret == PERF_ERRNO_TARGET__SUCCESS)
+			ret = PERF_ERRNO_TARGET__PID_OVERRIDE_CPU;
 	}
 
 	/* UID and PID are mutually exclusive */
 	if (target->tid && target->uid_str) {
-		ui__warning("PID/TID switch overriding UID\n");
-		sleep(1);
 		target->uid_str = NULL;
+		if (ret == PERF_ERRNO_TARGET__SUCCESS)
+			ret = PERF_ERRNO_TARGET__PID_OVERRIDE_UID;
 	}
 
 	/* UID and CPU are mutually exclusive */
 	if (target->uid_str && target->cpu_list) {
-		ui__warning("UID switch overriding CPU\n");
-		sleep(1);
 		target->cpu_list = NULL;
+		if (ret == PERF_ERRNO_TARGET__SUCCESS)
+			ret = PERF_ERRNO_TARGET__UID_OVERRIDE_CPU;
 	}
 
-	/* PID/UID and SYSTEM are mutually exclusive */
-	if ((target->tid || target->uid_str) && target->system_wide) {
-		ui__warning("PID/TID/UID switch overriding CPU\n");
-		sleep(1);
+	/* PID and SYSTEM are mutually exclusive */
+	if (target->tid && target->system_wide) {
 		target->system_wide = false;
+		if (ret == PERF_ERRNO_TARGET__SUCCESS)
+			ret = PERF_ERRNO_TARGET__PID_OVERRIDE_SYSTEM;
 	}
+
+	/* UID and SYSTEM are mutually exclusive */
+	if (target->uid_str && target->system_wide) {
+		target->system_wide = false;
+		if (ret == PERF_ERRNO_TARGET__SUCCESS)
+			ret = PERF_ERRNO_TARGET__UID_OVERRIDE_SYSTEM;
+	}
+
+	return ret;
 }
diff --git a/tools/perf/util/target.h b/tools/perf/util/target.h
index 218291f..eb0d210 100644
--- a/tools/perf/util/target.h
+++ b/tools/perf/util/target.h
@@ -13,6 +13,29 @@ struct perf_target {
 	bool	     system_wide;
 };
 
-void perf_target__validate(struct perf_target *target);
+enum perf_target_errno {
+	PERF_ERRNO_TARGET__SUCCESS		= 0,
+
+	/*
+	 * Choose an arbitrary negative big number not to clash with standard
+	 * errno since SUS requires the errno has distinct positive values.
+	 * See 'Issue 6' in the link below.
+	 *
+	 * http://pubs.opengroup.org/onlinepubs/9699919799/basedefs/errno.h.html
+	 */
+	__PERF_ERRNO_TARGET__START		= -10000,
+
+
+	/* for perf_target__validate() */
+	PERF_ERRNO_TARGET__PID_OVERRIDE_CPU	= __PERF_ERRNO_TARGET__START,
+	PERF_ERRNO_TARGET__PID_OVERRIDE_UID,
+	PERF_ERRNO_TARGET__UID_OVERRIDE_CPU,
+	PERF_ERRNO_TARGET__PID_OVERRIDE_SYSTEM,
+	PERF_ERRNO_TARGET__UID_OVERRIDE_SYSTEM,
+
+	__PERF_ERRNO_TARGET__END,
+};
+
+enum perf_target_errno perf_target__validate(struct perf_target *target);
 
 #endif /* _PERF_TARGET_H */
-- 
1.7.9.2.358.g22243

[PATCH 05/16] perf target: Introduce perf_target__parse_uid()

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

Add and use the modern perf_target__parse_uid() and get rid of the old
parse_target_uid().

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-5-git-send-email-namhyung.kim@lge.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |    4 +---
 tools/perf/builtin-top.c    |    3 +--
 tools/perf/util/target.c    |   35 +++++++++++++++++++++++++++++++++++
 tools/perf/util/target.h    |    5 +++++
 tools/perf/util/usage.c     |   31 -------------------------------
 tools/perf/util/util.h      |    3 ---
 6 files changed, 42 insertions(+), 39 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index d165909..d26a279 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -886,9 +886,7 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
 
 	perf_target__validate(&rec->opts.target);
 
-	rec->opts.target.uid = parse_target_uid(rec->opts.target.uid_str);
-	if (rec->opts.target.uid_str != NULL &&
-	    rec->opts.target.uid == UINT_MAX - 1)
+	if (perf_target__parse_uid(&rec->opts.target) < 0)
 		goto out_free_fd;
 
 	if (perf_evlist__create_maps(evsel_list, &rec->opts.target) < 0)
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index e40f86e..c9137ba 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1254,8 +1254,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
 
 	perf_target__validate(&top.target);
 
-	top.target.uid = parse_target_uid(top.target.uid_str);
-	if (top.target.uid_str != NULL && top.target.uid == UINT_MAX - 1)
+	if (perf_target__parse_uid(&top.target) < 0)
 		goto out_delete_evlist;
 
 	if (top.target.tid == 0 && top.target.pid == 0 &&
diff --git a/tools/perf/util/target.c b/tools/perf/util/target.c
index 5c59dcf..02a6bed 100644
--- a/tools/perf/util/target.c
+++ b/tools/perf/util/target.c
@@ -9,6 +9,8 @@
 #include "target.h"
 #include "debug.h"
 
+#include <pwd.h>
+
 
 enum perf_target_errno perf_target__validate(struct perf_target *target)
 {
@@ -54,3 +56,36 @@ enum perf_target_errno perf_target__validate(struct perf_target *target)
 
 	return ret;
 }
+
+enum perf_target_errno perf_target__parse_uid(struct perf_target *target)
+{
+	struct passwd pwd, *result;
+	char buf[1024];
+	const char *str = target->uid_str;
+
+	target->uid = UINT_MAX;
+	if (str == NULL)
+		return PERF_ERRNO_TARGET__SUCCESS;
+
+	/* Try user name first */
+	getpwnam_r(str, &pwd, buf, sizeof(buf), &result);
+
+	if (result == NULL) {
+		/*
+		 * The user name not found. Maybe it's a UID number.
+		 */
+		char *endptr;
+		int uid = strtol(str, &endptr, 10);
+
+		if (*endptr != '\0')
+			return PERF_ERRNO_TARGET__INVALID_UID;
+
+		getpwuid_r(uid, &pwd, buf, sizeof(buf), &result);
+
+		if (result == NULL)
+			return PERF_ERRNO_TARGET__USER_NOT_FOUND;
+	}
+
+	target->uid = result->pw_uid;
+	return PERF_ERRNO_TARGET__SUCCESS;
+}
diff --git a/tools/perf/util/target.h b/tools/perf/util/target.h
index eb0d210..d4aabda 100644
--- a/tools/perf/util/target.h
+++ b/tools/perf/util/target.h
@@ -33,9 +33,14 @@ enum perf_target_errno {
 	PERF_ERRNO_TARGET__PID_OVERRIDE_SYSTEM,
 	PERF_ERRNO_TARGET__UID_OVERRIDE_SYSTEM,
 
+	/* for perf_target__parse_uid() */
+	PERF_ERRNO_TARGET__INVALID_UID,
+	PERF_ERRNO_TARGET__USER_NOT_FOUND,
+
 	__PERF_ERRNO_TARGET__END,
 };
 
 enum perf_target_errno perf_target__validate(struct perf_target *target);
+enum perf_target_errno perf_target__parse_uid(struct perf_target *target);
 
 #endif /* _PERF_TARGET_H */
diff --git a/tools/perf/util/usage.c b/tools/perf/util/usage.c
index e851abc..4007aca 100644
--- a/tools/perf/util/usage.c
+++ b/tools/perf/util/usage.c
@@ -82,34 +82,3 @@ void warning(const char *warn, ...)
 	warn_routine(warn, params);
 	va_end(params);
 }
-
-uid_t parse_target_uid(const char *str)
-{
-	struct passwd pwd, *result;
-	char buf[1024];
-
-	if (str == NULL)
-		return UINT_MAX;
-
-	getpwnam_r(str, &pwd, buf, sizeof(buf), &result);
-
-	if (result == NULL) {
-		char *endptr;
-		int uid = strtol(str, &endptr, 10);
-
-		if (*endptr != '\0') {
-			ui__error("Invalid user %s\n", str);
-			return UINT_MAX - 1;
-		}
-
-		getpwuid_r(uid, &pwd, buf, sizeof(buf), &result);
-
-		if (result == NULL) {
-			ui__error("Problems obtaining information for user %s\n",
-				  str);
-			return UINT_MAX - 1;
-		}
-	}
-
-	return result->pw_uid;
-}
diff --git a/tools/perf/util/util.h b/tools/perf/util/util.h
index 52be74c..27a11a7 100644
--- a/tools/perf/util/util.h
+++ b/tools/perf/util/util.h
@@ -74,7 +74,6 @@
 #include <netinet/tcp.h>
 #include <arpa/inet.h>
 #include <netdb.h>
-#include <pwd.h>
 #include <inttypes.h>
 #include "../../../include/linux/magic.h"
 #include "types.h"
@@ -249,8 +248,6 @@ struct perf_event_attr;
 
 void event_attr_init(struct perf_event_attr *attr);
 
-uid_t parse_target_uid(const char *str);
-
 #define _STR(x) #x
 #define STR(x) _STR(x)
 
-- 
1.7.9.2.358.g22243

[PATCH 06/16] perf tools: Introduce perf_target__strerror()

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

The perf_target__strerror() sets @buf to a string that describes the
(perf_target-specific) error condition that is passed via @errnum.

This is similar to strerror_r() and does same thing if @errnum has a
standard errno value.

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Suggested-by: Arnaldo Carvalho de Melo <acme@ghostprotocols.net>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-6-git-send-email-namhyung.kim@lge.com
[ committer note: No need to use PERF_ERRNO_TARGET__SUCCESS, use shorter idiom ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |   18 +++++++++++++--
 tools/perf/builtin-top.c    |   19 +++++++++++++---
 tools/perf/util/debug.c     |    1 +
 tools/perf/util/target.c    |   51 +++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/target.h    |    3 +++
 5 files changed, 87 insertions(+), 5 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index d26a279..c8bf6ea 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -831,6 +831,7 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
 	struct perf_evsel *pos;
 	struct perf_evlist *evsel_list;
 	struct perf_record *rec = &record;
+	char errbuf[BUFSIZ];
 
 	perf_header__set_cmdline(argc, argv);
 
@@ -884,11 +885,24 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
 		goto out_symbol_exit;
 	}
 
-	perf_target__validate(&rec->opts.target);
+	err = perf_target__validate(&rec->opts.target);
+	if (err) {
+		perf_target__strerror(&rec->opts.target, err, errbuf, BUFSIZ);
+		ui__warning("%s", errbuf);
+	}
+
+	err = perf_target__parse_uid(&rec->opts.target);
+	if (err) {
+		int saved_errno = errno;
 
-	if (perf_target__parse_uid(&rec->opts.target) < 0)
+		perf_target__strerror(&rec->opts.target, err, errbuf, BUFSIZ);
+		ui__warning("%s", errbuf);
+
+		err = -saved_errno;
 		goto out_free_fd;
+	}
 
+	err = -ENOMEM;
 	if (perf_evlist__create_maps(evsel_list, &rec->opts.target) < 0)
 		usage_with_options(record_usage, record_options);
 
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index c9137ba..7ba0f03 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1150,7 +1150,8 @@ static const char * const top_usage[] = {
 int cmd_top(int argc, const char **argv, const char *prefix __used)
 {
 	struct perf_evsel *pos;
-	int status = -ENOMEM;
+	int status;
+	char errbuf[BUFSIZ];
 	struct perf_top top = {
 		.count_filter	     = 5,
 		.delay_secs	     = 2,
@@ -1252,10 +1253,22 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
 
 	setup_browser(false);
 
-	perf_target__validate(&top.target);
+	status = perf_target__validate(&top.target);
+	if (status) {
+		perf_target__strerror(&top.target, status, errbuf, BUFSIZ);
+		ui__warning("%s", errbuf);
+	}
+
+	status = perf_target__parse_uid(&top.target);
+	if (status) {
+		int saved_errno = errno;
 
-	if (perf_target__parse_uid(&top.target) < 0)
+		perf_target__strerror(&top.target, status, errbuf, BUFSIZ);
+		ui__warning("%s", errbuf);
+
+		status = -saved_errno;
 		goto out_delete_evlist;
+	}
 
 	if (top.target.tid == 0 && top.target.pid == 0 &&
 	    top.target.uid_str == NULL)
diff --git a/tools/perf/util/debug.c b/tools/perf/util/debug.c
index 26817da..efb1fce 100644
--- a/tools/perf/util/debug.c
+++ b/tools/perf/util/debug.c
@@ -11,6 +11,7 @@
 #include "event.h"
 #include "debug.h"
 #include "util.h"
+#include "target.h"
 
 int verbose;
 bool dump_trace = false, quiet = false;
diff --git a/tools/perf/util/target.c b/tools/perf/util/target.c
index 02a6bed..1064d5b 100644
--- a/tools/perf/util/target.c
+++ b/tools/perf/util/target.c
@@ -10,6 +10,7 @@
 #include "debug.h"
 
 #include <pwd.h>
+#include <string.h>
 
 
 enum perf_target_errno perf_target__validate(struct perf_target *target)
@@ -89,3 +90,53 @@ enum perf_target_errno perf_target__parse_uid(struct perf_target *target)
 	target->uid = result->pw_uid;
 	return PERF_ERRNO_TARGET__SUCCESS;
 }
+
+/*
+ * This must have a same ordering as the enum perf_target_errno.
+ */
+static const char *perf_target__error_str[] = {
+	"PID/TID switch overriding CPU",
+	"PID/TID switch overriding UID",
+	"UID switch overriding CPU",
+	"PID/TID switch overriding SYSTEM",
+	"UID switch overriding SYSTEM",
+	"Invalid User: %s",
+	"Problems obtaining information for user %s",
+};
+
+int perf_target__strerror(struct perf_target *target, int errnum,
+			  char *buf, size_t buflen)
+{
+	int idx;
+	const char *msg;
+
+	if (errnum >= 0) {
+		strerror_r(errnum, buf, buflen);
+		return 0;
+	}
+
+	if (errnum <  __PERF_ERRNO_TARGET__START ||
+	    errnum >= __PERF_ERRNO_TARGET__END)
+		return -1;
+
+	idx = errnum - __PERF_ERRNO_TARGET__START;
+	msg = perf_target__error_str[idx];
+
+	switch (errnum) {
+	case PERF_ERRNO_TARGET__PID_OVERRIDE_CPU
+	 ... PERF_ERRNO_TARGET__UID_OVERRIDE_SYSTEM:
+		snprintf(buf, buflen, "%s", msg);
+		break;
+
+	case PERF_ERRNO_TARGET__INVALID_UID:
+	case PERF_ERRNO_TARGET__USER_NOT_FOUND:
+		snprintf(buf, buflen, msg, target->uid_str);
+		break;
+
+	default:
+		/* cannot reach here */
+		break;
+	}
+
+	return 0;
+}
diff --git a/tools/perf/util/target.h b/tools/perf/util/target.h
index d4aabda..6fcd01c 100644
--- a/tools/perf/util/target.h
+++ b/tools/perf/util/target.h
@@ -43,4 +43,7 @@ enum perf_target_errno {
 enum perf_target_errno perf_target__validate(struct perf_target *target);
 enum perf_target_errno perf_target__parse_uid(struct perf_target *target);
 
+int perf_target__strerror(struct perf_target *target, int errnum, char *buf,
+			  size_t buflen);
+
 #endif /* _PERF_TARGET_H */
-- 
1.7.9.2.358.g22243

[PATCH 07/16] perf target: Consolidate target task/cpu checking

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

There are places that check whether target task/cpu is given or not and
some of them didn't check newly introduced uid or cpu list. Add and use
three of helper functions to treat them properly.

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-7-git-send-email-namhyung.kim@lge.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |    4 +---
 tools/perf/builtin-stat.c   |   12 ++++++------
 tools/perf/builtin-top.c    |    2 +-
 tools/perf/util/evlist.c    |   10 ++++------
 tools/perf/util/evsel.c     |    8 ++++----
 tools/perf/util/target.h    |   15 +++++++++++++++
 6 files changed, 31 insertions(+), 20 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index c8bf6ea..42e2414 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -843,9 +843,7 @@ int cmd_record(int argc, const char **argv, const char *prefix __used)
 
 	argc = parse_options(argc, argv, record_options, record_usage,
 			    PARSE_OPT_STOP_AT_NON_OPTION);
-	if (!argc && !rec->opts.target.pid && !rec->opts.target.tid &&
-	    !rec->opts.target.system_wide && !rec->opts.target.cpu_list &&
-	    !rec->opts.target.uid_str)
+	if (!argc && perf_target__none(&rec->opts.target))
 		usage_with_options(record_usage, record_options);
 
 	if (rec->force && rec->append_file) {
diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index bb77232..d9ff246 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -290,10 +290,10 @@ static int create_perf_stat_counter(struct perf_evsel *evsel,
 
 	attr->inherit = !no_inherit;
 
-	if (target.system_wide)
+	if (!perf_target__no_cpu(&target))
 		return perf_evsel__open_per_cpu(evsel, evsel_list->cpus,
 						group, group_fd);
-	if (!target.pid && !target.tid && (!group || evsel == first)) {
+	if (perf_target__no_task(&target) && (!group || evsel == first)) {
 		attr->disabled = 1;
 		attr->enable_on_exec = 1;
 	}
@@ -443,7 +443,7 @@ static int run_perf_stat(int argc __used, const char **argv)
 			exit(-1);
 		}
 
-		if (!target.tid && !target.pid && !target.system_wide)
+		if (perf_target__none(&target))
 			evsel_list->threads->map[0] = child_pid;
 
 		/*
@@ -965,7 +965,7 @@ static void print_stat(int argc, const char **argv)
 	if (!csv_output) {
 		fprintf(output, "\n");
 		fprintf(output, " Performance counter stats for ");
-		if (!target.pid && !target.tid) {
+		if (perf_target__no_task(&target)) {
 			fprintf(output, "\'%s", argv[0]);
 			for (i = 1; i < argc; i++)
 				fprintf(output, " %s", argv[i]);
@@ -1187,13 +1187,13 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 	} else if (big_num_opt == 0) /* User passed --no-big-num */
 		big_num = false;
 
-	if (!argc && !target.pid && !target.tid)
+	if (!argc && perf_target__no_task(&target))
 		usage_with_options(stat_usage, options);
 	if (run_count <= 0)
 		usage_with_options(stat_usage, options);
 
 	/* no_aggr, cgroup are for system-wide only */
-	if ((no_aggr || nr_cgroups) && !target.system_wide) {
+	if ((no_aggr || nr_cgroups) && perf_target__no_cpu(&target)) {
 		fprintf(stderr, "both cgroup and no-aggregation "
 			"modes only available in system-wide mode\n");
 
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 7ba0f03..e4ca827 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1016,7 +1016,7 @@ static int __cmd_top(struct perf_top *top)
 	if (ret)
 		goto out_delete;
 
-	if (top->target.tid || top->target.uid != UINT_MAX)
+	if (!perf_target__no_task(&top->target))
 		perf_event__synthesize_thread_map(&top->tool, top->evlist->threads,
 						  perf_event__process,
 						  &top->session->host_machine);
diff --git a/tools/perf/util/evlist.c b/tools/perf/util/evlist.c
index 183b199..1201daf 100644
--- a/tools/perf/util/evlist.c
+++ b/tools/perf/util/evlist.c
@@ -609,12 +609,10 @@ int perf_evlist__create_maps(struct perf_evlist *evlist,
 	if (evlist->threads == NULL)
 		return -1;
 
-	if (target->uid != UINT_MAX || target->tid)
-		evlist->cpus = cpu_map__dummy_new();
-	else if (!target->system_wide && target->cpu_list == NULL)
-		evlist->cpus = cpu_map__dummy_new();
-	else
+	if (!perf_target__no_cpu(target))
 		evlist->cpus = cpu_map__new(target->cpu_list);
+	else
+		evlist->cpus = cpu_map__dummy_new();
 
 	if (evlist->cpus == NULL)
 		goto out_delete_threads;
@@ -831,7 +829,7 @@ int perf_evlist__prepare_workload(struct perf_evlist *evlist,
 		exit(-1);
 	}
 
-	if (!opts->target.system_wide && !opts->target.tid && !opts->target.pid)
+	if (perf_target__none(&opts->target))
 		evlist->threads->map[0] = evlist->workload.pid;
 
 	close(child_ready_pipe[1]);
diff --git a/tools/perf/util/evsel.c b/tools/perf/util/evsel.c
index bb785a0..21eaab2 100644
--- a/tools/perf/util/evsel.c
+++ b/tools/perf/util/evsel.c
@@ -114,8 +114,8 @@ void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
 		attr->sample_type	|= PERF_SAMPLE_PERIOD;
 
 	if (!opts->sample_id_all_missing &&
-	    (opts->sample_time || opts->target.system_wide ||
-	     !opts->no_inherit || opts->target.cpu_list))
+	    (opts->sample_time || !opts->no_inherit ||
+	     !perf_target__no_cpu(&opts->target)))
 		attr->sample_type	|= PERF_SAMPLE_TIME;
 
 	if (opts->raw_samples) {
@@ -136,8 +136,8 @@ void perf_evsel__config(struct perf_evsel *evsel, struct perf_record_opts *opts,
 	attr->mmap = track;
 	attr->comm = track;
 
-	if (!opts->target.pid && !opts->target.tid &&
-	    !opts->target.system_wide && (!opts->group || evsel == first)) {
+	if (perf_target__none(&opts->target) &&
+	    (!opts->group || evsel == first)) {
 		attr->disabled = 1;
 		attr->enable_on_exec = 1;
 	}
diff --git a/tools/perf/util/target.h b/tools/perf/util/target.h
index 6fcd01c..127cff3 100644
--- a/tools/perf/util/target.h
+++ b/tools/perf/util/target.h
@@ -46,4 +46,19 @@ enum perf_target_errno perf_target__parse_uid(struct perf_target *target);
 int perf_target__strerror(struct perf_target *target, int errnum, char *buf,
 			  size_t buflen);
 
+static inline bool perf_target__no_task(struct perf_target *target)
+{
+	return !target->pid && !target->tid && !target->uid_str;
+}
+
+static inline bool perf_target__no_cpu(struct perf_target *target)
+{
+	return !target->system_wide && !target->cpu_list;
+}
+
+static inline bool perf_target__none(struct perf_target *target)
+{
+	return perf_target__no_task(target) && perf_target__no_cpu(target);
+}
+
 #endif /* _PERF_TARGET_H */
-- 
1.7.9.2.358.g22243

[PATCH 08/16] perf stat: Use perf_evlist__create_maps

[Arnaldo Carvalho de Melo]

From: Namhyung Kim <namhyung.kim@lge.com>

Use same function with perf record and top to share the code checks
combinations of different switches.

Signed-off-by: Namhyung Kim <namhyung.kim@lge.com>
Reviewed-by: David Ahern <dsahern@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Namhyung Kim <namhyung@gmail.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Link: http://lkml.kernel.org/r/1336367344-28071-8-git-send-email-namhyung.kim@lge.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c |   22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index d9ff246..e720ba7 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -175,7 +175,9 @@ static struct perf_event_attr very_very_detailed_attrs[] = {
 
 static struct perf_evlist	*evsel_list;
 
-static struct perf_target	target;
+static struct perf_target	target = {
+	.uid	= UINT_MAX,
+};
 
 static int			run_idx				=  0;
 static int			run_count			=  1;
@@ -1205,20 +1207,12 @@ int cmd_stat(int argc, const char **argv, const char *prefix __used)
 
 	perf_target__validate(&target);
 
-	evsel_list->threads = thread_map__new_str(target.pid,
-						  target.tid, UINT_MAX);
-	if (evsel_list->threads == NULL) {
-		pr_err("Problems finding threads of monitor\n");
-		usage_with_options(stat_usage, options);
-	}
-
-	if (target.system_wide)
-		evsel_list->cpus = cpu_map__new(target.cpu_list);
-	else
-		evsel_list->cpus = cpu_map__dummy_new();
+	if (perf_evlist__create_maps(evsel_list, &target) < 0) {
+		if (!perf_target__no_task(&target))
+			pr_err("Problems finding threads of monitor\n");
+		if (!perf_target__no_cpu(&target))
+			perror("failed to parse CPUs map");
 
-	if (evsel_list->cpus == NULL) {
-		perror("failed to parse CPUs map");
 		usage_with_options(stat_usage, options);
 		return -1;
 	}
-- 
1.7.9.2.358.g22243

[PATCH 09/16] perf top: Default to system wide using perf_target methods

[Arnaldo Carvalho de Melo]

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

Additionally we were not checking if a cpu list had been provided by the
user. Fix that.

Reported-by: David Ahern <dsahern@gmail.com>
Reported-by: Namhyung Kim <namhyung@gmail.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
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-ao3zrouylwmt7h9ikj0krubi@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-top.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index e4ca827..c53cdab 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1270,8 +1270,7 @@ int cmd_top(int argc, const char **argv, const char *prefix __used)
 		goto out_delete_evlist;
 	}
 
-	if (top.target.tid == 0 && top.target.pid == 0 &&
-	    top.target.uid_str == NULL)
+	if (perf_target__none(&top.target))
 		top.target.system_wide = true;
 
 	if (perf_evlist__create_maps(top.evlist, &top.target) < 0)
-- 
1.7.9.2.358.g22243

[PATCH 10/16] perf report: Fix format string for x86-32 compilation

[Arnaldo Carvalho de Melo]

From: Jiri Olsa <jolsa@redhat.com>

Using PRIu64 for printing out u64 nr_events to fix compilation
for x86 32 bits.

Cc: Arun Sharma <asharma@fb.com>
Cc: Corey Ashford <cjashfor@linux.vnet.ibm.com>
Cc: Cyrill Gorcunov <gorcunov@openvz.org>
Cc: Frank C. Eigler <fche@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Robert Richter <robert.richter@amd.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
Cc: Ulrich Drepper <drepper@gmail.com>
Link: http://lkml.kernel.org/r/1335958638-5160-7-git-send-email-jolsa@redhat.com
Signed-off-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-report.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 5df829f..7477655 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -304,7 +304,7 @@ static size_t hists__fprintf_nr_sample_events(struct hists *self,
 	if (evname != NULL)
 		ret += fprintf(fp, " of event '%s'", evname);
 
-	ret += fprintf(fp, "\n# Event count (approx.): %lu", nr_events);
+	ret += fprintf(fp, "\n# Event count (approx.): %" PRIu64, nr_events);
 	return ret + fprintf(fp, "\n#\n");
 }
 
-- 
1.7.9.2.358.g22243

[PATCH 11/16] perf record: Fix fallback to cpu-clock on ppc

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

perf-record on PPC is not falling back to cpu-clock:

$ perf record -ag -fo /tmp/perf.data -- sleep 1

  Error: sys_perf_event_open() syscall returned with 6 (No such device or address).  /bin/dmesg may provide additional information.

  Fatal: No CONFIG_PERF_EVENTS=y kernel support configured?

The problem is that until 2.6.37 (behavior changed with commit b0a873e)
perf on PPC returns ENXIO when hw_perf_event_init() fails. With this
patch we get the expected behavior:

$ perf record -ag -fo /tmp/perf.data -v -- sleep 1
Old kernel, cannot exclude guest or host samples.
The cycles event is not supported, trying to fall back to cpu-clock-ticks
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.151 MB /tmp/perf.data (~6592 samples) ]

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1336490937-57106-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |    8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 42e2414..1a9098c 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -242,9 +242,13 @@ try_again:
 			/*
 			 * If it's cycles then fall back to hrtimer
 			 * based cpu-clock-tick sw counter, which
-			 * is always available even if no PMU support:
+			 * is always available even if no PMU support.
+			 *
+			 * PPC returns ENXIO until 2.6.37 (behavior changed
+			 * with commit b0a873e).
 			 */
-			if (err == ENOENT && attr->type == PERF_TYPE_HARDWARE
+			if ((err == ENOENT || err == ENXIO)
+					&& attr->type == PERF_TYPE_HARDWARE
 					&& attr->config == PERF_COUNT_HW_CPU_CYCLES) {
 
 				if (verbose)
-- 
1.7.9.2.358.g22243

[PATCH 12/16] perf stat: handle ENXIO error for perf_event_open

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

perf stat on PPC currently fails to run:

$ perf stat -- sleep 1
  Error: open_counter returned with 6 (No such device or address). /bin/dmesg may provide additional information.

  Fatal: Not all events could be opened.

The problem is that until 2.6.37 (behavior changed with commit b0a873e)
perf on PPC returns ENXIO when hw_perf_event_init() fails. With this
patch we get the expected behavior:

$ perf stat -v -- sleep 1
cycles event is not supported by the kernel.
stalled-cycles-frontend event is not supported by the kernel.
stalled-cycles-backend event is not supported by the kernel.
instructions event is not supported by the kernel.
branches event is not supported by the kernel.
branch-misses event is not supported by the kernel.

...

Signed-off-by: David Ahern <dsahern@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1336490956-57145-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-stat.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index e720ba7..d060568 100644
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -462,8 +462,13 @@ static int run_perf_stat(int argc __used, const char **argv)
 
 	list_for_each_entry(counter, &evsel_list->entries, node) {
 		if (create_perf_stat_counter(counter, first) < 0) {
+			/*
+			 * PPC returns ENXIO for HW counters until 2.6.37
+			 * (behavior changed with commit b0a873e).
+			 */
 			if (errno == EINVAL || errno == ENOSYS ||
-			    errno == ENOENT || errno == EOPNOTSUPP) {
+			    errno == ENOENT || errno == EOPNOTSUPP ||
+			    errno == ENXIO) {
 				if (verbose)
 					ui__warning("%s event is not supported by the kernel.\n",
 						    event_name(counter));
-- 
1.7.9.2.358.g22243

[PATCH 13/16] perf top: Update event name when falling back to cpu-clock

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

The 'perf top' command falls back to cpu-clock if the H/W cycles event
is not supported, but the event name is not updated leading to a
misleading header:

PerfTop: 8 irqs/sec  kernel:75.0%  exact:  0.0% [1000Hz cycles],  ...

Update the event name when the event type is changed so that the
header displays correctly:

PerfTop: 794 irqs/sec  kernel:100.0%  exact:  0.0% [1000Hz cpu-clock], ...

Signed-off-by: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/1336495789-58420-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-top.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index c53cdab..4eb6171 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -948,6 +948,10 @@ try_again:
 
 				attr->type = PERF_TYPE_SOFTWARE;
 				attr->config = PERF_COUNT_SW_CPU_CLOCK;
+				if (counter->name) {
+					free(counter->name);
+					counter->name = strdup(event_name(counter));
+				}
 				goto try_again;
 			}
 
-- 
1.7.9.2.358.g22243

[PATCH 14/16] perf record: Reset event name when falling back to cpu-clock

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

perf-record defaults to the H/W cycles event and if it is not supported
falls back to cpu-clock. Reset the event name as well.

Signed-off-by: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/1336495811-58461-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/builtin-record.c |    4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/perf/builtin-record.c b/tools/perf/builtin-record.c
index 1a9098c..d19058a 100644
--- a/tools/perf/builtin-record.c
+++ b/tools/perf/builtin-record.c
@@ -256,6 +256,10 @@ try_again:
 						    "trying to fall back to cpu-clock-ticks\n");
 				attr->type = PERF_TYPE_SOFTWARE;
 				attr->config = PERF_COUNT_SW_CPU_CLOCK;
+				if (pos->name) {
+					free(pos->name);
+					pos->name = NULL;
+				}
 				goto try_again;
 			}
 
-- 
1.7.9.2.358.g22243

[PATCH 15/16] perf annotate: shorten helpline so it fits in visible space

[Arnaldo Carvalho de Melo]

From: David Ahern <dsahern@gmail.com>

Additional toggles have pushed the help line out of view on a modestly
sized terminal (120 columns wide). Shorten it to just reminders.

Signed-off-by: David Ahern <dsahern@gmail.com>
Link: http://lkml.kernel.org/r/1336510879-64610-1-git-send-email-dsahern@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/ui/browsers/annotate.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/tools/perf/ui/browsers/annotate.c b/tools/perf/ui/browsers/annotate.c
index 4db5186..a299290 100644
--- a/tools/perf/ui/browsers/annotate.c
+++ b/tools/perf/ui/browsers/annotate.c
@@ -491,9 +491,9 @@ static int annotate_browser__run(struct annotate_browser *self, int evidx,
 	struct map_symbol *ms = self->b.priv;
 	struct symbol *sym = ms->sym;
 	const char *help = "<-/ESC: Exit, TAB/shift+TAB: Cycle hot lines, "
-			   "H: Go to hottest line, ->/ENTER: Line action, "
-			   "O: Toggle offset view, "
-			   "S: Toggle source code view";
+			   "H: Hottest line, ->/ENTER: Line action, "
+			   "O: Offset view, "
+			   "S: Source view";
 	int key;
 
 	if (ui_browser__show(&self->b, sym->name, help) < 0)
-- 
1.7.9.2.358.g22243

[PATCH 16/16] perf hists browser: Use '/' for search/filter instead of 's'

[Arnaldo Carvalho de Melo]

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

That is what is used in vi and mutt, and as well on the 'annotate'
browser.

Eventually we can have keymappings to make people used to other key
associations more confortable.

Suggested-by: Ingo Molnar <mingo@kernel.org>
Cc: David Ahern <dsahern@gmail.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Namhyung Kim <namhyung@gmail.com>
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-fyln9286b8gx5q4n277l0djs@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
---
 tools/perf/ui/browsers/hists.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
index 466827e..a372a4b 100644
--- a/tools/perf/ui/browsers/hists.c
+++ b/tools/perf/ui/browsers/hists.c
@@ -941,7 +941,7 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events,
 			goto zoom_dso;
 		case 't':
 			goto zoom_thread;
-		case 's':
+		case '/':
 			if (ui_browser__input_window("Symbol to show",
 					"Please enter the name of symbol you want to see",
 					buf, "ENTER: OK, ESC: Cancel",
@@ -969,7 +969,7 @@ static int perf_evsel__hists_browse(struct perf_evsel *evsel, int nr_events,
 					"E             Expand all callchains\n"
 					"d             Zoom into current DSO\n"
 					"t             Zoom into current Thread\n"
-					"s             Filter symbol by name");
+					"/             Filter symbol by name");
 			continue;
 		case K_ENTER:
 		case K_RIGHT:
-- 
1.7.9.2.358.g22243

Re: [GIT PULL 00/16] perf/core fixes and improvements

[Ingo Molnar]

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

> Hi Ingo,
> 
> 	This is on top of a 15 patches long series still outstanding on my
> perf/core branch, please consider pulling,
> 
> - Arnaldo
> 
> The following changes since commit dc41b9b8f02dbe2228ae787d525dac43beebb7fa:
> 
>   perf ui: Change fallback policy of setup_browser() (2012-05-02 16:17:37 -0300)
> 
> 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 5a5626b1b4bf8467891c9297ffda979db97ed5ec:
> 
>   perf hists browser: Use '/' for search/filter instead of 's' (2012-05-10 13:07:59 -0300)
> 
> ----------------------------------------------------------------
> Fixes and improvements for perf/core:
> 
> . perf_target: abstraction for --uid, --pid, --tid, --cpu, --all-cpus handling,
>   eliminating code duplicated in the tools, having constraints that apply to
>   all of them, from Namhyung Kim
> 
> . Fixes for handling fallback to cpu-clock on PPC, from David Ahern
> 
> . Fix for processing events with unknown size, from Jiri Olsa
> 
> . Compilation fix on 32-bit, from Jiri Olsa
> 
> Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
> 
> ----------------------------------------------------------------
> Arnaldo Carvalho de Melo (3):
>       perf top: Set target.system_wide
>       perf top: Default to system wide using perf_target methods
>       perf hists browser: Use '/' for search/filter instead of 's'
> 
> David Ahern (5):
>       perf record: Fix fallback to cpu-clock on ppc
>       perf stat: handle ENXIO error for perf_event_open
>       perf top: Update event name when falling back to cpu-clock
>       perf record: Reset event name when falling back to cpu-clock
>       perf annotate: shorten helpline so it fits in visible space
> 
> Jiri Olsa (2):
>       perf session: Fail on processing event with unknown size
>       perf report: Fix format string for x86-32 compilation
> 
> Namhyung Kim (6):
>       perf evlist: Fix creation of cpu map
>       perf target: Introduce perf_target_errno
>       perf target: Introduce perf_target__parse_uid()
>       perf tools: Introduce perf_target__strerror()
>       perf target: Consolidate target task/cpu checking
>       perf stat: Use perf_evlist__create_maps
> 
>  tools/perf/builtin-record.c       |   36 ++++++++---
>  tools/perf/builtin-report.c       |    2 +-
>  tools/perf/builtin-stat.c         |   41 +++++++------
>  tools/perf/builtin-top.c          |   29 +++++++--
>  tools/perf/ui/browsers/annotate.c |    6 +-
>  tools/perf/ui/browsers/hists.c    |    4 +-
>  tools/perf/util/debug.c           |    1 +
>  tools/perf/util/evlist.c          |    9 ++-
>  tools/perf/util/evsel.c           |    8 +--
>  tools/perf/util/session.c         |   30 +++-------
>  tools/perf/util/target.c          |  119 +++++++++++++++++++++++++++++++++----
>  tools/perf/util/target.h          |   48 ++++++++++++++-
>  tools/perf/util/usage.c           |   31 ----------
>  tools/perf/util/util.h            |    3 -
>  14 files changed, 250 insertions(+), 117 deletions(-)

Pulled, thanks Arnaldo!

	Ingo