[PATCHSET 00/26] perf tools: Add support to accumulate hist periods (v10)

[PATCHSET 00/26] perf tools: Add support to accumulate hist periods (v10)

[Namhyung Kim]

Hello,

This is a new attempt to implement cumulative hist period report.
This work begins from Arun's SORT_INCLUSIVE patch [1] but I completely
rewrote it from scratch.

This patchset basically adds period in a sample to every node in the
callchain.  A hist_entry now has an additional fields to keep the
cumulative period if --children option is given on perf report.

I changed the option as a separate --children and added a new
"Children" column (and renamed the default "Overhead" column into
"Self").  The output will be sorted by children (cumulative) overhead
for now.  It also adds 'overhead_children' field to -F/--fields option
to be selected by user and it'll show you "N/A" if the cumulation is
not supported (due to missing callchain).

I added Tested-by from Rodrigo Campos since this version is basically
rebase of previous series + bugfix.  But it still needs to be tested
more intensively IMHO.  Also note that, this will change default
behavior of perf report/top if callchain is recorded, so might confuse
old users.  Let's see how many of them come to shout. :)  I think we
need to merge Jiri's TUI column header patch at least.


 * changes in v10:
  - fix callchain display bug in stdio, gtk
  - add a testcase
  - add Tested-by from Rodrigo

 * changes in v9:
  - support output field option
  - add Acked-by tags from Jiri

 * changes in v8:
  - not depends on --percentage patchkit
  - fix callchain resolving bug (Jiri)
  - convert to sample__resolve_{mem,bstack}
  - eliminate 'event' field from hist_entry_iter

 * changes in v7:
  - add Tested-by tags from Arun
  - rebase onto current acme/perf/core

 * changes in v6:
  - separate struct hist_iter_ops (Jiri)
  - check iter->he before calling ->add_entry_cb (Jiri)
  - fix locking issue on perf top (Jiri)

 * changes in v5:
  - support both of --children and --call-graph (Arun)
  - refactor hist_entry_iter to share with perf top (Jiri)
  - various cleanups and fixes (Jiri)
  - add ack's from Jiri

 * changes in v4:
  - change to --children option (Ingo)
  - rebased on new annotation change (Arnaldo)
  - support perf top also
  - enable --children option by default (Ingo)

 * changes in v3:
  - change to --cumulate option
  - fix a couple of bugs (Jiri, Rodrigo)
  - rename some help functions (Arnaldo)
  - cache previous hist entries rathen than just symbol and dso
  - add some preparatory cleanups
  - add report.cumulate config option


Let me show you an example:

  $ cat abc.c
  #define barrier() asm volatile("" ::: "memory")

  void a(void)
  {
  	int i;
  	for (i = 0; i < 1000000; i++)
  		barrier();
  }
  void b(void)
  {
  	a();
  }
  void c(void)
  {
  	b();
  }
  int main(void)
  {
  	c();
  	return 0;
  }

With this simple program I ran perf record and report:

  $ perf record -g -e cycles:u ./abc


Case 1.

  $ perf report --stdio --no-call-graph --no-children

  # Overhead  Command      Shared Object          Symbol
  # ........  .......  .................  ..............
  #
      91.50%      abc  abc                [.] a         
       8.18%      abc  ld-2.17.so         [.] strlen    
       0.31%      abc  [kernel.kallsyms]  [k] page_fault
       0.01%      abc  ld-2.17.so         [.] _start    


Case 2. (current default behavior)

  $ perf report --stdio --call-graph --no-children

  # Overhead  Command      Shared Object          Symbol
  # ........  .......  .................  ..............
  #
      91.50%      abc  abc                [.] a         
                  |
                  --- a
                      b
                      c
                      main
                      __libc_start_main

       8.18%      abc  ld-2.17.so         [.] strlen    
                  |
                  --- strlen
                      _dl_sysdep_start

       0.31%      abc  [kernel.kallsyms]  [k] page_fault
                  |
                  --- page_fault
                      _start

       0.01%      abc  ld-2.17.so         [.] _start    
                  |
                  --- _start


Case 3.

  $ perf report --no-call-graph --children --stdio

  #     Self  Children  Command      Shared Object                 Symbol
  # ........  ........  .......  .................  .....................
  #
       0.00%    91.50%      abc  libc-2.17.so       [.] __libc_start_main
       0.00%    91.50%      abc  abc                [.] main             
       0.00%    91.50%      abc  abc                [.] c                
       0.00%    91.50%      abc  abc                [.] b                
      91.50%    91.50%      abc  abc                [.] a                
       0.00%     8.18%      abc  ld-2.17.so         [.] _dl_sysdep_start 
       8.18%     8.18%      abc  ld-2.17.so         [.] strlen           
       0.01%     0.33%      abc  ld-2.17.so         [.] _start           
       0.31%     0.31%      abc  [kernel.kallsyms]  [k] page_fault       

As you can see __libc_start_main -> main -> c -> b -> a callchain show
up in the output.

Finally, it looks like below with both option enabled:

Case 4. (default behavior?)

  $ perf report --call-graph --children --stdio

  #     Self  Children  Command      Shared Object                 Symbol
  # ........  ........  .......  .................  .....................
  #
       0.00%    91.50%      abc  libc-2.17.so       [.] __libc_start_main
                  |
                  --- __libc_start_main

       0.00%    91.50%      abc  abc                [.] main             
                  |
                  --- main
                      __libc_start_main

       0.00%    91.50%      abc  abc                [.] c                
                  |
                  --- c
                      main
                      __libc_start_main

       0.00%    91.50%      abc  abc                [.] b                
                  |
                  --- b
                      c
                      main
                      __libc_start_main

      91.50%    91.50%      abc  abc                [.] a                
                  |
                  --- a
                      b
                      c
                      main
                      __libc_start_main
  ...


Currently the perf enables both of --call-graph and --children when it
finds callchains in the samples.  While this is useful for TUI or GTK,
I'm not sure for stdio as it'd consume so much lines.

It does not handle all kind of cases like event annotation yet, but I
really want to release it and get reviews.

You can also get this series on 'perf/cumulate-v10' branch in my tree at:

  git://git.kernel.org/pub/scm/linux/kernel/git/namhyung/linux-perf.git


Any comments are welcome, thanks.
Namhyung


Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>

[1] https://lkml.org/lkml/2012/3/31/6
[2] https://lkml.org/lkml/2014/3/10/48
[3] https://lkml.org/lkml/2014/3/19/689


Namhyung Kim (26):
  perf tools: Introduce struct hist_entry_iter
  perf hists: Add support for accumulated stat of hist entry
  perf hists: Check if accumulated when adding a hist entry
  perf hists: Accumulate hist entry stat based on the callchain
  perf tools: Update cpumode for each cumulative entry
  perf report: Cache cumulative callchains
  perf callchain: Add callchain_cursor_snapshot()
  perf tools: Save callchain info for each cumulative entry
  perf ui/hist: Add support to accumulated hist stat
  perf ui/browser: Add support to accumulated hist stat
  perf ui/gtk: Add support to accumulated hist stat
  perf tools: Apply percent-limit to cumulative percentage
  perf tools: Add more hpp helper functions
  perf report: Add --children option
  perf report: Add report.children config option
  perf tools: Do not auto-remove Children column if --fields given
  perf tools: Add callback function to hist_entry_iter
  perf top: Convert to hist_entry_iter
  perf top: Add --children option
  perf top: Add top.children config option
  perf tools: Enable --children option by default
  perf ui/stdio: Fix invalid percentage value of cumulated hist entries
  perf ui/gtk: Fix callchain display
  perf tools: Reset output/sort order to default
  perf tests: Define and use symbolic names for fake symbols
  perf tests: Add a test case for cumulating callchains

 tools/perf/Documentation/perf-report.txt |   7 +-
 tools/perf/Documentation/perf-top.txt    |   8 +-
 tools/perf/Makefile.perf                 |   1 +
 tools/perf/builtin-annotate.c            |   3 +-
 tools/perf/builtin-diff.c                |   2 +-
 tools/perf/builtin-report.c              | 186 ++------
 tools/perf/builtin-top.c                 |  95 ++--
 tools/perf/tests/builtin-test.c          |   4 +
 tools/perf/tests/hists_common.c          |  52 ++-
 tools/perf/tests/hists_common.h          |  32 +-
 tools/perf/tests/hists_cumulate.c        | 726 +++++++++++++++++++++++++++++++
 tools/perf/tests/hists_filter.c          |  59 ++-
 tools/perf/tests/hists_link.c            |  36 +-
 tools/perf/tests/hists_output.c          |  31 +-
 tools/perf/tests/tests.h                 |   1 +
 tools/perf/ui/browsers/hists.c           |  66 +--
 tools/perf/ui/gtk/hists.c                |  33 +-
 tools/perf/ui/hist.c                     | 119 +++++
 tools/perf/ui/stdio/hist.c               |   8 +-
 tools/perf/util/callchain.c              |  45 +-
 tools/perf/util/callchain.h              |  11 +
 tools/perf/util/hist.c                   | 490 ++++++++++++++++++++-
 tools/perf/util/hist.h                   |  49 ++-
 tools/perf/util/sort.c                   |   4 +
 tools/perf/util/sort.h                   |  18 +-
 tools/perf/util/symbol.c                 |  11 +-
 tools/perf/util/symbol.h                 |   1 +
 27 files changed, 1782 insertions(+), 316 deletions(-)
 create mode 100644 tools/perf/tests/hists_cumulate.c

-- 
1.9.2

[PATCH 01/26] perf tools: Introduce struct hist_entry_iter

[Namhyung Kim]

There're some duplicate code when adding hist entries.  They are
different in that some have branch info or mem info but generally do
same thing.  So introduce new struct hist_entry_iter and add callbacks
to customize each case in general way.

The new perf_evsel__add_entry() function will look like:

  iter->prepare_entry();
  iter->add_single_entry();

  while (iter->next_entry())
    iter->add_next_entry();

  iter->finish_entry();

This will help further work like the cumulative callchain patchset.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Stephane Eranian <eranian@google.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-report.c     | 194 +++----------------------
 tools/perf/tests/hists_filter.c |  18 +--
 tools/perf/tests/hists_output.c |  11 +-
 tools/perf/util/hist.c          | 303 ++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/hist.h          |  33 +++++
 5 files changed, 371 insertions(+), 188 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index bc0eec1ce4be..d224f5961faa 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -76,165 +76,6 @@ static int report__config(const char *var, const char *value, void *cb)
 	return perf_default_config(var, value, cb);
 }
 
-static void report__inc_stats(struct report *rep, struct hist_entry *he)
-{
-	/*
-	 * The @he is either of a newly created one or an existing one
-	 * merging current sample.  We only want to count a new one so
-	 * checking ->nr_events being 1.
-	 */
-	if (he->stat.nr_events == 1)
-		rep->nr_entries++;
-
-	/*
-	 * Only counts number of samples at this stage as it's more
-	 * natural to do it here and non-sample events are also
-	 * counted in perf_session_deliver_event().  The dump_trace
-	 * requires this info is ready before going to the output tree.
-	 */
-	hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
-	if (!he->filtered)
-		he->hists->stats.nr_non_filtered_samples++;
-}
-
-static int report__add_mem_hist_entry(struct report *rep, struct addr_location *al,
-				      struct perf_sample *sample, struct perf_evsel *evsel)
-{
-	struct symbol *parent = NULL;
-	struct hist_entry *he;
-	struct mem_info *mi, *mx;
-	uint64_t cost;
-	int err = sample__resolve_callchain(sample, &parent, evsel, al, rep->max_stack);
-
-	if (err)
-		return err;
-
-	mi = sample__resolve_mem(sample, al);
-	if (!mi)
-		return -ENOMEM;
-
-	if (rep->hide_unresolved && !al->sym)
-		return 0;
-
-	cost = sample->weight;
-	if (!cost)
-		cost = 1;
-
-	/*
-	 * must pass period=weight in order to get the correct
-	 * sorting from hists__collapse_resort() which is solely
-	 * based on periods. We want sorting be done on nr_events * weight
-	 * and this is indirectly achieved by passing period=weight here
-	 * and the he_stat__add_period() function.
-	 */
-	he = __hists__add_entry(&evsel->hists, al, parent, NULL, mi,
-				cost, cost, 0);
-	if (!he)
-		return -ENOMEM;
-
-	if (ui__has_annotation()) {
-		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
-		if (err)
-			goto out;
-
-		mx = he->mem_info;
-		err = addr_map_symbol__inc_samples(&mx->daddr, evsel->idx);
-		if (err)
-			goto out;
-	}
-
-	report__inc_stats(rep, he);
-
-	err = hist_entry__append_callchain(he, sample);
-out:
-	return err;
-}
-
-static int report__add_branch_hist_entry(struct report *rep, struct addr_location *al,
-					 struct perf_sample *sample, struct perf_evsel *evsel)
-{
-	struct symbol *parent = NULL;
-	unsigned i;
-	struct hist_entry *he;
-	struct branch_info *bi, *bx;
-	int err = sample__resolve_callchain(sample, &parent, evsel, al, rep->max_stack);
-
-	if (err)
-		return err;
-
-	bi = sample__resolve_bstack(sample, al);
-	if (!bi)
-		return -ENOMEM;
-
-	for (i = 0; i < sample->branch_stack->nr; i++) {
-		if (rep->hide_unresolved && !(bi[i].from.sym && bi[i].to.sym))
-			continue;
-
-		err = -ENOMEM;
-
-		/* overwrite the 'al' to branch-to info */
-		al->map = bi[i].to.map;
-		al->sym = bi[i].to.sym;
-		al->addr = bi[i].to.addr;
-		/*
-		 * The report shows the percentage of total branches captured
-		 * and not events sampled. Thus we use a pseudo period of 1.
-		 */
-		he = __hists__add_entry(&evsel->hists, al, parent, &bi[i], NULL,
-					1, 1, 0);
-		if (he) {
-			if (ui__has_annotation()) {
-				bx = he->branch_info;
-				err = addr_map_symbol__inc_samples(&bx->from,
-								   evsel->idx);
-				if (err)
-					goto out;
-
-				err = addr_map_symbol__inc_samples(&bx->to,
-								   evsel->idx);
-				if (err)
-					goto out;
-			}
-			report__inc_stats(rep, he);
-		} else
-			goto out;
-	}
-	err = 0;
-out:
-	free(bi);
-	return err;
-}
-
-static int report__add_hist_entry(struct report *rep, struct perf_evsel *evsel,
-				  struct addr_location *al, struct perf_sample *sample)
-{
-	struct symbol *parent = NULL;
-	struct hist_entry *he;
-	int err = sample__resolve_callchain(sample, &parent, evsel, al, rep->max_stack);
-
-	if (err)
-		return err;
-
-	he = __hists__add_entry(&evsel->hists, al, parent, NULL, NULL,
-				sample->period, sample->weight,
-				sample->transaction);
-	if (he == NULL)
-		return -ENOMEM;
-
-	err = hist_entry__append_callchain(he, sample);
-	if (err)
-		goto out;
-
-	if (ui__has_annotation())
-		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
-
-	report__inc_stats(rep, he);
-
-out:
-	return err;
-}
-
-
 static int process_sample_event(struct perf_tool *tool,
 				union perf_event *event,
 				struct perf_sample *sample,
@@ -243,6 +84,9 @@ static int process_sample_event(struct perf_tool *tool,
 {
 	struct report *rep = container_of(tool, struct report, tool);
 	struct addr_location al;
+	struct hist_entry_iter iter = {
+		.hide_unresolved = rep->hide_unresolved,
+	};
 	int ret;
 
 	if (perf_event__preprocess_sample(event, machine, &al, sample) < 0) {
@@ -257,22 +101,22 @@ static int process_sample_event(struct perf_tool *tool,
 	if (rep->cpu_list && !test_bit(sample->cpu, rep->cpu_bitmap))
 		return 0;
 
-	if (sort__mode == SORT_MODE__BRANCH) {
-		ret = report__add_branch_hist_entry(rep, &al, sample, evsel);
-		if (ret < 0)
-			pr_debug("problem adding lbr entry, skipping event\n");
-	} else if (rep->mem_mode == 1) {
-		ret = report__add_mem_hist_entry(rep, &al, sample, evsel);
-		if (ret < 0)
-			pr_debug("problem adding mem entry, skipping event\n");
-	} else {
-		if (al.map != NULL)
-			al.map->dso->hit = 1;
-
-		ret = report__add_hist_entry(rep, evsel, &al, sample);
-		if (ret < 0)
-			pr_debug("problem incrementing symbol period, skipping event\n");
-	}
+	if (sort__mode == SORT_MODE__BRANCH)
+		iter.ops = &hist_iter_branch;
+	else if (rep->mem_mode)
+		iter.ops = &hist_iter_mem;
+	else
+		iter.ops = &hist_iter_normal;
+
+	if (al.map != NULL)
+		al.map->dso->hit = 1;
+
+	rep->nr_entries++;
+
+	ret = hist_entry_iter__add(&iter, &al, evsel, sample, rep->max_stack);
+	if (ret < 0)
+		pr_debug("problem adding hist entry, skipping event\n");
+
 	return ret;
 }
 
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index c5ba924a3581..76b02e1de701 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -42,11 +42,11 @@ static struct sample fake_samples[] = {
 	{ .pid = 300, .ip = 0xf0000 + 800, },
 };
 
-static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
+static int add_hist_entries(struct perf_evlist *evlist,
+			    struct machine *machine __maybe_unused)
 {
 	struct perf_evsel *evsel;
 	struct addr_location al;
-	struct hist_entry *he;
 	struct perf_sample sample = { .cpu = 0, };
 	size_t i;
 
@@ -62,6 +62,10 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 					.misc = PERF_RECORD_MISC_USER,
 				},
 			};
+			struct hist_entry_iter iter = {
+				.ops = &hist_iter_normal,
+				.hide_unresolved = false,
+			};
 
 			/* make sure it has no filter at first */
 			evsel->hists.thread_filter = NULL;
@@ -71,23 +75,19 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 			sample.pid = fake_samples[i].pid;
 			sample.tid = fake_samples[i].pid;
 			sample.ip = fake_samples[i].ip;
+			sample.period = 100;
 
 			if (perf_event__preprocess_sample(&event, machine, &al,
 							  &sample) < 0)
 				goto out;
 
-			he = __hists__add_entry(&evsel->hists, &al, NULL,
-						NULL, NULL, 100, 1, 0);
-			if (he == NULL)
+			if (hist_entry_iter__add(&iter, &al, evsel, &sample,
+						 PERF_MAX_STACK_DEPTH) < 0)
 				goto out;
 
 			fake_samples[i].thread = al.thread;
 			fake_samples[i].map = al.map;
 			fake_samples[i].sym = al.sym;
-
-			hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
-			if (!he->filtered)
-				he->hists->stats.nr_non_filtered_samples++;
 		}
 	}
 
diff --git a/tools/perf/tests/hists_output.c b/tools/perf/tests/hists_output.c
index a16850551797..1308f88a9169 100644
--- a/tools/perf/tests/hists_output.c
+++ b/tools/perf/tests/hists_output.c
@@ -46,7 +46,7 @@ static struct sample fake_samples[] = {
 static int add_hist_entries(struct hists *hists, struct machine *machine)
 {
 	struct addr_location al;
-	struct hist_entry *he;
+	struct perf_evsel *evsel = hists_to_evsel(hists);
 	struct perf_sample sample = { .period = 100, };
 	size_t i;
 
@@ -56,6 +56,10 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
 				.misc = PERF_RECORD_MISC_USER,
 			},
 		};
+		struct hist_entry_iter iter = {
+			.ops = &hist_iter_normal,
+			.hide_unresolved = false,
+		};
 
 		sample.cpu = fake_samples[i].cpu;
 		sample.pid = fake_samples[i].pid;
@@ -66,9 +70,8 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
 						  &sample) < 0)
 			goto out;
 
-		he = __hists__add_entry(hists, &al, NULL, NULL, NULL,
-					sample.period, 1, 0);
-		if (he == NULL)
+		if (hist_entry_iter__add(&iter, &al, evsel, &sample,
+					 PERF_MAX_STACK_DEPTH) < 0)
 			goto out;
 
 		fake_samples[i].thread = al.thread;
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index b262b44b7a65..61cd2569dc2a 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -4,6 +4,7 @@
 #include "session.h"
 #include "sort.h"
 #include "evsel.h"
+#include "annotate.h"
 #include <math.h>
 
 static bool hists__filter_entry_by_dso(struct hists *hists,
@@ -429,6 +430,308 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 	return add_hist_entry(hists, &entry, al);
 }
 
+static int
+iter_next_nop_entry(struct hist_entry_iter *iter __maybe_unused,
+		    struct addr_location *al __maybe_unused)
+{
+	return 0;
+}
+
+static int
+iter_add_next_nop_entry(struct hist_entry_iter *iter __maybe_unused,
+			struct addr_location *al __maybe_unused)
+{
+	return 0;
+}
+
+static int
+iter_prepare_mem_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct perf_sample *sample = iter->sample;
+	struct mem_info *mi;
+
+	mi = sample__resolve_mem(sample, al);
+	if (mi == NULL)
+		return -ENOMEM;
+
+	iter->priv = mi;
+	return 0;
+}
+
+static int
+iter_add_single_mem_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	u64 cost;
+	struct mem_info *mi = iter->priv;
+	struct hist_entry *he;
+
+	if (mi == NULL)
+		return -EINVAL;
+
+	cost = iter->sample->weight;
+	if (!cost)
+		cost = 1;
+
+	/*
+	 * must pass period=weight in order to get the correct
+	 * sorting from hists__collapse_resort() which is solely
+	 * based on periods. We want sorting be done on nr_events * weight
+	 * and this is indirectly achieved by passing period=weight here
+	 * and the he_stat__add_period() function.
+	 */
+	he = __hists__add_entry(&iter->evsel->hists, al, iter->parent, NULL, mi,
+				cost, cost, 0);
+	if (!he)
+		return -ENOMEM;
+
+	iter->he = he;
+	return 0;
+}
+
+static int
+iter_finish_mem_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct perf_evsel *evsel = iter->evsel;
+	struct hist_entry *he = iter->he;
+	struct mem_info *mx;
+	int err = -EINVAL;
+
+	if (he == NULL)
+		goto out;
+
+	if (ui__has_annotation()) {
+		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+		if (err)
+			goto out;
+
+		mx = he->mem_info;
+		err = addr_map_symbol__inc_samples(&mx->daddr, evsel->idx);
+		if (err)
+			goto out;
+	}
+
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		evsel->hists.stats.nr_non_filtered_samples++;
+
+	err = hist_entry__append_callchain(he, iter->sample);
+
+out:
+	/*
+	 * We don't need to free iter->priv (mem_info) here since
+	 * the mem info was either already freed in add_hist_entry() or
+	 * passed to a new hist entry by hist_entry__new().
+	 */
+	iter->priv = NULL;
+
+	iter->he = NULL;
+	return err;
+}
+
+static int
+iter_prepare_branch_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct branch_info *bi;
+	struct perf_sample *sample = iter->sample;
+
+	bi = sample__resolve_bstack(sample, al);
+	if (!bi)
+		return -ENOMEM;
+
+	iter->curr = 0;
+	iter->total = sample->branch_stack->nr;
+
+	iter->priv = bi;
+	return 0;
+}
+
+static int
+iter_add_single_branch_entry(struct hist_entry_iter *iter __maybe_unused,
+			     struct addr_location *al __maybe_unused)
+{
+	return 0;
+}
+
+static int
+iter_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct branch_info *bi = iter->priv;
+	int i = iter->curr;
+
+	if (bi == NULL)
+		return 0;
+
+	if (iter->curr >= iter->total)
+		return 0;
+
+	al->map = bi[i].to.map;
+	al->sym = bi[i].to.sym;
+	al->addr = bi[i].to.addr;
+	return 1;
+}
+
+static int
+iter_add_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct branch_info *bi, *bx;
+	struct perf_evsel *evsel = iter->evsel;
+	struct hist_entry *he;
+	int i = iter->curr;
+	int err = 0;
+
+	bi = iter->priv;
+
+	if (iter->hide_unresolved && !(bi[i].from.sym && bi[i].to.sym))
+		goto out;
+
+	/*
+	 * The report shows the percentage of total branches captured
+	 * and not events sampled. Thus we use a pseudo period of 1.
+	 */
+	he = __hists__add_entry(&evsel->hists, al, iter->parent, &bi[i], NULL,
+				1, 1, 0);
+	if (he == NULL)
+		return -ENOMEM;
+
+	if (ui__has_annotation()) {
+		bx = he->branch_info;
+		err = addr_map_symbol__inc_samples(&bx->from, evsel->idx);
+		if (err)
+			goto out;
+
+		err = addr_map_symbol__inc_samples(&bx->to, evsel->idx);
+		if (err)
+			goto out;
+	}
+
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		evsel->hists.stats.nr_non_filtered_samples++;
+
+out:
+	iter->curr++;
+	return err;
+}
+
+static int
+iter_finish_branch_entry(struct hist_entry_iter *iter,
+			 struct addr_location *al __maybe_unused)
+{
+	zfree(&iter->priv);
+
+	return iter->curr >= iter->total ? 0 : -1;
+}
+
+static int
+iter_prepare_normal_entry(struct hist_entry_iter *iter __maybe_unused,
+			  struct addr_location *al __maybe_unused)
+{
+	return 0;
+}
+
+static int
+iter_add_single_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	struct perf_evsel *evsel = iter->evsel;
+	struct perf_sample *sample = iter->sample;
+	struct hist_entry *he;
+
+	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
+				sample->period, sample->weight,
+				sample->transaction);
+	if (he == NULL)
+		return -ENOMEM;
+
+	iter->he = he;
+	return 0;
+}
+
+static int
+iter_finish_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
+{
+	int err;
+	struct hist_entry *he = iter->he;
+	struct perf_evsel *evsel = iter->evsel;
+	struct perf_sample *sample = iter->sample;
+
+	if (he == NULL)
+		return 0;
+
+	iter->he = NULL;
+
+	if (ui__has_annotation()) {
+		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+		if (err)
+			return err;
+	}
+
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		evsel->hists.stats.nr_non_filtered_samples++;
+
+	return hist_entry__append_callchain(he, sample);
+}
+
+const struct hist_iter_ops hist_iter_mem = {
+	.prepare_entry 		= iter_prepare_mem_entry,
+	.add_single_entry 	= iter_add_single_mem_entry,
+	.next_entry 		= iter_next_nop_entry,
+	.add_next_entry 	= iter_add_next_nop_entry,
+	.finish_entry 		= iter_finish_mem_entry,
+};
+
+const struct hist_iter_ops hist_iter_branch = {
+	.prepare_entry 		= iter_prepare_branch_entry,
+	.add_single_entry 	= iter_add_single_branch_entry,
+	.next_entry 		= iter_next_branch_entry,
+	.add_next_entry 	= iter_add_next_branch_entry,
+	.finish_entry 		= iter_finish_branch_entry,
+};
+
+const struct hist_iter_ops hist_iter_normal = {
+	.prepare_entry 		= iter_prepare_normal_entry,
+	.add_single_entry 	= iter_add_single_normal_entry,
+	.next_entry 		= iter_next_nop_entry,
+	.add_next_entry 	= iter_add_next_nop_entry,
+	.finish_entry 		= iter_finish_normal_entry,
+};
+
+int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
+			 struct perf_evsel *evsel, struct perf_sample *sample,
+			 int max_stack_depth)
+{
+	int err, err2;
+
+	err = sample__resolve_callchain(sample, &iter->parent, evsel, al,
+					max_stack_depth);
+	if (err)
+		return err;
+
+	iter->evsel = evsel;
+	iter->sample = sample;
+
+	err = iter->ops->prepare_entry(iter, al);
+	if (err)
+		goto out;
+
+	err = iter->ops->add_single_entry(iter, al);
+	if (err)
+		goto out;
+
+	while (iter->ops->next_entry(iter, al)) {
+		err = iter->ops->add_next_entry(iter, al);
+		if (err)
+			break;
+	}
+
+out:
+	err2 = iter->ops->finish_entry(iter, al);
+	if (!err)
+		err = err2;
+
+	return err;
+}
+
 int64_t
 hist_entry__cmp(struct hist_entry *left, struct hist_entry *right)
 {
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index a8418d19808d..05fc6e7497c6 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -96,12 +96,45 @@ struct hists {
 	u16			col_len[HISTC_NR_COLS];
 };
 
+struct hist_entry_iter;
+
+struct hist_iter_ops {
+	int (*prepare_entry)(struct hist_entry_iter *, struct addr_location *);
+	int (*add_single_entry)(struct hist_entry_iter *, struct addr_location *);
+	int (*next_entry)(struct hist_entry_iter *, struct addr_location *);
+	int (*add_next_entry)(struct hist_entry_iter *, struct addr_location *);
+	int (*finish_entry)(struct hist_entry_iter *, struct addr_location *);
+};
+
+struct hist_entry_iter {
+	int total;
+	int curr;
+
+	bool hide_unresolved;
+
+	struct perf_evsel *evsel;
+	struct perf_sample *sample;
+	struct hist_entry *he;
+	struct symbol *parent;
+	void *priv;
+
+	const struct hist_iter_ops *ops;
+};
+
+extern const struct hist_iter_ops hist_iter_normal;
+extern const struct hist_iter_ops hist_iter_branch;
+extern const struct hist_iter_ops hist_iter_mem;
+
 struct hist_entry *__hists__add_entry(struct hists *hists,
 				      struct addr_location *al,
 				      struct symbol *parent,
 				      struct branch_info *bi,
 				      struct mem_info *mi, u64 period,
 				      u64 weight, u64 transaction);
+int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
+			 struct perf_evsel *evsel, struct perf_sample *sample,
+			 int max_stack_depth);
+
 int64_t hist_entry__cmp(struct hist_entry *left, struct hist_entry *right);
 int64_t hist_entry__collapse(struct hist_entry *left, struct hist_entry *right);
 int hist_entry__transaction_len(void);
-- 
1.9.2

[PATCH 02/26] perf hists: Add support for accumulated stat of hist entry

[Namhyung Kim]

Maintain accumulated stat information in hist_entry->stat_acc if
symbol_conf.cumulate_callchain is set.  Fields in ->stat_acc have same
vaules initially, and will be updated as callchain is processed later.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/hist.c   | 28 ++++++++++++++++++++++++++--
 tools/perf/util/sort.h   |  1 +
 tools/perf/util/symbol.h |  1 +
 3 files changed, 28 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 61cd2569dc2a..f231255116b0 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -232,6 +232,8 @@ static bool hists__decay_entry(struct hists *hists, struct hist_entry *he)
 		return true;
 
 	he_stat__decay(&he->stat);
+	if (symbol_conf.cumulate_callchain)
+		he_stat__decay(he->stat_acc);
 
 	diff = prev_period - he->stat.period;
 
@@ -279,12 +281,26 @@ void hists__decay_entries(struct hists *hists, bool zap_user, bool zap_kernel)
 
 static struct hist_entry *hist_entry__new(struct hist_entry *template)
 {
-	size_t callchain_size = symbol_conf.use_callchain ? sizeof(struct callchain_root) : 0;
-	struct hist_entry *he = zalloc(sizeof(*he) + callchain_size);
+	size_t callchain_size = 0;
+	struct hist_entry *he;
+
+	if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain)
+		callchain_size = sizeof(struct callchain_root);
+
+	he = zalloc(sizeof(*he) + callchain_size);
 
 	if (he != NULL) {
 		*he = *template;
 
+		if (symbol_conf.cumulate_callchain) {
+			he->stat_acc = malloc(sizeof(he->stat));
+			if (he->stat_acc == NULL) {
+				free(he);
+				return NULL;
+			}
+			memcpy(he->stat_acc, &he->stat, sizeof(he->stat));
+		}
+
 		if (he->ms.map)
 			he->ms.map->referenced = true;
 
@@ -296,6 +312,7 @@ static struct hist_entry *hist_entry__new(struct hist_entry *template)
 			 */
 			he->branch_info = malloc(sizeof(*he->branch_info));
 			if (he->branch_info == NULL) {
+				free(he->stat_acc);
 				free(he);
 				return NULL;
 			}
@@ -359,6 +376,8 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 
 		if (!cmp) {
 			he_stat__add_period(&he->stat, period, weight);
+			if (symbol_conf.cumulate_callchain)
+				he_stat__add_period(he->stat_acc, period, weight);
 
 			/*
 			 * This mem info was allocated from sample__resolve_mem
@@ -394,6 +413,8 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 	rb_insert_color(&he->rb_node_in, hists->entries_in);
 out:
 	he_stat__add_cpumode_period(&he->stat, al->cpumode, period);
+	if (symbol_conf.cumulate_callchain)
+		he_stat__add_cpumode_period(he->stat_acc, al->cpumode, period);
 	return he;
 }
 
@@ -772,6 +793,7 @@ void hist_entry__free(struct hist_entry *he)
 {
 	zfree(&he->branch_info);
 	zfree(&he->mem_info);
+	zfree(&he->stat_acc);
 	free_srcline(he->srcline);
 	free(he);
 }
@@ -797,6 +819,8 @@ static bool hists__collapse_insert_entry(struct hists *hists __maybe_unused,
 
 		if (!cmp) {
 			he_stat__add_stat(&iter->stat, &he->stat);
+			if (symbol_conf.cumulate_callchain)
+				he_stat__add_stat(iter->stat_acc, he->stat_acc);
 
 			if (symbol_conf.use_callchain) {
 				callchain_cursor_reset(&callchain_cursor);
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index 5f38d925e92f..c9ffa031becd 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -82,6 +82,7 @@ struct hist_entry {
 		struct list_head head;
 	} pairs;
 	struct he_stat		stat;
+	struct he_stat		*stat_acc;
 	struct map_symbol	ms;
 	struct thread		*thread;
 	struct comm		*comm;
diff --git a/tools/perf/util/symbol.h b/tools/perf/util/symbol.h
index 33ede53fa6b9..615c752dd767 100644
--- a/tools/perf/util/symbol.h
+++ b/tools/perf/util/symbol.h
@@ -109,6 +109,7 @@ struct symbol_conf {
 			show_nr_samples,
 			show_total_period,
 			use_callchain,
+			cumulate_callchain,
 			exclude_other,
 			show_cpu_utilization,
 			initialized,
-- 
1.9.2

[PATCH 03/26] perf hists: Check if accumulated when adding a hist entry

[Namhyung Kim]

To support callchain accumulation, @entry should be recognized if it's
accumulated or not when add_hist_entry() called.  The period of an
accumulated entry should be added to ->stat_acc but not ->stat. Add
@sample_self arg for that.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-annotate.c |  3 ++-
 tools/perf/builtin-diff.c     |  2 +-
 tools/perf/builtin-top.c      |  2 +-
 tools/perf/tests/hists_link.c |  4 ++--
 tools/perf/util/hist.c        | 29 ++++++++++++++++++-----------
 tools/perf/util/hist.h        |  3 ++-
 6 files changed, 26 insertions(+), 17 deletions(-)

diff --git a/tools/perf/builtin-annotate.c b/tools/perf/builtin-annotate.c
index d30d2c2e2a7a..9590ec4ad484 100644
--- a/tools/perf/builtin-annotate.c
+++ b/tools/perf/builtin-annotate.c
@@ -65,7 +65,8 @@ static int perf_evsel__add_sample(struct perf_evsel *evsel,
 		return 0;
 	}
 
-	he = __hists__add_entry(&evsel->hists, al, NULL, NULL, NULL, 1, 1, 0);
+	he = __hists__add_entry(&evsel->hists, al, NULL, NULL, NULL, 1, 1, 0,
+				true);
 	if (he == NULL)
 		return -ENOMEM;
 
diff --git a/tools/perf/builtin-diff.c b/tools/perf/builtin-diff.c
index 8bff543acaab..9a5a035cb426 100644
--- a/tools/perf/builtin-diff.c
+++ b/tools/perf/builtin-diff.c
@@ -315,7 +315,7 @@ static int hists__add_entry(struct hists *hists,
 			    u64 weight, u64 transaction)
 {
 	if (__hists__add_entry(hists, al, NULL, NULL, NULL, period, weight,
-			       transaction) != NULL)
+			       transaction, true) != NULL)
 		return 0;
 	return -ENOMEM;
 }
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 5b389ce4cd15..e7a0f8bc2fa7 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -247,7 +247,7 @@ static struct hist_entry *perf_evsel__add_hist_entry(struct perf_evsel *evsel,
 	pthread_mutex_lock(&evsel->hists.lock);
 	he = __hists__add_entry(&evsel->hists, al, NULL, NULL, NULL,
 				sample->period, sample->weight,
-				sample->transaction);
+				sample->transaction, true);
 	pthread_mutex_unlock(&evsel->hists.lock);
 	if (he == NULL)
 		return NULL;
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index 5ffa2c3eb77d..ca6693b37cd7 100644
--- a/tools/perf/tests/hists_link.c
+++ b/tools/perf/tests/hists_link.c
@@ -88,7 +88,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 				goto out;
 
 			he = __hists__add_entry(&evsel->hists, &al, NULL,
-						NULL, NULL, 1, 1, 0);
+						NULL, NULL, 1, 1, 0, true);
 			if (he == NULL)
 				goto out;
 
@@ -112,7 +112,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 				goto out;
 
 			he = __hists__add_entry(&evsel->hists, &al, NULL,
-						NULL, NULL, 1, 1, 0);
+						NULL, NULL, 1, 1, 0, true);
 			if (he == NULL)
 				goto out;
 
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index f231255116b0..8a79cc7f7ffb 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -279,7 +279,8 @@ void hists__decay_entries(struct hists *hists, bool zap_user, bool zap_kernel)
  * histogram, sorted on item, collects periods
  */
 
-static struct hist_entry *hist_entry__new(struct hist_entry *template)
+static struct hist_entry *hist_entry__new(struct hist_entry *template,
+					  bool sample_self)
 {
 	size_t callchain_size = 0;
 	struct hist_entry *he;
@@ -299,6 +300,8 @@ static struct hist_entry *hist_entry__new(struct hist_entry *template)
 				return NULL;
 			}
 			memcpy(he->stat_acc, &he->stat, sizeof(he->stat));
+			if (!sample_self)
+				memset(&he->stat, 0, sizeof(he->stat));
 		}
 
 		if (he->ms.map)
@@ -351,7 +354,8 @@ static u8 symbol__parent_filter(const struct symbol *parent)
 
 static struct hist_entry *add_hist_entry(struct hists *hists,
 					 struct hist_entry *entry,
-					 struct addr_location *al)
+					 struct addr_location *al,
+					 bool sample_self)
 {
 	struct rb_node **p;
 	struct rb_node *parent = NULL;
@@ -375,7 +379,8 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 		cmp = hist_entry__cmp(he, entry);
 
 		if (!cmp) {
-			he_stat__add_period(&he->stat, period, weight);
+			if (sample_self)
+				he_stat__add_period(&he->stat, period, weight);
 			if (symbol_conf.cumulate_callchain)
 				he_stat__add_period(he->stat_acc, period, weight);
 
@@ -405,14 +410,15 @@ static struct hist_entry *add_hist_entry(struct hists *hists,
 			p = &(*p)->rb_right;
 	}
 
-	he = hist_entry__new(entry);
+	he = hist_entry__new(entry, sample_self);
 	if (!he)
 		return NULL;
 
 	rb_link_node(&he->rb_node_in, parent, p);
 	rb_insert_color(&he->rb_node_in, hists->entries_in);
 out:
-	he_stat__add_cpumode_period(&he->stat, al->cpumode, period);
+	if (sample_self)
+		he_stat__add_cpumode_period(&he->stat, al->cpumode, period);
 	if (symbol_conf.cumulate_callchain)
 		he_stat__add_cpumode_period(he->stat_acc, al->cpumode, period);
 	return he;
@@ -423,7 +429,8 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 				      struct symbol *sym_parent,
 				      struct branch_info *bi,
 				      struct mem_info *mi,
-				      u64 period, u64 weight, u64 transaction)
+				      u64 period, u64 weight, u64 transaction,
+				      bool sample_self)
 {
 	struct hist_entry entry = {
 		.thread	= al->thread,
@@ -448,7 +455,7 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 		.transaction = transaction,
 	};
 
-	return add_hist_entry(hists, &entry, al);
+	return add_hist_entry(hists, &entry, al, sample_self);
 }
 
 static int
@@ -501,7 +508,7 @@ iter_add_single_mem_entry(struct hist_entry_iter *iter, struct addr_location *al
 	 * and the he_stat__add_period() function.
 	 */
 	he = __hists__add_entry(&iter->evsel->hists, al, iter->parent, NULL, mi,
-				cost, cost, 0);
+				cost, cost, 0, true);
 	if (!he)
 		return -ENOMEM;
 
@@ -610,7 +617,7 @@ iter_add_next_branch_entry(struct hist_entry_iter *iter, struct addr_location *a
 	 * and not events sampled. Thus we use a pseudo period of 1.
 	 */
 	he = __hists__add_entry(&evsel->hists, al, iter->parent, &bi[i], NULL,
-				1, 1, 0);
+				1, 1, 0, true);
 	if (he == NULL)
 		return -ENOMEM;
 
@@ -659,7 +666,7 @@ iter_add_single_normal_entry(struct hist_entry_iter *iter, struct addr_location
 
 	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
 				sample->period, sample->weight,
-				sample->transaction);
+				sample->transaction, true);
 	if (he == NULL)
 		return -ENOMEM;
 
@@ -1158,7 +1165,7 @@ static struct hist_entry *hists__add_dummy_entry(struct hists *hists,
 			p = &(*p)->rb_right;
 	}
 
-	he = hist_entry__new(pair);
+	he = hist_entry__new(pair, true);
 	if (he) {
 		memset(&he->stat, 0, sizeof(he->stat));
 		he->hists = hists;
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index 05fc6e7497c6..e655e87a829e 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -130,7 +130,8 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 				      struct symbol *parent,
 				      struct branch_info *bi,
 				      struct mem_info *mi, u64 period,
-				      u64 weight, u64 transaction);
+				      u64 weight, u64 transaction,
+				      bool sample_self);
 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
 			 struct perf_evsel *evsel, struct perf_sample *sample,
 			 int max_stack_depth);
-- 
1.9.2

[PATCH 04/26] perf hists: Accumulate hist entry stat based on the callchain

[Namhyung Kim]

Call __hists__add_entry() for each callchain node to get an
accumulated stat for an entry.  Introduce new cumulative_iter ops to
process them properly.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-report.c |  2 +
 tools/perf/util/callchain.c |  3 +-
 tools/perf/util/hist.c      | 94 +++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/hist.h      |  1 +
 4 files changed, 99 insertions(+), 1 deletion(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index d224f5961faa..a783135388ba 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -105,6 +105,8 @@ static int process_sample_event(struct perf_tool *tool,
 		iter.ops = &hist_iter_branch;
 	else if (rep->mem_mode)
 		iter.ops = &hist_iter_mem;
+	else if (symbol_conf.cumulate_callchain)
+		iter.ops = &hist_iter_cumulative;
 	else
 		iter.ops = &hist_iter_normal;
 
diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
index 9a42382b3921..2af69c47b725 100644
--- a/tools/perf/util/callchain.c
+++ b/tools/perf/util/callchain.c
@@ -616,7 +616,8 @@ int sample__resolve_callchain(struct perf_sample *sample, struct symbol **parent
 	if (sample->callchain == NULL)
 		return 0;
 
-	if (symbol_conf.use_callchain || sort__has_parent) {
+	if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain ||
+	    sort__has_parent) {
 		return machine__resolve_callchain(al->machine, evsel, al->thread,
 						  sample, parent, al, max_stack);
 	}
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 8a79cc7f7ffb..ddd487fc6c6e 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -700,6 +700,92 @@ iter_finish_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
 	return hist_entry__append_callchain(he, sample);
 }
 
+static int
+iter_prepare_cumulative_entry(struct hist_entry_iter *iter __maybe_unused,
+			      struct addr_location *al __maybe_unused)
+{
+	callchain_cursor_commit(&callchain_cursor);
+	return 0;
+}
+
+static int
+iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
+				 struct addr_location *al)
+{
+	struct perf_evsel *evsel = iter->evsel;
+	struct perf_sample *sample = iter->sample;
+	struct hist_entry *he;
+
+	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
+				sample->period, sample->weight,
+				sample->transaction, true);
+	if (he == NULL)
+		return -ENOMEM;
+
+	iter->he = he;
+	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+}
+
+static int
+iter_next_cumulative_entry(struct hist_entry_iter *iter,
+			   struct addr_location *al)
+{
+	struct callchain_cursor_node *node;
+
+	node = callchain_cursor_current(&callchain_cursor);
+	if (node == NULL)
+		return 0;
+
+	al->map = node->map;
+	al->sym = node->sym;
+	if (node->map)
+		al->addr = node->map->map_ip(node->map, node->ip);
+	else
+		al->addr = node->ip;
+
+	if (iter->hide_unresolved && al->sym == NULL)
+		return 0;
+
+	callchain_cursor_advance(&callchain_cursor);
+	return 1;
+}
+
+static int
+iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
+			       struct addr_location *al)
+{
+	struct perf_evsel *evsel = iter->evsel;
+	struct perf_sample *sample = iter->sample;
+	struct hist_entry *he;
+
+	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
+				sample->period, sample->weight,
+				sample->transaction, false);
+	if (he == NULL)
+		return -ENOMEM;
+
+	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+}
+
+static int
+iter_finish_cumulative_entry(struct hist_entry_iter *iter,
+			     struct addr_location *al __maybe_unused)
+{
+	struct perf_evsel *evsel = iter->evsel;
+	struct hist_entry *he = iter->he;
+
+	if (he == NULL)
+		return 0;
+
+	iter->he = NULL;
+
+	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		evsel->hists.stats.nr_non_filtered_samples++;
+
+	return 0;
+}
+
 const struct hist_iter_ops hist_iter_mem = {
 	.prepare_entry 		= iter_prepare_mem_entry,
 	.add_single_entry 	= iter_add_single_mem_entry,
@@ -724,6 +810,14 @@ const struct hist_iter_ops hist_iter_normal = {
 	.finish_entry 		= iter_finish_normal_entry,
 };
 
+const struct hist_iter_ops hist_iter_cumulative = {
+	.prepare_entry 		= iter_prepare_cumulative_entry,
+	.add_single_entry 	= iter_add_single_cumulative_entry,
+	.next_entry 		= iter_next_cumulative_entry,
+	.add_next_entry 	= iter_add_next_cumulative_entry,
+	.finish_entry 		= iter_finish_cumulative_entry,
+};
+
 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
 			 struct perf_evsel *evsel, struct perf_sample *sample,
 			 int max_stack_depth)
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index e655e87a829e..6dfbbaef65cc 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -124,6 +124,7 @@ struct hist_entry_iter {
 extern const struct hist_iter_ops hist_iter_normal;
 extern const struct hist_iter_ops hist_iter_branch;
 extern const struct hist_iter_ops hist_iter_mem;
+extern const struct hist_iter_ops hist_iter_cumulative;
 
 struct hist_entry *__hists__add_entry(struct hists *hists,
 				      struct addr_location *al,
-- 
1.9.2

[PATCH 05/26] perf tools: Update cpumode for each cumulative entry

[Namhyung Kim]

The cpumode and level in struct addr_localtion was set for a sample
and but updated as cumulative callchains were added.  This led to have
non-matching symbol and cpumode in the output.

Update it accordingly based on the fact whether the map is a part of
the kernel or not.  This is a reverse of what thread__find_addr_map()
does.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/callchain.c | 42 ++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/callchain.h |  2 ++
 tools/perf/util/hist.c      | 13 ++-----------
 3 files changed, 46 insertions(+), 11 deletions(-)

diff --git a/tools/perf/util/callchain.c b/tools/perf/util/callchain.c
index 2af69c47b725..48b6d3f50012 100644
--- a/tools/perf/util/callchain.c
+++ b/tools/perf/util/callchain.c
@@ -630,3 +630,45 @@ int hist_entry__append_callchain(struct hist_entry *he, struct perf_sample *samp
 		return 0;
 	return callchain_append(he->callchain, &callchain_cursor, sample->period);
 }
+
+int fill_callchain_info(struct addr_location *al, struct callchain_cursor_node *node,
+			bool hide_unresolved)
+{
+	al->map = node->map;
+	al->sym = node->sym;
+	if (node->map)
+		al->addr = node->map->map_ip(node->map, node->ip);
+	else
+		al->addr = node->ip;
+
+	if (al->sym == NULL) {
+		if (hide_unresolved)
+			return 0;
+		if (al->map == NULL)
+			goto out;
+	}
+
+	if (al->map->groups == &al->machine->kmaps) {
+		if (machine__is_host(al->machine)) {
+			al->cpumode = PERF_RECORD_MISC_KERNEL;
+			al->level = 'k';
+		} else {
+			al->cpumode = PERF_RECORD_MISC_GUEST_KERNEL;
+			al->level = 'g';
+		}
+	} else {
+		if (machine__is_host(al->machine)) {
+			al->cpumode = PERF_RECORD_MISC_USER;
+			al->level = '.';
+		} else if (perf_guest) {
+			al->cpumode = PERF_RECORD_MISC_GUEST_USER;
+			al->level = 'u';
+		} else {
+			al->cpumode = PERF_RECORD_MISC_HYPERVISOR;
+			al->level = 'H';
+		}
+	}
+
+out:
+	return 1;
+}
diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h
index bde2b0cc24cf..24a53d562d0a 100644
--- a/tools/perf/util/callchain.h
+++ b/tools/perf/util/callchain.h
@@ -162,6 +162,8 @@ int sample__resolve_callchain(struct perf_sample *sample, struct symbol **parent
 			      struct perf_evsel *evsel, struct addr_location *al,
 			      int max_stack);
 int hist_entry__append_callchain(struct hist_entry *he, struct perf_sample *sample);
+int fill_callchain_info(struct addr_location *al, struct callchain_cursor_node *node,
+			bool hide_unresolved);
 
 extern const char record_callchain_help[];
 int parse_callchain_report_opt(const char *arg);
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index ddd487fc6c6e..802af8495f83 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -736,18 +736,9 @@ iter_next_cumulative_entry(struct hist_entry_iter *iter,
 	if (node == NULL)
 		return 0;
 
-	al->map = node->map;
-	al->sym = node->sym;
-	if (node->map)
-		al->addr = node->map->map_ip(node->map, node->ip);
-	else
-		al->addr = node->ip;
-
-	if (iter->hide_unresolved && al->sym == NULL)
-		return 0;
-
 	callchain_cursor_advance(&callchain_cursor);
-	return 1;
+
+	return fill_callchain_info(al, node, iter->hide_unresolved);
 }
 
 static int
-- 
1.9.2

[PATCH 06/26] perf report: Cache cumulative callchains

[Namhyung Kim]

It is possble that a callchain has cycles or recursive calls.  In that
case it'll end up having entries more than 100% overhead in the
output.  In order to prevent such entries, cache each callchain node
and skip if same entry already cumulated.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/hist.c | 44 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 44 insertions(+)

diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index 802af8495f83..b3faabf3d2c2 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -704,7 +704,22 @@ static int
 iter_prepare_cumulative_entry(struct hist_entry_iter *iter __maybe_unused,
 			      struct addr_location *al __maybe_unused)
 {
+	struct hist_entry **he_cache;
+
 	callchain_cursor_commit(&callchain_cursor);
+
+	/*
+	 * This is for detecting cycles or recursions so that they're
+	 * cumulated only one time to prevent entries more than 100%
+	 * overhead.
+	 */
+	he_cache = malloc(sizeof(*he_cache) * (PERF_MAX_STACK_DEPTH + 1));
+	if (he_cache == NULL)
+		return -ENOMEM;
+
+	iter->priv = he_cache;
+	iter->curr = 0;
+
 	return 0;
 }
 
@@ -714,6 +729,7 @@ iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
 {
 	struct perf_evsel *evsel = iter->evsel;
 	struct perf_sample *sample = iter->sample;
+	struct hist_entry **he_cache = iter->priv;
 	struct hist_entry *he;
 
 	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
@@ -723,6 +739,8 @@ iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
 		return -ENOMEM;
 
 	iter->he = he;
+	he_cache[iter->curr++] = he;
+
 	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
 }
 
@@ -747,7 +765,29 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 {
 	struct perf_evsel *evsel = iter->evsel;
 	struct perf_sample *sample = iter->sample;
+	struct hist_entry **he_cache = iter->priv;
 	struct hist_entry *he;
+	struct hist_entry he_tmp = {
+		.cpu = al->cpu,
+		.thread = al->thread,
+		.comm = thread__comm(al->thread),
+		.ip = al->addr,
+		.ms = {
+			.map = al->map,
+			.sym = al->sym,
+		},
+		.parent = iter->parent,
+	};
+	int i;
+
+	/*
+	 * Check if there's duplicate entries in the callchain.
+	 * It's possible that it has cycles or recursive calls.
+	 */
+	for (i = 0; i < iter->curr; i++) {
+		if (hist_entry__cmp(he_cache[i], &he_tmp) == 0)
+			return 0;
+	}
 
 	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
 				sample->period, sample->weight,
@@ -755,6 +795,8 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 	if (he == NULL)
 		return -ENOMEM;
 
+	he_cache[iter->curr++] = he;
+
 	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
 }
 
@@ -765,6 +807,8 @@ iter_finish_cumulative_entry(struct hist_entry_iter *iter,
 	struct perf_evsel *evsel = iter->evsel;
 	struct hist_entry *he = iter->he;
 
+	zfree(&iter->priv);
+
 	if (he == NULL)
 		return 0;
 
-- 
1.9.2

[PATCH 07/26] perf callchain: Add callchain_cursor_snapshot()

[Namhyung Kim]

The callchain_cursor_snapshot() is for saving current status of the
callchain.  It'll be used to accumulate callchain information for each node.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/callchain.h | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/tools/perf/util/callchain.h b/tools/perf/util/callchain.h
index 24a53d562d0a..8f84423a75da 100644
--- a/tools/perf/util/callchain.h
+++ b/tools/perf/util/callchain.h
@@ -167,4 +167,13 @@ int fill_callchain_info(struct addr_location *al, struct callchain_cursor_node *
 
 extern const char record_callchain_help[];
 int parse_callchain_report_opt(const char *arg);
+
+static inline void callchain_cursor_snapshot(struct callchain_cursor *dest,
+					     struct callchain_cursor *src)
+{
+	*dest = *src;
+
+	dest->first = src->curr;
+	dest->nr -= src->pos;
+}
 #endif	/* __PERF_CALLCHAIN_H */
-- 
1.9.2

[PATCH 08/26] perf tools: Save callchain info for each cumulative entry

[Namhyung Kim]

When accumulating callchain entry, also save current snapshot of the
chain so that it can show the rest of the chain.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/hist.c | 17 +++++++++++++++--
 1 file changed, 15 insertions(+), 2 deletions(-)

diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index b3faabf3d2c2..de13f0a12940 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -741,6 +741,14 @@ iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
 	iter->he = he;
 	he_cache[iter->curr++] = he;
 
+	callchain_append(he->callchain, &callchain_cursor, sample->period);
+
+	/*
+	 * We need to re-initialize the cursor since callchain_append()
+	 * advanced the cursor to the end.
+	 */
+	callchain_cursor_commit(&callchain_cursor);
+
 	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
 }
 
@@ -754,8 +762,6 @@ iter_next_cumulative_entry(struct hist_entry_iter *iter,
 	if (node == NULL)
 		return 0;
 
-	callchain_cursor_advance(&callchain_cursor);
-
 	return fill_callchain_info(al, node, iter->hide_unresolved);
 }
 
@@ -779,6 +785,11 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 		.parent = iter->parent,
 	};
 	int i;
+	struct callchain_cursor cursor;
+
+	callchain_cursor_snapshot(&cursor, &callchain_cursor);
+
+	callchain_cursor_advance(&callchain_cursor);
 
 	/*
 	 * Check if there's duplicate entries in the callchain.
@@ -797,6 +808,8 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 
 	he_cache[iter->curr++] = he;
 
+	callchain_append(he->callchain, &cursor, sample->period);
+
 	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
 }
 
-- 
1.9.2

[PATCH 09/26] perf ui/hist: Add support to accumulated hist stat

[Namhyung Kim]

Print accumulated stat of a hist entry if requested.

To do that, add new HPP_PERCENT_ACC_FNS macro and generate a
perf_hpp_fmt using it.  The __hpp__sort_acc() function sorts entries
by accumulated period value.  When accumulated periods of two entries
are same (i.e. single path callchain) put the caller above since
accumulation tends to put callers on higher position for obvious
reason.

Also add "overhead_children" output field to be selected by user.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/hist.c   | 99 ++++++++++++++++++++++++++++++++++++++++++++++++++
 tools/perf/util/hist.h |  4 ++
 tools/perf/util/sort.c |  1 +
 3 files changed, 104 insertions(+)

diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 4484f5bd1b14..0ce3e79b2ca7 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -104,6 +104,18 @@ int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
 	return ret;
 }
 
+int __hpp__fmt_acc(struct perf_hpp *hpp, struct hist_entry *he,
+		   hpp_field_fn get_field, const char *fmt,
+		   hpp_snprint_fn print_fn, bool fmt_percent)
+{
+	if (!symbol_conf.cumulate_callchain) {
+		return snprintf(hpp->buf, hpp->size, "%*s",
+				fmt_percent ? 8 : 12, "N/A");
+	}
+
+	return __hpp__fmt(hpp, he, get_field, fmt, print_fn, fmt_percent);
+}
+
 static int field_cmp(u64 field_a, u64 field_b)
 {
 	if (field_a > field_b)
@@ -160,6 +172,24 @@ out:
 	return ret;
 }
 
+static int __hpp__sort_acc(struct hist_entry *a, struct hist_entry *b,
+			   hpp_field_fn get_field)
+{
+	s64 ret = 0;
+
+	if (symbol_conf.cumulate_callchain) {
+		/*
+		 * Put caller above callee when they have equal period.
+		 */
+		ret = field_cmp(get_field(a), get_field(b));
+		if (ret)
+			return ret;
+
+		ret = b->callchain->max_depth - a->callchain->max_depth;
+	}
+	return ret;
+}
+
 #define __HPP_HEADER_FN(_type, _str, _min_width, _unit_width) 		\
 static int hpp__header_##_type(struct perf_hpp_fmt *fmt __maybe_unused,	\
 			       struct perf_hpp *hpp,			\
@@ -242,6 +272,34 @@ static int64_t hpp__sort_##_type(struct hist_entry *a, struct hist_entry *b)	\
 	return __hpp__sort(a, b, he_get_##_field);				\
 }
 
+#define __HPP_COLOR_ACC_PERCENT_FN(_type, _field)				\
+static u64 he_get_acc_##_field(struct hist_entry *he)				\
+{										\
+	return he->stat_acc->_field;						\
+}										\
+										\
+static int hpp__color_##_type(struct perf_hpp_fmt *fmt __maybe_unused,		\
+			      struct perf_hpp *hpp, struct hist_entry *he) 	\
+{										\
+	return __hpp__fmt_acc(hpp, he, he_get_acc_##_field, " %6.2f%%",		\
+			      hpp_color_scnprintf, true);			\
+}
+
+#define __HPP_ENTRY_ACC_PERCENT_FN(_type, _field)				\
+static int hpp__entry_##_type(struct perf_hpp_fmt *_fmt __maybe_unused,		\
+			      struct perf_hpp *hpp, struct hist_entry *he) 	\
+{										\
+	const char *fmt = symbol_conf.field_sep ? " %.2f" : " %6.2f%%";		\
+	return __hpp__fmt_acc(hpp, he, he_get_acc_##_field, fmt,		\
+			      hpp_entry_scnprintf, true);			\
+}
+
+#define __HPP_SORT_ACC_FN(_type, _field)					\
+static int64_t hpp__sort_##_type(struct hist_entry *a, struct hist_entry *b)	\
+{										\
+	return __hpp__sort_acc(a, b, he_get_acc_##_field);			\
+}
+
 #define __HPP_ENTRY_RAW_FN(_type, _field)					\
 static u64 he_get_raw_##_field(struct hist_entry *he)				\
 {										\
@@ -270,18 +328,27 @@ __HPP_COLOR_PERCENT_FN(_type, _field)					\
 __HPP_ENTRY_PERCENT_FN(_type, _field)					\
 __HPP_SORT_FN(_type, _field)
 
+#define HPP_PERCENT_ACC_FNS(_type, _str, _field, _min_width, _unit_width)\
+__HPP_HEADER_FN(_type, _str, _min_width, _unit_width)			\
+__HPP_WIDTH_FN(_type, _min_width, _unit_width)				\
+__HPP_COLOR_ACC_PERCENT_FN(_type, _field)				\
+__HPP_ENTRY_ACC_PERCENT_FN(_type, _field)				\
+__HPP_SORT_ACC_FN(_type, _field)
+
 #define HPP_RAW_FNS(_type, _str, _field, _min_width, _unit_width)	\
 __HPP_HEADER_FN(_type, _str, _min_width, _unit_width)			\
 __HPP_WIDTH_FN(_type, _min_width, _unit_width)				\
 __HPP_ENTRY_RAW_FN(_type, _field)					\
 __HPP_SORT_RAW_FN(_type, _field)
 
+__HPP_HEADER_FN(overhead_self, "Self", 8, 8)
 
 HPP_PERCENT_FNS(overhead, "Overhead", period, 8, 8)
 HPP_PERCENT_FNS(overhead_sys, "sys", period_sys, 8, 8)
 HPP_PERCENT_FNS(overhead_us, "usr", period_us, 8, 8)
 HPP_PERCENT_FNS(overhead_guest_sys, "guest sys", period_guest_sys, 9, 8)
 HPP_PERCENT_FNS(overhead_guest_us, "guest usr", period_guest_us, 9, 8)
+HPP_PERCENT_ACC_FNS(overhead_acc, "Children", period, 8, 8)
 
 HPP_RAW_FNS(samples, "Samples", nr_events, 12, 12)
 HPP_RAW_FNS(period, "Period", period, 12, 12)
@@ -303,6 +370,17 @@ static int64_t hpp__nop_cmp(struct hist_entry *a __maybe_unused,
 		.sort	= hpp__sort_ ## _name,		\
 	}
 
+#define HPP__COLOR_ACC_PRINT_FNS(_name)			\
+	{						\
+		.header	= hpp__header_ ## _name,	\
+		.width	= hpp__width_ ## _name,		\
+		.color	= hpp__color_ ## _name,		\
+		.entry	= hpp__entry_ ## _name,		\
+		.cmp	= hpp__nop_cmp,			\
+		.collapse = hpp__nop_cmp,		\
+		.sort	= hpp__sort_ ## _name,		\
+	}
+
 #define HPP__PRINT_FNS(_name)				\
 	{						\
 		.header	= hpp__header_ ## _name,	\
@@ -319,6 +397,7 @@ struct perf_hpp_fmt perf_hpp__format[] = {
 	HPP__COLOR_PRINT_FNS(overhead_us),
 	HPP__COLOR_PRINT_FNS(overhead_guest_sys),
 	HPP__COLOR_PRINT_FNS(overhead_guest_us),
+	HPP__COLOR_ACC_PRINT_FNS(overhead_acc),
 	HPP__PRINT_FNS(samples),
 	HPP__PRINT_FNS(period)
 };
@@ -328,16 +407,23 @@ LIST_HEAD(perf_hpp__sort_list);
 
 
 #undef HPP__COLOR_PRINT_FNS
+#undef HPP__COLOR_ACC_PRINT_FNS
 #undef HPP__PRINT_FNS
 
 #undef HPP_PERCENT_FNS
+#undef HPP_PERCENT_ACC_FNS
 #undef HPP_RAW_FNS
 
 #undef __HPP_HEADER_FN
 #undef __HPP_WIDTH_FN
 #undef __HPP_COLOR_PERCENT_FN
 #undef __HPP_ENTRY_PERCENT_FN
+#undef __HPP_COLOR_ACC_PERCENT_FN
+#undef __HPP_ENTRY_ACC_PERCENT_FN
 #undef __HPP_ENTRY_RAW_FN
+#undef __HPP_SORT_FN
+#undef __HPP_SORT_ACC_FN
+#undef __HPP_SORT_RAW_FN
 
 
 void perf_hpp__init(void)
@@ -361,6 +447,13 @@ void perf_hpp__init(void)
 	if (field_order)
 		return;
 
+	if (symbol_conf.cumulate_callchain) {
+		perf_hpp__column_enable(PERF_HPP__OVERHEAD_ACC);
+
+		perf_hpp__format[PERF_HPP__OVERHEAD].header =
+						hpp__header_overhead_self;
+	}
+
 	perf_hpp__column_enable(PERF_HPP__OVERHEAD);
 
 	if (symbol_conf.show_cpu_utilization) {
@@ -383,6 +476,12 @@ void perf_hpp__init(void)
 	list = &perf_hpp__format[PERF_HPP__OVERHEAD].sort_list;
 	if (list_empty(list))
 		list_add(list, &perf_hpp__sort_list);
+
+	if (symbol_conf.cumulate_callchain) {
+		list = &perf_hpp__format[PERF_HPP__OVERHEAD_ACC].sort_list;
+		if (list_empty(list))
+			list_add(list, &perf_hpp__sort_list);
+	}
 }
 
 void perf_hpp__column_register(struct perf_hpp_fmt *format)
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index 6dfbbaef65cc..ce37f3f8edd9 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -227,6 +227,7 @@ enum {
 	PERF_HPP__OVERHEAD_US,
 	PERF_HPP__OVERHEAD_GUEST_SYS,
 	PERF_HPP__OVERHEAD_GUEST_US,
+	PERF_HPP__OVERHEAD_ACC,
 	PERF_HPP__SAMPLES,
 	PERF_HPP__PERIOD,
 
@@ -253,6 +254,9 @@ typedef int (*hpp_snprint_fn)(struct perf_hpp *hpp, const char *fmt, ...);
 int __hpp__fmt(struct perf_hpp *hpp, struct hist_entry *he,
 	       hpp_field_fn get_field, const char *fmt,
 	       hpp_snprint_fn print_fn, bool fmt_percent);
+int __hpp__fmt_acc(struct perf_hpp *hpp, struct hist_entry *he,
+		   hpp_field_fn get_field, const char *fmt,
+		   hpp_snprint_fn print_fn, bool fmt_percent);
 
 static inline void advance_hpp(struct perf_hpp *hpp, int inc)
 {
diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index 901b9bece2ee..9da8931d2394 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -1061,6 +1061,7 @@ static struct hpp_dimension hpp_sort_dimensions[] = {
 	DIM(PERF_HPP__OVERHEAD_US, "overhead_us"),
 	DIM(PERF_HPP__OVERHEAD_GUEST_SYS, "overhead_guest_sys"),
 	DIM(PERF_HPP__OVERHEAD_GUEST_US, "overhead_guest_us"),
+	DIM(PERF_HPP__OVERHEAD_ACC, "overhead_children"),
 	DIM(PERF_HPP__SAMPLES, "sample"),
 	DIM(PERF_HPP__PERIOD, "period"),
 };
-- 
1.9.2

[PATCH 10/26] perf ui/browser: Add support to accumulated hist stat

[Namhyung Kim]

Print accumulated stat of a hist entry if requested.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/browsers/hists.c | 26 +++++++++++++++++++++++++-
 1 file changed, 25 insertions(+), 1 deletion(-)

diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
index 1c331b934ffc..1de7428a9771 100644
--- a/tools/perf/ui/browsers/hists.c
+++ b/tools/perf/ui/browsers/hists.c
@@ -632,7 +632,6 @@ static int __hpp__slsmg_color_printf(struct perf_hpp *hpp, const char *fmt, ...)
 	ret = scnprintf(hpp->buf, hpp->size, fmt, percent);
 	slsmg_printf("%s", hpp->buf);
 
-	advance_hpp(hpp, ret);
 	return ret;
 }
 
@@ -651,13 +650,36 @@ hist_browser__hpp_color_##_type(struct perf_hpp_fmt *fmt __maybe_unused,\
 			  __hpp__slsmg_color_printf, true);		\
 }
 
+#define __HPP_COLOR_ACC_PERCENT_FN(_type, _field)			\
+static u64 __hpp_get_acc_##_field(struct hist_entry *he)		\
+{									\
+	return he->stat_acc->_field;					\
+}									\
+									\
+static int								\
+hist_browser__hpp_color_##_type(struct perf_hpp_fmt *fmt __maybe_unused,\
+				struct perf_hpp *hpp,			\
+				struct hist_entry *he)			\
+{									\
+	if (!symbol_conf.cumulate_callchain) {				\
+		int ret = scnprintf(hpp->buf, hpp->size, "%8s", "N/A");	\
+		slsmg_printf("%s", hpp->buf);				\
+									\
+		return ret;						\
+	}								\
+	return __hpp__fmt(hpp, he, __hpp_get_acc_##_field, " %6.2f%%",	\
+			  __hpp__slsmg_color_printf, true);		\
+}
+
 __HPP_COLOR_PERCENT_FN(overhead, period)
 __HPP_COLOR_PERCENT_FN(overhead_sys, period_sys)
 __HPP_COLOR_PERCENT_FN(overhead_us, period_us)
 __HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys)
 __HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us)
+__HPP_COLOR_ACC_PERCENT_FN(overhead_acc, period)
 
 #undef __HPP_COLOR_PERCENT_FN
+#undef __HPP_COLOR_ACC_PERCENT_FN
 
 void hist_browser__init_hpp(void)
 {
@@ -671,6 +693,8 @@ void hist_browser__init_hpp(void)
 				hist_browser__hpp_color_overhead_guest_sys;
 	perf_hpp__format[PERF_HPP__OVERHEAD_GUEST_US].color =
 				hist_browser__hpp_color_overhead_guest_us;
+	perf_hpp__format[PERF_HPP__OVERHEAD_ACC].color =
+				hist_browser__hpp_color_overhead_acc;
 }
 
 static int hist_browser__show_entry(struct hist_browser *browser,
-- 
1.9.2

[PATCH 11/26] perf ui/gtk: Add support to accumulated hist stat

[Namhyung Kim]

Print accumulated stat of a hist entry if requested.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/gtk/hists.c | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/tools/perf/ui/gtk/hists.c b/tools/perf/ui/gtk/hists.c
index 9d90683914d4..7e5da4af98d8 100644
--- a/tools/perf/ui/gtk/hists.c
+++ b/tools/perf/ui/gtk/hists.c
@@ -47,11 +47,26 @@ static int perf_gtk__hpp_color_##_type(struct perf_hpp_fmt *fmt __maybe_unused,
 			  __percent_color_snprintf, true);			\
 }
 
+#define __HPP_COLOR_ACC_PERCENT_FN(_type, _field)				\
+static u64 he_get_acc_##_field(struct hist_entry *he)				\
+{										\
+	return he->stat_acc->_field;						\
+}										\
+										\
+static int perf_gtk__hpp_color_##_type(struct perf_hpp_fmt *fmt __maybe_unused,	\
+				       struct perf_hpp *hpp,			\
+				       struct hist_entry *he)			\
+{										\
+	return __hpp__fmt_acc(hpp, he, he_get_acc_##_field, " %6.2f%%",		\
+			      __percent_color_snprintf, true);			\
+}
+
 __HPP_COLOR_PERCENT_FN(overhead, period)
 __HPP_COLOR_PERCENT_FN(overhead_sys, period_sys)
 __HPP_COLOR_PERCENT_FN(overhead_us, period_us)
 __HPP_COLOR_PERCENT_FN(overhead_guest_sys, period_guest_sys)
 __HPP_COLOR_PERCENT_FN(overhead_guest_us, period_guest_us)
+__HPP_COLOR_ACC_PERCENT_FN(overhead_acc, period)
 
 #undef __HPP_COLOR_PERCENT_FN
 
@@ -68,6 +83,8 @@ void perf_gtk__init_hpp(void)
 				perf_gtk__hpp_color_overhead_guest_sys;
 	perf_hpp__format[PERF_HPP__OVERHEAD_GUEST_US].color =
 				perf_gtk__hpp_color_overhead_guest_us;
+	perf_hpp__format[PERF_HPP__OVERHEAD_ACC].color =
+				perf_gtk__hpp_color_overhead_acc;
 }
 
 static void callchain_list__sym_name(struct callchain_list *cl,
-- 
1.9.2

[PATCH 12/26] perf tools: Apply percent-limit to cumulative percentage

[Namhyung Kim]

If -g cumulative option is given, it needs to show entries which don't
have self overhead.  So apply percent-limit to accumulated overhead
percentage in this case.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/browsers/hists.c | 40 +++++++++++-----------------------------
 tools/perf/ui/gtk/hists.c      |  6 ++----
 tools/perf/ui/stdio/hist.c     |  4 ++--
 tools/perf/util/hist.h         |  1 +
 tools/perf/util/sort.h         | 17 ++++++++++++++++-
 5 files changed, 32 insertions(+), 36 deletions(-)

diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
index 1de7428a9771..efb160ac304d 100644
--- a/tools/perf/ui/browsers/hists.c
+++ b/tools/perf/ui/browsers/hists.c
@@ -37,7 +37,6 @@ static int hists__browser_title(struct hists *hists, char *bf, size_t size,
 static void hist_browser__update_nr_entries(struct hist_browser *hb);
 
 static struct rb_node *hists__filter_entries(struct rb_node *nd,
-					     struct hists *hists,
 					     float min_pcnt);
 
 static bool hist_browser__has_filter(struct hist_browser *hb)
@@ -319,7 +318,7 @@ __hist_browser__set_folding(struct hist_browser *browser, bool unfold)
 	struct hists *hists = browser->hists;
 
 	for (nd = rb_first(&hists->entries);
-	     (nd = hists__filter_entries(nd, hists, browser->min_pcnt)) != NULL;
+	     (nd = hists__filter_entries(nd, browser->min_pcnt)) != NULL;
 	     nd = rb_next(nd)) {
 		struct hist_entry *he = rb_entry(nd, struct hist_entry, rb_node);
 		hist_entry__set_folding(he, unfold);
@@ -807,15 +806,12 @@ static unsigned int hist_browser__refresh(struct ui_browser *browser)
 
 	for (nd = browser->top; nd; nd = rb_next(nd)) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
-		u64 total = hists__total_period(h->hists);
-		float percent = 0.0;
+		float percent;
 
 		if (h->filtered)
 			continue;
 
-		if (total)
-			percent = h->stat.period * 100.0 / total;
-
+		percent = hist_entry__get_percent_limit(h);
 		if (percent < hb->min_pcnt)
 			continue;
 
@@ -828,16 +824,11 @@ static unsigned int hist_browser__refresh(struct ui_browser *browser)
 }
 
 static struct rb_node *hists__filter_entries(struct rb_node *nd,
-					     struct hists *hists,
 					     float min_pcnt)
 {
 	while (nd != NULL) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
-		u64 total = hists__total_period(hists);
-		float percent = 0.0;
-
-		if (total)
-			percent = h->stat.period * 100.0 / total;
+		float percent = hist_entry__get_percent_limit(h);
 
 		if (!h->filtered && percent >= min_pcnt)
 			return nd;
@@ -849,16 +840,11 @@ static struct rb_node *hists__filter_entries(struct rb_node *nd,
 }
 
 static struct rb_node *hists__filter_prev_entries(struct rb_node *nd,
-						  struct hists *hists,
 						  float min_pcnt)
 {
 	while (nd != NULL) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
-		u64 total = hists__total_period(hists);
-		float percent = 0.0;
-
-		if (total)
-			percent = h->stat.period * 100.0 / total;
+		float percent = hist_entry__get_percent_limit(h);
 
 		if (!h->filtered && percent >= min_pcnt)
 			return nd;
@@ -887,14 +873,14 @@ static void ui_browser__hists_seek(struct ui_browser *browser,
 	switch (whence) {
 	case SEEK_SET:
 		nd = hists__filter_entries(rb_first(browser->entries),
-					   hb->hists, hb->min_pcnt);
+					   hb->min_pcnt);
 		break;
 	case SEEK_CUR:
 		nd = browser->top;
 		goto do_offset;
 	case SEEK_END:
 		nd = hists__filter_prev_entries(rb_last(browser->entries),
-						hb->hists, hb->min_pcnt);
+						hb->min_pcnt);
 		first = false;
 		break;
 	default:
@@ -937,8 +923,7 @@ do_offset:
 					break;
 				}
 			}
-			nd = hists__filter_entries(rb_next(nd), hb->hists,
-						   hb->min_pcnt);
+			nd = hists__filter_entries(rb_next(nd), hb->min_pcnt);
 			if (nd == NULL)
 				break;
 			--offset;
@@ -971,7 +956,7 @@ do_offset:
 				}
 			}
 
-			nd = hists__filter_prev_entries(rb_prev(nd), hb->hists,
+			nd = hists__filter_prev_entries(rb_prev(nd),
 							hb->min_pcnt);
 			if (nd == NULL)
 				break;
@@ -1150,7 +1135,6 @@ static int hist_browser__fprintf_entry(struct hist_browser *browser,
 static int hist_browser__fprintf(struct hist_browser *browser, FILE *fp)
 {
 	struct rb_node *nd = hists__filter_entries(rb_first(browser->b.entries),
-						   browser->hists,
 						   browser->min_pcnt);
 	int printed = 0;
 
@@ -1158,8 +1142,7 @@ static int hist_browser__fprintf(struct hist_browser *browser, FILE *fp)
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
 
 		printed += hist_browser__fprintf_entry(browser, h, fp);
-		nd = hists__filter_entries(rb_next(nd), browser->hists,
-					   browser->min_pcnt);
+		nd = hists__filter_entries(rb_next(nd), browser->min_pcnt);
 	}
 
 	return printed;
@@ -1396,8 +1379,7 @@ static void hist_browser__update_nr_entries(struct hist_browser *hb)
 		return;
 	}
 
-	while ((nd = hists__filter_entries(nd, hb->hists,
-					   hb->min_pcnt)) != NULL) {
+	while ((nd = hists__filter_entries(nd, hb->min_pcnt)) != NULL) {
 		nr_entries++;
 		nd = rb_next(nd);
 	}
diff --git a/tools/perf/ui/gtk/hists.c b/tools/perf/ui/gtk/hists.c
index 7e5da4af98d8..03d6812d25dd 100644
--- a/tools/perf/ui/gtk/hists.c
+++ b/tools/perf/ui/gtk/hists.c
@@ -226,14 +226,12 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists,
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
 		GtkTreeIter iter;
 		u64 total = hists__total_period(h->hists);
-		float percent = 0.0;
+		float percent;
 
 		if (h->filtered)
 			continue;
 
-		if (total)
-			percent = h->stat.period * 100.0 / total;
-
+		percent = hist_entry__get_percent_limit(h);
 		if (percent < min_pcnt)
 			continue;
 
diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c
index 9f57991025a9..475d2f5c7e16 100644
--- a/tools/perf/ui/stdio/hist.c
+++ b/tools/perf/ui/stdio/hist.c
@@ -461,12 +461,12 @@ print_entries:
 
 	for (nd = rb_first(&hists->entries); nd; nd = rb_next(nd)) {
 		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
-		float percent = h->stat.period * 100.0 /
-					hists->stats.total_period;
+		float percent;
 
 		if (h->filtered)
 			continue;
 
+		percent = hist_entry__get_percent_limit(h);
 		if (percent < min_pcnt)
 			continue;
 
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index ce37f3f8edd9..ad955fbd194f 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -144,6 +144,7 @@ int hist_entry__sort_snprintf(struct hist_entry *he, char *bf, size_t size,
 			      struct hists *hists);
 void hist_entry__free(struct hist_entry *);
 
+
 void hists__output_resort(struct hists *hists);
 void hists__collapse_resort(struct hists *hists, struct ui_progress *prog);
 
diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
index c9ffa031becd..426b873e16ff 100644
--- a/tools/perf/util/sort.h
+++ b/tools/perf/util/sort.h
@@ -20,7 +20,7 @@
 
 #include "parse-options.h"
 #include "parse-events.h"
-
+#include "hist.h"
 #include "thread.h"
 
 extern regex_t parent_regex;
@@ -131,6 +131,21 @@ static inline void hist_entry__add_pair(struct hist_entry *pair,
 	list_add_tail(&pair->pairs.node, &he->pairs.head);
 }
 
+static inline float hist_entry__get_percent_limit(struct hist_entry *he)
+{
+	u64 period = he->stat.period;
+	u64 total_period = hists__total_period(he->hists);
+
+	if (unlikely(total_period == 0))
+		return 0;
+
+	if (symbol_conf.cumulate_callchain)
+		period = he->stat_acc->period;
+
+	return period * 100.0 / total_period;
+}
+
+
 enum sort_mode {
 	SORT_MODE__NORMAL,
 	SORT_MODE__BRANCH,
-- 
1.9.2

[PATCH 13/26] perf tools: Add more hpp helper functions

[Namhyung Kim]

Sometimes it needs to disable some columns at runtime.  Add help
functions to support that.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/hist.c   | 17 +++++++++++++++++
 tools/perf/util/hist.h |  4 ++++
 2 files changed, 21 insertions(+)

diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 0ce3e79b2ca7..8ca638754acc 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -489,6 +489,11 @@ void perf_hpp__column_register(struct perf_hpp_fmt *format)
 	list_add_tail(&format->list, &perf_hpp__list);
 }
 
+void perf_hpp__column_unregister(struct perf_hpp_fmt *format)
+{
+	list_del(&format->list);
+}
+
 void perf_hpp__register_sort_field(struct perf_hpp_fmt *format)
 {
 	list_add_tail(&format->sort_list, &perf_hpp__sort_list);
@@ -500,6 +505,18 @@ void perf_hpp__column_enable(unsigned col)
 	perf_hpp__column_register(&perf_hpp__format[col]);
 }
 
+void perf_hpp__column_disable(unsigned col)
+{
+	BUG_ON(col >= PERF_HPP__MAX_INDEX);
+	perf_hpp__column_unregister(&perf_hpp__format[col]);
+}
+
+void perf_hpp__cancel_cumulate(void)
+{
+	perf_hpp__column_disable(PERF_HPP__OVERHEAD_ACC);
+	perf_hpp__format[PERF_HPP__OVERHEAD].header = hpp__header_overhead;
+}
+
 void perf_hpp__setup_output_field(void)
 {
 	struct perf_hpp_fmt *fmt;
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index ad955fbd194f..e8efb43b2355 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -237,7 +237,11 @@ enum {
 
 void perf_hpp__init(void);
 void perf_hpp__column_register(struct perf_hpp_fmt *format);
+void perf_hpp__column_unregister(struct perf_hpp_fmt *format);
 void perf_hpp__column_enable(unsigned col);
+void perf_hpp__column_disable(unsigned col);
+void perf_hpp__cancel_cumulate(void);
+
 void perf_hpp__register_sort_field(struct perf_hpp_fmt *format);
 void perf_hpp__setup_output_field(void);
 void perf_hpp__reset_output_field(void);
-- 
1.9.2

[PATCH 14/26] perf report: Add --children option

[Namhyung Kim]

The --children option is for showing accumulated overhead (period)
value as well as self overhead.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/Documentation/perf-report.txt |  7 ++++++-
 tools/perf/builtin-report.c              | 10 ++++++++++
 2 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
index a1b5185402d5..cefdf430d1b4 100644
--- a/tools/perf/Documentation/perf-report.txt
+++ b/tools/perf/Documentation/perf-report.txt
@@ -111,7 +111,7 @@ OPTIONS
 --fields=::
 	Specify output field - multiple keys can be specified in CSV format.
 	Following fields are available:
-	overhead, overhead_sys, overhead_us, sample and period.
+	overhead, overhead_sys, overhead_us, overhead_children, sample and period.
 	Also it can contain any sort key(s).
 
 	By default, every sort keys not specified in -F will be appended
@@ -163,6 +163,11 @@ OPTIONS
 
 	Default: fractal,0.5,callee,function.
 
+--children::
+	Accumulate callchain of children to parent entry so that then can
+	show up in the output.  The output will have a new "Children" column
+	and will be sorted on the data.  It requires callchains are recorded.
+
 --max-stack::
 	Set the stack depth limit when parsing the callchain, anything
 	beyond the specified depth will be ignored. This is a trade-off
diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index a783135388ba..6a9f0b38eb38 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -175,6 +175,14 @@ static int report__setup_sample_type(struct report *rep)
 			}
 	}
 
+	if (symbol_conf.cumulate_callchain) {
+		/* Silently ignore if callchain is missing */
+		if (!(sample_type & PERF_SAMPLE_CALLCHAIN)) {
+			symbol_conf.cumulate_callchain = false;
+			perf_hpp__cancel_cumulate();
+		}
+	}
+
 	if (sort__mode == SORT_MODE__BRANCH) {
 		if (!is_pipe &&
 		    !(sample_type & PERF_SAMPLE_BRANCH_STACK)) {
@@ -558,6 +566,8 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_CALLBACK_DEFAULT('g', "call-graph", &report, "output_type,min_percent[,print_limit],call_order",
 		     "Display callchains using output_type (graph, flat, fractal, or none) , min percent threshold, optional print limit, callchain order, key (function or address). "
 		     "Default: fractal,0.5,callee,function", &report_parse_callchain_opt, callchain_default_opt),
+	OPT_BOOLEAN(0, "children", &symbol_conf.cumulate_callchain,
+		    "Accumulate callchains of children and show total overhead as well"),
 	OPT_INTEGER(0, "max-stack", &report.max_stack,
 		    "Set the maximum stack depth when parsing the callchain, "
 		    "anything beyond the specified depth will be ignored. "
-- 
1.9.2

[PATCH 15/26] perf report: Add report.children config option

[Namhyung Kim]

Add report.children config option for setting default value of
callchain accumulation.  It affects the report output only if
perf.data contains callchain info.

A user can write .perfconfig file like below to enable accumulation
by default:

  $ cat ~/.perfconfig
  [report]
  children = true

And it can be disabled through command line:

  $ perf report --no-children

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-report.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index 6a9f0b38eb38..df7d431c0cbf 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -72,6 +72,10 @@ static int report__config(const char *var, const char *value, void *cb)
 		rep->min_percent = strtof(value, NULL);
 		return 0;
 	}
+	if (!strcmp(var, "report.children")) {
+		symbol_conf.cumulate_callchain = perf_config_bool(var, value);
+		return 0;
+	}
 
 	return perf_default_config(var, value, cb);
 }
-- 
1.9.2

[PATCH 16/26] perf tools: Do not auto-remove Children column if --fields given

[Namhyung Kim]

Depending on the configuration perf inserts/removes the Children
column in the output automatically.  But it might not be what user
wants if [s]he give --fields option explicitly.

Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/hist.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tools/perf/ui/hist.c b/tools/perf/ui/hist.c
index 8ca638754acc..498adb23c02e 100644
--- a/tools/perf/ui/hist.c
+++ b/tools/perf/ui/hist.c
@@ -513,6 +513,9 @@ void perf_hpp__column_disable(unsigned col)
 
 void perf_hpp__cancel_cumulate(void)
 {
+	if (field_order)
+		return;
+
 	perf_hpp__column_disable(PERF_HPP__OVERHEAD_ACC);
 	perf_hpp__format[PERF_HPP__OVERHEAD].header = hpp__header_overhead;
 }
-- 
1.9.2

[PATCH 17/26] perf tools: Add callback function to hist_entry_iter

[Namhyung Kim]

The new ->add_entry_cb() will be called after an entry was added to
the histogram.  It's used for code sharing between perf report and
perf top.  Note that ops->add_*_entry() should set iter->he properly
in order to call the ->add_entry_cb.

Also pass @arg to the callback function.  It'll be used by perf top
later.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-report.c     | 54 ++++++++++++++++++++++++++++++++++++++---
 tools/perf/tests/hists_filter.c | 22 ++++++++++++++++-
 tools/perf/tests/hists_output.c |  2 +-
 tools/perf/util/hist.c          | 50 +++++++++++++++++---------------------
 tools/perf/util/hist.h          |  5 +++-
 5 files changed, 98 insertions(+), 35 deletions(-)

diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
index df7d431c0cbf..c9d780d48aba 100644
--- a/tools/perf/builtin-report.c
+++ b/tools/perf/builtin-report.c
@@ -80,6 +80,47 @@ static int report__config(const char *var, const char *value, void *cb)
 	return perf_default_config(var, value, cb);
 }
 
+static void report__inc_stats(struct report *rep, struct hist_entry *he)
+{
+	/*
+	 * The @he is either of a newly created one or an existing one
+	 * merging current sample.  We only want to count a new one so
+	 * checking ->nr_events being 1.
+	 */
+	if (he->stat.nr_events == 1)
+		rep->nr_entries++;
+
+	/*
+	 * Only counts number of samples at this stage as it's more
+	 * natural to do it here and non-sample events are also
+	 * counted in perf_session_deliver_event().  The dump_trace
+	 * requires this info is ready before going to the output tree.
+	 */
+	hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		he->hists->stats.nr_non_filtered_samples++;
+}
+
+static int hist_iter_cb(struct hist_entry_iter *iter,
+			struct addr_location *al, bool single, void *arg)
+{
+	int err = 0;
+	struct hist_entry *he = iter->he;
+	struct perf_evsel *evsel = iter->evsel;
+	struct report *rep = arg;
+
+	if (ui__has_annotation())
+		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+
+	if (!single)
+		goto out;
+
+	report__inc_stats(rep, he);
+
+out:
+	return err;
+}
+
 static int process_sample_event(struct perf_tool *tool,
 				union perf_event *event,
 				struct perf_sample *sample,
@@ -109,17 +150,22 @@ static int process_sample_event(struct perf_tool *tool,
 		iter.ops = &hist_iter_branch;
 	else if (rep->mem_mode)
 		iter.ops = &hist_iter_mem;
-	else if (symbol_conf.cumulate_callchain)
+	else if (symbol_conf.cumulate_callchain) {
 		iter.ops = &hist_iter_cumulative;
-	else
+		iter.add_entry_cb = hist_iter_cb;
+	} else {
 		iter.ops = &hist_iter_normal;
+		iter.add_entry_cb = hist_iter_cb;
+	}
 
 	if (al.map != NULL)
 		al.map->dso->hit = 1;
 
-	rep->nr_entries++;
+	if (iter.add_entry_cb == NULL)
+		rep->nr_entries++;
 
-	ret = hist_entry_iter__add(&iter, &al, evsel, sample, rep->max_stack);
+	ret = hist_entry_iter__add(&iter, &al, evsel, sample, rep->max_stack,
+				   rep);
 	if (ret < 0)
 		pr_debug("problem adding hist entry, skipping event\n");
 
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index 76b02e1de701..7063824c3566 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -42,6 +42,25 @@ static struct sample fake_samples[] = {
 	{ .pid = 300, .ip = 0xf0000 + 800, },
 };
 
+static int count_sample(struct hist_entry_iter *iter,
+			struct addr_location *al __maybe_unused, bool single,
+			void *arg __maybe_unused)
+{
+	struct hist_entry *he = iter->he;
+
+	if (he == NULL)
+		return -EINVAL;
+
+	if (!single)
+		return 0;
+
+	hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
+	if (!he->filtered)
+		he->hists->stats.nr_non_filtered_samples++;
+
+	return 0;
+}
+
 static int add_hist_entries(struct perf_evlist *evlist,
 			    struct machine *machine __maybe_unused)
 {
@@ -65,6 +84,7 @@ static int add_hist_entries(struct perf_evlist *evlist,
 			struct hist_entry_iter iter = {
 				.ops = &hist_iter_normal,
 				.hide_unresolved = false,
+				.add_entry_cb = count_sample,
 			};
 
 			/* make sure it has no filter at first */
@@ -82,7 +102,7 @@ static int add_hist_entries(struct perf_evlist *evlist,
 				goto out;
 
 			if (hist_entry_iter__add(&iter, &al, evsel, &sample,
-						 PERF_MAX_STACK_DEPTH) < 0)
+						 PERF_MAX_STACK_DEPTH, NULL) < 0)
 				goto out;
 
 			fake_samples[i].thread = al.thread;
diff --git a/tools/perf/tests/hists_output.c b/tools/perf/tests/hists_output.c
index 1308f88a9169..d40461ecd210 100644
--- a/tools/perf/tests/hists_output.c
+++ b/tools/perf/tests/hists_output.c
@@ -71,7 +71,7 @@ static int add_hist_entries(struct hists *hists, struct machine *machine)
 			goto out;
 
 		if (hist_entry_iter__add(&iter, &al, evsel, &sample,
-					 PERF_MAX_STACK_DEPTH) < 0)
+					 PERF_MAX_STACK_DEPTH, NULL) < 0)
 			goto out;
 
 		fake_samples[i].thread = al.thread;
diff --git a/tools/perf/util/hist.c b/tools/perf/util/hist.c
index de13f0a12940..0732a376afb9 100644
--- a/tools/perf/util/hist.c
+++ b/tools/perf/util/hist.c
@@ -675,11 +675,10 @@ iter_add_single_normal_entry(struct hist_entry_iter *iter, struct addr_location
 }
 
 static int
-iter_finish_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
+iter_finish_normal_entry(struct hist_entry_iter *iter,
+			 struct addr_location *al __maybe_unused)
 {
-	int err;
 	struct hist_entry *he = iter->he;
-	struct perf_evsel *evsel = iter->evsel;
 	struct perf_sample *sample = iter->sample;
 
 	if (he == NULL)
@@ -687,16 +686,6 @@ iter_finish_normal_entry(struct hist_entry_iter *iter, struct addr_location *al)
 
 	iter->he = NULL;
 
-	if (ui__has_annotation()) {
-		err = hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
-		if (err)
-			return err;
-	}
-
-	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
-	if (!he->filtered)
-		evsel->hists.stats.nr_non_filtered_samples++;
-
 	return hist_entry__append_callchain(he, sample);
 }
 
@@ -749,7 +738,7 @@ iter_add_single_cumulative_entry(struct hist_entry_iter *iter,
 	 */
 	callchain_cursor_commit(&callchain_cursor);
 
-	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+	return 0;
 }
 
 static int
@@ -796,8 +785,11 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 	 * It's possible that it has cycles or recursive calls.
 	 */
 	for (i = 0; i < iter->curr; i++) {
-		if (hist_entry__cmp(he_cache[i], &he_tmp) == 0)
+		if (hist_entry__cmp(he_cache[i], &he_tmp) == 0) {
+			/* to avoid calling callback function */
+			iter->he = NULL;
 			return 0;
+		}
 	}
 
 	he = __hists__add_entry(&evsel->hists, al, iter->parent, NULL, NULL,
@@ -806,31 +798,21 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 	if (he == NULL)
 		return -ENOMEM;
 
+	iter->he = he;
 	he_cache[iter->curr++] = he;
 
 	callchain_append(he->callchain, &cursor, sample->period);
 
-	return hist_entry__inc_addr_samples(he, evsel->idx, al->addr);
+	return 0;
 }
 
 static int
 iter_finish_cumulative_entry(struct hist_entry_iter *iter,
 			     struct addr_location *al __maybe_unused)
 {
-	struct perf_evsel *evsel = iter->evsel;
-	struct hist_entry *he = iter->he;
-
 	zfree(&iter->priv);
-
-	if (he == NULL)
-		return 0;
-
 	iter->he = NULL;
 
-	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
-	if (!he->filtered)
-		evsel->hists.stats.nr_non_filtered_samples++;
-
 	return 0;
 }
 
@@ -868,7 +850,7 @@ const struct hist_iter_ops hist_iter_cumulative = {
 
 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
 			 struct perf_evsel *evsel, struct perf_sample *sample,
-			 int max_stack_depth)
+			 int max_stack_depth, void *arg)
 {
 	int err, err2;
 
@@ -888,10 +870,22 @@ int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
 	if (err)
 		goto out;
 
+	if (iter->he && iter->add_entry_cb) {
+		err = iter->add_entry_cb(iter, al, true, arg);
+		if (err)
+			goto out;
+	}
+
 	while (iter->ops->next_entry(iter, al)) {
 		err = iter->ops->add_next_entry(iter, al);
 		if (err)
 			break;
+
+		if (iter->he && iter->add_entry_cb) {
+			err = iter->add_entry_cb(iter, al, false, arg);
+			if (err)
+				goto out;
+		}
 	}
 
 out:
diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
index e8efb43b2355..b8a81adc63b7 100644
--- a/tools/perf/util/hist.h
+++ b/tools/perf/util/hist.h
@@ -119,6 +119,9 @@ struct hist_entry_iter {
 	void *priv;
 
 	const struct hist_iter_ops *ops;
+	/* user-defined callback function (optional) */
+	int (*add_entry_cb)(struct hist_entry_iter *iter,
+			    struct addr_location *al, bool single, void *arg);
 };
 
 extern const struct hist_iter_ops hist_iter_normal;
@@ -135,7 +138,7 @@ struct hist_entry *__hists__add_entry(struct hists *hists,
 				      bool sample_self);
 int hist_entry_iter__add(struct hist_entry_iter *iter, struct addr_location *al,
 			 struct perf_evsel *evsel, struct perf_sample *sample,
-			 int max_stack_depth);
+			 int max_stack_depth, void *arg);
 
 int64_t hist_entry__cmp(struct hist_entry *left, struct hist_entry *right);
 int64_t hist_entry__collapse(struct hist_entry *left, struct hist_entry *right);
-- 
1.9.2

[PATCH 18/26] perf top: Convert to hist_entry_iter

[Namhyung Kim]

Reuse hist_entry_iter__add() function to share the similar code with
perf report.  Note that it needs to be called with hists.lock so tweak
some internal functions not to deadlock or hold the lock too long.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/builtin-top.c | 84 ++++++++++++++++++++++++++----------------------
 1 file changed, 46 insertions(+), 38 deletions(-)

diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index e7a0f8bc2fa7..2465215eff39 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -196,6 +196,12 @@ static void perf_top__record_precise_ip(struct perf_top *top,
 
 	pthread_mutex_unlock(&notes->lock);
 
+	/*
+	 * This function is now called with he->hists->lock held.
+	 * Release it before going to sleep.
+	 */
+	pthread_mutex_unlock(&he->hists->lock);
+
 	if (err == -ERANGE && !he->ms.map->erange_warned)
 		ui__warn_map_erange(he->ms.map, sym, ip);
 	else if (err == -ENOMEM) {
@@ -203,6 +209,8 @@ static void perf_top__record_precise_ip(struct perf_top *top,
 		       sym->name);
 		sleep(1);
 	}
+
+	pthread_mutex_lock(&he->hists->lock);
 }
 
 static void perf_top__show_details(struct perf_top *top)
@@ -238,27 +246,6 @@ out_unlock:
 	pthread_mutex_unlock(&notes->lock);
 }
 
-static struct hist_entry *perf_evsel__add_hist_entry(struct perf_evsel *evsel,
-						     struct addr_location *al,
-						     struct perf_sample *sample)
-{
-	struct hist_entry *he;
-
-	pthread_mutex_lock(&evsel->hists.lock);
-	he = __hists__add_entry(&evsel->hists, al, NULL, NULL, NULL,
-				sample->period, sample->weight,
-				sample->transaction, true);
-	pthread_mutex_unlock(&evsel->hists.lock);
-	if (he == NULL)
-		return NULL;
-
-	hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
-	if (!he->filtered)
-		evsel->hists.stats.nr_non_filtered_samples++;
-
-	return he;
-}
-
 static void perf_top__print_sym_table(struct perf_top *top)
 {
 	char bf[160];
@@ -662,6 +649,31 @@ static int symbol_filter(struct map *map __maybe_unused, struct symbol *sym)
 	return 0;
 }
 
+static int hist_iter_cb(struct hist_entry_iter *iter, struct addr_location *al,
+			bool single, void *arg)
+{
+	struct perf_top *top = arg;
+	struct hist_entry *he = iter->he;
+	struct perf_evsel *evsel = iter->evsel;
+
+	if (single) {
+		hists__inc_nr_events(&evsel->hists, PERF_RECORD_SAMPLE);
+		if (!he->filtered)
+			evsel->hists.stats.nr_non_filtered_samples++;
+	}
+
+	if (sort__has_sym) {
+		u64 ip = al->addr;
+
+		if (al->map)
+			ip = al->map->unmap_ip(al->map, ip);
+
+		perf_top__record_precise_ip(top, he, evsel->idx, ip);
+	}
+
+	return 0;
+}
+
 static void perf_event__process_sample(struct perf_tool *tool,
 				       const union perf_event *event,
 				       struct perf_evsel *evsel,
@@ -669,8 +681,6 @@ static void perf_event__process_sample(struct perf_tool *tool,
 				       struct machine *machine)
 {
 	struct perf_top *top = container_of(tool, struct perf_top, tool);
-	struct symbol *parent = NULL;
-	u64 ip = sample->ip;
 	struct addr_location al;
 	int err;
 
@@ -745,25 +755,23 @@ static void perf_event__process_sample(struct perf_tool *tool,
 	}
 
 	if (al.sym == NULL || !al.sym->ignore) {
-		struct hist_entry *he;
+		struct hist_entry_iter iter = {
+			.add_entry_cb = hist_iter_cb,
+		};
 
-		err = sample__resolve_callchain(sample, &parent, evsel, &al,
-						top->max_stack);
-		if (err)
-			return;
+		if (symbol_conf.cumulate_callchain)
+			iter.ops = &hist_iter_cumulative;
+		else
+			iter.ops = &hist_iter_normal;
 
-		he = perf_evsel__add_hist_entry(evsel, &al, sample);
-		if (he == NULL) {
-			pr_err("Problem incrementing symbol period, skipping event\n");
-			return;
-		}
+		pthread_mutex_lock(&evsel->hists.lock);
 
-		err = hist_entry__append_callchain(he, sample);
-		if (err)
-			return;
+		err = hist_entry_iter__add(&iter, &al, evsel, sample,
+					   top->max_stack, top);
+		if (err < 0)
+			pr_err("Problem incrementing symbol period, skipping event\n");
 
-		if (sort__has_sym)
-			perf_top__record_precise_ip(top, he, evsel->idx, ip);
+		pthread_mutex_unlock(&evsel->hists.lock);
 	}
 
 	return;
-- 
1.9.2

[PATCH 19/26] perf top: Add --children option

[Namhyung Kim]

The --children option is for showing accumulated overhead (period)
value as well as self overhead.  It should be used with one of -g or
--call-graph option.

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/Documentation/perf-top.txt | 8 +++++++-
 tools/perf/builtin-top.c              | 7 +++++++
 2 files changed, 14 insertions(+), 1 deletion(-)

diff --git a/tools/perf/Documentation/perf-top.txt b/tools/perf/Documentation/perf-top.txt
index dcfa54c851e9..180ae02137a5 100644
--- a/tools/perf/Documentation/perf-top.txt
+++ b/tools/perf/Documentation/perf-top.txt
@@ -119,7 +119,7 @@ Default is to monitor all CPUS.
 --fields=::
 	Specify output field - multiple keys can be specified in CSV format.
 	Following fields are available:
-	overhead, overhead_sys, overhead_us, sample and period.
+	overhead, overhead_sys, overhead_us, overhead_children, sample and period.
 	Also it can contain any sort key(s).
 
 	By default, every sort keys not specified in --field will be appended
@@ -161,6 +161,12 @@ Default is to monitor all CPUS.
 	Setup and enable call-graph (stack chain/backtrace) recording,
 	implies -g.
 
+--children::
+	Accumulate callchain of children to parent entry so that then can
+	show up in the output.  The output will have a new "Children" column
+	and will be sorted on the data.  It requires -g/--call-graph option
+	enabled.
+
 --max-stack::
 	Set the stack depth limit when parsing the callchain, anything
 	beyond the specified depth will be ignored. This is a trade-off
diff --git a/tools/perf/builtin-top.c b/tools/perf/builtin-top.c
index 2465215eff39..3f581a074c3d 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1103,6 +1103,8 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 	OPT_CALLBACK(0, "call-graph", &top.record_opts,
 		     "mode[,dump_size]", record_callchain_help,
 		     &parse_callchain_opt),
+	OPT_BOOLEAN(0, "children", &symbol_conf.cumulate_callchain,
+		    "Accumulate callchains of children and show total overhead as well"),
 	OPT_INTEGER(0, "max-stack", &top.max_stack,
 		    "Set the maximum stack depth when parsing the callchain. "
 		    "Default: " __stringify(PERF_MAX_STACK_DEPTH)),
@@ -1208,6 +1210,11 @@ int cmd_top(int argc, const char **argv, const char *prefix __maybe_unused)
 
 	top.sym_evsel = perf_evlist__first(top.evlist);
 
+	if (!symbol_conf.use_callchain) {
+		symbol_conf.cumulate_callchain = false;
+		perf_hpp__cancel_cumulate();
+	}
+
 	symbol_conf.priv_size = sizeof(struct annotation);
 
 	symbol_conf.try_vmlinux_path = (symbol_conf.vmlinux_name == NULL);
-- 
1.9.2

[PATCH 20/26] perf top: Add top.children config option

[Namhyung Kim]

Add top.children config option for setting default value of
callchain accumulation.  It affects the output only if one of
-g or --call-graph option is given as well.

A user can write .perfconfig file like below to enable accumulation
by default:

  $ cat ~/.perfconfig
  [top]
  children = true

And it can be disabled through command line:

  $ perf top --no-children

Tested-by: Arun Sharma <asharma@fb.com>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 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 3f581a074c3d..ecf42417e497 100644
--- a/tools/perf/builtin-top.c
+++ b/tools/perf/builtin-top.c
@@ -1009,6 +1009,10 @@ static int perf_top_config(const char *var, const char *value, void *cb)
 
 	if (!strcmp(var, "top.call-graph"))
 		return record_parse_callchain(value, &top->record_opts);
+	if (!strcmp(var, "top.children")) {
+		symbol_conf.cumulate_callchain = perf_config_bool(var, value);
+		return 0;
+	}
 
 	return perf_default_config(var, value, cb);
 }
-- 
1.9.2

[PATCH 21/26] perf tools: Enable --children option by default

[Namhyung Kim]

Now perf top and perf report will show children column by default if
it has callchain information.

Requested-by: Ingo Molnar <mingo@kernel.org>
Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
Tested-by: Arun Sharma <asharma@fb.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/symbol.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/tools/perf/util/symbol.c b/tools/perf/util/symbol.c
index 95e249779931..7b9096f29cdb 100644
--- a/tools/perf/util/symbol.c
+++ b/tools/perf/util/symbol.c
@@ -29,11 +29,12 @@ int vmlinux_path__nr_entries;
 char **vmlinux_path;
 
 struct symbol_conf symbol_conf = {
-	.use_modules	  = true,
-	.try_vmlinux_path = true,
-	.annotate_src	  = true,
-	.demangle	  = true,
-	.symfs            = "",
+	.use_modules		= true,
+	.try_vmlinux_path	= true,
+	.annotate_src		= true,
+	.demangle		= true,
+	.cumulate_callchain	= true,
+	.symfs			= "",
 };
 
 static enum dso_binary_type binary_type_symtab[] = {
-- 
1.9.2

[PATCH 22/26] perf ui/stdio: Fix invalid percentage value of cumulated hist entries

[Namhyung Kim]

On stdio, there's a problem that it shows invalid values for
callchains in cumulated hist entries.  It's because it only cares
about the self period.  But with --children behavior, we always add
callchain info to the cumulated entries so it should use the value in
that case.

Before:

  # Children      Self  Command      Shared Object            Symbol
  # ........  ........  .......  .................  ................
  #
      61.22%     0.32%  swapper  [kernel.kallsyms]      [k] cpu_idle
                    |
                    --- cpu_idle
                       |
                       |--16530.76%-- start_secondary
                       |
                       |--2758.70%-- rest_init
                       |          start_kernel
                       |          x86_64_start_reservations
                       |          x86_64_start_kernel
                        --6837850969203030.00%-- [...]

After:

  # Children      Self  Command      Shared Object            Symbol
  # ........  ........  .......  .................  ................
  #
      61.22%     0.32%  swapper  [kernel.kallsyms]      [k] cpu_idle
                    |
                    --- cpu_idle
                       |
                       |--85.70%-- start_secondary
                       |
                        --14.30%-- rest_init
                                  start_kernel
                                  x86_64_start_reservations
                                  x86_64_start_kernel

Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/stdio/hist.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c
index 475d2f5c7e16..90122abd3721 100644
--- a/tools/perf/ui/stdio/hist.c
+++ b/tools/perf/ui/stdio/hist.c
@@ -271,7 +271,9 @@ static size_t hist_entry_callchain__fprintf(struct hist_entry *he,
 {
 	switch (callchain_param.mode) {
 	case CHAIN_GRAPH_REL:
-		return callchain__fprintf_graph(fp, &he->sorted_chain, he->stat.period,
+		return callchain__fprintf_graph(fp, &he->sorted_chain,
+						symbol_conf.cumulate_callchain ?
+						he->stat_acc->period : he->stat.period,
 						left_margin);
 		break;
 	case CHAIN_GRAPH_ABS:
-- 
1.9.2

[PATCH 23/26] perf ui/gtk: Fix callchain display

[Namhyung Kim]

With current output field change, GTK browser cannot display callchain
information correctly since it couldn't determine where the symbol
column is.  This is a problem - just for now I changed to use the last
column since it'll work for most cases.

Also it has a same problem of the percentage as stdio code.

Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/ui/gtk/hists.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

diff --git a/tools/perf/ui/gtk/hists.c b/tools/perf/ui/gtk/hists.c
index 03d6812d25dd..6ca60e482cdc 100644
--- a/tools/perf/ui/gtk/hists.c
+++ b/tools/perf/ui/gtk/hists.c
@@ -198,6 +198,13 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists,
 		if (perf_hpp__should_skip(fmt))
 			continue;
 
+		/*
+		 * XXX no way to determine where symcol column is..
+		 *     Just use last column for now.
+		 */
+		if (perf_hpp__is_sort_entry(fmt))
+			sym_col = col_idx;
+
 		fmt->header(fmt, &hpp, hists_to_evsel(hists));
 
 		gtk_tree_view_insert_column_with_attributes(GTK_TREE_VIEW(view),
@@ -253,7 +260,8 @@ static void perf_gtk__show_hists(GtkWidget *window, struct hists *hists,
 
 		if (symbol_conf.use_callchain && sort__has_sym) {
 			if (callchain_param.mode == CHAIN_GRAPH_REL)
-				total = h->stat.period;
+				total = symbol_conf.cumulate_callchain ?
+					h->stat_acc->period : h->stat.period;
 
 			perf_gtk__add_callchain(&h->sorted_chain, store, &iter,
 						sym_col, total);
-- 
1.9.2

[PATCH 24/26] perf tools: Reset output/sort order to default

[Namhyung Kim]

When reset_output_field() is called, also reset field/sort order to
NULL so that it can have the default values.  It's needed for testing.

CC: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/util/sort.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
index 9da8931d2394..254f583a52ab 100644
--- a/tools/perf/util/sort.c
+++ b/tools/perf/util/sort.c
@@ -1582,6 +1582,9 @@ void reset_output_field(void)
 	sort__has_sym = 0;
 	sort__has_dso = 0;
 
+	field_order = NULL;
+	sort_order = NULL;
+
 	reset_dimensions();
 	perf_hpp__reset_output_field();
 }
-- 
1.9.2

[PATCH 25/26] perf tests: Define and use symbolic names for fake symbols

[Namhyung Kim]

In various histogram test cases, fake symbols are used as raw numbers.
Define macros for each pid, map, symbols so that it can increase
readability somewhat.

Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/tests/hists_common.c | 47 ++++++++++++++++++++++-------------------
 tools/perf/tests/hists_common.h | 32 ++++++++++++++++++++++++++--
 tools/perf/tests/hists_filter.c | 23 ++++++++++----------
 tools/perf/tests/hists_link.c   | 32 ++++++++++++++--------------
 tools/perf/tests/hists_output.c | 20 +++++++++---------
 5 files changed, 92 insertions(+), 62 deletions(-)

diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
index e4e01aadc3be..e4e120d3c16f 100644
--- a/tools/perf/tests/hists_common.c
+++ b/tools/perf/tests/hists_common.c
@@ -12,9 +12,9 @@ static struct {
 	u32 pid;
 	const char *comm;
 } fake_threads[] = {
-	{ 100, "perf" },
-	{ 200, "perf" },
-	{ 300, "bash" },
+	{ FAKE_PID_PERF1, "perf" },
+	{ FAKE_PID_PERF2, "perf" },
+	{ FAKE_PID_BASH,  "bash" },
 };
 
 static struct {
@@ -22,15 +22,15 @@ static struct {
 	u64 start;
 	const char *filename;
 } fake_mmap_info[] = {
-	{ 100, 0x40000, "perf" },
-	{ 100, 0x50000, "libc" },
-	{ 100, 0xf0000, "[kernel]" },
-	{ 200, 0x40000, "perf" },
-	{ 200, 0x50000, "libc" },
-	{ 200, 0xf0000, "[kernel]" },
-	{ 300, 0x40000, "bash" },
-	{ 300, 0x50000, "libc" },
-	{ 300, 0xf0000, "[kernel]" },
+	{ FAKE_PID_PERF1, FAKE_MAP_PERF,   "perf" },
+	{ FAKE_PID_PERF1, FAKE_MAP_LIBC,   "libc" },
+	{ FAKE_PID_PERF1, FAKE_MAP_KERNEL, "[kernel]" },
+	{ FAKE_PID_PERF2, FAKE_MAP_PERF,   "perf" },
+	{ FAKE_PID_PERF2, FAKE_MAP_LIBC,   "libc" },
+	{ FAKE_PID_PERF2, FAKE_MAP_KERNEL, "[kernel]" },
+	{ FAKE_PID_BASH,  FAKE_MAP_BASH,   "bash" },
+	{ FAKE_PID_BASH,  FAKE_MAP_LIBC,   "libc" },
+	{ FAKE_PID_BASH,  FAKE_MAP_KERNEL, "[kernel]" },
 };
 
 struct fake_sym {
@@ -40,27 +40,30 @@ struct fake_sym {
 };
 
 static struct fake_sym perf_syms[] = {
-	{ 700, 100, "main" },
-	{ 800, 100, "run_command" },
-	{ 900, 100, "cmd_record" },
+	{ FAKE_SYM_OFFSET1, FAKE_SYM_LENGTH, "main" },
+	{ FAKE_SYM_OFFSET2, FAKE_SYM_LENGTH, "run_command" },
+	{ FAKE_SYM_OFFSET3, FAKE_SYM_LENGTH, "cmd_record" },
 };
 
 static struct fake_sym bash_syms[] = {
-	{ 700, 100, "main" },
-	{ 800, 100, "xmalloc" },
-	{ 900, 100, "xfree" },
+	{ FAKE_SYM_OFFSET1, FAKE_SYM_LENGTH, "main" },
+	{ FAKE_SYM_OFFSET2, FAKE_SYM_LENGTH, "xmalloc" },
+	{ FAKE_SYM_OFFSET3, FAKE_SYM_LENGTH, "xfree" },
 };
 
 static struct fake_sym libc_syms[] = {
 	{ 700, 100, "malloc" },
 	{ 800, 100, "free" },
 	{ 900, 100, "realloc" },
+	{ FAKE_SYM_OFFSET1, FAKE_SYM_LENGTH, "malloc" },
+	{ FAKE_SYM_OFFSET2, FAKE_SYM_LENGTH, "free" },
+	{ FAKE_SYM_OFFSET3, FAKE_SYM_LENGTH, "realloc" },
 };
 
 static struct fake_sym kernel_syms[] = {
-	{ 700, 100, "schedule" },
-	{ 800, 100, "page_fault" },
-	{ 900, 100, "sys_perf_event_open" },
+	{ FAKE_SYM_OFFSET1, FAKE_SYM_LENGTH, "schedule" },
+	{ FAKE_SYM_OFFSET2, FAKE_SYM_LENGTH, "page_fault" },
+	{ FAKE_SYM_OFFSET3, FAKE_SYM_LENGTH, "sys_perf_event_open" },
 };
 
 static struct {
@@ -102,7 +105,7 @@ struct machine *setup_fake_machine(struct machines *machines)
 				.pid = fake_mmap_info[i].pid,
 				.tid = fake_mmap_info[i].pid,
 				.start = fake_mmap_info[i].start,
-				.len = 0x1000ULL,
+				.len = FAKE_MAP_LENGTH,
 				.pgoff = 0ULL,
 			},
 		};
diff --git a/tools/perf/tests/hists_common.h b/tools/perf/tests/hists_common.h
index 1415ae69d7b6..888254e8665c 100644
--- a/tools/perf/tests/hists_common.h
+++ b/tools/perf/tests/hists_common.h
@@ -4,6 +4,34 @@
 struct machine;
 struct machines;
 
+#define FAKE_PID_PERF1  100
+#define FAKE_PID_PERF2  200
+#define FAKE_PID_BASH   300
+
+#define FAKE_MAP_PERF    0x400000
+#define FAKE_MAP_BASH    0x400000
+#define FAKE_MAP_LIBC    0x500000
+#define FAKE_MAP_KERNEL  0xf00000
+#define FAKE_MAP_LENGTH  0x100000
+
+#define FAKE_SYM_OFFSET1  700
+#define FAKE_SYM_OFFSET2  800
+#define FAKE_SYM_OFFSET3  900
+#define FAKE_SYM_LENGTH   100
+
+#define FAKE_IP_PERF_MAIN  FAKE_MAP_PERF + FAKE_SYM_OFFSET1
+#define FAKE_IP_PERF_RUN_COMMAND  FAKE_MAP_PERF + FAKE_SYM_OFFSET2
+#define FAKE_IP_PERF_CMD_RECORD  FAKE_MAP_PERF + FAKE_SYM_OFFSET3
+#define FAKE_IP_BASH_MAIN  FAKE_MAP_BASH + FAKE_SYM_OFFSET1
+#define FAKE_IP_BASH_XMALLOC  FAKE_MAP_BASH + FAKE_SYM_OFFSET2
+#define FAKE_IP_BASH_XFREE  FAKE_MAP_BASH + FAKE_SYM_OFFSET3
+#define FAKE_IP_LIBC_MALLOC  FAKE_MAP_LIBC + FAKE_SYM_OFFSET1
+#define FAKE_IP_LIBC_FREE  FAKE_MAP_LIBC + FAKE_SYM_OFFSET2
+#define FAKE_IP_LIBC_REALLOC  FAKE_MAP_LIBC + FAKE_SYM_OFFSET3
+#define FAKE_IP_KERNEL_SCHEDULE  FAKE_MAP_KERNEL + FAKE_SYM_OFFSET1
+#define FAKE_IP_KERNEL_PAGE_FAULT  FAKE_MAP_KERNEL + FAKE_SYM_OFFSET2
+#define FAKE_IP_KERNEL_SYS_PERF_EVENT_OPEN  FAKE_MAP_KERNEL + FAKE_SYM_OFFSET3
+
 /*
  * The setup_fake_machine() provides a test environment which consists
  * of 3 processes that have 3 mappings and in turn, have 3 symbols
@@ -13,7 +41,7 @@ struct machines;
  * .............  .............  ...................
  *    perf:  100           perf  main
  *    perf:  100           perf  run_command
- *    perf:  100           perf  comd_record
+ *    perf:  100           perf  cmd_record
  *    perf:  100           libc  malloc
  *    perf:  100           libc  free
  *    perf:  100           libc  realloc
@@ -22,7 +50,7 @@ struct machines;
  *    perf:  100       [kernel]  sys_perf_event_open
  *    perf:  200           perf  main
  *    perf:  200           perf  run_command
- *    perf:  200           perf  comd_record
+ *    perf:  200           perf  cmd_record
  *    perf:  200           libc  malloc
  *    perf:  200           libc  free
  *    perf:  200           libc  realloc
diff --git a/tools/perf/tests/hists_filter.c b/tools/perf/tests/hists_filter.c
index 7063824c3566..06a7352a1a3d 100644
--- a/tools/perf/tests/hists_filter.c
+++ b/tools/perf/tests/hists_filter.c
@@ -21,25 +21,25 @@ struct sample {
 /* For the numbers, see hists_common.c */
 static struct sample fake_samples[] = {
 	/* perf [kernel] schedule() */
-	{ .pid = 100, .ip = 0xf0000 + 700, },
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_KERNEL_SCHEDULE, },
 	/* perf [perf]   main() */
-	{ .pid = 100, .ip = 0x40000 + 700, },
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_MAIN, },
 	/* perf [libc]   malloc() */
-	{ .pid = 100, .ip = 0x50000 + 700, },
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_MALLOC, },
 	/* perf [perf]   main() */
-	{ .pid = 200, .ip = 0x40000 + 700, }, /* will be merged */
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_MAIN, }, /* will be merged */
 	/* perf [perf]   cmd_record() */
-	{ .pid = 200, .ip = 0x40000 + 900, },
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_CMD_RECORD, },
 	/* perf [kernel] page_fault() */
-	{ .pid = 200, .ip = 0xf0000 + 800, },
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 	/* bash [bash]   main() */
-	{ .pid = 300, .ip = 0x40000 + 700, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_MAIN, },
 	/* bash [bash]   xmalloc() */
-	{ .pid = 300, .ip = 0x40000 + 800, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_XMALLOC, },
 	/* bash [libc]   malloc() */
-	{ .pid = 300, .ip = 0x50000 + 700, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_LIBC_MALLOC, },
 	/* bash [kernel] page_fault() */
-	{ .pid = 300, .ip = 0xf0000 + 800, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 };
 
 static int count_sample(struct hist_entry_iter *iter,
@@ -66,7 +66,7 @@ static int add_hist_entries(struct perf_evlist *evlist,
 {
 	struct perf_evsel *evsel;
 	struct addr_location al;
-	struct perf_sample sample = { .cpu = 0, };
+	struct perf_sample sample = { .period = 100, };
 	size_t i;
 
 	/*
@@ -95,7 +95,6 @@ static int add_hist_entries(struct perf_evlist *evlist,
 			sample.pid = fake_samples[i].pid;
 			sample.tid = fake_samples[i].pid;
 			sample.ip = fake_samples[i].ip;
-			sample.period = 100;
 
 			if (perf_event__preprocess_sample(&event, machine, &al,
 							  &sample) < 0)
diff --git a/tools/perf/tests/hists_link.c b/tools/perf/tests/hists_link.c
index ca6693b37cd7..d4b34b0f50a2 100644
--- a/tools/perf/tests/hists_link.c
+++ b/tools/perf/tests/hists_link.c
@@ -21,41 +21,41 @@ struct sample {
 /* For the numbers, see hists_common.c */
 static struct sample fake_common_samples[] = {
 	/* perf [kernel] schedule() */
-	{ .pid = 100, .ip = 0xf0000 + 700, },
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_KERNEL_SCHEDULE, },
 	/* perf [perf]   main() */
-	{ .pid = 200, .ip = 0x40000 + 700, },
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_MAIN, },
 	/* perf [perf]   cmd_record() */
-	{ .pid = 200, .ip = 0x40000 + 900, },
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_CMD_RECORD, },
 	/* bash [bash]   xmalloc() */
-	{ .pid = 300, .ip = 0x40000 + 800, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_XMALLOC, },
 	/* bash [libc]   malloc() */
-	{ .pid = 300, .ip = 0x50000 + 700, },
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_LIBC_MALLOC, },
 };
 
 static struct sample fake_samples[][5] = {
 	{
 		/* perf [perf]   run_command() */
-		{ .pid = 100, .ip = 0x40000 + 800, },
+		{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_RUN_COMMAND, },
 		/* perf [libc]   malloc() */
-		{ .pid = 100, .ip = 0x50000 + 700, },
+		{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_MALLOC, },
 		/* perf [kernel] page_fault() */
-		{ .pid = 100, .ip = 0xf0000 + 800, },
+		{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 		/* perf [kernel] sys_perf_event_open() */
-		{ .pid = 200, .ip = 0xf0000 + 900, },
+		{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_KERNEL_SYS_PERF_EVENT_OPEN, },
 		/* bash [libc]   free() */
-		{ .pid = 300, .ip = 0x50000 + 800, },
+		{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_LIBC_FREE, },
 	},
 	{
 		/* perf [libc]   free() */
-		{ .pid = 200, .ip = 0x50000 + 800, },
+		{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_LIBC_FREE, },
 		/* bash [libc]   malloc() */
-		{ .pid = 300, .ip = 0x50000 + 700, }, /* will be merged */
+		{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_LIBC_MALLOC, }, /* will be merged */
 		/* bash [bash]   xfee() */
-		{ .pid = 300, .ip = 0x40000 + 900, },
+		{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_XFREE, },
 		/* bash [libc]   realloc() */
-		{ .pid = 300, .ip = 0x50000 + 900, },
+		{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_LIBC_REALLOC, },
 		/* bash [kernel] page_fault() */
-		{ .pid = 300, .ip = 0xf0000 + 800, },
+		{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 	},
 };
 
@@ -64,7 +64,7 @@ static int add_hist_entries(struct perf_evlist *evlist, struct machine *machine)
 	struct perf_evsel *evsel;
 	struct addr_location al;
 	struct hist_entry *he;
-	struct perf_sample sample = { .cpu = 0, };
+	struct perf_sample sample = { .period = 1, };
 	size_t i = 0, k;
 
 	/*
diff --git a/tools/perf/tests/hists_output.c b/tools/perf/tests/hists_output.c
index d40461ecd210..e3bbd6c54c1b 100644
--- a/tools/perf/tests/hists_output.c
+++ b/tools/perf/tests/hists_output.c
@@ -22,25 +22,25 @@ struct sample {
 /* For the numbers, see hists_common.c */
 static struct sample fake_samples[] = {
 	/* perf [kernel] schedule() */
-	{ .cpu = 0, .pid = 100, .ip = 0xf0000 + 700, },
+	{ .cpu = 0, .pid = FAKE_PID_PERF1, .ip = FAKE_IP_KERNEL_SCHEDULE, },
 	/* perf [perf]   main() */
-	{ .cpu = 1, .pid = 100, .ip = 0x40000 + 700, },
+	{ .cpu = 1, .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_MAIN, },
 	/* perf [perf]   cmd_record() */
-	{ .cpu = 1, .pid = 100, .ip = 0x40000 + 900, },
+	{ .cpu = 1, .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_CMD_RECORD, },
 	/* perf [libc]   malloc() */
-	{ .cpu = 1, .pid = 100, .ip = 0x50000 + 700, },
+	{ .cpu = 1, .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_MALLOC, },
 	/* perf [libc]   free() */
-	{ .cpu = 2, .pid = 100, .ip = 0x50000 + 800, },
+	{ .cpu = 2, .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_FREE, },
 	/* perf [perf]   main() */
-	{ .cpu = 2, .pid = 200, .ip = 0x40000 + 700, },
+	{ .cpu = 2, .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_MAIN, },
 	/* perf [kernel] page_fault() */
-	{ .cpu = 2, .pid = 200, .ip = 0xf0000 + 800, },
+	{ .cpu = 2, .pid = FAKE_PID_PERF2, .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 	/* bash [bash]   main() */
-	{ .cpu = 3, .pid = 300, .ip = 0x40000 + 700, },
+	{ .cpu = 3, .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_MAIN, },
 	/* bash [bash]   xmalloc() */
-	{ .cpu = 0, .pid = 300, .ip = 0x40000 + 800, },
+	{ .cpu = 0, .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_XMALLOC, },
 	/* bash [kernel] page_fault() */
-	{ .cpu = 1, .pid = 300, .ip = 0xf0000 + 800, },
+	{ .cpu = 1, .pid = FAKE_PID_BASH,  .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
 };
 
 static int add_hist_entries(struct hists *hists, struct machine *machine)
-- 
1.9.2

[PATCH 26/26] perf tests: Add a test case for cumulating callchains

[Namhyung Kim]

Now it adds a new testcase to verify --children option working
correctly.

Cc: Arun Sharma <asharma@fb.com>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
---
 tools/perf/Makefile.perf          |   1 +
 tools/perf/tests/builtin-test.c   |   4 +
 tools/perf/tests/hists_common.c   |   5 +-
 tools/perf/tests/hists_cumulate.c | 726 ++++++++++++++++++++++++++++++++++++++
 tools/perf/tests/tests.h          |   1 +
 5 files changed, 735 insertions(+), 2 deletions(-)
 create mode 100644 tools/perf/tests/hists_cumulate.c

diff --git a/tools/perf/Makefile.perf b/tools/perf/Makefile.perf
index 02f0a4dd1a80..67f7c0575b26 100644
--- a/tools/perf/Makefile.perf
+++ b/tools/perf/Makefile.perf
@@ -400,6 +400,7 @@ LIB_OBJS += $(OUTPUT)tests/hists_common.o
 LIB_OBJS += $(OUTPUT)tests/hists_link.o
 LIB_OBJS += $(OUTPUT)tests/hists_filter.o
 LIB_OBJS += $(OUTPUT)tests/hists_output.o
+LIB_OBJS += $(OUTPUT)tests/hists_cumulate.o
 LIB_OBJS += $(OUTPUT)tests/python-use.o
 LIB_OBJS += $(OUTPUT)tests/bp_signal.o
 LIB_OBJS += $(OUTPUT)tests/bp_signal_overflow.o
diff --git a/tools/perf/tests/builtin-test.c b/tools/perf/tests/builtin-test.c
index 831f52cae197..802e3cd50f6f 100644
--- a/tools/perf/tests/builtin-test.c
+++ b/tools/perf/tests/builtin-test.c
@@ -140,6 +140,10 @@ static struct test {
 		.func = test__hists_output,
 	},
 	{
+		.desc = "Test cumulation of child hist entries",
+		.func = test__hists_cumulate,
+	},
+	{
 		.func = NULL,
 	},
 };
diff --git a/tools/perf/tests/hists_common.c b/tools/perf/tests/hists_common.c
index e4e120d3c16f..a62c09134516 100644
--- a/tools/perf/tests/hists_common.c
+++ b/tools/perf/tests/hists_common.c
@@ -196,10 +196,11 @@ void print_hists_out(struct hists *hists)
 		he = rb_entry(node, struct hist_entry, rb_node);
 
 		if (!he->filtered) {
-			pr_info("%2d: entry: %8s:%5d [%-8s] %20s: period = %"PRIu64"\n",
+			pr_info("%2d: entry: %8s:%5d [%-8s] %20s: period = %"PRIu64"/%"PRIu64"\n",
 				i, thread__comm_str(he->thread), he->thread->tid,
 				he->ms.map->dso->short_name,
-				he->ms.sym->name, he->stat.period);
+				he->ms.sym->name, he->stat.period,
+				he->stat_acc ? he->stat_acc->period : 0);
 		}
 
 		i++;
diff --git a/tools/perf/tests/hists_cumulate.c b/tools/perf/tests/hists_cumulate.c
new file mode 100644
index 000000000000..0ac240db2e24
--- /dev/null
+++ b/tools/perf/tests/hists_cumulate.c
@@ -0,0 +1,726 @@
+#include "perf.h"
+#include "util/debug.h"
+#include "util/symbol.h"
+#include "util/sort.h"
+#include "util/evsel.h"
+#include "util/evlist.h"
+#include "util/machine.h"
+#include "util/thread.h"
+#include "util/parse-events.h"
+#include "tests/tests.h"
+#include "tests/hists_common.h"
+
+struct sample {
+	u32 pid;
+	u64 ip;
+	struct thread *thread;
+	struct map *map;
+	struct symbol *sym;
+};
+
+/* For the numbers, see hists_common.c */
+static struct sample fake_samples[] = {
+	/* perf [kernel] schedule() */
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_KERNEL_SCHEDULE, },
+	/* perf [perf]   main() */
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_MAIN, },
+	/* perf [perf]   cmd_record() */
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_PERF_CMD_RECORD, },
+	/* perf [libc]   malloc() */
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_MALLOC, },
+	/* perf [libc]   free() */
+	{ .pid = FAKE_PID_PERF1, .ip = FAKE_IP_LIBC_FREE, },
+	/* perf [perf]   main() */
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_PERF_MAIN, },
+	/* perf [kernel] page_fault() */
+	{ .pid = FAKE_PID_PERF2, .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
+	/* bash [bash]   main() */
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_MAIN, },
+	/* bash [bash]   xmalloc() */
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_BASH_XMALLOC, },
+	/* bash [kernel] page_fault() */
+	{ .pid = FAKE_PID_BASH,  .ip = FAKE_IP_KERNEL_PAGE_FAULT, },
+};
+
+/*
+ * Will be casted to struct ip_callchain which has all 64 bit entries
+ * of nr and ips[].
+ */
+static u64 fake_callchains[][10] = {
+	/*   schedule => run_command => main */
+	{ 3, FAKE_IP_KERNEL_SCHEDULE, FAKE_IP_PERF_RUN_COMMAND, FAKE_IP_PERF_MAIN, },
+	/*   main  */
+	{ 1, FAKE_IP_PERF_MAIN, },
+	/*   cmd_record => run_command => main */
+	{ 3, FAKE_IP_PERF_CMD_RECORD, FAKE_IP_PERF_RUN_COMMAND, FAKE_IP_PERF_MAIN, },
+	/*   malloc => cmd_record => run_command => main */
+	{ 4, FAKE_IP_LIBC_MALLOC, FAKE_IP_PERF_CMD_RECORD, FAKE_IP_PERF_RUN_COMMAND,
+	     FAKE_IP_PERF_MAIN, },
+	/*   free => cmd_record => run_command => main */
+	{ 4, FAKE_IP_LIBC_FREE, FAKE_IP_PERF_CMD_RECORD, FAKE_IP_PERF_RUN_COMMAND,
+	     FAKE_IP_PERF_MAIN, },
+	/*   main */
+	{ 1, FAKE_IP_PERF_MAIN, },
+	/*   page_fault => sys_perf_event_open => run_command => main */
+	{ 4, FAKE_IP_KERNEL_PAGE_FAULT, FAKE_IP_KERNEL_SYS_PERF_EVENT_OPEN,
+	     FAKE_IP_PERF_RUN_COMMAND, FAKE_IP_PERF_MAIN, },
+	/*   main */
+	{ 1, FAKE_IP_BASH_MAIN, },
+	/*   xmalloc => malloc => xmalloc => malloc => xmalloc => main */
+	{ 6, FAKE_IP_BASH_XMALLOC, FAKE_IP_LIBC_MALLOC, FAKE_IP_BASH_XMALLOC,
+	     FAKE_IP_LIBC_MALLOC, FAKE_IP_BASH_XMALLOC, FAKE_IP_BASH_MAIN, },
+	/*   page_fault => malloc => main */
+	{ 3, FAKE_IP_KERNEL_PAGE_FAULT, FAKE_IP_LIBC_MALLOC, FAKE_IP_BASH_MAIN, },
+};
+
+static int add_hist_entries(struct hists *hists, struct machine *machine)
+{
+	struct addr_location al;
+	struct perf_evsel *evsel = hists_to_evsel(hists);
+	struct perf_sample sample = { .period = 1000, };
+	size_t i;
+
+	for (i = 0; i < ARRAY_SIZE(fake_samples); i++) {
+		const union perf_event event = {
+			.header = {
+				.misc = PERF_RECORD_MISC_USER,
+			},
+		};
+		struct hist_entry_iter iter = {
+			.hide_unresolved = false,
+		};
+
+		if (symbol_conf.cumulate_callchain)
+			iter.ops = &hist_iter_cumulative;
+		else
+			iter.ops = &hist_iter_normal;
+
+		sample.pid = fake_samples[i].pid;
+		sample.tid = fake_samples[i].pid;
+		sample.ip = fake_samples[i].ip;
+		sample.callchain = (struct ip_callchain *)fake_callchains[i];
+
+		if (perf_event__preprocess_sample(&event, machine, &al,
+						  &sample) < 0)
+			goto out;
+
+		if (hist_entry_iter__add(&iter, &al, evsel, &sample,
+					 PERF_MAX_STACK_DEPTH, NULL) < 0)
+			goto out;
+
+		fake_samples[i].thread = al.thread;
+		fake_samples[i].map = al.map;
+		fake_samples[i].sym = al.sym;
+	}
+
+	return TEST_OK;
+
+out:
+	pr_debug("Not enough memory for adding a hist entry\n");
+	return TEST_FAIL;
+}
+
+static void del_hist_entries(struct hists *hists)
+{
+	struct hist_entry *he;
+	struct rb_root *root_in;
+	struct rb_root *root_out;
+	struct rb_node *node;
+
+	if (sort__need_collapse)
+		root_in = &hists->entries_collapsed;
+	else
+		root_in = hists->entries_in;
+
+	root_out = &hists->entries;
+
+	while (!RB_EMPTY_ROOT(root_out)) {
+		node = rb_first(root_out);
+
+		he = rb_entry(node, struct hist_entry, rb_node);
+		rb_erase(node, root_out);
+		rb_erase(&he->rb_node_in, root_in);
+		hist_entry__free(he);
+	}
+}
+
+typedef int (*test_fn_t)(struct perf_evsel *, struct machine *);
+
+#define COMM(he)  (thread__comm_str(he->thread))
+#define DSO(he)   (he->ms.map->dso->short_name)
+#define SYM(he)   (he->ms.sym->name)
+#define CPU(he)   (he->cpu)
+#define PID(he)   (he->thread->tid)
+#define DEPTH(he) (he->callchain->max_depth)
+#define CDSO(cl)  (cl->ms.map->dso->short_name)
+#define CSYM(cl)  (cl->ms.sym->name)
+
+struct result {
+	u64 children;
+	u64 self;
+	const char *comm;
+	const char *dso;
+	const char *sym;
+};
+
+struct callchain_result {
+	u64 nr;
+	struct {
+		const char *dso;
+		const char *sym;
+	} node[10];
+};
+
+static int do_test(struct hists *hists, struct result *expected, size_t nr_expected,
+		   struct callchain_result *expected_callchain, size_t nr_callchain)
+{
+	char buf[32];
+	size_t i, c;
+	struct hist_entry *he;
+	struct rb_root *root;
+	struct rb_node *node;
+	struct callchain_node *cnode;
+	struct callchain_list *clist;
+
+	/*
+	 * adding and deleting hist entries must be done outside of this
+	 * function since TEST_ASSERT_VAL() returns in case of failure.
+	 */
+	hists__collapse_resort(hists, NULL);
+	hists__output_resort(hists);
+
+	if (verbose > 2) {
+		pr_info("use callchain: %d, cumulate callchain: %d\n",
+			symbol_conf.use_callchain,
+			symbol_conf.cumulate_callchain);
+		print_hists_out(hists);
+	}
+
+	root = &hists->entries;
+	for (node = rb_first(root), i = 0;
+	     node && (he = rb_entry(node, struct hist_entry, rb_node));
+	     node = rb_next(node), i++) {
+		scnprintf(buf, sizeof(buf), "Invalid hist entry #%zd", i);
+
+		TEST_ASSERT_VAL("Incorrect number of hist entry",
+				i < nr_expected);
+		TEST_ASSERT_VAL(buf, he->stat.period == expected[i].self &&
+				!strcmp(COMM(he), expected[i].comm) &&
+				!strcmp(DSO(he), expected[i].dso) &&
+				!strcmp(SYM(he), expected[i].sym));
+
+		if (symbol_conf.cumulate_callchain)
+			TEST_ASSERT_VAL(buf, he->stat_acc->period == expected[i].children);
+
+		if (!symbol_conf.use_callchain)
+			continue;
+
+		/* check callchain entries */
+		root = &he->callchain->node.rb_root;
+		cnode = rb_entry(rb_first(root), struct callchain_node, rb_node);
+
+		c = 0;
+		list_for_each_entry(clist, &cnode->val, list) {
+			scnprintf(buf, sizeof(buf), "Invalid callchain entry #%zd/%zd", i, c);
+
+			TEST_ASSERT_VAL("Incorrect number of callchain entry",
+					c < expected_callchain[i].nr);
+			TEST_ASSERT_VAL(buf,
+				!strcmp(CDSO(clist), expected_callchain[i].node[c].dso) &&
+				!strcmp(CSYM(clist), expected_callchain[i].node[c].sym));
+			c++;
+		}
+		/* TODO: handle multiple child nodes properly */
+		TEST_ASSERT_VAL("Incorrect number of callchain entry",
+				c <= expected_callchain[i].nr);
+	}
+	TEST_ASSERT_VAL("Incorrect number of hist entry",
+			i == nr_expected);
+	TEST_ASSERT_VAL("Incorrect number of callchain entry",
+			!symbol_conf.use_callchain || nr_expected == nr_callchain);
+	return 0;
+}
+
+/* NO callchain + NO children */
+static int test1(struct perf_evsel *evsel, struct machine *machine)
+{
+	int err;
+	struct hists *hists = &evsel->hists;
+	/*
+	 * expected output:
+	 *
+	 * Overhead  Command  Shared Object          Symbol
+	 * ========  =======  =============  ==============
+	 *   20.00%     perf  perf           [.] main
+	 *   10.00%     bash  [kernel]       [k] page_fault
+	 *   10.00%     bash  bash           [.] main
+	 *   10.00%     bash  bash           [.] xmalloc
+	 *   10.00%     perf  [kernel]       [k] page_fault
+	 *   10.00%     perf  [kernel]       [k] schedule
+	 *   10.00%     perf  libc           [.] free
+	 *   10.00%     perf  libc           [.] malloc
+	 *   10.00%     perf  perf           [.] cmd_record
+	 */
+	struct result expected[] = {
+		{ 0, 2000, "perf", "perf",     "main" },
+		{ 0, 1000, "bash", "[kernel]", "page_fault" },
+		{ 0, 1000, "bash", "bash",     "main" },
+		{ 0, 1000, "bash", "bash",     "xmalloc" },
+		{ 0, 1000, "perf", "[kernel]", "page_fault" },
+		{ 0, 1000, "perf", "[kernel]", "schedule" },
+		{ 0, 1000, "perf", "libc",     "free" },
+		{ 0, 1000, "perf", "libc",     "malloc" },
+		{ 0, 1000, "perf", "perf",     "cmd_record" },
+	};
+
+	symbol_conf.use_callchain = false;
+	symbol_conf.cumulate_callchain = false;
+
+	setup_sorting();
+	callchain_register_param(&callchain_param);
+
+	err = add_hist_entries(hists, machine);
+	if (err < 0)
+		goto out;
+
+	err = do_test(hists, expected, ARRAY_SIZE(expected), NULL, 0);
+
+out:
+	del_hist_entries(hists);
+	reset_output_field();
+	return err;
+}
+
+/* callcain + NO children */
+static int test2(struct perf_evsel *evsel, struct machine *machine)
+{
+	int err;
+	struct hists *hists = &evsel->hists;
+	/*
+	 * expected output:
+	 *
+	 * Overhead  Command  Shared Object          Symbol
+	 * ========  =======  =============  ==============
+	 *   20.00%     perf  perf           [.] main
+	 *              |
+	 *              --- main
+	 *
+	 *   10.00%     bash  [kernel]       [k] page_fault
+	 *              |
+	 *              --- page_fault
+	 *                  malloc
+	 *                  main
+	 *
+	 *   10.00%     bash  bash           [.] main
+	 *              |
+	 *              --- main
+	 *
+	 *   10.00%     bash  bash           [.] xmalloc
+	 *              |
+	 *              --- xmalloc
+	 *                  malloc
+	 *                  xmalloc     <--- NOTE: there's a cycle
+	 *                  malloc
+	 *                  xmalloc
+	 *                  main
+	 *
+	 *   10.00%     perf  [kernel]       [k] page_fault
+	 *              |
+	 *              --- page_fault
+	 *                  sys_perf_event_open
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%     perf  [kernel]       [k] schedule
+	 *              |
+	 *              --- schedule
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%     perf  libc           [.] free
+	 *              |
+	 *              --- free
+	 *                  cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%     perf  libc           [.] malloc
+	 *              |
+	 *              --- malloc
+	 *                  cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%     perf  perf           [.] cmd_record
+	 *              |
+	 *              --- cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 */
+	struct result expected[] = {
+		{ 0, 2000, "perf", "perf",     "main" },
+		{ 0, 1000, "bash", "[kernel]", "page_fault" },
+		{ 0, 1000, "bash", "bash",     "main" },
+		{ 0, 1000, "bash", "bash",     "xmalloc" },
+		{ 0, 1000, "perf", "[kernel]", "page_fault" },
+		{ 0, 1000, "perf", "[kernel]", "schedule" },
+		{ 0, 1000, "perf", "libc",     "free" },
+		{ 0, 1000, "perf", "libc",     "malloc" },
+		{ 0, 1000, "perf", "perf",     "cmd_record" },
+	};
+	struct callchain_result expected_callchain[] = {
+		{
+			1, {	{ "perf",     "main" }, },
+		},
+		{
+			3, {	{ "[kernel]", "page_fault" },
+				{ "libc",     "malloc" },
+				{ "bash",     "main" }, },
+		},
+		{
+			1, {	{ "bash",     "main" }, },
+		},
+		{
+			6, {	{ "bash",     "xmalloc" },
+				{ "libc",     "malloc" },
+				{ "bash",     "xmalloc" },
+				{ "libc",     "malloc" },
+				{ "bash",     "xmalloc" },
+				{ "bash",     "main" }, },
+		},
+		{
+			4, {	{ "[kernel]", "page_fault" },
+				{ "[kernel]", "sys_perf_event_open" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			3, {	{ "[kernel]", "schedule" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "libc",     "free" },
+				{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "libc",     "malloc" },
+				{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			3, {	{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+	};
+
+	symbol_conf.use_callchain = true;
+	symbol_conf.cumulate_callchain = false;
+
+	setup_sorting();
+	callchain_register_param(&callchain_param);
+
+	err = add_hist_entries(hists, machine);
+	if (err < 0)
+		goto out;
+
+	err = do_test(hists, expected, ARRAY_SIZE(expected),
+		      expected_callchain, ARRAY_SIZE(expected_callchain));
+
+out:
+	del_hist_entries(hists);
+	reset_output_field();
+	return err;
+}
+
+/* NO callchain + children */
+static int test3(struct perf_evsel *evsel, struct machine *machine)
+{
+	int err;
+	struct hists *hists = &evsel->hists;
+	/*
+	 * expected output:
+	 *
+	 * Children      Self  Command  Shared Object                   Symbol
+	 * ========  ========  =======  =============  =======================
+	 *   70.00%    20.00%     perf  perf           [.] main
+	 *   50.00%     0.00%     perf  perf           [.] run_command
+	 *   30.00%    10.00%     bash  bash           [.] main
+	 *   30.00%    10.00%     perf  perf           [.] cmd_record
+	 *   20.00%     0.00%     bash  libc           [.] malloc
+	 *   10.00%    10.00%     bash  [kernel]       [k] page_fault
+	 *   10.00%    10.00%     perf  [kernel]       [k] schedule
+	 *   10.00%     0.00%     perf  [kernel]       [k] sys_perf_event_open
+	 *   10.00%    10.00%     perf  [kernel]       [k] page_fault
+	 *   10.00%    10.00%     perf  libc           [.] free
+	 *   10.00%    10.00%     perf  libc           [.] malloc
+	 *   10.00%    10.00%     bash  bash           [.] xmalloc
+	 */
+	struct result expected[] = {
+		{ 7000, 2000, "perf", "perf",     "main" },
+		{ 5000,    0, "perf", "perf",     "run_command" },
+		{ 3000, 1000, "bash", "bash",     "main" },
+		{ 3000, 1000, "perf", "perf",     "cmd_record" },
+		{ 2000,    0, "bash", "libc",     "malloc" },
+		{ 1000, 1000, "bash", "[kernel]", "page_fault" },
+		{ 1000, 1000, "perf", "[kernel]", "schedule" },
+		{ 1000,    0, "perf", "[kernel]", "sys_perf_event_open" },
+		{ 1000, 1000, "perf", "[kernel]", "page_fault" },
+		{ 1000, 1000, "perf", "libc",     "free" },
+		{ 1000, 1000, "perf", "libc",     "malloc" },
+		{ 1000, 1000, "bash", "bash",     "xmalloc" },
+	};
+
+	symbol_conf.use_callchain = false;
+	symbol_conf.cumulate_callchain = true;
+
+	setup_sorting();
+	callchain_register_param(&callchain_param);
+
+	err = add_hist_entries(hists, machine);
+	if (err < 0)
+		goto out;
+
+	err = do_test(hists, expected, ARRAY_SIZE(expected), NULL, 0);
+
+out:
+	del_hist_entries(hists);
+	reset_output_field();
+	return err;
+}
+
+/* callchain + children */
+static int test4(struct perf_evsel *evsel, struct machine *machine)
+{
+	int err;
+	struct hists *hists = &evsel->hists;
+	/*
+	 * expected output:
+	 *
+	 * Children      Self  Command  Shared Object                   Symbol
+	 * ========  ========  =======  =============  =======================
+	 *   70.00%    20.00%     perf  perf           [.] main
+	 *              |
+	 *              --- main
+	 *
+	 *   50.00%     0.00%     perf  perf           [.] run_command
+	 *              |
+	 *              --- run_command
+	 *                  main
+	 *
+	 *   30.00%    10.00%     bash  bash           [.] main
+	 *              |
+	 *              --- main
+	 *
+	 *   30.00%    10.00%     perf  perf           [.] cmd_record
+	 *              |
+	 *              --- cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 *   20.00%     0.00%     bash  libc           [.] malloc
+	 *              |
+	 *              --- malloc
+	 *                 |
+	 *                 |--50.00%-- xmalloc
+	 *                 |           main
+	 *                  --50.00%-- main
+	 *
+	 *   10.00%    10.00%     bash  [kernel]       [k] page_fault
+	 *              |
+	 *              --- page_fault
+	 *                  malloc
+	 *                  main
+	 *
+	 *   10.00%    10.00%     perf  [kernel]       [k] schedule
+	 *              |
+	 *              --- schedule
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%     0.00%     perf  [kernel]       [k] sys_perf_event_open
+	 *              |
+	 *              --- sys_perf_event_open
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%    10.00%     perf  [kernel]       [k] page_fault
+	 *              |
+	 *              --- page_fault
+	 *                  sys_perf_event_open
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%    10.00%     perf  libc           [.] free
+	 *              |
+	 *              --- free
+	 *                  cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%    10.00%     perf  libc           [.] malloc
+	 *              |
+	 *              --- malloc
+	 *                  cmd_record
+	 *                  run_command
+	 *                  main
+	 *
+	 *   10.00%    10.00%     bash  bash           [.] xmalloc
+	 *              |
+	 *              --- xmalloc
+	 *                  malloc
+	 *                  xmalloc     <--- NOTE: there's a cycle
+	 *                  malloc
+	 *                  xmalloc
+	 *                  main
+	 *
+	 */
+	struct result expected[] = {
+		{ 7000, 2000, "perf", "perf",     "main" },
+		{ 5000,    0, "perf", "perf",     "run_command" },
+		{ 3000, 1000, "bash", "bash",     "main" },
+		{ 3000, 1000, "perf", "perf",     "cmd_record" },
+		{ 2000,    0, "bash", "libc",     "malloc" },
+		{ 1000, 1000, "bash", "[kernel]", "page_fault" },
+		{ 1000, 1000, "perf", "[kernel]", "schedule" },
+		{ 1000,    0, "perf", "[kernel]", "sys_perf_event_open" },
+		{ 1000, 1000, "perf", "[kernel]", "page_fault" },
+		{ 1000, 1000, "perf", "libc",     "free" },
+		{ 1000, 1000, "perf", "libc",     "malloc" },
+		{ 1000, 1000, "bash", "bash",     "xmalloc" },
+	};
+	struct callchain_result expected_callchain[] = {
+		{
+			1, {	{ "perf",     "main" }, },
+		},
+		{
+			2, {	{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			1, {	{ "bash",     "main" }, },
+		},
+		{
+			3, {	{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "libc",     "malloc" },
+				{ "bash",     "xmalloc" },
+				{ "bash",     "main" },
+				{ "bash",     "main" }, },
+		},
+		{
+			3, {	{ "[kernel]", "page_fault" },
+				{ "libc",     "malloc" },
+				{ "bash",     "main" }, },
+		},
+		{
+			3, {	{ "[kernel]", "schedule" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			3, {	{ "[kernel]", "sys_perf_event_open" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "[kernel]", "page_fault" },
+				{ "[kernel]", "sys_perf_event_open" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "libc",     "free" },
+				{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			4, {	{ "libc",     "malloc" },
+				{ "perf",     "cmd_record" },
+				{ "perf",     "run_command" },
+				{ "perf",     "main" }, },
+		},
+		{
+			6, {	{ "bash",     "xmalloc" },
+				{ "libc",     "malloc" },
+				{ "bash",     "xmalloc" },
+				{ "libc",     "malloc" },
+				{ "bash",     "xmalloc" },
+				{ "bash",     "main" }, },
+		},
+	};
+
+	symbol_conf.use_callchain = true;
+	symbol_conf.cumulate_callchain = true;
+
+	setup_sorting();
+	callchain_register_param(&callchain_param);
+
+	err = add_hist_entries(hists, machine);
+	if (err < 0)
+		goto out;
+
+	err = do_test(hists, expected, ARRAY_SIZE(expected),
+		      expected_callchain, ARRAY_SIZE(expected_callchain));
+
+out:
+	del_hist_entries(hists);
+	reset_output_field();
+	return err;
+}
+
+int test__hists_cumulate(void)
+{
+	int err = TEST_FAIL;
+	struct machines machines;
+	struct machine *machine;
+	struct perf_evsel *evsel;
+	struct perf_evlist *evlist = perf_evlist__new();
+	size_t i;
+	test_fn_t testcases[] = {
+		test1,
+		test2,
+		test3,
+		test4,
+	};
+
+	TEST_ASSERT_VAL("No memory", evlist);
+
+	err = parse_events(evlist, "cpu-clock");
+	if (err)
+		goto out;
+
+	machines__init(&machines);
+
+	/* setup threads/dso/map/symbols also */
+	machine = setup_fake_machine(&machines);
+	if (!machine)
+		goto out;
+
+	if (verbose > 1)
+		machine__fprintf(machine, stderr);
+
+	evsel = perf_evlist__first(evlist);
+
+	for (i = 0; i < ARRAY_SIZE(testcases); i++) {
+		err = testcases[i](evsel, machine);
+		if (err < 0)
+			break;
+	}
+
+out:
+	/* tear down everything */
+	perf_evlist__delete(evlist);
+	machines__exit(&machines);
+
+	return err;
+}
diff --git a/tools/perf/tests/tests.h b/tools/perf/tests/tests.h
index d76c0e2e6635..022bb68fd9c7 100644
--- a/tools/perf/tests/tests.h
+++ b/tools/perf/tests/tests.h
@@ -45,6 +45,7 @@ int test__hists_filter(void);
 int test__mmap_thread_lookup(void);
 int test__thread_mg_share(void);
 int test__hists_output(void);
+int test__hists_cumulate(void);
 
 #if defined(__x86_64__) || defined(__i386__) || defined(__arm__)
 #ifdef HAVE_DWARF_UNWIND_SUPPORT
-- 
1.9.2

Re: [PATCH 10/26] perf ui/browser: Add support to accumulated hist stat

[Jiri Olsa]

On Fri, May 23, 2014 at 07:04:07PM +0900, Namhyung Kim wrote:
> Print accumulated stat of a hist entry if requested.
> 
> Tested-by: Arun Sharma <asharma@fb.com>
> Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
> Acked-by: Jiri Olsa <jolsa@redhat.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> ---
>  tools/perf/ui/browsers/hists.c | 26 +++++++++++++++++++++++++-
>  1 file changed, 25 insertions(+), 1 deletion(-)
> 
> diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
> index 1c331b934ffc..1de7428a9771 100644
> --- a/tools/perf/ui/browsers/hists.c
> +++ b/tools/perf/ui/browsers/hists.c
> @@ -632,7 +632,6 @@ static int __hpp__slsmg_color_printf(struct perf_hpp *hpp, const char *fmt, ...)
>  	ret = scnprintf(hpp->buf, hpp->size, fmt, percent);
>  	slsmg_printf("%s", hpp->buf);
>  
> -	advance_hpp(hpp, ret);
>  	return ret;

Whats the purpose of this one..? does not fit to what's in changelog 

thanks,
jirka

Re: [PATCH 12/26] perf tools: Apply percent-limit to cumulative percentage

[Jiri Olsa]

On Fri, May 23, 2014 at 07:04:09PM +0900, Namhyung Kim wrote:

SNIP

> diff --git a/tools/perf/ui/stdio/hist.c b/tools/perf/ui/stdio/hist.c
> index 9f57991025a9..475d2f5c7e16 100644
> --- a/tools/perf/ui/stdio/hist.c
> +++ b/tools/perf/ui/stdio/hist.c
> @@ -461,12 +461,12 @@ print_entries:
>  
>  	for (nd = rb_first(&hists->entries); nd; nd = rb_next(nd)) {
>  		struct hist_entry *h = rb_entry(nd, struct hist_entry, rb_node);
> -		float percent = h->stat.period * 100.0 /
> -					hists->stats.total_period;
> +		float percent;
>  
>  		if (h->filtered)
>  			continue;
>  
> +		percent = hist_entry__get_percent_limit(h);
>  		if (percent < min_pcnt)
>  			continue;
>  
> diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
> index ce37f3f8edd9..ad955fbd194f 100644
> --- a/tools/perf/util/hist.h
> +++ b/tools/perf/util/hist.h
> @@ -144,6 +144,7 @@ int hist_entry__sort_snprintf(struct hist_entry *he, char *bf, size_t size,
>  			      struct hists *hists);
>  void hist_entry__free(struct hist_entry *);
>  
> +
>  void hists__output_resort(struct hists *hists);
>  void hists__collapse_resort(struct hists *hists, struct ui_progress *prog);

nit.. extra hunk ;-) I'll squash it

jirka

Re: [PATCH 01/26] perf tools: Introduce struct hist_entry_iter

[Jiri Olsa]

On Fri, May 23, 2014 at 07:03:58PM +0900, Namhyung Kim wrote:

SNIP

>  static int process_sample_event(struct perf_tool *tool,
>  				union perf_event *event,
>  				struct perf_sample *sample,
> @@ -243,6 +84,9 @@ static int process_sample_event(struct perf_tool *tool,
>  {
>  	struct report *rep = container_of(tool, struct report, tool);
>  	struct addr_location al;
> +	struct hist_entry_iter iter = {
> +		.hide_unresolved = rep->hide_unresolved,
> +	};
>  	int ret;
>  
>  	if (perf_event__preprocess_sample(event, machine, &al, sample) < 0) {
> @@ -257,22 +101,22 @@ static int process_sample_event(struct perf_tool *tool,
>  	if (rep->cpu_list && !test_bit(sample->cpu, rep->cpu_bitmap))
>  		return 0;
>  
> -	if (sort__mode == SORT_MODE__BRANCH) {
> -		ret = report__add_branch_hist_entry(rep, &al, sample, evsel);
> -		if (ret < 0)
> -			pr_debug("problem adding lbr entry, skipping event\n");
> -	} else if (rep->mem_mode == 1) {
> -		ret = report__add_mem_hist_entry(rep, &al, sample, evsel);
> -		if (ret < 0)
> -			pr_debug("problem adding mem entry, skipping event\n");
> -	} else {
> -		if (al.map != NULL)
> -			al.map->dso->hit = 1;
> -
> -		ret = report__add_hist_entry(rep, evsel, &al, sample);
> -		if (ret < 0)
> -			pr_debug("problem incrementing symbol period, skipping event\n");
> -	}
> +	if (sort__mode == SORT_MODE__BRANCH)
> +		iter.ops = &hist_iter_branch;
> +	else if (rep->mem_mode)
> +		iter.ops = &hist_iter_mem;
> +	else
> +		iter.ops = &hist_iter_normal;
> +
> +	if (al.map != NULL)
> +		al.map->dso->hit = 1;
> +
> +	rep->nr_entries++;

hum, should we call report__inc_stats instead? aaand...

> +
> +	ret = hist_entry_iter__add(&iter, &al, evsel, sample, rep->max_stack);
> +	if (ret < 0)
> +		pr_debug("problem adding hist entry, skipping event\n");
> +

... in here? 

>  	return ret;
>  }


jirka

Re: [PATCH 01/26] perf tools: Introduce struct hist_entry_iter

[Jiri Olsa]

On Fri, May 23, 2014 at 07:03:58PM +0900, Namhyung Kim wrote:
> There're some duplicate code when adding hist entries.  They are
> different in that some have branch info or mem info but generally do
> same thing.  So introduce new struct hist_entry_iter and add callbacks
> to customize each case in general way.
> 
> The new perf_evsel__add_entry() function will look like:
> 
>   iter->prepare_entry();
>   iter->add_single_entry();
> 
>   while (iter->next_entry())
>     iter->add_next_entry();
> 
>   iter->finish_entry();
> 
> This will help further work like the cumulative callchain patchset.
> 
> Tested-by: Arun Sharma <asharma@fb.com>
> Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
> Acked-by: Jiri Olsa <jolsa@redhat.com>
> Cc: David Ahern <dsahern@gmail.com>
> Cc: Stephane Eranian <eranian@google.com>
> Cc: Frederic Weisbecker <fweisbec@gmail.com>
> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
> ---
>  tools/perf/builtin-report.c     | 194 +++----------------------
>  tools/perf/tests/hists_filter.c |  18 +--
>  tools/perf/tests/hists_output.c |  11 +-
>  tools/perf/util/hist.c          | 303 ++++++++++++++++++++++++++++++++++++++++
>  tools/perf/util/hist.h          |  33 +++++
>  5 files changed, 371 insertions(+), 188 deletions(-)
> 
> diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
> index bc0eec1ce4be..d224f5961faa 100644
> --- a/tools/perf/builtin-report.c
> +++ b/tools/perf/builtin-report.c
> @@ -76,165 +76,6 @@ static int report__config(const char *var, const char *value, void *cb)
>  	return perf_default_config(var, value, cb);
>  }
>  
> -static void report__inc_stats(struct report *rep, struct hist_entry *he)
> -{
> -	/*
> -	 * The @he is either of a newly created one or an existing one
> -	 * merging current sample.  We only want to count a new one so
> -	 * checking ->nr_events being 1.
> -	 */
> -	if (he->stat.nr_events == 1)
> -		rep->nr_entries++;
> -

hm.. above is report specific counter update

while below code is repeated for each iterator in finish_entry,
maybe we should separated this and put below into a function
called from generic part

> -	/*
> -	 * Only counts number of samples at this stage as it's more
> -	 * natural to do it here and non-sample events are also
> -	 * counted in perf_session_deliver_event().  The dump_trace
> -	 * requires this info is ready before going to the output tree.
> -	 */
> -	hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
> -	if (!he->filtered)
> -		he->hists->stats.nr_non_filtered_samples++;
> -}

jirka

Re: [PATCH 01/26] perf tools: Introduce struct hist_entry_iter

[Namhyung Kim]

Hi Jiri,

On Mon, 26 May 2014 20:44:25 +0200, Jiri Olsa wrote:
> On Fri, May 23, 2014 at 07:03:58PM +0900, Namhyung Kim wrote:
>> There're some duplicate code when adding hist entries.  They are
>> different in that some have branch info or mem info but generally do
>> same thing.  So introduce new struct hist_entry_iter and add callbacks
>> to customize each case in general way.
>> 
>> The new perf_evsel__add_entry() function will look like:
>> 
>>   iter->prepare_entry();
>>   iter->add_single_entry();
>> 
>>   while (iter->next_entry())
>>     iter->add_next_entry();
>> 
>>   iter->finish_entry();
>> 
>> This will help further work like the cumulative callchain patchset.
>> 
>> Tested-by: Arun Sharma <asharma@fb.com>
>> Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
>> Acked-by: Jiri Olsa <jolsa@redhat.com>
>> Cc: David Ahern <dsahern@gmail.com>
>> Cc: Stephane Eranian <eranian@google.com>
>> Cc: Frederic Weisbecker <fweisbec@gmail.com>
>> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
>> ---
>>  tools/perf/builtin-report.c     | 194 +++----------------------
>>  tools/perf/tests/hists_filter.c |  18 +--
>>  tools/perf/tests/hists_output.c |  11 +-
>>  tools/perf/util/hist.c          | 303 ++++++++++++++++++++++++++++++++++++++++
>>  tools/perf/util/hist.h          |  33 +++++
>>  5 files changed, 371 insertions(+), 188 deletions(-)
>> 
>> diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
>> index bc0eec1ce4be..d224f5961faa 100644
>> --- a/tools/perf/builtin-report.c
>> +++ b/tools/perf/builtin-report.c
>> @@ -76,165 +76,6 @@ static int report__config(const char *var, const char *value, void *cb)
>>  	return perf_default_config(var, value, cb);
>>  }
>>  
>> -static void report__inc_stats(struct report *rep, struct hist_entry *he)
>> -{
>> -	/*
>> -	 * The @he is either of a newly created one or an existing one
>> -	 * merging current sample.  We only want to count a new one so
>> -	 * checking ->nr_events being 1.
>> -	 */
>> -	if (he->stat.nr_events == 1)
>> -		rep->nr_entries++;
>> -
>
> hm.. above is report specific counter update
>
> while below code is repeated for each iterator in finish_entry,
> maybe we should separated this and put below into a function
> called from generic part

Right.  I'll separate the generic part into a function.

The problem of calling report__inc_stats() in process_sample_event() is
that the hist_entry is no longer available.  And each mode has slightly
different behavior when accounting entries so it's inaccurate at this
stage.  Once we have the callback mechanism later in this series, it'll
be solved.

Thanks,
Namhyung

>
>> -	/*
>> -	 * Only counts number of samples at this stage as it's more
>> -	 * natural to do it here and non-sample events are also
>> -	 * counted in perf_session_deliver_event().  The dump_trace
>> -	 * requires this info is ready before going to the output tree.
>> -	 */
>> -	hists__inc_nr_events(he->hists, PERF_RECORD_SAMPLE);
>> -	if (!he->filtered)
>> -		he->hists->stats.nr_non_filtered_samples++;
>> -}
>
> jirka

Re: [PATCH 10/26] perf ui/browser: Add support to accumulated hist stat

[Namhyung Kim]

On Mon, 26 May 2014 19:52:26 +0200, Jiri Olsa wrote:
> On Fri, May 23, 2014 at 07:04:07PM +0900, Namhyung Kim wrote:
>> Print accumulated stat of a hist entry if requested.
>> 
>> Tested-by: Arun Sharma <asharma@fb.com>
>> Tested-by: Rodrigo Campos <rodrigo@sdfg.com.ar>
>> Acked-by: Jiri Olsa <jolsa@redhat.com>
>> Cc: Frederic Weisbecker <fweisbec@gmail.com>
>> Signed-off-by: Namhyung Kim <namhyung@kernel.org>
>> ---
>>  tools/perf/ui/browsers/hists.c | 26 +++++++++++++++++++++++++-
>>  1 file changed, 25 insertions(+), 1 deletion(-)
>> 
>> diff --git a/tools/perf/ui/browsers/hists.c b/tools/perf/ui/browsers/hists.c
>> index 1c331b934ffc..1de7428a9771 100644
>> --- a/tools/perf/ui/browsers/hists.c
>> +++ b/tools/perf/ui/browsers/hists.c
>> @@ -632,7 +632,6 @@ static int __hpp__slsmg_color_printf(struct perf_hpp *hpp, const char *fmt, ...)
>>  	ret = scnprintf(hpp->buf, hpp->size, fmt, percent);
>>  	slsmg_printf("%s", hpp->buf);
>>  
>> -	advance_hpp(hpp, ret);
>>  	return ret;
>
> Whats the purpose of this one..? does not fit to what's in changelog 

Hmm.. looks like a rebase artifact.  It's not a part of this patch.
I'll remove it.

Thanks,
Namhyung