Skip to contentSkip to conversation

perf/urgent fixes and improvements

GIT PULL New Arnaldo Carvalho de Melo 2 participants 5 messages Dec 23, 2014 – Jan 1, 2015 perf events

lore
Share & tools
lore

Arnaldo Carvalho de Melo

author maintainer
#0 lore raw

Hi Ingo,

Please consider pulling,

- Arnaldo

The following changes since commit ac931f87a647ca156f65a4c00e7297165e4fa2d8:

perf: Fix building warning on ARM 32 (2014-12-19 13:09:43 +0100)

are available in the git repository at:

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

for you to fetch changes up to 82aa019e0098a1e0801df94345c0297448323126:

perf callchain: Append callchains only when requested (2014-12-23 12:06:39 -0300)

----------------------------------------------------------------
perf/urgent fixes and improvements:

User visible:

- Show progress bar in more places while doing histogram processing in the hists browser (Namhyung Kim)

- Print backtrace symbols when segfault occurs in 'report' (Namhyung Kim)

Infrastructure:

- Append callchains only when requested (Namhyung Kim)

Signed-off-by Arnaldo Carvalho de Melo <[email protected]>

----------------------------------------------------------------
Namhyung Kim (3):
perf report: Show progress bar for output resorting perf ui/tui: Print backtrace symbols when segfault occurs perf callchain: Append callchains only when requested

 tools/perf/builtin-annotate.c     |  2 +-
 tools/perf/builtin-diff.c         |  2 +-
 tools/perf/builtin-report.c       | 24 ++++++++++++++++++++++--
 tools/perf/builtin-top.c          |  4 ++--
 tools/perf/tests/hists_cumulate.c |  2 +-
 tools/perf/tests/hists_filter.c   |  2 +-
 tools/perf/tests/hists_output.c   | 10 +++++-----
 tools/perf/ui/tui/setup.c         | 26 ++++++++++++++++++++++++--
 tools/perf/util/hist.c            | 17 +++++++++++++----
 tools/perf/util/hist.h            |  2 +-
 10 files changed, 71 insertions(+), 20 deletions(-)
author maintainer
#1 lore raw

PATCH 1/3 perf report: Show progress bar for output resorting

From: Namhyung Kim <[email protected]>

Sometimes it takes a long time to resort hist entries for output in case of a large data file. Show a progress bar window and inform user.

Signed-off-by Namhyung Kim <[email protected]>
Cc Adrian Hunter <[email protected]>
Cc David Ahern <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Stephane Eranian <[email protected]>
Link http://lkml.kernel.org/r/[email protected] open here
Signed-off-by Arnaldo Carvalho de Melo <[email protected]>
---
 tools/perf/builtin-annotate.c     |  2 +-
 tools/perf/builtin-diff.c         |  2 +-
 tools/perf/builtin-report.c       | 24 ++++++++++++++++++++++--
 tools/perf/builtin-top.c          |  4 ++--
 tools/perf/tests/hists_cumulate.c |  2 +-
 tools/perf/tests/hists_filter.c   |  2 +-
 tools/perf/tests/hists_output.c   | 10 +++++-----
 tools/perf/util/hist.c            | 10 +++++++++-
 tools/perf/util/hist.h            |  2 +-
 9 files changed, 43 insertions(+), 15 deletions(-)
tools/perf/builtin-annotate.c +1 -1
index e7417fe97a97..747f86103599 100644
@@ -232,7 +232,7 @@ static int __cmd_annotate(struct perf_annotate *ann)
 		if (nr_samples > 0) {
 			total_nr_samples += nr_samples;
 			hists__collapse_resort(hists, NULL);
-			hists__output_resort(hists);
+			hists__output_resort(hists, NULL);
 
 			if (symbol_conf.event_group &&
 			    !perf_evsel__is_group_leader(pos))
tools/perf/builtin-diff.c +1 -1
index 1ce425d101a9..303c1e151dcf 100644
@@ -605,7 +605,7 @@ static void hists__process(struct hists *hists)
 		hists__precompute(hists);
 		hists__compute_resort(hists);
 	} else {
-		hists__output_resort(hists);
+		hists__output_resort(hists, NULL);
 	}
 
 	hists__fprintf(hists, true, 0, 0, 0, stdout);
tools/perf/builtin-report.c +22 -2
index 39367609c707..072ae8ad67fc 100644
@@ -457,6 +457,19 @@ static void report__collapse_hists(struct report *rep)
 	ui_progress__finish();
 }
 
+static void report__output_resort(struct report *rep)
+{
+	struct ui_progress prog;
+	struct perf_evsel *pos;
+
+	ui_progress__init(&prog, rep->nr_entries, "Sorting events for output...");
+
+	evlist__for_each(rep->session->evlist, pos)
+		hists__output_resort(evsel__hists(pos), &prog);
+
+	ui_progress__finish();
+}
+
 static int __cmd_report(struct report *rep)
 {
 	int ret;

@@ -505,13 +518,20 @@ static int __cmd_report(struct report *rep)
 	if (session_done())
 		return 0;
 
+	/*
+	 * recalculate number of entries after collapsing since it
+	 * might be changed during the collapse phase.
+	 */
+	rep->nr_entries = 0;
+	evlist__for_each(session->evlist, pos)
+		rep->nr_entries += evsel__hists(pos)->nr_entries;
+
 	if (rep->nr_entries == 0) {
 		ui__error("The %s file has no samples!\n", file->path);
 		return 0;
 	}
 
-	evlist__for_each(session->evlist, pos)
-		hists__output_resort(evsel__hists(pos));
+	report__output_resort(rep);
 
 	return report__browse_hists(rep);
 }
tools/perf/builtin-top.c +2 -2
index 0aa7747ff139..961cea183a83 100644
@@ -285,7 +285,7 @@ static void perf_top__print_sym_table(struct perf_top *top)
 	}
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	hists__output_recalc_col_len(hists, top->print_entries - printed);
 	putchar('\n');

@@ -554,7 +554,7 @@ static void perf_top__sort_new_samples(void *arg)
 	}
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 }
 
 static void *display_thread_tui(void *arg)
tools/perf/tests/hists_cumulate.c +1 -1
index 614d5c4978ab..4b8226e19a91 100644
@@ -187,7 +187,7 @@ static int do_test(struct hists *hists, struct result *expected, size_t nr_expec
 	 * function since TEST_ASSERT_VAL() returns in case of failure.
 	 */
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("use callchain: %d, cumulate callchain: %d\n",
tools/perf/tests/hists_filter.c +1 -1
index 74f257a81265..59e53db7914c 100644
@@ -138,7 +138,7 @@ int test__hists_filter(void)
 		struct hists *hists = evsel__hists(evsel);
 
 		hists__collapse_resort(hists, NULL);
-		hists__output_resort(hists);
+		hists__output_resort(hists, NULL);
 
 		if (verbose > 2) {
 			pr_info("Normal histogram\n");
tools/perf/tests/hists_output.c +5 -5
index a748f2be1222..f5547610da02 100644
@@ -152,7 +152,7 @@ static int test1(struct perf_evsel *evsel, struct machine *machine)
 		goto out;
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("[fields = %s, sort = %s]\n", field_order, sort_order);

@@ -252,7 +252,7 @@ static int test2(struct perf_evsel *evsel, struct machine *machine)
 		goto out;
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("[fields = %s, sort = %s]\n", field_order, sort_order);

@@ -306,7 +306,7 @@ static int test3(struct perf_evsel *evsel, struct machine *machine)
 		goto out;
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("[fields = %s, sort = %s]\n", field_order, sort_order);

@@ -384,7 +384,7 @@ static int test4(struct perf_evsel *evsel, struct machine *machine)
 		goto out;
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("[fields = %s, sort = %s]\n", field_order, sort_order);

@@ -487,7 +487,7 @@ static int test5(struct perf_evsel *evsel, struct machine *machine)
 		goto out;
 
 	hists__collapse_resort(hists, NULL);
-	hists__output_resort(hists);
+	hists__output_resort(hists, NULL);
 
 	if (verbose > 2) {
 		pr_info("[fields = %s, sort = %s]\n", field_order, sort_order);
tools/perf/util/hist.c +9 -1
index 6e88b9e395df..1cc6ea4b7065 100644
@@ -6,6 +6,7 @@
 #include "evlist.h"
 #include "evsel.h"
 #include "annotate.h"
+#include "ui/progress.h"
 #include <math.h>
 
 static bool hists__filter_entry_by_dso(struct hists *hists,

@@ -987,6 +988,7 @@ static bool hists__collapse_insert_entry(struct hists *hists __maybe_unused,
 		else
 			p = &(*p)->rb_right;
 	}
+	hists->nr_entries++;
 
 	rb_link_node(&he->rb_node_in, parent, p);
 	rb_insert_color(&he->rb_node_in, root);

@@ -1024,7 +1026,10 @@ void hists__collapse_resort(struct hists *hists, struct ui_progress *prog)
 	if (!sort__need_collapse)
 		return;
 
+	hists->nr_entries = 0;
+
 	root = hists__get_rotate_entries_in(hists);
+
 	next = rb_first(root);
 
 	while (next) {

@@ -1119,7 +1124,7 @@ static void __hists__insert_output_entry(struct rb_root *entries,
 	rb_insert_color(&he->rb_node, entries);
 }
 
-void hists__output_resort(struct hists *hists)
+void hists__output_resort(struct hists *hists, struct ui_progress *prog)
 {
 	struct rb_root *root;
 	struct rb_node *next;

@@ -1148,6 +1153,9 @@ void hists__output_resort(struct hists *hists)
 
 		if (!n->filtered)
 			hists__calc_col_len(hists, n);
+
+		if (prog)
+			ui_progress__update(prog, 1);
 	}
 }
tools/perf/util/hist.h +1 -1
index d0ef9a19a744..46bd50344f85 100644
@@ -121,7 +121,7 @@ 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__output_resort(struct hists *hists, struct ui_progress *prog);
 void hists__collapse_resort(struct hists *hists, struct ui_progress *prog);
 
 void hists__decay_entries(struct hists *hists, bool zap_user, bool zap_kernel);
Signature · 1 line
1.9.3
author maintainer
#2 lore raw

PATCH 2/3 perf ui/tui: Print backtrace symbols when segfault occurs

In reply to Arnaldo Carvalho de Melo

From: Namhyung Kim <[email protected]>

The output will look like below. (I added an error into ui__init() for the test).

  $ perf report
  perf: Segmentation fault
  -------- backtrace --------
  perf[0x503781]
  /usr/lib/libc.so.6(+0x33b20)[0x7f1a14f04b20]
  perf(ui__init+0xd5)[0x503645]
  perf(setup_browser+0x97)[0x4ce4e7]
  perf(cmd_report+0xcea)[0x4392ba]
  perf[0x428493]
  perf(main+0x60a)[0x427c0a]
  /usr/lib/libc.so.6(__libc_start_main+0xf0)[0x7f1a14ef1040]
  perf[0x427d29]
  [0x0]
Signed-off-by Namhyung Kim <[email protected]>
Cc Adrian Hunter <[email protected]>
Cc David Ahern <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Stephane Eranian <[email protected]>
Link http://lkml.kernel.org/r/[email protected] open here
Signed-off-by Arnaldo Carvalho de Melo <[email protected]>
---
 tools/perf/ui/tui/setup.c | 26 ++++++++++++++++++++++++--
 1 file changed, 24 insertions(+), 2 deletions(-)
tools/perf/ui/tui/setup.c +24 -2
index 2f612562978c..3c38f25b1695 100644
@@ -1,5 +1,8 @@
 #include <signal.h>
 #include <stdbool.h>
+#ifdef HAVE_BACKTRACE_SUPPORT
+#include <execinfo.h>
+#endif
 
 #include "../../util/cache.h"
 #include "../../util/debug.h"

@@ -88,6 +91,25 @@ int ui__getch(int delay_secs)
 	return SLkp_getkey();
 }
 
+#ifdef HAVE_BACKTRACE_SUPPORT
+static void ui__signal_backtrace(int sig)
+{
+	void *stackdump[32];
+	size_t size;
+
+	ui__exit(false);
+	psignal(sig, "perf");
+
+	printf("-------- backtrace --------\n");
+	size = backtrace(stackdump, ARRAY_SIZE(stackdump));
+	backtrace_symbols_fd(stackdump, size, STDOUT_FILENO);
+
+	exit(0);
+}
+#else
+# define ui__signal_backtrace  ui__signal
+#endif
+
 static void ui__signal(int sig)
 {
 	ui__exit(false);

@@ -122,8 +144,8 @@ int ui__init(void)
 	ui_browser__init();
 	tui_progress__init();
 
-	signal(SIGSEGV, ui__signal);
-	signal(SIGFPE, ui__signal);
+	signal(SIGSEGV, ui__signal_backtrace);
+	signal(SIGFPE, ui__signal_backtrace);
 	signal(SIGINT, ui__signal);
 	signal(SIGQUIT, ui__signal);
 	signal(SIGTERM, ui__signal);
Signature · 1 line
1.9.3
author maintainer
#3 lore raw

PATCH 3/3 perf callchain: Append callchains only when requested

In reply to Arnaldo Carvalho de Melo

From: Namhyung Kim <[email protected]>

The perf report --children can be called with callchain disabled so no need to append callchains. Actually the root of callchain tree is not initialized properly in this case.

Signed-off-by Namhyung Kim <[email protected]>
Cc Adrian Hunter <[email protected]>
Cc David Ahern <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Peter Zijlstra <[email protected]>
Cc Stephane Eranian <[email protected]>
Link http://lkml.kernel.org/r/[email protected] open here
Signed-off-by Arnaldo Carvalho de Melo <[email protected]>
---
 tools/perf/util/hist.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)
tools/perf/util/hist.c +4 -3
index 1cc6ea4b7065..0ced178ce306 100644
@@ -304,7 +304,7 @@ static struct hist_entry *hist_entry__new(struct hist_entry *template,
 	size_t callchain_size = 0;
 	struct hist_entry *he;
 
-	if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain)
+	if (symbol_conf.use_callchain)
 		callchain_size = sizeof(struct callchain_root);
 
 	he = zalloc(sizeof(*he) + callchain_size);

@@ -737,7 +737,7 @@ 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);
+	hist_entry__append_callchain(he, sample);
 
 	/*
 	 * We need to re-initialize the cursor since callchain_append()

@@ -810,7 +810,8 @@ iter_add_next_cumulative_entry(struct hist_entry_iter *iter,
 	iter->he = he;
 	he_cache[iter->curr++] = he;
 
-	callchain_append(he->callchain, &cursor, sample->period);
+	if (symbol_conf.use_callchain)
+		callchain_append(he->callchain, &cursor, sample->period);
 	return 0;
 }
Signature · 1 line
1.9.3
#4 lore raw
In reply to Arnaldo Carvalho de Melo
Quoted 40 lines · Arnaldo Carvalho de Melo

Hi Ingo,

Please consider pulling,

- Arnaldo

The following changes since commit ac931f87a647ca156f65a4c00e7297165e4fa2d8:

perf: Fix building warning on ARM 32 (2014-12-19 13:09:43 +0100)

are available in the git repository at:

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

for you to fetch changes up to 82aa019e0098a1e0801df94345c0297448323126:

perf callchain: Append callchains only when requested (2014-12-23 12:06:39 -0300)

----------------------------------------------------------------
perf/urgent fixes and improvements:

User visible:

- Show progress bar in more places while doing histogram processing in the hists browser (Namhyung Kim)

- Print backtrace symbols when segfault occurs in 'report' (Namhyung Kim)

Infrastructure:

- Append callchains only when requested (Namhyung Kim)

Signed-off-by Arnaldo Carvalho de Melo <[email protected]>

----------------------------------------------------------------
Namhyung Kim (3):
perf report: Show progress bar for output resorting perf ui/tui: Print backtrace symbols when segfault occurs perf callchain: Append callchains only when requested

Quoted 11 lines
 tools/perf/builtin-annotate.c     |  2 +-
 tools/perf/builtin-diff.c         |  2 +-
 tools/perf/builtin-report.c       | 24 ++++++++++++++++++++++--
 tools/perf/builtin-top.c          |  4 ++--
 tools/perf/tests/hists_cumulate.c |  2 +-
 tools/perf/tests/hists_filter.c   |  2 +-
 tools/perf/tests/hists_output.c   | 10 +++++-----
 tools/perf/ui/tui/setup.c         | 26 ++++++++++++++++++++++++--
 tools/perf/util/hist.c            | 17 +++++++++++++----
 tools/perf/util/hist.h            |  2 +-
 10 files changed, 71 insertions(+), 20 deletions(-)

Pulled into tip:perf/urgent, thanks a lot Arnaldo!

Ingo