perf callchain: Allow disabling call graphs per event

This patch introduce "call-graph=no" to disable per-event callgraph.

Here is an example.

  perf record -e 'cpu/cpu-cycles,call-graph=fp/,cpu/instructions,call-graph=no/' sleep 1

  perf report --stdio

  # To display the perf.data header info, please use
  --header/--header-only options.
  #
  #
  # Total Lost Samples: 0
  #
  # Samples: 6  of event 'cpu/cpu-cycles,call-graph=fp/'
  # Event count (approx.): 774218
  #
  # Children      Self  Command  Shared Object     Symbol
  # ........  ........  .......  ................  ........................................
  #
    61.94%     0.00%  sleep    [kernel.vmlinux]  [k] entry_SYSCALL_64_fastpath
              |
              ---entry_SYSCALL_64_fastpath
                 |
                 |--97.30%-- __brk
                 |
                  --2.70%-- mmap64
                            _dl_check_map_versions
                            _dl_check_all_versions

    61.94%     0.00%  sleep    [kernel.vmlinux]  [k] perf_event_mmap
              |
              ---perf_event_mmap
                 |
                 |--97.30%-- do_brk
                 |          sys_brk
                 |          entry_SYSCALL_64_fastpath
                 |          __brk
                 |
                  --2.70%-- mmap_region
                            do_mmap_pgoff
                            vm_mmap_pgoff
                            sys_mmap_pgoff
                            sys_mmap
                            entry_SYSCALL_64_fastpath
                            mmap64
                            _dl_check_map_versions
                            _dl_check_all_versions
  ......

  # Samples: 6  of event 'cpu/instructions,call-graph=no/'
  # Event count (approx.): 359692
  #
  # Children      Self  Command  Shared Object     Symbol
  # ........  ........  .......  ................  .................................
  #
     89.03%     0.00%  sleep    [unknown]         [.] 0xffff6598ffff6598
     89.03%     0.00%  sleep    ld-2.17.so        [.] _dl_resolve_conflicts
     89.03%     0.00%  sleep    [kernel.vmlinux]  [k] page_fault

Signed-off-by: Kan Liang <kan.liang@intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lkml.kernel.org/r/1439289050-40510-2-git-send-email-kan.liang@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

tools/perf/Documentation/perf-record.txt

@@ -54,7 +54,8 @@ OPTIONS
 		    enabling time stamping. 0 for disabling time stamping.
 		    The default is 1.
 	  - 'call-graph': Disable/enable callgraph. Acceptable str are "fp" for
-			 FP mode, "dwarf" for DWARF mode, "lbr" for LBR mode.
+			 FP mode, "dwarf" for DWARF mode, "lbr" for LBR mode and
+			 "no" for disable callgraph.
 	  - 'stack-size': user stack size for dwarf mode
 	  Note: If user explicitly sets options which conflict with the params,
 	  the value set by the params will be overridden.

tools/perf/builtin-annotate.c

@@ -239,6 +239,8 @@ static int __cmd_annotate(struct perf_annotate *ann)
 		if (nr_samples > 0) {
 			total_nr_samples += nr_samples;
 			hists__collapse_resort(hists, NULL);
+			/* Don't sort callchain */
+			perf_evsel__reset_sample_bit(pos, CALLCHAIN);
 			hists__output_resort(hists, NULL);
 
 			if (symbol_conf.event_group &&

tools/perf/builtin-diff.c

@@ -722,6 +722,9 @@ static void data_process(void)
 		if (verbose || data__files_cnt > 2)
 			data__fprintf();
 
+		/* Don't sort callchain for perf diff */
+		perf_evsel__reset_sample_bit(evsel_base, CALLCHAIN);
+
 		hists__process(hists_base);
 	}
 }

tools/perf/tests/hists_cumulate.c

@@ -279,6 +279,7 @@ static int test1(struct perf_evsel *evsel, struct machine *machine)
 
 	symbol_conf.use_callchain = false;
 	symbol_conf.cumulate_callchain = false;
+	perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
 
 	setup_sorting();
 	callchain_register_param(&callchain_param);
@@ -425,6 +426,7 @@ static int test2(struct perf_evsel *evsel, struct machine *machine)
 
 	symbol_conf.use_callchain = true;
 	symbol_conf.cumulate_callchain = false;
+	perf_evsel__set_sample_bit(evsel, CALLCHAIN);
 
 	setup_sorting();
 	callchain_register_param(&callchain_param);
@@ -482,6 +484,7 @@ static int test3(struct perf_evsel *evsel, struct machine *machine)
 
 	symbol_conf.use_callchain = false;
 	symbol_conf.cumulate_callchain = true;
+	perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
 
 	setup_sorting();
 	callchain_register_param(&callchain_param);
@@ -665,6 +668,7 @@ static int test4(struct perf_evsel *evsel, struct machine *machine)
 
 	symbol_conf.use_callchain = true;
 	symbol_conf.cumulate_callchain = true;
+	perf_evsel__set_sample_bit(evsel, CALLCHAIN);
 
 	setup_sorting();
 	callchain_register_param(&callchain_param);

tools/perf/util/evsel.c

@@ -651,6 +651,10 @@ static void apply_config_terms(struct perf_evsel *evsel,
 
 		/* parse callgraph parameters */
 		if (callgraph_buf != NULL) {
+			if (!strcmp(callgraph_buf, "no")) {
+				param.enabled = false;
+				param.record_mode = CALLCHAIN_NONE;
+			} else {
 				param.enabled = true;
 				if (parse_callchain_record(callgraph_buf, &param)) {
 					pr_err("per-event callgraph setting for %s failed. "
@@ -659,6 +663,7 @@ static void apply_config_terms(struct perf_evsel *evsel,
 					return;
 				}
 			}
+		}
 		if (dump_size > 0) {
 			dump_size = round_up(dump_size, sizeof(u64));
 			param.dump_size = dump_size;

tools/perf/util/hist.c

@@ -1109,13 +1109,14 @@ void hists__inc_stats(struct hists *hists, struct hist_entry *h)
 
 static void __hists__insert_output_entry(struct rb_root *entries,
 					 struct hist_entry *he,
-					 u64 min_callchain_hits)
+					 u64 min_callchain_hits,
+					 bool use_callchain)
 {
 	struct rb_node **p = &entries->rb_node;
 	struct rb_node *parent = NULL;
 	struct hist_entry *iter;
 
-	if (symbol_conf.use_callchain)
+	if (use_callchain)
 		callchain_param.sort(&he->sorted_chain, he->callchain,
 				      min_callchain_hits, &callchain_param);
 
@@ -1139,6 +1140,8 @@ void hists__output_resort(struct hists *hists, struct ui_progress *prog)
 	struct rb_node *next;
 	struct hist_entry *n;
 	u64 min_callchain_hits;
+	struct perf_evsel *evsel = hists_to_evsel(hists);
+	bool use_callchain = evsel ? (evsel->attr.sample_type & PERF_SAMPLE_CALLCHAIN) : symbol_conf.use_callchain;
 
 	min_callchain_hits = hists->stats.total_period * (callchain_param.min_percent / 100);
 
@@ -1157,7 +1160,7 @@ void hists__output_resort(struct hists *hists, struct ui_progress *prog)
 		n = rb_entry(next, struct hist_entry, rb_node_in);
 		next = rb_next(&n->rb_node_in);
 
-		__hists__insert_output_entry(&hists->entries, n, min_callchain_hits);
+		__hists__insert_output_entry(&hists->entries, n, min_callchain_hits, use_callchain);
 		hists__inc_stats(hists, n);
 
 		if (!n->filtered)