perf callchain: Per-event type selection support

This patchkit adds the ability to set callgraph mode (fp, dwarf, lbr) per
event. This in term can reduce sampling overhead and the size of the
perf.data.

Here is an example.

perf record -e 'cpu/cpu-cycles,period=1000,call-graph=fp,time=1/,cpu/instructions,call-graph=lbr/' sleep 1

perf evlist -v
cpu/cpu-cycles,period=1000,call-graph=fp,time=1/: type: 4, size: 112,
config: 0x3c, { sample_period, sample_freq }: 1000, sample_type:
IP|TID|TIME|CALLCHAIN|PERIOD|IDENTIFIER, read_format: ID, disabled: 1,
inherit: 1, mmap: 1, comm: 1, enable_on_exec: 1, task: 1, sample_id_all:
1, exclude_guest: 1, mmap2: 1, comm_exec: 1
cpu/instructions,call-graph=lbr/: type: 4, size: 112, config: 0xc0, {
sample_period, sample_freq }: 4000, sample_type:
IP|TID|TIME|CALLCHAIN|PERIOD|BRANCH_STACK|IDENTIFIER, read_format: ID,
disabled: 1, inherit: 1, freq: 1, enable_on_exec: 1, sample_id_all: 1,
exclude_guest: 1

Signed-off-by: Kan Liang <[e-mail]>
Tested-by: Arnaldo Carvalho de Melo <[e-mail]>
Cc: Andi Kleen <[e-mail]>
Cc: Jiri Olsa <[e-mail]>
Cc: Namhyung Kim <[e-mail]>
Link: http://lkml.kernel.org/r/[e-mail]
Signed-off-by: Arnaldo Carvalho de Melo <[e-mail]>

tools/perf/Documentation/perf-record.txt

···
 	  - 'time': Disable/enable time stamping. Acceptable values are 1 for
 		    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.
+	  - '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/util/evsel.c

···
 	}
 }
 
-static void apply_config_terms(struct perf_evsel *evsel)
+static void
+perf_evsel__reset_callgraph(struct perf_evsel *evsel,
+			    struct callchain_param *param)
+{
+	struct perf_event_attr *attr = &evsel->attr;
+
+	perf_evsel__reset_sample_bit(evsel, CALLCHAIN);
+	if (param->record_mode == CALLCHAIN_LBR) {
+		perf_evsel__reset_sample_bit(evsel, BRANCH_STACK);
+		attr->branch_sample_type &= ~(PERF_SAMPLE_BRANCH_USER |
+					      PERF_SAMPLE_BRANCH_CALL_STACK);
+	}
+	if (param->record_mode == CALLCHAIN_DWARF) {
+		perf_evsel__reset_sample_bit(evsel, REGS_USER);
+		perf_evsel__reset_sample_bit(evsel, STACK_USER);
+	}
+}
+
+static void apply_config_terms(struct perf_evsel *evsel,
+			       struct record_opts *opts)
 {
 	struct perf_evsel_config_term *term;
 	struct list_head *config_terms = &evsel->config_terms;
 	struct perf_event_attr *attr = &evsel->attr;
+	struct callchain_param param;
+	u32 dump_size = 0;
+	char *callgraph_buf = NULL;
+
+	/* callgraph default */
+	param.record_mode = callchain_param.record_mode;
 
 	list_for_each_entry(term, config_terms, list) {
 		switch (term->type) {
···
 			else
 				perf_evsel__reset_sample_bit(evsel, TIME);
 			break;
+		case PERF_EVSEL__CONFIG_TERM_CALLGRAPH:
+			callgraph_buf = term->val.callgraph;
+			break;
+		case PERF_EVSEL__CONFIG_TERM_STACK_USER:
+			dump_size = term->val.stack_user;
+			break;
 		default:
 			break;
 		}
+	}
+
+	/* User explicitly set per-event callgraph, clear the old setting and reset. */
+	if ((callgraph_buf != NULL) || (dump_size > 0)) {
+
+		/* parse callgraph parameters */
+		if (callgraph_buf != NULL) {
+			param.enabled = true;
+			if (parse_callchain_record(callgraph_buf, &param)) {
+				pr_err("per-event callgraph setting for %s failed. "
+				       "Apply callgraph global setting for it\n",
+				       evsel->name);
+				return;
+			}
+		}
+		if (dump_size > 0) {
+			dump_size = round_up(dump_size, sizeof(u64));
+			param.dump_size = dump_size;
+		}
+
+		/* If global callgraph set, clear it */
+		if (callchain_param.enabled)
+			perf_evsel__reset_callgraph(evsel, &callchain_param);
+
+		/* set perf-event callgraph */
+		if (param.enabled)
+			perf_evsel__config_callgraph(evsel, opts, &param);
 	}
 }
 
···
 	 * Apply event specific term settings,
 	 * it overloads any global configuration.
 	 */
-	apply_config_terms(evsel);
+	apply_config_terms(evsel, opts);
 }
 
 static int perf_evsel__alloc_fd(struct perf_evsel *evsel, int ncpus, int nthreads)

tools/perf/util/evsel.h

···
 	PERF_EVSEL__CONFIG_TERM_PERIOD,
 	PERF_EVSEL__CONFIG_TERM_FREQ,
 	PERF_EVSEL__CONFIG_TERM_TIME,
+	PERF_EVSEL__CONFIG_TERM_CALLGRAPH,
+	PERF_EVSEL__CONFIG_TERM_STACK_USER,
 	PERF_EVSEL__CONFIG_TERM_MAX,
 };
 
···
 		u64	period;
 		u64	freq;
 		bool	time;
+		char	*callgraph;
+		u64	stack_user;
 	} val;
 };
 

tools/perf/util/parse-events.c

···
 			return -EINVAL;
 		}
 		break;
+	case PARSE_EVENTS__TERM_TYPE_CALLGRAPH:
+		CHECK_TYPE_VAL(STR);
+		break;
+	case PARSE_EVENTS__TERM_TYPE_STACKSIZE:
+		CHECK_TYPE_VAL(NUM);
+		break;
 	case PARSE_EVENTS__TERM_TYPE_NAME:
 		CHECK_TYPE_VAL(STR);
 		break;
···
 			break;
 		case PARSE_EVENTS__TERM_TYPE_TIME:
 			ADD_CONFIG_TERM(TIME, time, term->val.num);
+			break;
+		case PARSE_EVENTS__TERM_TYPE_CALLGRAPH:
+			ADD_CONFIG_TERM(CALLGRAPH, callgraph, term->val.str);
+			break;
+		case PARSE_EVENTS__TERM_TYPE_STACKSIZE:
+			ADD_CONFIG_TERM(STACK_USER, stack_user, term->val.num);
 			break;
 		default:
 			break;

tools/perf/util/parse-events.h

···
 	PARSE_EVENTS__TERM_TYPE_SAMPLE_FREQ,
 	PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE,
 	PARSE_EVENTS__TERM_TYPE_TIME,
+	PARSE_EVENTS__TERM_TYPE_CALLGRAPH,
+	PARSE_EVENTS__TERM_TYPE_STACKSIZE,
 };
 
 struct parse_events_term {

tools/perf/util/parse-events.l

···
 freq			{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_SAMPLE_FREQ); }
 branch_type		{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_BRANCH_SAMPLE_TYPE); }
 time			{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_TIME); }
+call-graph		{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_CALLGRAPH); }
+stack-size		{ return term(yyscanner, PARSE_EVENTS__TERM_TYPE_STACKSIZE); }
 ,			{ return ','; }
 "/"			{ BEGIN(INITIAL); return '/'; }
 {name_minus}		{ return str(yyscanner, PE_NAME); }

tools/perf/util/pmu.c

···
 {
 	struct perf_pmu_format *format;
 	char *err, *str;
-	static const char *static_terms = "config,config1,config2,name,period,freq,branch_type,time\n";
+	static const char *static_terms = "config,config1,config2,name,"
+					  "period,freq,branch_type,time,"
+					  "call-graph,stack-size\n";
 	unsigned i = 0;
 
 	if (!asprintf(&str, "valid terms:"))