perf stat: Choose the most disaggregate command line option

When multiple aggregation options are passed to perf stat the behavior
isn't clear. Consider "perf stat -A --per-socket .." and "perf stat
--per-socket -A ..", the first won't aggregate at all while the second
will do per-socket aggregation, even though the same options were
passed.

Rather than set an enum value, gather the options in a struct and
process them from most to least aggregate. This ensures the least
aggregate option always applies, so no aggregation if "-A" is passed.

Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Stephane Eranian <eranian@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Link: https://lore.kernel.org/r/20240605063828.195700-2-irogers@google.com

diff --git a/tools/perf/builtin-stat.c b/tools/perf/builtin-stat.c
index 6227b25d14468a71384c489b23ebda6beda5067a..603a9684153d80b0e294f74321b4eb700dca03d9 100644 (file)
--- a/tools/perf/builtin-stat.c
+++ b/tools/perf/builtin-stat.c
@@ -164,6 +164,35 @@ static struct perf_stat_config stat_config = {
        .iostat_run             = false,
 };
 
+/* Options set from the command line. */
+struct opt_aggr_mode {
+       bool node, socket, die, cluster, cache, core, thread, no_aggr;
+};
+
+/* Turn command line option into most generic aggregation mode setting. */
+static enum aggr_mode opt_aggr_mode_to_aggr_mode(struct opt_aggr_mode *opt_mode)
+{
+       enum aggr_mode mode = AGGR_GLOBAL;
+
+       if (opt_mode->node)
+               mode = AGGR_NODE;
+       if (opt_mode->socket)
+               mode = AGGR_SOCKET;
+       if (opt_mode->die)
+               mode = AGGR_DIE;
+       if (opt_mode->cluster)
+               mode = AGGR_CLUSTER;
+       if (opt_mode->cache)
+               mode = AGGR_CACHE;
+       if (opt_mode->core)
+               mode = AGGR_CORE;
+       if (opt_mode->thread)
+               mode = AGGR_THREAD;
+       if (opt_mode->no_aggr)
+               mode = AGGR_NONE;
+       return mode;
+}
+
 static void evlist__check_cpu_maps(struct evlist *evlist)
 {
        struct evsel *evsel, *warned_leader = NULL;
@@ -1096,7 +1125,7 @@ static int parse_cache_level(const struct option *opt,
                             int unset __maybe_unused)
 {
        int level;
-       u32 *aggr_mode = (u32 *)opt->value;
+       struct opt_aggr_mode *opt_aggr_mode = (struct opt_aggr_mode *)opt->value;
        u32 *aggr_level = (u32 *)opt->data;
 
        /*
@@ -1135,7 +1164,7 @@ static int parse_cache_level(const struct option *opt,
                return -EINVAL;
        }
 out:
-       *aggr_mode = AGGR_CACHE;
+       opt_aggr_mode->cache = true;
        *aggr_level = level;
        return 0;
 }
@@ -2101,13 +2130,15 @@ static void init_features(struct perf_session *session)
        perf_header__clear_feat(&session->header, HEADER_AUXTRACE);
 }
 
-static int __cmd_record(const struct option stat_options[], int argc, const char **argv)
+static int __cmd_record(const struct option stat_options[], struct opt_aggr_mode *opt_mode,
+                       int argc, const char **argv)
 {
        struct perf_session *session;
        struct perf_data *data = &perf_stat.data;
 
        argc = parse_options(argc, argv, stat_options, stat_record_usage,
                             PARSE_OPT_STOP_AT_NON_OPTION);
+       stat_config.aggr_mode = opt_aggr_mode_to_aggr_mode(opt_mode);
 
        if (output_name)
                data->path = output_name;
@@ -2350,6 +2381,7 @@ static void setup_system_wide(int forks)
 
 int cmd_stat(int argc, const char **argv)
 {
+       struct opt_aggr_mode opt_mode = {};
        struct option stat_options[] = {
                OPT_BOOLEAN('T', "transaction", &transaction_run,
                        "hardware transaction statistics"),
@@ -2393,10 +2425,10 @@ int cmd_stat(int argc, const char **argv)
                                stat__set_big_num),
                OPT_STRING('C', "cpu", &target.cpu_list, "cpu",
                        "list of cpus to monitor in system-wide"),
-               OPT_SET_UINT('A', "no-aggr", &stat_config.aggr_mode,
-                       "disable aggregation across CPUs or PMUs", AGGR_NONE),
-               OPT_SET_UINT(0, "no-merge", &stat_config.aggr_mode,
-                       "disable aggregation the same as -A or -no-aggr", AGGR_NONE),
+               OPT_BOOLEAN('A', "no-aggr", &opt_mode.no_aggr,
+                       "disable aggregation across CPUs or PMUs"),
+               OPT_BOOLEAN(0, "no-merge", &opt_mode.no_aggr,
+                       "disable aggregation the same as -A or -no-aggr"),
                OPT_BOOLEAN(0, "hybrid-merge", &stat_config.hybrid_merge,
                        "Merge identical named hybrid events"),
                OPT_STRING('x', "field-separator", &stat_config.csv_sep, "separator",
@@ -2424,21 +2456,18 @@ int cmd_stat(int argc, const char **argv)
                        "clear screen in between new interval"),
                OPT_UINTEGER(0, "timeout", &stat_config.timeout,
                        "stop workload and print counts after a timeout period in ms (>= 10ms)"),
-               OPT_SET_UINT(0, "per-socket", &stat_config.aggr_mode,
-                       "aggregate counts per processor socket", AGGR_SOCKET),
-               OPT_SET_UINT(0, "per-die", &stat_config.aggr_mode,
-                       "aggregate counts per processor die", AGGR_DIE),
-               OPT_SET_UINT(0, "per-cluster", &stat_config.aggr_mode,
-                       "aggregate counts per processor cluster", AGGR_CLUSTER),
-               OPT_CALLBACK_OPTARG(0, "per-cache", &stat_config.aggr_mode, &stat_config.aggr_level,
+               OPT_BOOLEAN(0, "per-socket", &opt_mode.socket,
+                       "aggregate counts per processor socket"),
+               OPT_BOOLEAN(0, "per-die", &opt_mode.die, "aggregate counts per processor die"),
+               OPT_BOOLEAN(0, "per-cluster", &opt_mode.cluster,
+                       "aggregate counts per processor cluster"),
+               OPT_CALLBACK_OPTARG(0, "per-cache", &opt_mode, &stat_config.aggr_level,
                                "cache level", "aggregate count at this cache level (Default: LLC)",
                                parse_cache_level),
-               OPT_SET_UINT(0, "per-core", &stat_config.aggr_mode,
-                       "aggregate counts per physical processor core", AGGR_CORE),
-               OPT_SET_UINT(0, "per-thread", &stat_config.aggr_mode,
-                       "aggregate counts per thread", AGGR_THREAD),
-               OPT_SET_UINT(0, "per-node", &stat_config.aggr_mode,
-                       "aggregate counts per numa node", AGGR_NODE),
+               OPT_BOOLEAN(0, "per-core", &opt_mode.core,
+                       "aggregate counts per physical processor core"),
+               OPT_BOOLEAN(0, "per-thread", &opt_mode.thread, "aggregate counts per thread"),
+               OPT_BOOLEAN(0, "per-node", &opt_mode.node, "aggregate counts per numa node"),
                OPT_INTEGER('D', "delay", &target.initial_delay,
                        "ms to wait before starting measurement after program start (-1: start with events disabled)"),
                OPT_CALLBACK_NOOPT(0, "metric-only", &stat_config.metric_only, NULL,
@@ -2521,6 +2550,8 @@ int cmd_stat(int argc, const char **argv)
                                        (const char **) stat_usage,
                                        PARSE_OPT_STOP_AT_NON_OPTION);
 
+       stat_config.aggr_mode = opt_aggr_mode_to_aggr_mode(&opt_mode);
+
        if (stat_config.csv_sep) {
                stat_config.csv_output = true;
                if (!strcmp(stat_config.csv_sep, "\\t"))
@@ -2529,7 +2560,7 @@ int cmd_stat(int argc, const char **argv)
                stat_config.csv_sep = DEFAULT_SEPARATOR;
 
        if (argc && strlen(argv[0]) > 2 && strstarts("record", argv[0])) {
-               argc = __cmd_record(stat_options, argc, argv);
+               argc = __cmd_record(stat_options, &opt_mode, argc, argv);
                if (argc < 0)
                        return -1;
        } else if (argc && strlen(argv[0]) > 2 && strstarts("report", argv[0]))