perf trace: Setup struct syscall_tp for syscalls:sys_{enter,exit}_NAME events
authorArnaldo Carvalho de Melo <acme@redhat.com>
Thu, 2 Aug 2018 18:09:24 +0000 (15:09 -0300)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Thu, 2 Aug 2018 18:38:57 +0000 (15:38 -0300)
Mapping "__syscall_nr" to "id" and setting up "args" from the offset of
"__syscall_nr" + sizeof(u64), as the payload for syscalls:* is the same
as for raw_syscalls:*, just the fields have different names.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: https://lkml.kernel.org/n/tip-ogeenrpviwcpwl3oy1l55f3m@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/builtin-trace.c

index 7fca844ced0b2e4ed997b0f7ad69db2f7510420b..41799596c04567e430b3cc379714dbd43e992761 100644 (file)
@@ -247,6 +247,22 @@ static void perf_evsel__delete_priv(struct perf_evsel *evsel)
        perf_evsel__delete(evsel);
 }
 
+static int perf_evsel__init_syscall_tp(struct perf_evsel *evsel)
+{
+       struct syscall_tp *sc = evsel->priv = malloc(sizeof(struct syscall_tp));
+
+       if (evsel->priv != NULL) {
+               if (perf_evsel__init_tp_uint_field(evsel, &sc->id, "__syscall_nr"))
+                       goto out_delete;
+               return 0;
+       }
+
+       return -ENOMEM;
+out_delete:
+       zfree(&evsel->priv);
+       return -ENOENT;
+}
+
 static int perf_evsel__init_raw_syscall_tp(struct perf_evsel *evsel, void *handler)
 {
        evsel->priv = malloc(sizeof(struct syscall_tp));
@@ -2977,6 +2993,36 @@ static void evlist__set_evsel_handler(struct perf_evlist *evlist, void *handler)
                evsel->handler = handler;
 }
 
+static int evlist__set_syscall_tp_fields(struct perf_evlist *evlist)
+{
+       struct perf_evsel *evsel;
+
+       evlist__for_each_entry(evlist, evsel) {
+               if (evsel->priv || !evsel->tp_format)
+                       continue;
+
+               if (strcmp(evsel->tp_format->system, "syscalls"))
+                       continue;
+
+               if (perf_evsel__init_syscall_tp(evsel))
+                       return -1;
+
+               if (!strncmp(evsel->tp_format->name, "sys_enter_", 10)) {
+                       struct syscall_tp *sc = evsel->priv;
+
+                       if (__tp_field__init_ptr(&sc->args, sc->id.offset + sizeof(u64)))
+                               return -1;
+               } else if (!strncmp(evsel->tp_format->name, "sys_exit_", 9)) {
+                       struct syscall_tp *sc = evsel->priv;
+
+                       if (__tp_field__init_uint(&sc->ret, sizeof(u64), sc->id.offset + sizeof(u64), evsel->needs_swap))
+                               return -1;
+               }
+       }
+
+       return 0;
+}
+
 /*
  * XXX: Hackish, just splitting the combined -e+--event (syscalls
  * (raw_syscalls:{sys_{enter,exit}} + events (tracepoints, HW, SW, etc) to use
@@ -3236,8 +3282,13 @@ int cmd_trace(int argc, const char **argv)
                symbol_conf.use_callchain = true;
        }
 
-       if (trace.evlist->nr_entries > 0)
+       if (trace.evlist->nr_entries > 0) {
                evlist__set_evsel_handler(trace.evlist, trace__event_handler);
+               if (evlist__set_syscall_tp_fields(trace.evlist)) {
+                       perror("failed to set syscalls:* tracepoint fields");
+                       goto out;
+               }
+       }
 
        if ((argc >= 1) && (strcmp(argv[0], "record") == 0))
                return trace__record(&trace, argc-1, &argv[1]);