perf test: Add test for counting open syscalls
Arnaldo Carvalho de Melo [Tue, 4 Jan 2011 02:16:20 +0000 (00:16 -0200)]
To test the use of the perf_evsel class on something other than
the tools from where we refactored code to create it.

It calls open() N times and then checks if the event created to
monitor it returns N events.

[acme@felicio linux]$ perf test
 1: vmlinux symtab matches kallsyms: Ok
 2: detect open syscall event: Ok
[acme@felicio linux]$

It does.

Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Han Pingtian <phan@redhat.com>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Paul Mackerras <paulus@samba.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Tom Zanussi <tzanussi@gmail.com>
LKML-Reference: <new-submission>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

tools/perf/builtin-test.c

index e0c3f47..6c99152 100644 (file)
@@ -234,6 +234,85 @@ out:
        return err;
 }
 
+#include "util/evsel.h"
+#include <sys/types.h>
+
+static int trace_event__id(const char *event_name)
+{
+       char *filename;
+       int err = -1, fd;
+
+       if (asprintf(&filename,
+                    "/sys/kernel/debug/tracing/events/syscalls/%s/id",
+                    event_name) < 0)
+               return -1;
+
+       fd = open(filename, O_RDONLY);
+       if (fd >= 0) {
+               char id[16];
+               if (read(fd, id, sizeof(id)) > 0)
+                       err = atoi(id);
+               close(fd);
+       }
+
+       free(filename);
+       return err;
+}
+
+static int test__open_syscall_event(void)
+{
+       int err = -1, fd;
+       struct thread_map *threads;
+       struct perf_evsel *evsel;
+       unsigned int nr_open_calls = 111, i;
+       int id = trace_event__id("sys_enter_open");
+
+       if (id < 0) {
+               pr_debug("trace_event__id(\"sys_enter_open\") ");
+               return -1;
+       }
+
+       threads = thread_map__new(-1, getpid());
+       if (threads == NULL) {
+               pr_debug("thread_map__new ");
+               return -1;
+       }
+
+       evsel = perf_evsel__new(PERF_TYPE_TRACEPOINT, id, 0);
+       if (evsel == NULL) {
+               pr_debug("perf_evsel__new ");
+               goto out_thread_map_delete;
+       }
+
+       if (perf_evsel__open_per_thread(evsel, threads) < 0) {
+               pr_debug("perf_evsel__open_per_thread ");
+               goto out_evsel_delete;
+       }
+
+       for (i = 0; i < nr_open_calls; ++i) {
+               fd = open("/etc/passwd", O_RDONLY);
+               close(fd);
+       }
+
+       if (perf_evsel__read_on_cpu(evsel, 0, 0) < 0) {
+               pr_debug("perf_evsel__open_read_on_cpu ");
+               goto out_close_fd;
+       }
+
+       if (evsel->counts->cpu[0].val != nr_open_calls)
+               pr_debug("perf_evsel__read_on_cpu: expected to intercept %d calls, got %Ld ",
+                        nr_open_calls, evsel->counts->cpu[0].val);
+       
+       err = 0;
+out_close_fd:
+       perf_evsel__close_fd(evsel, 1, threads->nr);
+out_evsel_delete:
+       perf_evsel__delete(evsel);
+out_thread_map_delete:
+       thread_map__delete(threads);
+       return err;
+}
+
 static struct test {
        const char *desc;
        int (*func)(void);
@@ -243,6 +322,10 @@ static struct test {
                .func = test__vmlinux_matches_kallsyms,
        },
        {
+               .desc = "detect open syscall event",
+               .func = test__open_syscall_event,
+       },
+       {
                .func = NULL,
        },
 };