perf trace: Do not print raw args list for syscalls with no args
authorArnaldo Carvalho de Melo <acme@redhat.com>
Fri, 6 May 2016 02:38:05 +0000 (23:38 -0300)
committerArnaldo Carvalho de Melo <acme@redhat.com>
Fri, 6 May 2016 11:44:30 +0000 (08:44 -0300)
The test to check if the arg format had been read from the
syscall:sys_enter_name/format file was looking at the list of non-commom
fields, and if that is empty, it would think it had failed to read it,
because it doesn't exist, for instance, for the clone() syscall.

So instead before dumping the raw syscall args list check
IS_ERR(sc->tp_format), if that is true, then an attempt was made to read
the format file and failed, in which case dump the raw arg list values.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Milian Wolff <milian.wolff@kdab.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/n/tip-ls7pmdqb2xy9339vdburwvnk@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
tools/perf/builtin-trace.c

index 66aa2a00414bd1cedd89aa83f7271cafe8d8fa15..9d23edc06698558231a1da49c465cbf092bdc194 100644 (file)
@@ -1606,7 +1606,12 @@ static size_t syscall__scnprintf_args(struct syscall *sc, char *bf, size_t size,
                                                     "%ld", val);
                        }
                }
-       } else {
+       } else if (IS_ERR(sc->tp_format)) {
+               /*
+                * If we managed to read the tracepoint /format file, then we
+                * may end up not having any args, like with gettid(), so only
+                * print the raw args when we didn't manage to read it.
+                */
                int i = 0;
 
                while (i < 6) {