perf evsel: Move bpf_counter__install_pe() to success path in evsel__open_cpu()

I don't see why bpf_counter__install_pe() should get called even if
fd = -1, so I'm moving it to the success path.

This will be useful in following patches to separate the actual open and
the related operations from the fallback mechanisms.
Signed-off-by: NRiccardo Mancini <rickyman7@gmail.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <songliubraving@fb.com>
Link: http://lore.kernel.org/lkml/64f8a1b0a838a6e6049cd43c1beafd432999ae57.1629490974.git.rickyman7@gmail.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>

tools/perf/util/evsel.c

@@ -2009,8 +2009,6 @@ static int evsel__open_cpu(struct evsel *evsel, struct perf_cpu_map *cpus,
 
 			FD(evsel, cpu, thread) = fd;
 
-			bpf_counter__install_pe(evsel, cpu, fd);
-
 			if (fd < 0) {
 				err = -errno;
 
@@ -2019,6 +2017,8 @@ static int evsel__open_cpu(struct evsel *evsel, struct perf_cpu_map *cpus,
 				goto try_fallback;
 			}
 
+			bpf_counter__install_pe(evsel, cpu, fd);
+
 			if (unlikely(test_attr__enabled)) {
 				test_attr__open(&evsel->core.attr, pid, cpus->map[cpu],
 						fd, group_fd, evsel->open_flags);