[PATCH v2] perf test 39 (Session topology) dumps core on s390
From: Thomas Richter
Date: Mon May 28 2018 - 03:37:14 EST
perf test 39 fails with core dump on s390. The root cause
is a NULL pointer reference in function check_cpu_topology()
line 76 (or line 82 without -v).
The session->header.env.cpu variable is NULL because on s390
function process_cpu_topology() returns with error:
socket_id number is too big.
You may need to upgrade the perf tool.
and releases the env.cpu variable via zfree() and sets it to NULL.
Here is the gdb output:
(gdb) n
76 pr_debug("CPU %d, core %d, socket %d\n", i,
(gdb) n
Program received signal SIGSEGV, Segmentation fault.
0x00000000010f4d9e in check_cpu_topology (path=0x3ffffffd6c8
"/tmp/perf-test-J6CHMa", map=0x14a1740) at tests/topology.c:76
76 pr_debug("CPU %d, core %d, socket %d\n", i,
(gdb)
Make sure the env.cpu variable is not used when its NULL.
Test for NULL pointer and return TEST_SKIP if so.
Output before:
[root@p23lp27 perf]# ./perf test -F 39
39: Session topology :Segmentation fault (core dumped)
[root@p23lp27 perf]#
Output after:
[root@p23lp27 perf]# ./perf test -vF 39
39: Session topology :
--- start ---
templ file: /tmp/perf-test-Ajx59D
socket_id number is too big.You may need to upgrade the perf tool.
---- end ----
Session topology: Skip
[root@p23lp27 perf]#
Signed-off-by: Thomas Richter <tmricht@xxxxxxxxxxxxx>
---
---
tools/perf/tests/topology.c | 30 ++++++++++++++++++++++++------
1 file changed, 24 insertions(+), 6 deletions(-)
diff --git a/tools/perf/tests/topology.c b/tools/perf/tests/topology.c
index 17cb1bb3448c..40e30a26b23c 100644
--- a/tools/perf/tests/topology.c
+++ b/tools/perf/tests/topology.c
@@ -70,6 +70,27 @@ static int check_cpu_topology(char *path, struct cpu_map *map)
session = perf_session__new(&data, false, NULL);
TEST_ASSERT_VAL("can't get session", session);
+ /* On platforms with large numbers of CPUs process_cpu_topology()
+ * might issue an error while reading the perf.data file section
+ * HEADER_CPU_TOPOLOGY and the cpu_topology_map pointed to by member
+ * cpu is a NULL pointer.
+ * Example: On s390
+ * CPU 0 is on core_id 0 and physical_package_id 6
+ * CPU 1 is on core_id 1 and physical_package_id 3
+ *
+ * Core_id and physical_package_id are platform and architecture
+ * dependend and might have higher numbers than the CPU id.
+ * This actually depends on the configuration.
+ *
+ * In this case process_cpu_topology() prints error message:
+ * "socket_id number is too big. You may need to upgrade the
+ * perf tool."
+ *
+ * This is the reason why this test might be skipped.
+ */
+ if (!session->header.env.cpu)
+ return TEST_SKIP;
+
for (i = 0; i < session->header.env.nr_cpus_avail; i++) {
if (!cpu_map__has(map, i))
continue;
@@ -95,7 +116,7 @@ int test__session_topology(struct test *test __maybe_unused, int subtest __maybe
{
char path[PATH_MAX];
struct cpu_map *map;
- int ret = -1;
+ int ret = TEST_FAIL;
TEST_ASSERT_VAL("can't get templ file", !get_temp(path));
@@ -110,12 +131,9 @@ int test__session_topology(struct test *test __maybe_unused, int subtest __maybe
goto free_path;
}
- if (check_cpu_topology(path, map))
- goto free_map;
- ret = 0;
-
-free_map:
+ ret = check_cpu_topology(path, map);
cpu_map__put(map);
+
free_path:
unlink(path);
return ret;
--
2.14.3