selftests/bpf: fix "ctx:write sysctl:write read ok" on s390
authorIlya Leoshkevich <iii@linux.ibm.com>
Fri, 30 Aug 2019 11:07:30 +0000 (13:07 +0200)
committerDaniel Borkmann <daniel@iogearbox.net>
Tue, 3 Sep 2019 19:01:52 +0000 (21:01 +0200)
"ctx:write sysctl:write read ok" fails on s390 because it reads the
first byte of an int assuming it's the least-significant one, which
is not the case on big-endian arches. Since we are not testing narrow
accesses here (there is e.g. "ctx:file_pos sysctl:read read ok narrow"
for that), simply read the whole int.

Fixes: 1f5fa9ab6e2e ("selftests/bpf: Test BPF_CGROUP_SYSCTL")
Signed-off-by: Ilya Leoshkevich <iii@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
tools/testing/selftests/bpf/test_sysctl.c

index a3bebd7c68ddc6262c5eaf6021bd8c8304b6a006..908f327839d5c9ee81ce9b8e089feeaa985691fa 100644 (file)
@@ -100,7 +100,7 @@ static struct sysctl_test tests[] = {
                .descr = "ctx:write sysctl:write read ok",
                .insns = {
                        /* If (write) */
-                       BPF_LDX_MEM(BPF_B, BPF_REG_7, BPF_REG_1,
+                       BPF_LDX_MEM(BPF_W, BPF_REG_7, BPF_REG_1,
                                    offsetof(struct bpf_sysctl, write)),
                        BPF_JMP_IMM(BPF_JNE, BPF_REG_7, 1, 2),