bpf_base_test - unconditionally run TestSocketFilterSetUp

We require 4.14 kernel so the condition is always true, and
thus this test is never actually skipped.

Additionally this test is outright incorrect, because BpfHandler::initMaps() requires the uid permission map open
to succeed, and furthermore BpfHandler::hasUpdateDeviceStatsPermission
requires it for things to even work, as this is used as part
of tagSocket() permission checks.

(side note: it's ugly that this netd test is reaching into
mainline stuff... but fixing that is a different TODO)

Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I4ea401e8bd73f9c3fce4498582dbcdf247a9453b
(cherry picked from commit 92ac128dc27a91b6a3ae90918edefe61eecda5f4)
Merged-In: I4ea401e8bd73f9c3fce4498582dbcdf247a9453b
diff --git a/tests/bpf_base_test.cpp b/tests/bpf_base_test.cpp
index e402e08..60741b7 100644
--- a/tests/bpf_base_test.cpp
+++ b/tests/bpf_base_test.cpp
@@ -76,8 +76,6 @@
 }
 
 TEST_F(BpfBasicTest, TestSocketFilterSetUp) {
-    SKIP_IF_EXTENDED_BPF_NOT_SUPPORTED;
-
     ASSERT_EQ(0, access(CGROUP_SOCKET_PROG_PATH, R_OK));
     ASSERT_EQ(0, access(UID_PERMISSION_MAP_PATH, R_OK));
 }