These patches try to test the fix made in commit e2f7fc0ac695 ("bpf:
fix undefined behavior in narrow load handling"). The problem existed
in the generated BPF bytecode that was doing a 32bit narrow read of a
64bit field, so to test it the code would need to be executed.
Currently the only such field exists in BPF_PROG_TYPE_PERF_EVENT,
which is not supported by bpf_prog_test_run().
First commit adds the test, but since I can't run it, I'm not sure if
it is even valid (because endianness, for example). Second commit adds
a message to test_verifier when it couldn't run the program because of
lack permissions or program type being not supported. Third commit
fixes a possible problem with errno.
With these patches, I get the following output on my test:
/kernel/tools/testing/selftests/bpf$ sudo ./test_verifier 920 920
#920/p 32bit loads of a 64bit field (both least and most significant words) Did not run the program (not supported) OK
Summary: 1 PASSED, 0 SKIPPED, 0 FAILED
So it looks like I need to pick a different approach.
Krzesimir Nowak (3):
selftests/bpf: Test correctness of narrow 32bit read on 64bit field
selftests/bpf: Print a message when tester could not run a program
selftests/bpf: Avoid a clobbering of errno
tools/testing/selftests/bpf/test_verifier.c | 19 +++++++++++++++----
.../testing/selftests/bpf/verifier/var_off.c | 15 +++++++++++++++
2 files changed, 30 insertions(+), 4 deletions(-)
--
2.20.1
