]> git.ipfire.org Git - thirdparty/kernel/linux.git/commit
selftests/bpf: Use -4095 as the bad address for bits iterator
authorHou Tao <houtao1@huawei.com>
Tue, 5 Nov 2024 04:30:57 +0000 (12:30 +0800)
committerAlexei Starovoitov <ast@kernel.org>
Tue, 5 Nov 2024 22:02:08 +0000 (14:02 -0800)
commit6801cf7890f2ed8fcc14859b47501f8ee7a58ec7
tree8f25d0aa85e4024adbf8a65c5bbdf1cdb306d3d4
parent2e1b3cc9d7f790145a80cb705b168f05dab65df2
selftests/bpf: Use -4095 as the bad address for bits iterator

As reported by Byeonguk, the bad_words test in verifier_bits_iter.c
occasionally fails on s390 host. Quoting Ilya's explanation:

  s390 kernel runs in a completely separate address space, there is no
  user/kernel split at TASK_SIZE. The same address may be valid in both
  the kernel and the user address spaces, there is no way to tell by
  looking at it. The config option related to this property is
  ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE.

  Also, unfortunately, 0 is a valid address in the s390 kernel address
  space.

Fix the issue by using -4095 as the bad address for bits iterator, as
suggested by Ilya. Verify that bpf_iter_bits_new() returns -EINVAL for
NULL address and -EFAULT for bad address.

Fixes: ebafc1e535db ("selftests/bpf: Add three test cases for bits_iter")
Reported-by: Byeonguk Jeong <jungbu2855@gmail.com>
Closes: https://lore.kernel.org/bpf/ZycSXwjH4UTvx-Cn@ub22/
Signed-off-by: Hou Tao <houtao1@huawei.com>
Acked-by: Ilya Leoshkevich <iii@linux.ibm.com>
Link: https://lore.kernel.org/r/20241105043057.3371482-1-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
tools/testing/selftests/bpf/progs/verifier_bits_iter.c