^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 1) ==================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 2) BPF Selftest Notes
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 3) ==================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 4) General instructions on running selftests can be found in
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 5) `Documentation/bpf/bpf_devel_QA.rst`_.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 6)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 7) Additional information about selftest failures are
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 8) documented here.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 9)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 10) profiler[23] test failures with clang/llvm <12.0.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 11) ==================================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 12)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 13) With clang/llvm <12.0.0, the profiler[23] test may fail.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 14) The symptom looks like
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 15)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 16) .. code-block:: c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 17)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 18) // r9 is a pointer to map_value
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 19) // r7 is a scalar
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 20) 17: bf 96 00 00 00 00 00 00 r6 = r9
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 21) 18: 0f 76 00 00 00 00 00 00 r6 += r7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 22) math between map_value pointer and register with unbounded min value is not allowed
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 23)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 24) // the instructions below will not be seen in the verifier log
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 25) 19: a5 07 01 00 01 01 00 00 if r7 < 257 goto +1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 26) 20: bf 96 00 00 00 00 00 00 r6 = r9
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 27) // r6 is used here
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 28)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 29) The verifier will reject such code with above error.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 30) At insn 18 the r7 is indeed unbounded. The later insn 19 checks the bounds and
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 31) the insn 20 undoes map_value addition. It is currently impossible for the
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 32) verifier to understand such speculative pointer arithmetic.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 33) Hence
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 34) https://reviews.llvm.org/D85570
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 35) addresses it on the compiler side. It was committed on llvm 12.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 36)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 37) The corresponding C code
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 38) .. code-block:: c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 39)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 40) for (int i = 0; i < MAX_CGROUPS_PATH_DEPTH; i++) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 41) filepart_length = bpf_probe_read_str(payload, ...);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 42) if (filepart_length <= MAX_PATH) {
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 43) barrier_var(filepart_length); // workaround
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 44) payload += filepart_length;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 45) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 46) }
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 47)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 48) bpf_iter test failures with clang/llvm 10.0.0
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 49) =============================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 50)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 51) With clang/llvm 10.0.0, the following two bpf_iter tests failed:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 52) * ``bpf_iter/ipv6_route``
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 53) * ``bpf_iter/netlink``
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 54)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 55) The symptom for ``bpf_iter/ipv6_route`` looks like
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 56)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 57) .. code-block:: c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 58)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 59) 2: (79) r8 = *(u64 *)(r1 +8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 60) ...
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 61) 14: (bf) r2 = r8
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 62) 15: (0f) r2 += r1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 63) ; BPF_SEQ_PRINTF(seq, "%pi6 %02x ", &rt->fib6_dst.addr, rt->fib6_dst.plen);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 64) 16: (7b) *(u64 *)(r8 +64) = r2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 65) only read is supported
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 66)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 67) The symptom for ``bpf_iter/netlink`` looks like
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 68)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 69) .. code-block:: c
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 70)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 71) ; struct netlink_sock *nlk = ctx->sk;
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 72) 2: (79) r7 = *(u64 *)(r1 +8)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 73) ...
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 74) 15: (bf) r2 = r7
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 75) 16: (0f) r2 += r1
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 76) ; BPF_SEQ_PRINTF(seq, "%pK %-3d ", s, s->sk_protocol);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 77) 17: (7b) *(u64 *)(r7 +0) = r2
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 78) only read is supported
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 79)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 80) This is due to a llvm BPF backend bug. The fix
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 81) https://reviews.llvm.org/D78466
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 82) has been pushed to llvm 10.x release branch and will be
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 83) available in 10.0.1. The fix is available in llvm 11.0.0 trunk.
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 84)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 85) BPF CO-RE-based tests and Clang version
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 86) =======================================
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 87)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 88) A set of selftests use BPF target-specific built-ins, which might require
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 89) bleeding-edge Clang versions (Clang 12 nightly at this time).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 90)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 91) Few sub-tests of core_reloc test suit (part of test_progs test runner) require
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 92) the following built-ins, listed with corresponding Clang diffs introducing
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 93) them to Clang/LLVM. These sub-tests are going to be skipped if Clang is too
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 94) old to support them, they shouldn't cause build failures or runtime test
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 95) failures:
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 96)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 97) - __builtin_btf_type_id() ([0], [1], [2]);
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 98) - __builtin_preserve_type_info(), __builtin_preserve_enum_value() ([3], [4]).
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 99)
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 100) [0] https://reviews.llvm.org/D74572
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 101) [1] https://reviews.llvm.org/D74668
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 102) [2] https://reviews.llvm.org/D85174
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 103) [3] https://reviews.llvm.org/D83878
^8f3ce5b39 (kx 2023-10-28 12:00:06 +0300 104) [4] https://reviews.llvm.org/D83242
Orange Pi5 kernel
Deprecated Linux kernel 5.10.110 for OrangePi 5/5B/5+ boards
3 Commits
0 Branches
0 Tags