"ctx:file_pos sysctl:read read ok narrow" works on s390 by accident: it reads the wrong byte, which happens to have the expected value of 0. Improve the test by seeking to the 4th byte and expecting 4 instead of 0. This makes the latent problem apparent: the test attempts to read the first byte of bpf_sysctl.file_pos, assuming this is the least-significant byte, which is not the case on big-endian machines: a non-zero offset is needed. The point of the test is to verify narrow loads, so we cannot cheat our way out by simply using BPF_W. The existence of the test means that such loads have to be supported, most likely because llvm can generate them. Fix the test by adding a big-endian variant, which uses an offset to access the least-significant byte of bpf_sysctl.file_pos. This reveals the final problem: verifier rejects accesses to bpf_sysctl fields with offset > 0. Such accesses are already allowed for a wide range of structs: __sk_buff, bpf_sock_addr and sk_msg_md to name a few. Extend this support to bpf_sysctl by using bpf_ctx_range instead of offsetof when matching field offsets.
Unified · Split
| @@ -1311,12 +1311,12 @@ static bool sysctl_is_valid_access(int off, int size, enum bpf_access_type type, · 1310 unmodified lines | ||
| 1311 | 1311 | return false; |
| 1312 | 1312 | |
| 1313 | 1313 | switch (off) { |
| 1314 | − case offsetof(struct bpf_sysctl, write): | |
| 1314 | + case bpf_ctx_range(struct bpf_sysctl, write): | |
| 1315 | 1315 | if (type != BPF_READ) |
| 1316 | 1316 | return false; |
| 1317 | 1317 | bpf_ctx_record_field_size(info, size_default); |
| 1318 | 1318 | return bpf_ctx_narrow_access_ok(off, size, size_default); |
| 1319 | − case offsetof(struct bpf_sysctl, file_pos): | |
| 1319 | + case bpf_ctx_range(struct bpf_sysctl, file_pos): | |
| 1320 | 1320 | if (type == BPF_READ) { |
| 1321 | 1321 | bpf_ctx_record_field_size(info, size_default); |
| 1322 | 1322 | return bpf_ctx_narrow_access_ok(off, size, size_default); |
| @@ -161,9 +161,14 @@ static struct sysctl_test tests[] = { · 160 unmodified lines | ||
| 161 | 161 | .descr = "ctx:file_pos sysctl:read read ok narrow", |
| 162 | 162 | .insns = { |
| 163 | 163 | /* If (file_pos == X) */ |
| 164 | +#if __BYTE_ORDER == __LITTLE_ENDIAN | |
| 164 | 165 | BPF_LDX_MEM(BPF_B, BPF_REG_7, BPF_REG_1, |
| 165 | 166 | offsetof(struct bpf_sysctl, file_pos)), |
| 166 | − BPF_JMP_IMM(BPF_JNE, BPF_REG_7, 0, 2), | |
| 167 | +#else | |
| 168 | + BPF_LDX_MEM(BPF_B, BPF_REG_7, BPF_REG_1, | |
| 169 | + offsetof(struct bpf_sysctl, file_pos) + 3), | |
| 170 | +#endif | |
| 171 | + BPF_JMP_IMM(BPF_JNE, BPF_REG_7, 4, 2), | |
| 167 | 172 | |
| 168 | 173 | /* return ALLOW; */ |
| 169 | 174 | BPF_MOV64_IMM(BPF_REG_0, 1), |
| @@ -176,6 +181,7 @@ static struct sysctl_test tests[] = { · 6 unmodified lines | ||
| 176 | 181 | .attach_type = BPF_CGROUP_SYSCTL, |
| 177 | 182 | .sysctl = "kernel/ostype", |
| 178 | 183 | .open_flags = O_RDONLY, |
| 184 | + .seek = 4, | |
| 179 | 185 | .result = SUCCESS, |
| 180 | 186 | }, |
| 181 | 187 | { |