Messages in this thread | | | Subject | Re: An invalid memory access was discovered by a fuzz test | From | Daniel Borkmann <> | Date | Tue, 19 Dec 2023 16:06:59 +0100 |
| |
On 12/19/23 3:15 PM, Xin Liu wrote: > Hi all: > > The issue occurred while reading an ELF file in libbpf.c during fuzzing > > Using host libthread_db library "/usr/lib64/libthread_db.so.1". > 0.000243187s DEBUG total counters = 7816 > 0.000346533s DEBUG binary maps to 400000-155f280, len = 18215552 > 0.000765462s DEBUG init_fuzzer:run_seed: running initial seed path="crash-sigsegv-b905489aaeb39555ff1245117f1efd1677195b9ac1437bfb18b8d2d04099704b" > > Program received signal SIGSEGV, Segmentation fault. > 0x0000000000958e97 in bpf_object.collect_prog_relos () at libbpf.c:4206 > 4206 in libbpf.c > (gdb) bt > #0 0x0000000000958e97 in bpf_object.collect_prog_relos () at libbpf.c:4206 > #1 0x000000000094f9d6 in bpf_object.collect_relos () at libbpf.c:6706 > #2 0x000000000092bef3 in bpf_object_open () at libbpf.c:7437 > #3 0x000000000092c046 in bpf_object.open_mem () at libbpf.c:7497 > #4 0x0000000000924afa in LLVMFuzzerTestOneInput () at fuzz/bpf-object-fuzzer.c:16 > #5 0x000000000060be11 in testblitz_engine::fuzzer::Fuzzer::run_one () > #6 0x000000000087ad92 in tracing::span::Span::in_scope () > #7 0x00000000006078aa in testblitz_engine::fuzzer::util::walkdir () > #8 0x00000000005f3217 in testblitz_engine::entrypoint::main::{{closure}} () > #9 0x00000000005f2601 in main () > (gdb) > > then, I checked the code and found that scn_data was null at this code(tools/lib/bpf/src/libbpf.c): > > if (rel->r_offset % BPF_INSN_SZ || rel->r_offset >= scn_data->d_size) { > > The scn_data is derived from the code above: > > scn = elf_sec_by_idx(obj, sec_idx); > scn_data = elf_sec_data(obj, scn); > > relo_sec_name = elf_sec_str(obj, shdr->sh_name); > sec_name = elf_sec_name(obj, scn); > if (!relo_sec_name || !sec_name) // don't check whether scn_data is NULL > return -EINVAL; > > Do sec_data and sec_name always occur together? Is it possible that scn_data is NULL but sec_name > is not NULL? libbpf uses sec_name to determine if it’s a null pointer, Maybe we should do some > check here.
Weird, is this based on a malformed elf given sec_idx comes from shdr->sh_info? It probably makes sense to NULL check and then return with -LIBBPF_ERRNO__FORMAT as we do elsewhere. Do you want to send a fix?
Thanks, Daniel
| |