RFC bpf, x64: Fix tailcall infinite loop
Get series with b4
b4 am [email protected]Versions newest first · linked by author and title
- v4 RFC Latest4 patches 11 repliesIn review
- v3 RFC2 patches 5 repliesSuperseded
- v2 RFC2 patches 5 repliesSuperseded
Earlier versions · 1
| Version | Posted | Patches | Replies | Status |
|---|---|---|---|---|
| v1 RFC | 2 patches | 7 replies | Superseded |
Changes in v4 from the cover letter
* Suggestions from Maciej:
* Separate tail_call_cnt initialisation to a single patch.
* Unnecessary to check subprog.
Patches in v4 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/4 | bpf, x64: Comment tail_call_cnt initialisationarch/ | +4 | 2 replies | |
| 2/4 | bpf, x64: Fix tailcall infinite looparch/ include/ kernel/ +1 more | +32 | 1 reply | |
| 3/4 | selftests/bpf: Correct map_fd to data_fd in tailcallstools/ | +6 | 2 replies | |
| 4/4 | selftests/bpf: Add testcases for tailcall infinite loop fixingtools/ tools/ tools/ | +335 | 2 replies |
Reviews trailers per patch
1 of 4 patches has a Reviewed-by or Acked-by.
| Patch | Maciej Fijalkowski |
|---|---|
| 1/4 | |
| 2/4 | R-b |
| 3/4 | |
| 4/4 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [RFC PATCH bpf-next v4 0/4] bpf, x64: Fix tailcall infinite loop
This patch series fixes a tailcall infinite loop on x64.
From commit ebf7d1f508a73871 ("bpf, x64: rework pro/epilogue and tailcall handling in JIT"), the tailcall on x64 works better than before.
From commit e411901c0b775a3a ("bpf: allow for tailcalls in BPF subprograms for x64 JIT"), tailcall is able to run in BPF subprograms on x64.
From commit 5b92a28aae4dd0f8 ("bpf: Support attaching tracing BPF program to other BPF programs"), BPF program is able to trace other BPF programs.
How about combining them all together?
1. FENTRY/FEXIT on a BPF subprogram.
2. A tailcall runs in the BPF subprogram.
3. The tailcall calls the subprogram's caller.
As a result, a tailcall infinite loop comes up. And the loop would halt the machine.
As we know, in tail call context, the tail_call_cnt propagates by stack and rax register between BPF subprograms. So do in trampolines.
How did I discover the bug?
From commit 7f6e4312e15a5c37 ("bpf: Limit caller's stack depth 256 for subprogs with tailcalls"), the total stack size limits to around 8KiB. Then, I write some bpf progs to validate the stack consuming, that are tailcalls running in bpf2bpf and FENTRY/FEXIT tracing on bpf2bpf[1].
At that time, accidently, I made a tailcall loop. And then the loop halted my VM. Without the loop, the bpf progs would consume over 8KiB stack size. But the _stack-overflow_ did not halt my VM.
With bpf_printk(), I confirmed that the tailcall count limit did not work expectedly. Next, read the code and fix it.
Finally, unfortunately, I only fix it on x64 but other arches. As a result, CI tests failed because this bug hasn't been fixed on s390x.
Some helps on s390x are requested.
v3 -> v4:
* Suggestions from Maciej:
* Separate tail_call_cnt initialisation to a single patch.
* Unnecessary to check subprog.
v2 -> v3:
* Suggestions from Alexei:
* Fix comment style.
* Remove FIXME comment.
* Remove arch_prepare_bpf_trampoline() function comment update.
* Remove the unnecessary 'tgt_prog->aux->func[subprog]->is_func' check.
[1]: https://github.com/Asphaltt/learn-by-example/tree/main/ebpf/tailcall-stackoverflow
Leon Hwang (4):
bpf, x64: Comment tail_call_cnt initialisation
bpf, x64: Fix tailcall infinite loop
selftests/bpf: Correct map_fd to data_fd in tailcalls selftests/bpf: Add testcases for tailcall infinite loop fixing
arch/x86/net/bpf_jit_comp.c | 32 +- include/linux/bpf.h | 5 + kernel/bpf/trampoline.c | 4 +- kernel/bpf/verifier.c | 3 + .../selftests/bpf/prog_tests/tailcalls.c | 311 +++++++++++++++++- .../bpf/progs/tailcall_bpf2bpf_fentry.c | 18 + .../bpf/progs/tailcall_bpf2bpf_fexit.c | 18 + 7 files changed, 377 insertions(+), 14 deletions(-) create mode 100644 tools/testing/selftests/bpf/progs/tailcall_bpf2bpf_fentry.c create mode 100644 tools/testing/selftests/bpf/progs/tailcall_bpf2bpf_fexit.c
base-commit: 9930e4af4b509bcf6f060b09b16884f26102d110
Signature · 1 line
2.41.0