On Mon, Mar 4, 2024 at 8:45 AM <dthaler1968@xxxxxxxxxxxxxx> wrote: > > Alexei Starovoitov <alexei.starovoitov@xxxxxxxxx> writes: > [...] > > Alexei Starovoitov (4): > > bpf: Introduce may_goto instruction > > bpf: Recognize that two registers are safe when their ranges match > > bpf: Add cond_break macro > > selftests/bpf: Test may_goto > > > > include/linux/bpf_verifier.h | 2 + > > include/uapi/linux/bpf.h | 1 + > > kernel/bpf/core.c | 1 + > > kernel/bpf/disasm.c | 3 + > > kernel/bpf/verifier.c | 280 +++++++++++++----- > > tools/include/uapi/linux/bpf.h | 1 + > > tools/testing/selftests/bpf/DENYLIST.s390x | 1 + > > .../testing/selftests/bpf/bpf_experimental.h | 12 + > > .../bpf/progs/verifier_iterating_callbacks.c | 103 ++++++- > > 9 files changed, 330 insertions(+), 74 deletions(-) > > Don't we also need to add may_goto to instruction-set.rst? If it was a real insn it would be too early to update the standard, since it's not in any kernel release yet. But more importantly it's a pseudo insn. CPUs/JITs won't see it. I think we need a different document for such pseudo instructions. Same thing for upcoming nop_or_goto pseudo insn.
