Re: [PATCH bpf-next v5 09/11] selftests/bpf: Add C tests for 16-byte returns in R0:R2

On Thu, 2026-08-13 at 13:02 -0700, Yonghong Song wrote:
> Add selftests that exercise a 16-byte return value passed in the R0:R2
> register pair, written in C so that they depend on the compiler lowering
> the register-pair return. Covered are an __int128 return, a 16-byte struct
> return (from a static and from a global subprogram) and a 16-byte union
> return, plus __int128 and 16-byte struct returns from a kfunc. The union
> program shares an object with the struct ones.
> 
> The R0:R2 convention is only emitted by LLVM 23 and newer, and a by-value
> aggregate return does not compile at all before that, so the programs sit
> behind a __clang_major__ guard. An older compiler builds the dummy test in
> the #else branch instead, which keeps each object non-empty and says in
> its description why nothing was exercised.
> 
> The kfunc tests are tagged __arch_x86_64/__arch_arm64 and skip elsewhere.
> Those are the architectures whose JIT advertises
> bpf_jit_supports_kfunc_ret_reg_pair(), which bpf_add_kfunc_call() requires
> before it accepts a kfunc returning more than 8 bytes, and they are also
> the only ones building the kfuncs.
> 
> Signed-off-by: Yonghong Song <[email protected]>
> ---

Acked-by: Eduard Zingerman <[email protected]>

>  .../selftests/bpf/prog_tests/aggregate_ret.c  |  13 ++
>  .../bpf/progs/aggregate_ret_int128_c.c        |  49 ++++++++
>  .../bpf/progs/aggregate_ret_kfunc_c.c         |  65 ++++++++++
>  .../bpf/progs/aggregate_ret_struct_c.c        | 114 ++++++++++++++++++

Nit: I'd merge all three files into one and move loader to prog_tests/verifier.c.
     Also not sure about union vs struct tests.

> diff --git a/tools/testing/selftests/bpf/progs/aggregate_ret_int128_c.c b/tools/testing/selftests/bpf/progs/aggregate_ret_int128_c.c
> new file mode 100644
> index 000000000000..913cc374215d
> --- /dev/null
> +++ b/tools/testing/selftests/bpf/progs/aggregate_ret_int128_c.c

...

> +#if defined(__clang_major__) && __clang_major__ >= 23

...

> +#else
> +
> +SEC("socket")
> +__description("aggregate_ret_int128_c: needs LLVM 23, dummy test")
> +__success
> +int dummy_test(void)
> +{
> +	return 0;
> +}

Q: what would happen w/o #else branch? Is loading an empty object an error?

> +#endif

...