Hi, Hengqi,
On Wed, Sep 3, 2025 at 8:06 PM Hengqi Chen <hengqi.chen@xxxxxxxxx> wrote:
>
> The current implementation of bpf_arch_text_poke() requires 5 nops
> at patch site which is not applicable for kernel/module functions.
> With CONFIG_DYNAMIC_FTRACE_WITH_DIRECT_CALLS=y, this can be done
> by ftrace instead.
Does this mean BPF trampoline can only work with FTRACE enabled?
Huacai
>
> See the following commit for details:
> * commit b91e014f078e ("bpf: Make BPF trampoline use register_ftrace_direct() API")
> * commit 9cdc3b6a299c ("LoongArch: ftrace: Add direct call support")
>
> Signed-off-by: Hengqi Chen <hengqi.chen@xxxxxxxxx>
> ---
> arch/loongarch/net/bpf_jit.c | 7 +++++--
> 1 file changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/arch/loongarch/net/bpf_jit.c b/arch/loongarch/net/bpf_jit.c
> index 7b7e449b9ea9..35b13d91a979 100644
> --- a/arch/loongarch/net/bpf_jit.c
> +++ b/arch/loongarch/net/bpf_jit.c
> @@ -1294,8 +1294,11 @@ int bpf_arch_text_poke(void *ip, enum bpf_text_poke_type poke_type,
> u32 old_insns[LOONGARCH_LONG_JUMP_NINSNS] = {[0 ... 4] = INSN_NOP};
> u32 new_insns[LOONGARCH_LONG_JUMP_NINSNS] = {[0 ... 4] = INSN_NOP};
>
> - if (!is_kernel_text((unsigned long)ip) &&
> - !is_bpf_text_address((unsigned long)ip))
> + if (!is_bpf_text_address((unsigned long)ip))
> + /* Only poking bpf text is supported. Since kernel function
> + * entry is set up by ftrace, we reply on ftrace to poke kernel
> + * functions.
> + */
> return -ENOTSUPP;
>
> ret = emit_jump_or_nops(old_addr, ip, old_insns, is_call);
> --
> 2.43.5
>
