Re: [PATCH] LoongArch: BPF: Optimize sign-extention mov instructions
From: Hengqi Chen
Date: Tue Aug 26 2025 - 21:27:57 EST
On Tue, Aug 26, 2025 at 2:49 PM Tiezhu Yang <yangtiezhu@xxxxxxxxxxx> wrote:
>
> For 8-bit and 16-bit sign-extention mov instructions, it can use the native
> instructions ext.w.b and ext.w.h directly, no need to use the temporary t1
> register, just remove the redundant operations.
>
> Here are the test results:
>
> # modprobe test_bpf test_range=81,84
> # dmesg -t | tail -5
> test_bpf: #81 ALU_MOVSX | BPF_B jited:1 5 PASS
> test_bpf: #82 ALU_MOVSX | BPF_H jited:1 5 PASS
> test_bpf: #83 ALU64_MOVSX | BPF_B jited:1 5 PASS
> test_bpf: #84 ALU64_MOVSX | BPF_H jited:1 5 PASS
> test_bpf: Summary: 4 PASSED, 0 FAILED, [4/4 JIT'ed]
>
> Signed-off-by: Tiezhu Yang <yangtiezhu@xxxxxxxxxxx>
> ---
> arch/loongarch/net/bpf_jit.c | 6 ++----
> 1 file changed, 2 insertions(+), 4 deletions(-)
>
> diff --git a/arch/loongarch/net/bpf_jit.c b/arch/loongarch/net/bpf_jit.c
> index abfdb6bb5c38..7072db18c6cd 100644
> --- a/arch/loongarch/net/bpf_jit.c
> +++ b/arch/loongarch/net/bpf_jit.c
> @@ -527,13 +527,11 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, bool ext
> emit_zext_32(ctx, dst, is32);
> break;
> case 8:
> - move_reg(ctx, t1, src);
> - emit_insn(ctx, extwb, dst, t1);
> + emit_insn(ctx, extwb, dst, src);
> emit_zext_32(ctx, dst, is32);
> break;
> case 16:
> - move_reg(ctx, t1, src);
> - emit_insn(ctx, extwh, dst, t1);
> + emit_insn(ctx, extwh, dst, src);
> emit_zext_32(ctx, dst, is32);
> break;
> case 32:
> --
Acked-by: Hengqi Chen <hengqi.chen@xxxxxxxxx>
> 2.42.0
>