From patchwork Thu Mar 17 14:02:41 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Xu Kuohai X-Patchwork-Id: 12784044 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from bombadil.infradead.org (bombadil.infradead.org [198.137.202.133]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.lore.kernel.org (Postfix) with ESMTPS id 96612C433EF for ; Thu, 17 Mar 2022 13:52:59 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; q=dns/txt; c=relaxed/relaxed; d=lists.infradead.org; s=bombadil.20210309; h=Sender: Content-Transfer-Encoding:Content-Type:List-Subscribe:List-Help:List-Post: List-Archive:List-Unsubscribe:List-Id:MIME-Version:References:In-Reply-To: Message-ID:Date:Subject:CC:To:From:Reply-To:Content-ID:Content-Description: Resent-Date:Resent-From:Resent-Sender:Resent-To:Resent-Cc:Resent-Message-ID: List-Owner; bh=nO9CQF9l4AC83WYBVVXDPh5S972xupsweaLvN/LKd2c=; b=vyvkPl5omZaW++ z5ZlaUF3Jp6UsnxHawzLMyL059mvKDhqY9Nrq2gs9zyQKTXw0yyPOFOhFJFKOuMtBhMZ1+mREDdOB 4H6ITQkdGD44C66wNg3VXEySitNZFOQYpi6yfSjBO03cP57Ey0MrbtqL/B+QYLApwtCqIJs4slVWu 6XnCUq+FGfAQSLaLv3SiqbOtjAq6V3IPzdpuzk7OgL5jAf3Vhs1bGVegNDx8/W3EzeWf7yri79J+B jOs9slFVmcha639wQCqOAvMzxpSoQZfaRwsAbJJEh/4KDWTS/JrMAq+S4vTaqjuSUKX91b7RiVheM GkpZEiiqdRdTnNl3BLkQ==; Received: from localhost ([::1] helo=bombadil.infradead.org) by bombadil.infradead.org with esmtp (Exim 4.94.2 #2 (Red Hat Linux)) id 1nUqXh-00GIj6-7N; Thu, 17 Mar 2022 13:51:53 +0000 Received: from szxga08-in.huawei.com ([45.249.212.255]) by bombadil.infradead.org with esmtps (Exim 4.94.2 #2 (Red Hat Linux)) id 1nUqXM-00GIf2-Cz for linux-arm-kernel@lists.infradead.org; Thu, 17 Mar 2022 13:51:34 +0000 Received: from kwepemi500013.china.huawei.com (unknown [172.30.72.57]) by szxga08-in.huawei.com (SkyGuard) with ESMTP id 4KK7hz2W5pz1GC0D; Thu, 17 Mar 2022 21:46:27 +0800 (CST) Received: from huawei.com (10.67.174.197) by kwepemi500013.china.huawei.com (7.221.188.120) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256) id 15.1.2308.21; Thu, 17 Mar 2022 21:51:26 +0800 From: Xu Kuohai To: , CC: Catalin Marinas , Will Deacon , Daniel Borkmann , Alexei Starovoitov , Zi Shen Lim , Andrii Nakryiko , Martin KaFai Lau , Song Liu , Yonghong Song , John Fastabend , KP Singh , Julien Thierry , Mark Rutland , Hou Tao , Fuad Tabba , James Morse Subject: [PATCH bpf-next v4 2/4] bpf, arm64: Optimize BPF store/load using str/ldr with immediate offset Date: Thu, 17 Mar 2022 10:02:41 -0400 Message-ID: <20220317140243.212509-3-xukuohai@huawei.com> X-Mailer: git-send-email 2.30.2 In-Reply-To: <20220317140243.212509-1-xukuohai@huawei.com> References: <20220317140243.212509-1-xukuohai@huawei.com> MIME-Version: 1.0 X-Originating-IP: [10.67.174.197] X-ClientProxiedBy: dggems705-chm.china.huawei.com (10.3.19.182) To kwepemi500013.china.huawei.com (7.221.188.120) X-CFilter-Loop: Reflected X-CRM114-Version: 20100106-BlameMichelson ( TRE 0.8.0 (BSD) ) MR-646709E3 X-CRM114-CacheID: sfid-20220317_065132_862192_C2A652B6 X-CRM114-Status: GOOD ( 13.58 ) X-BeenThere: linux-arm-kernel@lists.infradead.org X-Mailman-Version: 2.1.34 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=archiver.kernel.org@lists.infradead.org The current BPF store/load instruction is translated by the JIT into two instructions. The first instruction moves the immediate offset into a temporary register. The second instruction uses this temporary register to do the real store/load. In fact, arm64 supports addressing with immediate offsets. So This patch introduces optimization that uses arm64 str/ldr instruction with immediate offset when the offset fits. Example of generated instuction for r2 = *(u64 *)(r1 + 0): without optimization: mov x10, 0 ldr x1, [x0, x10] with optimization: ldr x1, [x0, 0] If the offset is negative, or is not aligned correctly, or exceeds max value, rollback to the use of temporary register. Signed-off-by: Xu Kuohai --- arch/arm64/net/bpf_jit.h | 14 ++++ arch/arm64/net/bpf_jit_comp.c | 128 ++++++++++++++++++++++++++++++---- 2 files changed, 127 insertions(+), 15 deletions(-) diff --git a/arch/arm64/net/bpf_jit.h b/arch/arm64/net/bpf_jit.h index dd59b5ad8fe4..3920213244f0 100644 --- a/arch/arm64/net/bpf_jit.h +++ b/arch/arm64/net/bpf_jit.h @@ -66,6 +66,20 @@ #define A64_STR64(Xt, Xn, Xm) A64_LS_REG(Xt, Xn, Xm, 64, STORE) #define A64_LDR64(Xt, Xn, Xm) A64_LS_REG(Xt, Xn, Xm, 64, LOAD) +/* Load/store register (immediate offset) */ +#define A64_LS_IMM(Rt, Rn, imm, size, type) \ + aarch64_insn_gen_load_store_imm(Rt, Rn, imm, \ + AARCH64_INSN_SIZE_##size, \ + AARCH64_INSN_LDST_##type##_IMM_OFFSET) +#define A64_STRBI(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 8, STORE) +#define A64_LDRBI(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 8, LOAD) +#define A64_STRHI(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 16, STORE) +#define A64_LDRHI(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 16, LOAD) +#define A64_STR32I(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 32, STORE) +#define A64_LDR32I(Wt, Xn, imm) A64_LS_IMM(Wt, Xn, imm, 32, LOAD) +#define A64_STR64I(Xt, Xn, imm) A64_LS_IMM(Xt, Xn, imm, 64, STORE) +#define A64_LDR64I(Xt, Xn, imm) A64_LS_IMM(Xt, Xn, imm, 64, LOAD) + /* Load/store register pair */ #define A64_LS_PAIR(Rt, Rt2, Rn, offset, ls, type) \ aarch64_insn_gen_load_store_pair(Rt, Rt2, Rn, offset, \ diff --git a/arch/arm64/net/bpf_jit_comp.c b/arch/arm64/net/bpf_jit_comp.c index e850c69e128c..06037607c04a 100644 --- a/arch/arm64/net/bpf_jit_comp.c +++ b/arch/arm64/net/bpf_jit_comp.c @@ -191,6 +191,47 @@ static bool is_addsub_imm(u32 imm) return !(imm & ~0xfff) || !(imm & ~0xfff000); } +/* + * There are 3 types of AArch64 LDR/STR (immediate) instruction: + * Post-index, Pre-index, Unsigned offset. + * + * For BPF ldr/str, the "unsigned offset" type is sufficient. + * + * "Unsigned offset" type LDR(immediate) format: + * + * 3 2 1 0 + * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * |x x|1 1 1 0 0 1 0 1| imm12 | Rn | Rt | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * scale + * + * "Unsigned offset" type STR(immediate) format: + * 3 2 1 0 + * 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 9 8 7 6 5 4 3 2 1 0 + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * |x x|1 1 1 0 0 1 0 0| imm12 | Rn | Rt | + * +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ + * scale + * + * The offset is calculated from imm12 and scale in the following way: + * + * offset = (u64)imm12 << scale + */ +static bool is_lsi_offset(s16 offset, int scale) +{ + if (offset < 0) + return false; + + if (offset > (0xFFF << scale)) + return false; + + if (offset & ((1 << scale) - 1)) + return false; + + return true; +} + /* Tail call offset to jump into */ #if IS_ENABLED(CONFIG_ARM64_BTI_KERNEL) #define PROLOGUE_OFFSET 8 @@ -971,19 +1012,38 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, case BPF_LDX | BPF_PROBE_MEM | BPF_W: case BPF_LDX | BPF_PROBE_MEM | BPF_H: case BPF_LDX | BPF_PROBE_MEM | BPF_B: - emit_a64_mov_i(1, tmp, off, ctx); switch (BPF_SIZE(code)) { case BPF_W: - emit(A64_LDR32(dst, src, tmp), ctx); + if (is_lsi_offset(off, 2)) { + emit(A64_LDR32I(dst, src, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_LDR32(dst, src, tmp), ctx); + } break; case BPF_H: - emit(A64_LDRH(dst, src, tmp), ctx); + if (is_lsi_offset(off, 1)) { + emit(A64_LDRHI(dst, src, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_LDRH(dst, src, tmp), ctx); + } break; case BPF_B: - emit(A64_LDRB(dst, src, tmp), ctx); + if (is_lsi_offset(off, 0)) { + emit(A64_LDRBI(dst, src, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_LDRB(dst, src, tmp), ctx); + } break; case BPF_DW: - emit(A64_LDR64(dst, src, tmp), ctx); + if (is_lsi_offset(off, 3)) { + emit(A64_LDR64I(dst, src, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_LDR64(dst, src, tmp), ctx); + } break; } @@ -1011,20 +1071,39 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, case BPF_ST | BPF_MEM | BPF_B: case BPF_ST | BPF_MEM | BPF_DW: /* Load imm to a register then store it */ - emit_a64_mov_i(1, tmp2, off, ctx); emit_a64_mov_i(1, tmp, imm, ctx); switch (BPF_SIZE(code)) { case BPF_W: - emit(A64_STR32(tmp, dst, tmp2), ctx); + if (is_lsi_offset(off, 2)) { + emit(A64_STR32I(tmp, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp2, off, ctx); + emit(A64_STR32(tmp, dst, tmp2), ctx); + } break; case BPF_H: - emit(A64_STRH(tmp, dst, tmp2), ctx); + if (is_lsi_offset(off, 1)) { + emit(A64_STRHI(tmp, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp2, off, ctx); + emit(A64_STRH(tmp, dst, tmp2), ctx); + } break; case BPF_B: - emit(A64_STRB(tmp, dst, tmp2), ctx); + if (is_lsi_offset(off, 0)) { + emit(A64_STRBI(tmp, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp2, off, ctx); + emit(A64_STRB(tmp, dst, tmp2), ctx); + } break; case BPF_DW: - emit(A64_STR64(tmp, dst, tmp2), ctx); + if (is_lsi_offset(off, 3)) { + emit(A64_STR64I(tmp, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp2, off, ctx); + emit(A64_STR64(tmp, dst, tmp2), ctx); + } break; } break; @@ -1034,19 +1113,38 @@ static int build_insn(const struct bpf_insn *insn, struct jit_ctx *ctx, case BPF_STX | BPF_MEM | BPF_H: case BPF_STX | BPF_MEM | BPF_B: case BPF_STX | BPF_MEM | BPF_DW: - emit_a64_mov_i(1, tmp, off, ctx); switch (BPF_SIZE(code)) { case BPF_W: - emit(A64_STR32(src, dst, tmp), ctx); + if (is_lsi_offset(off, 2)) { + emit(A64_STR32I(src, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_STR32(src, dst, tmp), ctx); + } break; case BPF_H: - emit(A64_STRH(src, dst, tmp), ctx); + if (is_lsi_offset(off, 1)) { + emit(A64_STRHI(src, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_STRH(src, dst, tmp), ctx); + } break; case BPF_B: - emit(A64_STRB(src, dst, tmp), ctx); + if (is_lsi_offset(off, 0)) { + emit(A64_STRBI(src, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_STRB(src, dst, tmp), ctx); + } break; case BPF_DW: - emit(A64_STR64(src, dst, tmp), ctx); + if (is_lsi_offset(off, 3)) { + emit(A64_STR64I(src, dst, off), ctx); + } else { + emit_a64_mov_i(1, tmp, off, ctx); + emit(A64_STR64(src, dst, tmp), ctx); + } break; } break;