// Copyright 2018 The Go Authors. All rights reserved. // Use of this source code is governed by a BSD-style // license that can be found in the LICENSE file. /* Package arm64 implements an ARM64 assembler. Go assembly syntax is different from GNU ARM64 syntax, but we can still follow the general rules to map between them. # Instructions mnemonics mapping rules 1. Most instructions use width suffixes of instruction names to indicate operand width rather than using different register names. Examples: ADC R24, R14, R12 <=> adc x12, x14, x24 ADDW R26->24, R21, R15 <=> add w15, w21, w26, asr #24 FCMPS F2, F3 <=> fcmp s3, s2 FCMPD F2, F3 <=> fcmp d3, d2 FCVTDH F2, F3 <=> fcvt h3, d2 2. Go uses .P and .W suffixes to indicate post-increment and pre-increment. Examples: MOVD.P -8(R10), R8 <=> ldr x8, [x10],#-8 MOVB.W 16(R16), R10 <=> ldrsb x10, [x16,#16]! MOVBU.W 16(R16), R10 <=> ldrb x10, [x16,#16]! 3. Go uses a series of MOV instructions as load and store. 64-bit variant ldr, str, stur => MOVD; 32-bit variant str, stur, ldrsw => MOVW; 32-bit variant ldr => MOVWU; ldrb => MOVBU; ldrh => MOVHU; ldrsb, sturb, strb => MOVB; ldrsh, sturh, strh => MOVH. 4. Go moves conditions into opcode suffix, like BLT. 5. Go adds a V prefix for most floating-point and SIMD instructions, except cryptographic extension instructions and floating-point(scalar) instructions. Examples: VADD V5.H8, V18.H8, V9.H8 <=> add v9.8h, v18.8h, v5.8h VLD1.P (R6)(R11), [V31.D1] <=> ld1 {v31.1d}, [x6], x11 VFMLA V29.S2, V20.S2, V14.S2 <=> fmla v14.2s, v20.2s, v29.2s AESD V22.B16, V19.B16 <=> aesd v19.16b, v22.16b SCVTFWS R3, F16 <=> scvtf s17, w6 6. Align directive Go asm supports the PCALIGN directive, which indicates that the next instruction should be aligned to a specified boundary by padding with NOOP instruction. The alignment value supported on arm64 must be a power of 2 and in the range of [8, 2048]. Examples: PCALIGN $16 MOVD $2, R0 // This instruction is aligned with 16 bytes. PCALIGN $1024 MOVD $3, R1 // This instruction is aligned with 1024 bytes. PCALIGN also changes the function alignment. If a function has one or more PCALIGN directives, its address will be aligned to the same or coarser boundary, which is the maximum of all the alignment values. In the following example, the function Add is aligned with 128 bytes. Examples: TEXT ·Add(SB),$40-16 MOVD $2, R0 PCALIGN $32 MOVD $4, R1 PCALIGN $128 MOVD $8, R2 RET On arm64, functions in Go are aligned to 16 bytes by default, we can also use PCALIGN to set the function alignment. The functions that need to be aligned are preferably using NOFRAME and NOSPLIT to avoid the impact of the prologues inserted by the assembler, so that the function address will have the same alignment as the first hand-written instruction. In the following example, PCALIGN at the entry of the function Add will align its address to 2048 bytes. Examples: TEXT ·Add(SB),NOSPLIT|NOFRAME,$0 PCALIGN $2048 MOVD $1, R0 MOVD $1, R1 RET 7. Move large constants to vector registers. Go asm uses VMOVQ/VMOVD/VMOVS to move 128-bit, 64-bit and 32-bit constants into vector registers, respectively. And for a 128-bit integer, it take two 64-bit operands, for the low and high parts separately. Examples: VMOVS $0x11223344, V0 VMOVD $0x1122334455667788, V1 VMOVQ $0x1122334455667788, $0x99aabbccddeeff00, V2 // V2=0x99aabbccddeeff001122334455667788 8. Move an optionally-shifted 16-bit immediate value to a register. The instructions are MOVK(W), MOVZ(W) and MOVN(W), the assembly syntax is "op $(uimm16<". The is the 16-bit unsigned immediate, in the range 0 to 65535; For the 32-bit variant, the is 0 or 16, for the 64-bit variant, the is 0, 16, 32 or 48. The current Go assembler does not accept zero shifts, such as "op $0, Rd" and "op $(0<<(16|32|48)), Rd" instructions. Examples: MOVK $(10<<32), R20 <=> movk x20, #10, lsl #32 MOVZW $(20<<16), R8 <=> movz w8, #20, lsl #16 MOVK $(0<<16), R10 will be reported as an error by the assembler. Special Cases. (1) umov is written as VMOV. (2) br is renamed JMP, blr is renamed CALL. (3) No need to add "W" suffix: LDARB, LDARH, LDAXRB, LDAXRH, LDTRH, LDXRB, LDXRH. (4) In Go assembly syntax, NOP is a zero-width pseudo-instruction serves generic purpose, nothing related to real ARM64 instruction. NOOP serves for the hardware nop instruction. NOOP is an alias of HINT $0. Examples: VMOV V13.B[1], R20 <=> mov x20, v13.b[1] VMOV V13.H[1], R20 <=> mov w20, v13.h[1] JMP (R3) <=> br x3 CALL (R17) <=> blr x17 LDAXRB (R19), R16 <=> ldaxrb w16, [x19] NOOP <=> nop # Register mapping rules 1. All basic register names are written as Rn. 2. Go uses ZR as the zero register and RSP as the stack pointer. 3. Bn, Hn, Dn, Sn and Qn instructions are written as Fn in floating-point instructions and as Vn in SIMD instructions. # Argument mapping rules 1. The operands appear in left-to-right assignment order. Go reverses the arguments of most instructions. Examples: ADD R11.SXTB<<1, RSP, R25 <=> add x25, sp, w11, sxtb #1 VADD V16, V19, V14 <=> add d14, d19, d16 Special Cases. (1) Argument order is the same as in the GNU ARM64 syntax: cbz, cbnz and some store instructions, such as str, stur, strb, sturb, strh, sturh stlr, stlrb. stlrh, st1. Examples: MOVD R29, 384(R19) <=> str x29, [x19,#384] MOVB.P R30, 30(R4) <=> strb w30, [x4],#30 STLRH R21, (R19) <=> stlrh w21, [x19] (2) MADD, MADDW, MSUB, MSUBW, SMADDL, SMSUBL, UMADDL, UMSUBL , , , Examples: MADD R2, R30, R22, R6 <=> madd x6, x22, x2, x30 SMSUBL R10, R3, R17, R27 <=> smsubl x27, w17, w10, x3 (3) FMADDD, FMADDS, FMSUBD, FMSUBS, FNMADDD, FNMADDS, FNMSUBD, FNMSUBS , , , Examples: FMADDD F30, F20, F3, F29 <=> fmadd d29, d3, d30, d20 FNMSUBS F7, F25, F7, F22 <=> fnmsub s22, s7, s7, s25 (4) BFI, BFXIL, SBFIZ, SBFX, UBFIZ, UBFX $, , $, Examples: BFIW $16, R20, $6, R0 <=> bfi w0, w20, #16, #6 UBFIZ $34, R26, $5, R20 <=> ubfiz x20, x26, #34, #5 (5) FCCMPD, FCCMPS, FCCMPED, FCCMPES , Fm. Fn, $ Examples: FCCMPD AL, F8, F26, $0 <=> fccmp d26, d8, #0x0, al FCCMPS VS, F29, F4, $4 <=> fccmp s4, s29, #0x4, vs FCCMPED LE, F20, F5, $13 <=> fccmpe d5, d20, #0xd, le FCCMPES NE, F26, F10, $0 <=> fccmpe s10, s26, #0x0, ne (6) CCMN, CCMNW, CCMP, CCMPW , , $, $ Examples: CCMP MI, R22, $12, $13 <=> ccmp x22, #0xc, #0xd, mi CCMNW AL, R1, $11, $8 <=> ccmn w1, #0xb, #0x8, al (7) CCMN, CCMNW, CCMP, CCMPW , , , $ Examples: CCMN VS, R13, R22, $10 <=> ccmn x13, x22, #0xa, vs CCMPW HS, R19, R14, $11 <=> ccmp w19, w14, #0xb, cs (9) CSEL, CSELW, CSNEG, CSNEGW, CSINC, CSINCW , , , ; FCSELD, FCSELS , , , Examples: CSEL GT, R0, R19, R1 <=> csel x1, x0, x19, gt CSNEGW GT, R7, R17, R8 <=> csneg w8, w7, w17, gt FCSELD EQ, F15, F18, F16 <=> fcsel d16, d15, d18, eq (10) TBNZ, TBZ $, ,