// Derived from Inferno's libkern/memmove-386.s (adapted for amd64) // https://bitbucket.org/inferno-os/inferno-os/src/default/libkern/memmove-386.s // // Copyright © 1994-1999 Lucent Technologies Inc. All rights reserved. // Revisions Copyright © 2000-2007 Vita Nuova Holdings Limited (www.vitanuova.com). All rights reserved. // Portions Copyright 2009 The Go Authors. All rights reserved. // // Permission is hereby granted, free of charge, to any person obtaining a copy // of this software and associated documentation files (the "Software"), to deal // in the Software without restriction, including without limitation the rights // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell // copies of the Software, and to permit persons to whom the Software is // furnished to do so, subject to the following conditions: // // The above copyright notice and this permission notice shall be included in // all copies or substantial portions of the Software. // // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE // AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN // THE SOFTWARE. // +build !plan9 #include "textflag.h" // void runtime·memmove(void*, void*, uintptr) TEXT runtime·memmove(SB), NOSPLIT, $0-24 MOVQ to+0(FP), DI MOVQ from+8(FP), SI MOVQ n+16(FP), BX // REP instructions have a high startup cost, so we handle small sizes // with some straightline code. The REP MOVSQ instruction is really fast // for large sizes. The cutover is approximately 2K. tail: // move_129through256 or smaller work whether or not the source and the // destination memory regions overlap because they load all data into // registers before writing it back. move_256through2048 on the other // hand can be used only when the memory regions don't overlap or the copy // direction is forward. TESTQ BX, BX JEQ move_0 CMPQ BX, $2 JBE move_1or2 CMPQ BX, $4 JBE move_3or4 CMPQ BX, $8 JB move_5through7 JE move_8 CMPQ BX, $16 JBE move_9through16 CMPQ BX, $32 JBE move_17through32 CMPQ BX, $64 JBE move_33through64 CMPQ BX, $128 JBE move_65through128 CMPQ BX, $256 JBE move_129through256 // TODO: use branch table and BSR to make this just a single dispatch TESTB $1, runtime·useAVXmemmove(SB) JNZ avxUnaligned /* * check and set for backwards */ CMPQ SI, DI JLS back /* * forward copy loop */ forward: CMPQ BX, $2048 JLS move_256through2048 // If REP MOVSB isn't fast, don't use it CMPB runtime·support_erms(SB), $1 // enhanced REP MOVSB/STOSB JNE fwdBy8 // Check alignment MOVL SI, AX ORL DI, AX TESTL $7, AX JEQ fwdBy8 // Do 1 byte at a time MOVQ BX, CX REP; MOVSB RET fwdBy8: // Do 8 bytes at a time MOVQ BX, CX SHRQ $3, CX ANDQ $7, BX REP; MOVSQ JMP tail back: /* * check overlap */ MOVQ SI, CX ADDQ BX, CX CMPQ CX, DI JLS forward /* * whole thing backwards has * adjusted addresses */ ADDQ BX, DI ADDQ BX, SI STD /* * copy */ MOVQ BX, CX SHRQ $3, CX ANDQ $7, BX SUBQ $8, DI SUBQ $8, SI REP; MOVSQ CLD ADDQ $8, DI ADDQ $8, SI SUBQ BX, DI SUBQ BX, SI JMP tail move_1or2: MOVB (SI), AX MOVB -1(SI)(BX*1), CX MOVB AX, (DI) MOVB CX, -1(DI)(BX*1) RET move_0: RET move_3or4: CMPQ BX, $4 JB move_3 MOVL (SI), AX MOVL AX, (DI) RET move_3: MOVW (SI), AX MOVB 2(SI), CX MOVW AX, (DI) MOVB CX, 2(DI) RET move_5through7: MOVL (SI), AX MOVL -4(SI)(BX*1), CX MOVL AX, (DI) MOVL CX, -4(DI)(BX*1) RET move_8: // We need a separate case for 8 to make sure we write pointers atomically. MOVQ (SI), AX MOVQ AX, (DI) RET move_9through16: MOVQ (SI), AX MOVQ -8(SI)(BX*1), CX MOVQ AX, (DI) MOVQ CX, -8(DI)(BX*1) RET move_17through32: MOVOU (SI), X0 MOVOU -16(SI)(BX*1), X1 MOVOU X0, (DI) MOVOU X1, -16(DI)(BX*1) RET move_33through64: MOVOU (SI), X0 MOVOU 16(SI), X1 MOVOU -32(SI)(BX*1), X2 MOVOU -16(SI)(BX*1), X3 MOVOU X0, (DI) MOVOU X1, 16(DI) MOVOU X2, -32(DI)(BX*1) MOVOU X3, -16(DI)(BX*1) RET move_65through128: MOVOU (SI), X0 MOVOU 16(SI), X1 MOVOU 32(SI), X2 MOVOU 48(SI), X3 MOVOU -64(SI)(BX*1), X4 MOVOU -48(SI)(BX*1), X5 MOVOU -32(SI)(BX*1), X6 MOVOU -16(SI)(BX*1), X7 MOVOU X0, (DI) MOVOU X1, 16(DI) MOVOU X2, 32(DI) MOVOU X3, 48(DI) MOVOU X4, -64(DI)(BX*1) MOVOU X5, -48(DI)(BX*1) MOVOU X6, -32(DI)(BX*1) MOVOU X7, -16(DI)(BX*1) RET move_129through256: MOVOU (SI), X0 MOVOU 16(SI), X1 MOVOU 32(SI), X2 MOVOU 48(SI), X3 MOVOU 64(SI), X4 MOVOU 80(SI), X5 MOVOU 96(SI), X6 MOVOU 112(SI), X7 MOVOU -128(SI)(BX*1), X8 MOVOU -112(SI)(BX*1), X9 MOVOU -96(SI)(BX*1), X10 MOVOU -80(SI)(BX*1), X11 MOVOU -64(SI)(BX*1), X12 MOVOU -48(SI)(BX*1), X13 MOVOU -32(SI)(BX*1), X14 MOVOU -16(SI)(BX*1), X15 MOVOU X0, (DI) MOVOU X1, 16(DI) MOVOU X2, 32(DI) MOVOU X3, 48(DI) MOVOU X4, 64(DI) MOVOU X5, 80(DI) MOVOU X6, 96(DI) MOVOU X7, 112(DI) MOVOU X8, -128(DI)(BX*1) MOVOU X9, -112(DI)(BX*1) MOVOU X10, -96(DI)(BX*1) MOVOU X11, -80(DI)(BX*1) MOVOU X12, -64(DI)(BX*1) MOVOU X13, -48(DI)(BX*1) MOVOU X14, -32(DI)(BX*1) MOVOU X15, -16(DI)(BX*1) RET move_256through2048: SUBQ $256, BX MOVOU (SI), X0 MOVOU 16(SI), X1 MOVOU 32(SI), X2 MOVOU 48(SI), X3 MOVOU 64(SI), X4 MOVOU 80(SI), X5 MOVOU 96(SI), X6 MOVOU 112(SI), X7 MOVOU 128(SI), X8 MOVOU 144(SI), X9 MOVOU 160(SI), X10 MOVOU 176(SI), X11 MOVOU 192(SI), X12 MOVOU 208(SI), X13 MOVOU 224(SI), X14 MOVOU 240(SI), X15 MOVOU X0, (DI) MOVOU X1, 16(DI) MOVOU X2, 32(DI) MOVOU X3, 48(DI) MOVOU X4, 64(DI) MOVOU X5, 80(DI) MOVOU X6, 96(DI) MOVOU X7, 112(DI) MOVOU X8, 128(DI) MOVOU X9, 144(DI) MOVOU X10, 160(DI) MOVOU X11, 176(DI) MOVOU X12, 192(DI) MOVOU X13, 208(DI) MOVOU X14, 224(DI) MOVOU X15, 240(DI) CMPQ BX, $256 LEAQ 256(SI), SI LEAQ 256(DI), DI JGE move_256through2048 JMP tail avxUnaligned: // There are two implementations of move algorithm. // The first one for non-ovelapped memory regions. It uses forward copying. // The second one for overlapped regions. It uses backward copying MOVQ DI, CX SUBQ SI, CX // Now CX contains distance between SRC and DEST CMPQ CX, BX // If the distance lesser than region length it means that regions are overlapped JC copy_backward // Non-temporal copy would be better for big sizes. CMPQ BX, $0x100000 JAE gobble_big_data_fwd // Memory layout on the source side // SI CX // |<---------BX before correction--------->| // | |<--BX corrected-->| | // | | |<--- AX --->| // |<-R11->| |<-128 bytes->| // +----------------------------------------+ // | Head | Body | Tail | // +-------+------------------+-------------+ // ^ ^ ^ // | | | // Save head into Y4 Save tail into X5..X12 // | // SI+R11, where R11 = ((DI & -32) + 32) - DI // Algorithm: // 1. Unaligned save of the tail's 128 bytes // 2. Unaligned save of the head's 32 bytes // 3. Destination-aligned copying of body (128 bytes per iteration) // 4. Put head on the new place // 5. Put the tail on the new place // It can be important to satisfy processor's pipeline requirements for // small sizes as the cost of unaligned memory region copying is // comparable with the cost of main loop. So code is slightly messed there. // There is more clean implementation of that algorithm for bigger sizes // where the cost of unaligned part copying is negligible. // You can see it after gobble_big_data_fwd label. LEAQ (SI)(BX*1), CX MOVQ DI, R10 // CX points to the end of buffer so we need go back slightly. We will use negative offsets there. MOVOU -0x80(CX), X5 MOVOU -0x70(CX), X6 MOVQ $0x80, AX // Align destination address ANDQ $-32, DI ADDQ $32, DI // Continue tail saving. MOVOU -0x60(CX), X7 MOVOU -0x50(CX), X8 // Make R11 delta between aligned and unaligned destination addresses. MOVQ DI, R11 SUBQ R10, R11 // Continue tail saving. MOVOU -0x40(CX), X9 MOVOU -0x30(CX), X10 // Let's make bytes-to-copy value adjusted as we've prepared unaligned part for copying. SUBQ R11, BX // Continue tail saving. MOVOU -0x20(CX), X11 MOVOU -0x10(CX), X12 // The tail will be put on it's place after main body copying. // It's time for the unaligned heading part. VMOVDQU (SI), Y4 // Adjust source address to point past head. ADDQ R11, SI SUBQ AX, BX // Aligned memory copying there gobble_128_loop: VMOVDQU (SI), Y0 VMOVDQU 0x20(SI), Y1 VMOVDQU 0x40(SI), Y2 VMOVDQU 0x60(SI), Y3 ADDQ AX, SI VMOVDQA Y0, (DI) VMOVDQA Y1, 0x20(DI) VMOVDQA Y2, 0x40(DI) VMOVDQA Y3, 0x60(DI) ADDQ AX, DI SUBQ AX, BX JA gobble_128_loop // Now we can store unaligned parts. ADDQ AX, BX ADDQ DI, BX VMOVDQU Y4, (R10) VZEROUPPER MOVOU X5, -0x80(BX) MOVOU X6, -0x70(BX) MOVOU X7, -0x60(BX) MOVOU X8, -0x50(BX) MOVOU X9, -0x40(BX) MOVOU X10, -0x30(BX) MOVOU X11, -0x20(BX) MOVOU X12, -0x10(BX) RET gobble_big_data_fwd: // There is forward copying for big regions. // It uses non-temporal mov instructions. // Details of this algorithm are commented previously for small sizes. LEAQ (SI)(BX*1), CX MOVOU -0x80(SI)(BX*1), X5 MOVOU -0x70(CX), X6 MOVOU -0x60(CX), X7 MOVOU -0x50(CX), X8 MOVOU -0x40(CX), X9 MOVOU -0x30(CX), X10 MOVOU -0x20(CX), X11 MOVOU -0x10(CX), X12 VMOVDQU (SI), Y4 MOVQ DI, R8 ANDQ $-32, DI ADDQ $32, DI MOVQ DI, R10 SUBQ R8, R10 SUBQ R10, BX ADDQ R10, SI LEAQ (DI)(BX*1), CX SUBQ $0x80, BX gobble_mem_fwd_loop: PREFETCHNTA 0x1C0(SI) PREFETCHNTA 0x280(SI) // Prefetch values were chosen empirically. // Approach for prefetch usage as in 7.6.6 of [1] // [1] 64-ia-32-architectures-optimization-manual.pdf // http://www.intel.ru/content/dam/www/public/us/en/documents/manuals/64-ia-32-architectures-optimization-manual.pdf VMOVDQU (SI), Y0 VMOVDQU 0x20(SI), Y1 VMOVDQU 0x40(SI), Y2 VMOVDQU 0x60(SI), Y3 ADDQ $0x80, SI VMOVNTDQ Y0, (DI) VMOVNTDQ Y1, 0x20(DI) VMOVNTDQ Y2, 0x40(DI) VMOVNTDQ Y3, 0x60(DI) ADDQ $0x80, DI SUBQ $0x80, BX JA gobble_mem_fwd_loop // NT instructions don't follow the normal cache-coherency rules. // We need SFENCE there to make copied data available timely. SFENCE VMOVDQU Y4, (R8) VZEROUPPER MOVOU X5, -0x80(CX) MOVOU X6, -0x70(CX) MOVOU X7, -0x60(CX) MOVOU X8, -0x50(CX) MOVOU X9, -0x40(CX) MOVOU X10, -0x30(CX) MOVOU X11, -0x20(CX) MOVOU X12, -0x10(CX) RET copy_backward: MOVQ DI, AX // Backward copying is about the same as the forward one. // Firstly we load unaligned tail in the beginning of region. MOVOU (SI), X5 MOVOU 0x10(SI), X6 ADDQ BX, DI MOVOU 0x20(SI), X7 MOVOU 0x30(SI), X8 LEAQ -0x20(DI), R10 MOVQ DI, R11 MOVOU 0x40(SI), X9 MOVOU 0x50(SI), X10 ANDQ $0x1F, R11 MOVOU 0x60(SI), X11 MOVOU 0x70(SI), X12 XORQ R11, DI // Let's point SI to the end of region ADDQ BX, SI // and load unaligned head into X4. VMOVDQU -0x20(SI), Y4 SUBQ R11, SI SUBQ R11, BX // If there is enough data for non-temporal moves go to special loop CMPQ BX, $0x100000 JA gobble_big_data_bwd SUBQ $0x80, BX gobble_mem_bwd_loop: VMOVDQU -0x20(SI), Y0 VMOVDQU -0x40(SI), Y1 VMOVDQU -0x60(SI), Y2 VMOVDQU -0x80(SI), Y3 SUBQ $0x80, SI VMOVDQA Y0, -0x20(DI) VMOVDQA Y1, -0x40(DI) VMOVDQA Y2, -0x60(DI) VMOVDQA Y3, -0x80(DI) SUBQ $0x80, DI SUBQ $0x80, BX JA gobble_mem_bwd_loop // Let's store unaligned data VMOVDQU Y4, (R10) VZEROUPPER MOVOU X5, (AX) MOVOU X6, 0x10(AX) MOVOU X7, 0x20(AX) MOVOU X8, 0x30(AX) MOVOU X9, 0x40(AX) MOVOU X10, 0x50(AX) MOVOU X11, 0x60(AX) MOVOU X12, 0x70(AX) RET gobble_big_data_bwd: SUBQ $0x80, BX gobble_big_mem_bwd_loop: PREFETCHNTA -0x1C0(SI) PREFETCHNTA -0x280(SI) VMOVDQU -0x20(SI), Y0 VMOVDQU -0x40(SI), Y1 VMOVDQU -0x60(SI), Y2 VMOVDQU -0x80(SI), Y3 SUBQ $0x80, SI VMOVNTDQ Y0, -0x20(DI) VMOVNTDQ Y1, -0x40(DI) VMOVNTDQ Y2, -0x60(DI) VMOVNTDQ Y3, -0x80(DI) SUBQ $0x80, DI SUBQ $0x80, BX JA gobble_big_mem_bwd_loop SFENCE VMOVDQU Y4, (R10) VZEROUPPER MOVOU X5, (AX) MOVOU X6, 0x10(AX) MOVOU X7, 0x20(AX) MOVOU X8, 0x30(AX) MOVOU X9, 0x40(AX) MOVOU X10, 0x50(AX) MOVOU X11, 0x60(AX) MOVOU X12, 0x70(AX) RET