fanbird2008的专栏

http://cs.lmu.edu/~ray/notes/nasmtutorial/NASM TutorialYep, it's a tutorial.Scope of the TutorialThis tutorial will show you how to write assembly language programs for Linux on the x86-64

%macro WELS_AbsW 2    pxor        %2, %2    psubw       %2, %1    pmaxsw      %1, %2%endmacroWELS_AbsW  xmm0, xmm3展开如下：pxor xmm3, xmm3     => xmm3 = 0psubw xmm3, xmm0  => x

;*!;* \copy;*     Copyright (c)  2009-2013, Cisco Systems;*     All rights reserved.;*;*     Redistribution and use in source and binary forms, with or without;*     modification, are

;*!;* \copy;*     Copyright (c)  2009-2013, Cisco Systems;*     All rights reserved.;*;*     Redistribution and use in source and binary forms, with or without;*     modification, are perm

1.   addr.S        extern printf  ; the C function to be called%macro pabc 1   ; a "simple" print macro section .data.str db %1,0  ; %1 is first actual in macro call section .text

#include #include int64_t testregs(int a, int b, int c, int d, int e, int f, int g, int h);int64_t testregs(int a, int b, int c, int d, int e, int f, int g, int h){    int64_t result;    r

1.  print1.S; Declare external function        extern  printf          ; the C function, to be called        SECTION .data           ; Data section, initialized variables        a:      dq

; testreg_64.asm   test what register names can be used;; compile: nasm -f elf64 -l testreg_64.lst  testasm_64.asm; link: gcc -o testreg_64  testreg_64.o; run: ./testreg  # may get segfault or

; loopint_64.asm  code loopint.c for nasm ; /* loopint_64.c a very simple loop that will be coded for nasm */; #include ; int main(); {;   long int dd1[100]; // 100 could be 3 gigabytes;

; horner_64.asm  Horners method of evaluating polynomials;; given a polynomial  Y = a_n X^n + a_n-1 X^n-1 + ... a_1 X + a_0; a_n is the coefficient 'a' with subscript n. X^n is X to nth power;

%ifdef WIN64 ; Windows x64 ;************************************DEFAULT RELBITS 64%define arg1 rcx%define arg2 rdx%define arg3 r8%define arg4 r9%define arg5 [rsp + push_num*8 + 40]%def

http://eli.thegreenplace.net/2011/09/06/stack-frame-layout-on-x86-64/Stack frame layout on x86-64 September 06, 2011 at 20:13Tags Articles , Assembly , LinuxA few months ag

1.   file add.S.globl my_addmy_add:  pushq %rbp  movq  %rsp, %rbp  addq %rsi, %rdi  movq %rdi, %rax  leaveq  retq2.    file max.S.globl  my_maxmy_max:  pushq    %rb

; in: xmm1, xmm2, xmm3, xmm4  pOut:  xmm1, xmm4, xmm5, mm3%macro SSE2_Trans4x4D 5    SSE2_XSawp dq,  %1, %2, %5    SSE2_XSawp dq,  %3, %4, %2    SSE2_XSawp qdq, %1, %3, %4    SSE2_XSawp qdq,

;copy a dw into a xmm for 8 times%macro SSE2_Copy8Times 2    movd    %1, %2    punpcklwd %1, %1    pshufd  %1,     %1,     0%endmacro    SSE2_Copy8Times xmm1, r2d   ; xmm1 = b,b,b,b,b,

extern printf%define arg1 rdi%define arg2 rsi%define arg3 rdx%define arg4 rcx%define arg5 r8%define arg6 r9%define arg7 [rsp + push_num*8 + 8]%define arg8 [rsp + push_num*8 + 16]%def

%macro SSE2_Store4x8p 6    SSE2_XSawp qdq, %2, %3, %6    SSE2_XSawp qdq, %4, %5, %3    MOVDQ    [%1+0x00], %2    MOVDQ    [%1+0x10], %4    MOVDQ    [%1+0x20], %6    MOVDQ    [%1+0x30], %3

%macro SSE2_Load4x8p 6    MOVDQ    %2,    [%1+0x00]    MOVDQ    %4,    [%1+0x10]    MOVDQ    %6,    [%1+0x20]    MOVDQ    %3,    [%1+0x30]    SSE2_XSawp qdq, %4, %3, %5          ; 为什么要做这两步？？

;***********************************************************************; void WelsIDctRecI16x16Dc_sse2(uint8_t *rec, int32_t stride, uint8_t *pred, int32_t pred_stride, int16_t *dct_dc);*********

;***********************************************************************; void WelsIDctFourT4Rec_sse2(uint8_t *rec, int32_t stride, uint8_t *pred, int32_t pred_stride, int16_t *rs);;**************

;***********************************************************************; void WelsDctFourT4_sse2(int16_t *pDct, uint8_t *pix1, int32_t i_pix1, uint8_t *pix2, int32_t i_pix2 );********************

;***********************************************************************;void WelsHadamardT4Dc_sse2( int16_t *luma_dc, int16_t *pDct);**************************************************************

;***********************************************************************; void WelsDctFourT4_sse2(int16_t *pDct, uint8_t *pix1, int32_t i_pix1, uint8_t *pix2, int32_t i_pix2 );********************

%macro SSE2_StoreDiff8p 6    paddw       %1, %3    psraw       %1, $06    movq        %2, %6    punpcklbw   %2, %4    paddsw      %2, %1    packuswb    %2, %2    movq        %5, %2%end

; pOut mm1, mm4, mm5, mm3%macro MMX_Trans4x4W 5    MMX_XSwap wd, %1, %2, %5    MMX_XSwap wd, %3, %4, %2    MMX_XSwap dq, %1, %3, %4    MMX_XSwap dq, %5, %2, %3%endmacro

;in: xmm0, xmm1, xmm2, xmm3  pOut:  xmm0, xmm1, xmm3, xmm4%macro SSE2_TransTwo4x4W 5    SSE2_XSawp wd,  %1, %2, %5    SSE2_XSawp wd,  %3, %4, %2    SSE2_XSawp dq,  %1, %3, %4    SSE2_XSawp d

%macro SSE2_Load8DC 6    movdqa      %1,     %6      ; %1 = dc0 dc1    paddw       %1,     %5    psraw       %1,     $06     ; (dc + 32) >> 6    movdqa      %2,     %1    psrldq      %2,

;copy a db into a xmm for 16 times%macro SSE2_Copy16Times 2    movd            %1, %2    pshuflw         %1, %1, 0    punpcklqdq      %1, %1    packuswb        %1,     %1%endmacro

extern printf%macro prt1 1  section .data.str db %1, 0  section .text  mov rdi, fmt1  mov rsi, .str  mov rdx, rax  mov rax, 0  call printf%endmacro%macro prt2 1  section .data

; ifint_64.asm  code ifint_64.c for nasm ; /* ifint_64.c an 'if' statement that will be coded for nasm */; #include ; int main(); {;   long int a=1;;   long int b=2;;   long int c=3;;

movslq %eax,%rbx movsx rbx, eax40056e:       8b 45 d0                mov    -0x30(%rbp),%eax400571:       48 63 d8                movslq %eax,%rbx400574:       48 8b 45 f0             mov    -0

http://cs.lmu.edu/~ray/notes/nasmtutorial/Register OperandsIn this tutorial we only care about the integer registers and the xmm registers. You should already know what the registers are,

http://www.tortall.net/projects/yasm/manual/html/nasm-multi-line-macros.htmlMulti-line macros are much more like the type of macro seen in MASM and TASM: a multi-line macro definition in NASM lo

;****************************************************************************************************;   void WelsCPUId( int32_t uiIndex, int32_t *pFeatureA, int32_t *pFeatureB, int32_t *pFeatureC,

From a function viewpoint, function may be implmented by inline assembly way, or by non-inline assembly way.Example:inline assembly functionint inline_add(int x, int y){  asm volatile ("ad

Intel and AT&T Syntax.Intel and AT&T syntax Assembly language are very different from each other in appearance, and this will lead to confusion when one first comes across AT&T

/* areatest.c - An example of using floating point regs */#include int main(){ int radius = 10; float area; asm(“fild %1\n\t” “fimul %1\n\t” “fldpi\n\t” “fmul %%st(1), %%st(0)” : “

http://www.ibm.com/developerworks/cn/linux/l-gas-nasm.html本文解释两种最流行的 Linux® 汇编器 —— GNU Assembler（GAS）和 Netwide Assembler（NASM） —— 之间一些比较重要的语法差异和语义差异，包括基本语法、变量和内存访问、宏处理、函数和外部例程、堆栈处理以及重复执行代码块的技术方面

1. As far as following inline assembly#include int main(){    float arg1, arg2, add, sub, mul, div ;    printf( "Enter two numbers : " );    scanf( "%f%f", &arg1, &arg2 );    /* Perform

int gcd(int a, int b){    int result ;    asm volatile (    "movl %1, %%eax;"    "movl %2, %%ebx;"    "CONTD: cmpl $0, %%ebx;"    "je DONE;"    "xorl %%edx, %%edx;"    "idivl %%ebx;"