disasm.c

汇编编译器的最新版本的源码.买了自己动手写操作系统这本书的人一定要下

/* disasm.c   where all the _work_ gets done in the Netwide Disassembler
 *
 * The Netwide Assembler is copyright (C) 1996 Simon Tatham and
 * Julian Hall. All rights reserved. The software is
 * redistributable under the license given in the file "LICENSE"
 * distributed in the NASM archive.
 *
 * initial version 27/iii/95 by Simon Tatham
 */

#include "compiler.h"

#include <stdio.h>
#include <string.h>
#include <limits.h>
#include <inttypes.h>

#include "nasm.h"
#include "disasm.h"
#include "sync.h"
#include "insns.h"
#include "tables.h"
#include "regdis.h"

/*
 * Flags that go into the `segment' field of `insn' structures
 * during disassembly.
 */
#define SEG_RELATIVE	  1
#define SEG_32BIT	  2
#define SEG_RMREG	  4
#define SEG_DISP8	  8
#define SEG_DISP16	 16
#define SEG_DISP32	 32
#define SEG_NODISP	 64
#define SEG_SIGNED	128
#define SEG_64BIT	256

/*
 * Prefix information
 */
struct prefix_info {
    uint8_t osize;		/* Operand size */
    uint8_t asize;		/* Address size */
    uint8_t osp;		/* Operand size prefix present */
    uint8_t asp;		/* Address size prefix present */
    uint8_t rep;		/* Rep prefix present */
    uint8_t seg;		/* Segment override prefix present */
    uint8_t lock;		/* Lock prefix present */
    uint8_t vex[3];		/* VEX prefix present */
    uint8_t vex_m;		/* VEX.M field */
    uint8_t vex_v;
    uint8_t vex_lp;		/* VEX.LP fields */
    uint32_t rex;		/* REX prefix present */
};

#define getu8(x) (*(uint8_t *)(x))
#if X86_MEMORY
/* Littleendian CPU which can handle unaligned references */
#define getu16(x) (*(uint16_t *)(x))
#define getu32(x) (*(uint32_t *)(x))
#define getu64(x) (*(uint64_t *)(x))
#else
static uint16_t getu16(uint8_t *data)
{
    return (uint16_t)data[0] + ((uint16_t)data[1] << 8);
}
static uint32_t getu32(uint8_t *data)
{
    return (uint32_t)getu16(data) + ((uint32_t)getu16(data+2) << 16);
}
static uint64_t getu64(uint8_t *data)
{
    return (uint64_t)getu32(data) + ((uint64_t)getu32(data+4) << 32);
}
#endif

#define gets8(x) ((int8_t)getu8(x))
#define gets16(x) ((int16_t)getu16(x))
#define gets32(x) ((int32_t)getu32(x))
#define gets64(x) ((int64_t)getu64(x))

/* Important: regval must already have been adjusted for rex extensions */
static enum reg_enum whichreg(int32_t regflags, int regval, int rex)
{
    if (!(regflags & (REGISTER|REGMEM)))
	return 0;		/* Registers not permissible?! */

    regflags |= REGISTER;

    if (!(REG_AL & ~regflags))
        return R_AL;
    if (!(REG_AX & ~regflags))
        return R_AX;
    if (!(REG_EAX & ~regflags))
        return R_EAX;
    if (!(REG_RAX & ~regflags))
	return R_RAX;
    if (!(REG_DL & ~regflags))
        return R_DL;
    if (!(REG_DX & ~regflags))
        return R_DX;
    if (!(REG_EDX & ~regflags))
        return R_EDX;
    if (!(REG_RDX & ~regflags))
        return R_RDX;
    if (!(REG_CL & ~regflags))
        return R_CL;
    if (!(REG_CX & ~regflags))
        return R_CX;
    if (!(REG_ECX & ~regflags))
        return R_ECX;
    if (!(REG_RCX & ~regflags))
        return R_RCX;
    if (!(FPU0 & ~regflags))
        return R_ST0;
    if (!(XMM0 & ~regflags))
	return R_XMM0;
    if (!(YMM0 & ~regflags))
	return R_YMM0;
    if (!(REG_CS & ~regflags))
        return (regval == 1) ? R_CS : 0;
    if (!(REG_DESS & ~regflags))
        return (regval == 0 || regval == 2
                || regval == 3 ? nasm_rd_sreg[regval] : 0);
    if (!(REG_FSGS & ~regflags))
        return (regval == 4 || regval == 5 ? nasm_rd_sreg[regval] : 0);
    if (!(REG_SEG67 & ~regflags))
        return (regval == 6 || regval == 7 ? nasm_rd_sreg[regval] : 0);

    /* All the entries below look up regval in an 16-entry array */
    if (regval < 0 || regval > 15)
        return 0;

    if (!(REG8 & ~regflags)) {
	if (rex & REX_P)
	    return nasm_rd_reg8_rex[regval];
	else
	    return nasm_rd_reg8[regval];
    }
    if (!(REG16 & ~regflags))
        return nasm_rd_reg16[regval];
    if (!(REG32 & ~regflags))
        return nasm_rd_reg32[regval];
    if (!(REG64 & ~regflags))
        return nasm_rd_reg64[regval];
    if (!(REG_SREG & ~regflags))
        return nasm_rd_sreg[regval & 7]; /* Ignore REX */
    if (!(REG_CREG & ~regflags))
        return nasm_rd_creg[regval];
    if (!(REG_DREG & ~regflags))
        return nasm_rd_dreg[regval];
    if (!(REG_TREG & ~regflags)) {
	if (rex & REX_P)
	    return 0;		/* TR registers are ill-defined with rex */
        return nasm_rd_treg[regval];
    }
    if (!(FPUREG & ~regflags))
        return nasm_rd_fpureg[regval & 7]; /* Ignore REX */
    if (!(MMXREG & ~regflags))
        return nasm_rd_mmxreg[regval & 7]; /* Ignore REX */
    if (!(XMMREG & ~regflags))
        return nasm_rd_xmmreg[regval];
    if (!(YMMREG & ~regflags))
        return nasm_rd_ymmreg[regval];

    return 0;
}

/*
 * Process a DREX suffix
 */
static uint8_t *do_drex(uint8_t *data, insn *ins)
{
    uint8_t drex = *data++;
    operand *dst = &ins->oprs[ins->drexdst];

    if ((drex & 8) != ((ins->rex & REX_OC) ? 8 : 0))
	return NULL;	/* OC0 mismatch */
    ins->rex = (ins->rex & ~7) | (drex & 7);

    dst->segment = SEG_RMREG;
    dst->basereg = drex >> 4;
    return data;
}


/*
 * Process an effective address (ModRM) specification.
 */
static uint8_t *do_ea(uint8_t *data, int modrm, int asize,
		      int segsize, operand * op, insn *ins)
{
    int mod, rm, scale, index, base;
    int rex;
    uint8_t sib = 0;

    mod = (modrm >> 6) & 03;
    rm = modrm & 07;

    if (mod != 3 && rm == 4 && asize != 16)
	sib = *data++;

    if (ins->rex & REX_D) {
	data = do_drex(data, ins);
	if (!data)
	    return NULL;
    }
    rex = ins->rex;

    if (mod == 3) {             /* pure register version */
        op->basereg = rm+(rex & REX_B ? 8 : 0);
        op->segment |= SEG_RMREG;
        return data;
    }

    op->disp_size = 0;
    op->eaflags = 0;

    if (asize == 16) {
        /*
         * <mod> specifies the displacement size (none, byte or
         * word), and <rm> specifies the register combination.
         * Exception: mod=0,rm=6 does not specify [BP] as one might
         * expect, but instead specifies [disp16].
         */
        op->indexreg = op->basereg = -1;
        op->scale = 1;          /* always, in 16 bits */
        switch (rm) {
        case 0:
            op->basereg = R_BX;
            op->indexreg = R_SI;
            break;
        case 1:
            op->basereg = R_BX;
            op->indexreg = R_DI;
            break;
        case 2:
            op->basereg = R_BP;
            op->indexreg = R_SI;
            break;
        case 3:
            op->basereg = R_BP;
            op->indexreg = R_DI;
            break;
        case 4:
            op->basereg = R_SI;
            break;
        case 5:
            op->basereg = R_DI;
            break;
        case 6:
            op->basereg = R_BP;
            break;
        case 7:
            op->basereg = R_BX;
            break;
        }
        if (rm == 6 && mod == 0) {      /* special case */
            op->basereg = -1;
            if (segsize != 16)
                op->disp_size = 16;
            mod = 2;            /* fake disp16 */
        }
        switch (mod) {
        case 0:
            op->segment |= SEG_NODISP;
            break;
        case 1:
            op->segment |= SEG_DISP8;
            op->offset = (int8_t)*data++;
            break;
        case 2:
            op->segment |= SEG_DISP16;
            op->offset = *data++;
            op->offset |= ((unsigned)*data++) << 8;
            break;
        }
        return data;
    } else {
        /*
         * Once again, <mod> specifies displacement size (this time
         * none, byte or *dword*), while <rm> specifies the base
         * register. Again, [EBP] is missing, replaced by a pure
         * disp32 (this time that's mod=0,rm=*5*) in 32-bit mode,
	 * and RIP-relative addressing in 64-bit mode.
	 *
	 * However, rm=4
         * indicates not a single base register, but instead the
         * presence of a SIB byte...
         */
	int a64 = asize == 64;

        op->indexreg = -1;

	if (a64)
	    op->basereg = nasm_rd_reg64[rm | ((rex & REX_B) ? 8 : 0)];
	else
	    op->basereg = nasm_rd_reg32[rm | ((rex & REX_B) ? 8 : 0)];

        if (rm == 5 && mod == 0) {
	    if (segsize == 64) {
		op->eaflags |= EAF_REL;
		op->segment |= SEG_RELATIVE;
		mod = 2;	/* fake disp32 */
	    }

	    if (asize != 64)
		op->disp_size = asize;

	    op->basereg = -1;
	    mod = 2;            /* fake disp32 */
        }

        if (rm == 4) {          /* process SIB */
            scale = (sib >> 6) & 03;
            index = (sib >> 3) & 07;
            base = sib & 07;

            op->scale = 1 << scale;

	    if (index == 4 && !(rex & REX_X))
		op->indexreg = -1; /* ESP/RSP cannot be an index */
	    else if (a64)
		op->indexreg = nasm_rd_reg64[index | ((rex & REX_X) ? 8 : 0)];
	    else
		op->indexreg = nasm_rd_reg32[index | ((rex & REX_X) ? 8 : 0)];

	    if (base == 5 && mod == 0) {
		op->basereg = -1;
		mod = 2;	/* Fake disp32 */
	    } else if (a64)
		op->basereg = nasm_rd_reg64[base | ((rex & REX_B) ? 8 : 0)];
	    else
		op->basereg = nasm_rd_reg32[base | ((rex & REX_B) ? 8 : 0)];

	    if (segsize == 16)
		op->disp_size = 32;
        }

        switch (mod) {
        case 0:
            op->segment |= SEG_NODISP;
            break;
        case 1:
            op->segment |= SEG_DISP8;
            op->offset = gets8(data);
	    data++;
            break;
        case 2:
            op->segment |= SEG_DISP32;
            op->offset = gets32(data);
	    data += 4;
            break;
        }
        return data;
    }
}

/*
 * Determine whether the instruction template in t corresponds to the data
 * stream in data. Return the number of bytes matched if so.
 */
#define case4(x) case (x): case (x)+1: case (x)+2: case (x)+3

static int matches(const struct itemplate *t, uint8_t *data,
		   const struct prefix_info *prefix, int segsize, insn *ins)
{
    uint8_t *r = (uint8_t *)(t->code);
    uint8_t *origdata = data;
    bool a_used = false, o_used = false;
    enum prefixes drep = 0;
    uint8_t lock = prefix->lock;
    int osize = prefix->osize;
    int asize = prefix->asize;
    int i, c;
    struct operand *opx;
    int s_field_for = -1;	/* No 144/154 series code encountered */
    bool vex_ok = false;
    int regmask = (segsize == 64) ? 15 : 7;

    for (i = 0; i < MAX_OPERANDS; i++) {
	ins->oprs[i].segment = ins->oprs[i].disp_size =
	    (segsize == 64 ? SEG_64BIT : segsize == 32 ? SEG_32BIT : 0);
    }
    ins->condition = -1;
    ins->rex = prefix->rex;
    memset(ins->prefixes, 0, sizeof ins->prefixes);

    if (t->flags & (segsize == 64 ? IF_NOLONG : IF_LONG))
        return false;

    if (prefix->rep == 0xF2)
        drep = P_REPNE;
    else if (prefix->rep == 0xF3)
        drep = P_REP;

    while ((c = *r++) != 0) {
	opx = &ins->oprs[c & 3];

	switch (c) {
	case 01:
	case 02:
	case 03:
            while (c--)
                if (*r++ != *data++)
                    return false;
	    break;

	case4(010):
	{
            int t = *r++, d = *data++;
            if (d < t || d > t + 7)
                return false;
            else {
                opx->basereg = (d-t)+
		    (ins->rex & REX_B ? 8 : 0);
                opx->segment |= SEG_RMREG;
            }
	    break;
	}

	case4(014):
	case4(0274):
            opx->offset = (int8_t)*data++;
            opx->segment |= SEG_SIGNED;
	    break;

	case4(020):
            opx->offset = *data++;
	    break;

	case4(024):
            opx->offset = *data++;
	    break;

	case4(030):
            opx->offset = getu16(data);
	    data += 2;
	    break;

	case4(034):
	    if (osize == 32) {
		opx->offset = getu32(data);
		data += 4;
	    } else {
		opx->offset = getu16(data);
		data += 2;
	    }
            if (segsize != asize)
                opx->disp_size = asize;
	    break;

	case4(040):
	case4(0254):
            opx->offset = getu32(data);
	    data += 4;
	    break;

	case4(044):
	    switch (asize) {
	    case 16:
		opx->offset = getu16(data);
		data += 2;
		if (segsize != 16)
		    opx->disp_size = 16;
		break;
	    case 32:
		opx->offset = getu32(data);
		data += 4;
		if (segsize == 16)
		    opx->disp_size = 32;
		break;
	    case 64:
		opx->offset = getu64(data);
		opx->disp_size = 64;
		data += 8;
		break;