x86nasm.md

开源的nasm编译器源码,研究编译器原理很有帮且

	vmask[IREG] = 0;
	tmask[FREG] = 0xff;
	vmask[FREG] = 0;
	cseg = 0;
	quo = mkreg("eax", EAX, 1, IREG);
	quo->x.regnode->mask |= 1<<EDX;
	rem = mkreg("edx", EDX, 1, IREG);
	rem->x.regnode->mask |= 1<<EAX;
}
static void segment(n) int n; {
	if (n == cseg)
		return;
	cseg = n;
	if (cseg == CODE)
		print("[section .text]\n");
	else if (cseg == DATA || cseg == LIT)
		print("[section .data]\n");
	else if (cseg == BSS)
		print("[section .bss]\n");
}
static void progend() {

}
static void target(p) Node p; {
	assert(p);
	switch (p->op) {
	case RSHI: case RSHU: case LSHI: case LSHU:
		if (generic(p->kids[1]->op) != CNST
		&& !(   generic(p->kids[1]->op) == INDIR
		     && p->kids[1]->kids[0]->op == VREG+P
		     && p->kids[1]->syms[RX]->u.t.cse
		     && generic(p->kids[1]->syms[RX]->u.t.cse->op) == CNST
)) {
			rtarget(p, 1, intreg[ECX]);
			setreg(p, intreg[EAX]);
		}
		break;
	case MULU:
		setreg(p, quo);
		rtarget(p, 0, intreg[EAX]);
		break;
	case DIVI: case DIVU:
		setreg(p, quo);
		rtarget(p, 0, intreg[EAX]);
		rtarget(p, 1, intreg[ECX]);
		break;
	case MODI: case MODU:
		setreg(p, rem);
		rtarget(p, 0, intreg[EAX]);
		rtarget(p, 1, intreg[ECX]);
		break;
	case ASGNB:
		rtarget(p, 0, intreg[EDI]);
		rtarget(p->kids[1], 0, intreg[ESI]);
		break;
	case ARGB:
		rtarget(p->kids[0], 0, intreg[ESI]);
		break;
	case CALLI: case CALLV:
		setreg(p, intreg[EAX]);
		break;
	case RETI:
		rtarget(p, 0, intreg[EAX]);
		break;
	}
}

static void clobber(p) Node p; {
	static int nstack = 0;

	assert(p);
	nstack = ckstack(p, nstack);
	assert(p->count > 0 || nstack == 0);
	switch (p->op) {
	case ASGNB: case ARGB:
		spill(1<<ECX | 1<<ESI | 1<<EDI, IREG, p);
		break;
	case EQD: case LED: case GED: case LTD: case GTD: case NED:
	case EQF: case LEF: case GEF: case LTF: case GTF: case NEF:
		spill(1<<EAX, IREG, p);
		break;
	case CALLD: case CALLF:
		spill(1<<EDX | 1<<EAX, IREG, p);
		break;
	}
}
#define isfp(p) (optype((p)->op)==F || optype((p)->op)==D)

static int ckstack(p, n) Node p; int n; {
	int i;

	for (i = 0; i < NELEMS(p->x.kids) && p->x.kids[i]; i++)
		if (isfp(p->x.kids[i]))
			n--;
	if (isfp(p) && p->count > 0)
		n++;
	if (n > 8)
		error("expression too complicated\n");
	debug(fprint(2, "(ckstack(%x)=%d)\n", p, n));
	assert(n >= 0);
	return n;
}
static int memop(p) Node p; {
	assert(p);
	assert(generic(p->op) == ASGN);
	assert(p->kids[0]);
	assert(p->kids[1]);
	if (generic(p->kids[1]->kids[0]->op) == INDIR
	&& sametree(p->kids[0], p->kids[1]->kids[0]->kids[0]))
		return 3;
	else
		return LBURG_MAX;
}
static int sametree(p, q) Node p, q; {
	return p == NULL && q == NULL
	|| p && q && p->op == q->op && p->syms[0] == q->syms[0]
		&& sametree(p->kids[0], q->kids[0])
		&& sametree(p->kids[1], q->kids[1]);
}
static void emit2(p) Node p; {
#define preg(f) ((f)[getregnum(p->x.kids[0])]->x.name)

	if (p->op == CVCI)
		print("movsx %s,%s\n", p->syms[RX]->x.name
, preg(charreg));
	else if (p->op == CVCU)
		print("movzx %s,%s\n", p->syms[RX]->x.name
, preg(charreg));
	else if (p->op == CVSI)
		print("movsx %s,%s\n", p->syms[RX]->x.name
, preg(shortreg));
	else if (p->op == CVSU)
		print("movzx %s,%s\n", p->syms[RX]->x.name
, preg(shortreg));
	else if (p->op == CVIC || p->op == CVIS
	      || p->op == CVUC || p->op == CVUS) {
		char *dst = shortreg[getregnum(p)]->x.name;
		char *src = preg(shortreg);
		if (dst != src)
			print("mov %s,%s\n", dst, src);
	}
}

static void doarg(p) Node p; {
	assert(p && p->syms[0]);
	mkactual(4, p->syms[0]->u.c.v.i);
}
static void blkfetch(k, off, reg, tmp)
int k, off, reg, tmp; {}
static void blkstore(k, off, reg, tmp)
int k, off, reg, tmp; {}
static void blkloop(dreg, doff, sreg, soff, size, tmps)
int dreg, doff, sreg, soff, size, tmps[]; {}
static void local(p) Symbol p; {
	if (isfloat(p->type))
		p->sclass = AUTO;
	if (askregvar(p, rmap[ttob(p->type)]) == 0)
		mkauto(p);
}
static void function(f, caller, callee, n)
Symbol f, callee[], caller[]; int n; {
	int i;

	print("%s:\n", f->x.name);
	print("push ebx\n");
	print("push esi\n");
	print("push edi\n");
	print("push ebp\n");
	print("mov ebp,esp\n");
usedmask[0] = usedmask[1] = 0;
freemask[0] = freemask[1] = ~(unsigned)0;
	offset = 16 + 4;
	for (i = 0; callee[i]; i++) {
		Symbol p = callee[i];
		Symbol q = caller[i];
		assert(q);
		p->x.offset = q->x.offset = offset;
		p->x.name = q->x.name = stringf("%d", p->x.offset);
		p->sclass = q->sclass = AUTO;
		offset += roundup(q->type->size, 4);
	}
	assert(caller[i] == 0);
	offset = maxoffset = 0;
	gencode(caller, callee);
	framesize = roundup(maxoffset, 4);
	if (framesize > 0)
		print("sub esp,%d\n", framesize);
	emitcode();
	print("mov esp,ebp\n");
	print("pop ebp\n");
	print("pop edi\n");
	print("pop esi\n");
	print("pop ebx\n");
	print("ret\n");
}
static void defsymbol(p) Symbol p; {
	if (p->scope >= LOCAL && p->sclass == STATIC)
		p->x.name = stringf("L%d", genlabel(1));
	else if (p->generated)
		p->x.name = stringf("$L%s", p->name);
	else if (p->scope == GLOBAL || p->sclass == EXTERN)
	/* CHANGE THIS FOR a.out */
#if 0
		p->x.name = stringf("$_%s", p->name);
#else
		p->x.name = stringf("$%s", p->name);
#endif
	else if (p->scope == CONSTANTS
	&& (isint(p->type) || isptr(p->type))
	&& p->name[0] == '0' && p->name[1] == 'x')
		p->x.name = stringf("0%sH", &p->name[2]);
	else
		p->x.name = p->name;
}
static void address(q, p, n) Symbol q, p; int n; {
	if (p->scope == GLOBAL
	|| p->sclass == STATIC || p->sclass == EXTERN)
		q->x.name = stringf("%s%s%d",
			p->x.name, n >= 0 ? "+" : "", n);
	else {
		q->x.offset = p->x.offset + n;
		q->x.name = stringd(q->x.offset);
	}
}
static void defconst(ty, v) int ty; Value v; {
	switch (ty) {
		case C: print("db %d\n",   v.uc); return;
		case S: print("dw %d\n",   v.ss); return;
		case I: print("dd %d\n",   v.i ); return;
		case U: print("dd 0%xH\n", v.u ); return;
		case P: print("dd 0%xH\n", v.p ); return;
		case F:
			print("dd 0%xH\n", *(unsigned *)&v.f);
			return;
		case D: {
			unsigned *p = (unsigned *)&v.d;
			print("dd 0%xH,0%xH\n", p[swap], p[1 - swap]);
			return;
			}
	}
	assert(0);
}
static void defaddress(p) Symbol p; {
	print("dd %s\n", p->x.name);
}
static void defstring(n, str) int n; char *str; {
	char *s;
	int inquote = 1;

	print("db '");

	for (s = str; s < str + n; s++)
	{
		if ((*s & 0x7F) == *s && *s >= ' ' && *s != '\'') {
			if (!inquote){
				print(", '");
				inquote = 1;
			}
			print("%c",*s);
		}
		else
		{
			if (inquote){
				print("', ");
				inquote = 0;
			}
			else
				print(", ");
			print("%d",*s);
		}
	}
	if (inquote) print("'");
	print("\n");
}
static void export(p) Symbol p; {
	print("[global %s]\n", p->x.name);
}
static void import(p) Symbol p; {
	if (p->ref > 0) {
		print("[extern %s]\n", p->x.name);
	}
}
static void global(p) Symbol p; {
	int i;

	if (p->u.seg == BSS)
		print("resb ($-$$) & %d\n",
			p->type->align > 4 ? 3 : p->type->align-1);
	else
		print("times ($-$$) & %d nop\n",
			p->type->align > 4 ? 3 : p->type->align-1);
	print("%s:\n", p->x.name);
	if (p->u.seg == BSS)
		print("resb %d\n", p->type->size);
}
static void space(n) int n; {
	int i;

	if (cseg != BSS)
		print("times %d db 0\n", n);
}
Interface x86nasmIR = {
	1, 1, 0,  /* char */
	2, 2, 0,  /* short */
	4, 4, 0,  /* int */
	4, 4, 1,  /* float */
	8, 4, 1,  /* double */
	4, 4, 0,  /* T * */
	0, 4, 0,  /* struct; so that ARGB keeps stack aligned */
	1,        /* little_endian */
	0,        /* mulops_calls */
	0,        /* wants_callb */
	1,        /* wants_argb */
	0,        /* left_to_right */
	0,        /* wants_dag */
	address,
	blockbeg,
	blockend,
	defaddress,
	defconst,
	defstring,
	defsymbol,
	emit,
	export,
	function,
	gen,
	global,
	import,
	local,
	progbeg,
	progend,
	segment,
	space,
	0, 0, 0, 0, 0, 0, 0,
	{1, blkfetch, blkstore, blkloop,
	    _label,
	    _rule,
	    _nts,
	    _kids,
	    _opname,
	    _arity,
	    _string,
	    _templates,
	    _isinstruction,
	    _ntname,
	    emit2,
	    doarg,
	    target,
	    clobber,
}
};