TEXT _rt0_386_windows_lib_go(SB),NOSPLIT,$0
	PUSHL	$0
	PUSHL	$0
	JMP	runtime·rt0_go(SB)

TEXT _main(SB),NOSPLIT,$0
	// Remove the return address from the stack.
	// rt0_go doesn't expect it to be there.
	ADDL	$4, SP

		// binary ops
		{name: "ADDL", argLength: 2, reg: gp21sp, asm: "ADDL", commutative: true, clobberFlags: true},                // arg0 + arg1
		{name: "ADDLconst", argLength: 1, reg: gp11sp, asm: "ADDL", aux: "Int32", typ: "UInt32", clobberFlags: true}, // arg0 + auxint

TEXT runtime·_mul64by32(SB), NOSPLIT, $0
	MOVL	lo64+0(FP), CX
	MOVL	a_lo+4(FP), AX
	MULL	b+12(FP)
	MOVL	AX, 0(CX)
	MOVL	DX, BX
	MOVL	a_hi+8(FP), AX
	MULL	b+12(FP)
	ADDL	AX, BX
	ADCL	$0, DX
	MOVL	BX, 4(CX)
	MOVL	DX, AX
	MOVL	AX, hi32+16(FP)
	RET

TEXT runtime·_div64by32(SB), NOSPLIT, $0
	MOVL	r+12(FP), CX
	MOVL	a_lo+0(FP), AX
	MOVL	a_hi+4(FP), DX
	DIVL	b+8(FP)

// Atomically:
//	*val += delta;
//	return *val;
TEXT ·Xadd(SB), NOSPLIT, $0-12
	MOVL	ptr+0(FP), BX
	MOVL	delta+4(FP), AX
	MOVL	AX, CX
	LOCK
	XADDL	AX, 0(BX)
	ADDL	CX, AX
	MOVL	AX, ret+8(FP)
	RET

TEXT ·Xadd64(SB), NOSPLIT, $0-20
	NO_LOCAL_POINTERS
	// no XADDQ so use CMPXCHG8B loop
	MOVL	ptr+0(FP), BP
	TESTL	$7, BP
	JZ	2(PC)
	CALL	·panicUnaligned(SB)

add    eax, DWORD PTR 8[esp]
ret    0
_sumx    ENDP
_TEXT   ENDS
END
'''

    private static String i386GnuAsmSource = '''
    .text
    .globl  _sumx
_sumx:
    movl    8(%esp), %eax
    addl    4(%esp), %eax
    ret
'''

    private static String x64GnuAsmSource = '''
    .text
    .p2align 4,,15
.globl sumx
    .type   sumx, @function
sumx:
    leal    (%rsi,%rdi), %eax
    ret

	ADDQ 3395469782(AX), AX         // ERROR "offset too large"
	ADDL 3395469782(AX), AX         // ERROR "offset too large"
	ADDW 3395469782(AX), AX         // ERROR "offset too large"
	LEAQ 433954697820(AX), AX       // ERROR "offset too large"
	ADDQ 433954697820(AX), AX       // ERROR "offset too large"
	ADDL 433954697820(AX), AX       // ERROR "offset too large"

	MOVOU	X0, 112(DI)
	MOVOU	X0, 128(DI)
	MOVOU	X0, 144(DI)
	MOVOU	X0, 160(DI)
	MOVOU	X0, 176(DI)
	MOVOU	X0, 192(DI)
	MOVOU	X0, 208(DI)
	MOVOU	X0, 224(DI)
	MOVOU	X0, 240(DI)
	SUBL	$256, BX
	ADDL	$256, DI
	CMPL	BX, $256
	JAE	loop
	JMP	tail

_1or2:
	MOVB	AX, (DI)
	MOVB	AX, -1(DI)(BX*1)
	RET
_0:
	RET
_3:
	MOVW	AX, (DI)
	MOVB	AX, 2(DI)
	RET
_4:

	/* copy return values back */		\
	MOVL	stackArgsType+0(FP), DX;		\
	MOVL	stackArgs+8(FP), DI;		\
	MOVL	stackArgsSize+12(FP), CX;		\
	MOVL	stackRetOffset+16(FP), BX;		\
	MOVL	SP, SI;				\
	ADDL	BX, DI;				\
	ADDL	BX, SI;				\
	SUBL	BX, CX;				\
	CALL	callRet<>(SB);			\
	RET

// callRet copies return values back at the end of call*. This is a
// separate function so it can allocate stack space for the arguments

	MOVL	4(SP), CX
	MOVL	CX, m_vdsoSP(SI)
	MOVL	0(SP), CX
	MOVL	CX, m_vdsoPC(SI)

	// sec is in AX, nsec in BX
	// convert to DX:AX nsec
	MOVL	$1000000000, CX
	MULL	CX
	ADDL	BX, AX
	ADCL	$0, DX

	MOVL	AX, ret_lo+0(FP)
	MOVL	DX, ret_hi+4(FP)
	RET

TEXT runtime·rtsigprocmask(SB),NOSPLIT,$0
	MOVL	$SYS_rt_sigprocmask, AX
	MOVL	how+0(FP), BX
	MOVL	new+4(FP), CX

		{name: "ADDQ", argLength: 2, reg: gp21sp, asm: "ADDQ", commutative: true, clobberFlags: true},
		{name: "ADDL", argLength: 2, reg: gp21sp, asm: "ADDL", commutative: true, clobberFlags: true},
		{name: "ADDQconst", argLength: 1, reg: gp11sp, asm: "ADDQ", aux: "Int32", typ: "UInt64", clobberFlags: true},
		{name: "ADDLconst", argLength: 1, reg: gp11sp, asm: "ADDL", aux: "Int32", clobberFlags: true},