TEXT ·asyncPreempt(SB),NOSPLIT|NOFRAME,$0-0
	MOVW.W R14, -188(R13)
	MOVW R0, 4(R13)
	MOVW R1, 8(R13)
	MOVW R2, 12(R13)
	MOVW R3, 16(R13)
	MOVW R4, 20(R13)
	MOVW R5, 24(R13)
	MOVW R6, 28(R13)
	MOVW R7, 32(R13)
	MOVW R8, 36(R13)
	MOVW R9, 40(R13)
	MOVW R11, 44(R13)
	MOVW R12, 48(R13)
	MOVW CPSR, R0
	MOVW R0, 52(R13)
	MOVB ·goarmsoftfp(SB), R0
	CMP $0, R0
	BNE nofp
	MOVW FPCR, R0
	MOVW R0, 56(R13)
	MOVD F0, 60(R13)

TEXT	racecall<>(SB), NOSPLIT, $0-0
	BL	runtime·save_g(SB)		// Save g for callbacks.
	MOVD	R15, R7				// Save SP.
	MOVD	g_m(g), R8			// R8 = thread.
	MOVD	m_g0(R8), R8			// R8 = g0.
	CMPBEQ	R8, g, call			// Already on g0?
	MOVD	(g_sched+gobuf_sp)(R8), R15	// Switch SP to g0.
call:	SUB	$160, R15			// Allocate C frame.
	BL	R1				// Call C code.
	MOVD	R7, R15				// Restore SP.
	RET					// Return to Go.

//

// func rawVforkSyscall(trap, a1, a2, a3 uintptr) (r1, err uintptr)
TEXT ·rawVforkSyscall(SB),NOSPLIT|NOFRAME,$0-48
	MOVV	a1+8(FP), R4
	MOVV	a2+16(FP), R5
	MOVV	a3+24(FP), R6
	MOVV	R0, R7
	MOVV	R0, R8
	MOVV	R0, R9
	MOVV	trap+0(FP), R2	// syscall entry
	SYSCALL
	BEQ	R7, ok
	MOVV	$-1, R1
	MOVV	R1, r1+32(FP)	// r1
	MOVV	R2, err+40(FP)	// errno
	RET
ok:
	MOVV	R2, r1+32(FP)	// r1

	// Cannot grow signal stack (m->gsignal).
	MOVD	m_gsignal(R8), R4
	CMP	g, R4
	BNE	3(PC)
	BL	runtime·badmorestackgsignal(SB)
	B	runtime·abort(SB)

	// Called from f.
	// Set m->morebuf to f's callers.
	MOVD	R3, (m_morebuf+gobuf_pc)(R8)	// f's caller's PC
	MOVD	RSP, R0
	MOVD	R0, (m_morebuf+gobuf_sp)(R8)	// f's caller's RSP
	MOVD	g, (m_morebuf+gobuf_g)(R8)

	// Call newstack on m->g0's stack.
	MOVD	m_g0(R8), g

	MOVD	R1, (g_sched+gobuf_sp)(g)
	MOVD	LR, R8
	MOVD	R8, (g_sched+gobuf_pc)(g)
	MOVD	R5, (g_sched+gobuf_lr)(g)
	MOVD	R11, (g_sched+gobuf_ctxt)(g)

	// Cannot grow scheduler stack (m->g0).
	MOVD	g_m(g), R7
	MOVD	m_g0(R7), R8
	CMP	g, R8
	BNE	3(PC)
	BL	runtime·badmorestackg0(SB)
	BL	runtime·abort(SB)

	// Cannot grow signal stack (m->gsignal).
	MOVD	m_gsignal(R7), R8
	CMP	g, R8
	BNE	3(PC)

	MOVD	0(R14), R3		// arg 1 - addr
	MOVD	8(R14), R4		// arg 2 - len
	MOVW	16(R14), R5		// arg 3 - prot
	MOVW	20(R14), R6		// arg 4 - flags
	MOVW	24(R14), R7		// arg 5 - fid
	MOVW	28(R14), R8		// arg 6 - offset
	CALL	libc_mmap(SB)
	MOVD	$0, R4
	CMP	R3, $-1
	BNE	noerr
	CALL	libc_errno(SB)
	MOVW	(R3), R4		// errno
	MOVD	$0, R3
noerr:
	MOVD	R3, 32(R14)
	MOVD	R4, 40(R14)
	RET

	// R4: size of stack arguments (n-8)*8
	// R5: &args[8]
	// R6: loop counter, from 0 to (n-8)*8
	// R7: scratch
	// R8: copy of RSP - (R2)(RSP) assembles as (R2)(ZR)
	SUB	$8, R0, R4
	LSL	$3, R4
	ADD	$(8*8), R12, R5
	MOVD	$0, R6
	MOVD	RSP, R8
stackargs:
	MOVD	(R6)(R5), R7
	MOVD	R7, (R6)(R8)
	ADD	$8, R6
	CMP	R6, R4
	BNE	stackargs

_8args:
	MOVD	(7*8)(R12), R7
_7args:
	MOVD	(6*8)(R12), R6
_6args:

	MOVL CX, 0(SP)
	INCL CX
	MOVL CX, 4(SP)
	INCL CX
	MOVL CX, 8(SP)
	INCL CX
	MOVL CX, 12(SP)
	MOVOU 0(SP), X12

	// Load seed words into next two rows and into DI, SI, R8..R13
	SEED(0, DI, X4)
	SEED(1, SI, X5)
	SEED(2, R8, X6)
	SEED(3, R9, X7)
	SEED(4, R10, X8)
	SEED(5, R11, X9)
	SEED(6, R12, X10)
	SEED(7, R13, X11)

	// Zeros for remaining two matrix entries.

.globl __umodti3
__umodti3:
	# specialized to u128 % u64, so verify that
	test %rcx,%rcx
	jne 1f

	# save divisor
	movq %rdx, %r8

	# reduce top 64 bits mod divisor
	movq %rsi, %rax
	xorl %edx, %edx
	divq %r8

	# reduce full 128-bit mod divisor
	# quotient fits in 64 bits because top 64 bits have been reduced < divisor.

TEXT ·Syscall9(SB),NOSPLIT,$28-52
	NO_LOCAL_POINTERS
	JAL	runtime·entersyscall(SB)
	MOVW	a1+4(FP), R4
	MOVW	a2+8(FP), R5
	MOVW	a3+12(FP), R6
	MOVW	a4+16(FP), R7
	MOVW	a5+20(FP), R8
	MOVW	a6+24(FP), R9
	MOVW	a7+28(FP), R10
	MOVW	a8+32(FP), R11
	MOVW	R8, 16(R29)
	MOVW	R9, 20(R29)
	MOVW	R10, 24(R29)
	MOVW	R11, 28(R29)
	MOVW	trap+0(FP), R2	// syscall entry
	MOVW	R0, R3	// reset R3 to zero as 1-ret SYSCALL keeps it
	SYSCALL