MOV	$0x800000, A0                     // stacksize = 8192KB
	MOV	$_rt0_riscv64_freebsd_lib_go(SB), A1
	MOV	A0, 8(X2)
	MOV	A1, 16(X2)
	MOV	$runtime·newosproc0(SB), T0
	JALR	RA, T0

restore:
	// Restore callee-save registers, along with X1 (LR).
	MOV	(8*3)(X2), X1
	MOV	(8*4)(X2), X8
	MOV	(8*5)(X2), X9
	MOV	(8*6)(X2), X18
	MOV	(8*7)(X2), X19
	MOV	(8*8)(X2), X20
	MOV	(8*9)(X2), X21
	MOV	(8*10)(X2), X22
	MOV	(8*11)(X2), X23

	BNEZ	X5, f_align

f_loop_check:
	MOV	$16, X9
	BLT	X12, X9, f_loop8_check
	MOV	$32, X9
	BLT	X12, X9, f_loop16_check
	MOV	$64, X9
	BLT	X12, X9, f_loop32_check
f_loop64:
	MOV	0(X11), X14
	MOV	8(X11), X15
	MOV	16(X11), X16
	MOV	24(X11), X17
	MOV	32(X11), X18
	MOV	40(X11), X19
	MOV	48(X11), X20
	MOV	56(X11), X21
	MOV	X14, 0(X10)
	MOV	X15, 8(X10)
	MOV	X16, 16(X10)
	MOV	X17, 24(X10)
	MOV	X18, 32(X10)

// err | A2          | part of A0
TEXT ·Syscall6<ABIInternal>(SB),NOSPLIT,$0-80
	MOV	A0, A7
	MOV	A1, A0
	MOV	A2, A1
	MOV	A3, A2
	MOV	A4, A3
	MOV	A5, A4
	MOV	A6, A5
	ECALL
	MOV	$-4096, T0
	BLTU	T0, A0, err
	// r1 already in A0
	// r2 already in A1
	MOV	ZERO, A2 // errno
	RET
err:
	SUB	A0, ZERO, A2 // errno
	MOV	$-1, A0	     // r1
	MOV	ZERO, A1     // r2

	MOV	(0*8)(X20), X5
	MOV	(1*8)(X20), X6
	MOV	(2*8)(X20), X7
	MOV	(3*8)(X20), X8
	ADD	X5, X10		// H0 = a + H0
	ADD	X6, X11		// H1 = b + H1
	ADD	X7, X12		// H2 = c + H2
	ADD	X8, X13		// H3 = d + H3
	MOV	X10, (0*8)(X20)
	MOV	X11, (1*8)(X20)
	MOV	X12, (2*8)(X20)
	MOV	X13, (3*8)(X20)
	MOV	(4*8)(X20), X5
	MOV	(5*8)(X20), X6
	MOV	(6*8)(X20), X7
	MOV	(7*8)(X20), X8

finish:
	MOV	8(X2), T0	// sec
	MOV	16(X2), T1	// nsec
	// restore stack
	MOV	S2, X2
	MOV	24(X2), T2
	MOV	T2, m_vdsoPC(S3)

	MOV	32(X2), T2
	MOV	T2, m_vdsoSP(S3)
	// sec is in T0, nsec in T1
	// return nsec in T0
	MOV	$1000000000, T2
	MUL	T2, T0
	ADD	T1, T0
	MOV	T0, ret+0(FP)
	RET

fallback:
	MOV	$8(X2), A1
	MOV	$SYS_clock_gettime, A7
	ECALL
	MOV	8(X2), T0	// sec

TEXT runtime·usleep(SB),NOSPLIT,$24-4
	MOVWU	usec+0(FP), A0
	MOV	$1000, A1
	MUL	A1, A0, A0
	MOV	$1000000000, A1
	DIV	A1, A0, A2
	MOV	A2, 8(X2)
	REM	A1, A0, A3
	MOV	A3, 16(X2)
	ADD	$8, X2, A0
	MOV	ZERO, A1
	MOV	$SYS_nanosleep, T0
	ECALL
	RET

// func thr_self() thread
TEXT runtime·thr_self(SB),NOSPLIT,$8-8
	MOV	$ptr-8(SP), A0	// arg 1 &8(SP)
	MOV	$SYS_thr_self, T0
	ECALL
	MOV	ptr-8(SP), A0

	 * 8(X2).
	 */
	ADD	$(-8*29), X2
	MOV	X10, (8*1)(X2) // fn unsafe.Pointer
	MOV	X11, (8*2)(X2) // a unsafe.Pointer
	MOV	X13, (8*3)(X2) // ctxt uintptr
	MOV	X8, (8*4)(X2)
	MOV	X9, (8*5)(X2)
	MOV	X18, (8*6)(X2)
	MOV	X19, (8*7)(X2)
	MOV	X20, (8*8)(X2)
	MOV	X21, (8*9)(X2)
	MOV	X22, (8*10)(X2)
	MOV	X23, (8*11)(X2)
	MOV	X24, (8*12)(X2)
	MOV	X25, (8*13)(X2)
	MOV	X26, (8*14)(X2)
	MOV	g, (8*15)(X2)
	MOV	X1, (8*16)(X2)

	MOV	$64, X9
	BLT	X11, X9, zero32
loop64:
	MOV	ZERO, 0(X10)
	MOV	ZERO, 8(X10)
	MOV	ZERO, 16(X10)
	MOV	ZERO, 24(X10)
	MOV	ZERO, 32(X10)
	MOV	ZERO, 40(X10)
	MOV	ZERO, 48(X10)
	MOV	ZERO, 56(X10)
	ADD	$64, X10
	SUB	$64, X11
	BGE	X11, X9, loop64
	BEQZ	X11, done

check32:
	MOV	$32, X9
	BLT	X11, X9, check16
zero32:
	MOV	ZERO, 0(X10)
	MOV	ZERO, 8(X10)
	MOV	ZERO, 16(X10)
	MOV	ZERO, 24(X10)

			// amd64 doesn't support CMOV with byte registers
			return false
		}
		return true
	default:
		return false
	}
}

// elimIf converts the one-way branch starting at dom in f to a conditional move if possible.
// loadAddr is a set of values which are used to compute the address of a load.
// Those values are exempt from CMOV generation.

func cmovfloatmove(x, y int) float64 {
	a := 1.0
	if x <= y {
		a = 2.0
	}
	// amd64:-"CMOV"
	// arm64:-"CSEL"
	// ppc64x:-"ISEL"
	// wasm:-"Select"
	return a
}

// On amd64, the following patterns trigger comparison inversion.
// Test that we correctly invert the CMOV condition
var gsink int64
var gusink uint64

func cmovinvert1(x, y int64) int64 {
	if x < gsink {
		y = -y