// so this function is reentrant.
	MOVQ	m_vdsoPC(BX), CX
	MOVQ	m_vdsoSP(BX), DX
	MOVQ	CX, 0(SP)
	MOVQ	DX, 8(SP)

	LEAQ	sec+0(FP), DX
	MOVQ	-8(DX), CX	// Sets CX to function return address.
	MOVQ	CX, m_vdsoPC(BX)
	MOVQ	DX, m_vdsoSP(BX)

	CMPQ	R14, m_curg(BX)	// Only switch if on curg.
	JNE	noswitch

	MOVQ	m_g0(BX), DX
	MOVQ	(g_sched+gobuf_sp)(DX), SP	// Set SP to g0 stack

noswitch:

	// frame is specially handled in the runtime. See the comment above LOCAL_RETVALID.
	LEAQ	LOCAL_REGARGS(SP), R12
	CALL	runtime·spillArgs(SB)
	MOVQ	DX, 24(SP) // outside of moveMakeFuncArgPtrs's arg area
	MOVQ	DX, 0(SP)
	MOVQ	R12, 8(SP)
	CALL	·moveMakeFuncArgPtrs(SB)
	MOVQ	24(SP), DX
	MOVQ	DX, 0(SP)
	LEAQ	argframe+0(FP), CX
	MOVQ	CX, 8(SP)
	MOVB	$0, LOCAL_RETVALID(SP)
	LEAQ	LOCAL_RETVALID(SP), AX
	MOVQ	AX, 16(SP)

	// CX = a_cap  (unused)
	// DI = b_base (want in DI)
	// SI = b_len  (want in DX)
	// R8 = b_cap  (unused)
	MOVQ	SI, DX
	MOVQ	AX, SI
	JMP	cmpbody<>(SB)

TEXT runtime·cmpstring<ABIInternal>(SB),NOSPLIT,$0-40
	// AX = a_base (want in SI)
	// BX = a_len  (want in BX)
	// CX = b_base (want in DI)
	// DI = b_len  (want in DX)
	MOVQ	AX, SI
	MOVQ	DI, DX
	MOVQ	CX, DI
	JMP	cmpbody<>(SB)

// input:
//   SI = a

	CALL _expand_key_128<>(SB)
Lexp_dec:
	// dec
	SUBQ $16, BX
	MOVUPS (BX), X1
	MOVUPS X1, (DX)
	DECQ CX
Lexp_dec_loop:
	MOVUPS -16(BX), X1
	AESIMC X1, X0
	MOVUPS X0, 16(DX)
	SUBQ $16, BX
	ADDQ $16, DX
	DECQ CX
	JNZ Lexp_dec_loop
	MOVUPS -16(BX), X0
	MOVUPS X0, 16(DX)
	RET

TEXT _expand_key_128<>(SB),NOSPLIT,$0
	PSHUFD $0xff, X1, X1
	SHUFPS $0x10, X0, X4
	PXOR X4, X0

	LEAQ	8(SP), SI
	SYSCALL
	MOVQ	8(SP), AX	// sec
	MOVQ	16(SP), DX	// nsec

	// sec is in AX, nsec in DX
	// return nsec in AX
	IMULQ	$1000000000, AX
	ADDQ	DX, AX
	MOVQ	AX, ret+0(FP)
	RET

TEXT runtime·asmSigaction(SB),NOSPLIT,$0
	MOVQ	sig+0(FP), DI		// arg 1 sig
	MOVQ	new+8(FP), SI		// arg 2 act
	MOVQ	old+16(FP), DX		// arg 3 oact
	MOVL	$SYS_sigaction, AX
	SYSCALL
	JCC	2(PC)

	VGF2P8AFFINEINVQB $47, (DX), Z14, K2, Z28          // 62638d4acf222f
	VGF2P8AFFINEINVQB $47, Z3, Z28, K2, Z28            // 62639d42cfe32f
	VGF2P8AFFINEINVQB $47, Z12, Z28, K2, Z28           // 62439d42cfe42f
	VGF2P8AFFINEINVQB $47, 99(R15)(R15*1), Z28, K2, Z28 // 62039d42cfa43f630000002f
	VGF2P8AFFINEINVQB $47, (DX), Z28, K2, Z28          // 62639d42cf222f

	MOVL	4(BX), SI		// arg 2 cmd
	MOVL	8(BX), DX		// arg 3 arg
	XORL	AX, AX			// vararg: say "no float args"
	CALL	libc_fcntl(SB)
	XORL	DX, DX
	CMPL	AX, $-1
	JNE	noerr
	CALL	libc_errno(SB)
	MOVL	(AX), DX
	MOVL	$-1, AX
noerr:
	MOVL	AX, 12(BX)
	MOVL	DX, 16(BX)
	RET

TEXT runtime·sigaction_trampoline(SB),NOSPLIT,$0
	MOVQ	8(DI), SI		// arg 2 new
	MOVQ	16(DI), DX		// arg 3 old
	MOVL	0(DI), DI		// arg 1 sig

	get_tls(BX)
	MOVQ	DX, g(BX)
	MOVQ	DI, g_m(DX)

	// Layout new m scheduler stack on os stack.
	MOVQ	SP, AX
	MOVQ	AX, (g_stack+stack_hi)(DX)
	SUBQ	$(0x100000), AX		// stack size
	MOVQ	AX, (g_stack+stack_lo)(DX)
	ADDQ	$const_stackGuard, AX
	MOVQ	AX, g_stackguard0(DX)
	MOVQ	AX, g_stackguard1(DX)

	// Someday the convention will be D is always cleared.
	CLD

	VPCMPEQB	Y2, Y3, Y5
	VPAND	Y4, Y5, Y6
	VPMOVMSKB Y6, DX
	ADDQ	$64, SI
	ADDQ	$64, DI
	SUBQ	$64, BX
	CMPL	DX, $0xffffffff
	JEQ	hugeloop_avx2
	VZEROUPPER
	XORQ	AX, AX	// return 0
	RET

bigloop_avx2:
	VZEROUPPER

	// 8 bytes at a time using 64-bit register
	PCALIGN $16
bigloop:
	CMPQ	BX, $8
	JBE	leftover
	MOVQ	(SI), CX
	MOVQ	(DI), DX
	ADDQ	$8, SI
	ADDQ	$8, DI
	SUBQ	$8, BX

	VMOVDQU8 X11, K5, -7(CX)(DX*1)                     // 62717f0d7f9c11f9ffffff
	VMOVDQU8 X11, K5, -15(R14)(R15*4)                  // 62117f0d7f9cbef1ffffff
	VMOVDQU8 X24, K3, X31                              // 62017f0b7fc7
	VMOVDQU8 15(DX)(BX*1), K3, X31                     // 62617f0b6fbc1a0f000000
	VMOVDQU8 -7(CX)(DX*2), K3, X31                     // 62617f0b6fbc51f9ffffff