cmd/internal/obj/loong64: add [X]VSHUF4I.{B/H/W/D} instructions support

Go asm syntax:
	 VSHUF4I{B/H/W/V}	$1, V1, V2
	XVSHUF4I{B/H/W/V}	$2, X1, X2

Equivalent platform assembler syntax:
	 vshuf4i.{b/h/w/d}	v2, v1, $1
	xvshuf4i.{b/h/w/d}	x2, x1, $2

Change-Id: I6a847ccbd2c93432d87bd1390b5cf1508da06496
Reviewed-on: https://go-review.googlesource.com/c/go/+/658376
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: abner chenc <chenguoqi@loongson.cn>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>

src/cmd/asm/internal/asm/endtoend_test.go

@@ -470,6 +470,10 @@ func TestLOONG64Encoder(t *testing.T) {
 	testEndToEnd(t, "loong64", "loong64")
 }
 
+func TestLOONG64Errors(t *testing.T) {
+	testErrors(t, "loong64", "loong64error")
+}
+
 func TestPPC64EndToEnd(t *testing.T) {
 	defer func(old int) { buildcfg.GOPPC64 = old }(buildcfg.GOPPC64)
 	for _, goppc64 := range []int{8, 9, 10} {

src/cmd/asm/internal/asm/testdata/loong64enc1.s

@@ -911,3 +911,29 @@ lable2:
 	XVMULWODWHUH	X1, X2, X3	// 4384a274
 	XVMULWODVWUW	X1, X2, X3	// 4304a374
 	XVMULWODQVUV	X1, X2, X3	// 4384a374
+
+	// [X]VSHUF4I.{B/H/W/D} instructions
+	VSHUF4IB	$0, V2, V1	// 41009073
+	VSHUF4IB	$16, V2, V1	// 41409073
+	VSHUF4IB	$255, V2, V1	// 41fc9373
+	VSHUF4IH	$0, V2, V1	// 41009473
+	VSHUF4IH	$128, V2, V1	// 41009673
+	VSHUF4IH	$255, V2, V1	// 41fc9773
+	VSHUF4IW	$0, V2, V1	// 41009873
+	VSHUF4IW	$96, V2, V1	// 41809973
+	VSHUF4IW	$255, V2, V1	// 41fc9b73
+	VSHUF4IV	$0, V2, V1	// 41009c73
+	VSHUF4IV	$8, V2, V1	// 41209c73
+	VSHUF4IV	$15, V2, V1	// 413c9c73
+	XVSHUF4IB	$0, X1, X2	// 22009077
+	XVSHUF4IB	$16, X1, X2	// 22409077
+	XVSHUF4IB	$255, X1, X2	// 22fc9377
+	XVSHUF4IH	$0, X1, X2	// 22009477
+	XVSHUF4IH	$128, X1, X2	// 22009677
+	XVSHUF4IH	$255, X1, X2	// 22fc9777
+	XVSHUF4IW	$0, X1, X2	// 22009877
+	XVSHUF4IW	$96, X1, X2	// 22809977
+	XVSHUF4IW	$255, X1, X2	// 22fc9b77
+	XVSHUF4IV	$0, X1, X2	// 22009c77
+	XVSHUF4IV	$8, X1, X2	// 22209c77
+	XVSHUF4IV	$15, X1, X2	// 223c9c77

src/cmd/asm/internal/asm/testdata/loong64error.s

@@ -0,0 +1,7 @@
+// Copyright 2025 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+TEXT errors(SB),$0
+	VSHUF4IV	$16, V1, V2	// ERROR "operand out of range 0 to 15"
+	XVSHUF4IV	$16, X1, X2	// ERROR "operand out of range 0 to 15"

src/cmd/internal/obj/loong64/a.out.go

@@ -1010,6 +1010,15 @@ const (
 	AXVMULWODVWUW
 	AXVMULWODQVUV
 
+	AVSHUF4IB
+	AVSHUF4IH
+	AVSHUF4IW
+	AVSHUF4IV
+	AXVSHUF4IB
+	AXVSHUF4IH
+	AXVSHUF4IW
+	AXVSHUF4IV
+
 	ALAST
 
 	// aliases

src/cmd/internal/obj/loong64/anames.go

@@ -505,5 +505,13 @@ var Anames = []string{
 	"XVMULWODWHUH",
 	"XVMULWODVWUW",
 	"XVMULWODQVUV",
+	"VSHUF4IB",
+	"VSHUF4IH",
+	"VSHUF4IW",
+	"VSHUF4IV",
+	"XVSHUF4IB",
+	"XVSHUF4IH",
+	"XVSHUF4IW",
+	"XVSHUF4IV",
 	"LAST",
 }

src/cmd/internal/obj/loong64/asm.go

@@ -1677,11 +1677,19 @@ func buildop(ctxt *obj.Link) {
 			opset(AVORB, r0)
 			opset(AVXORB, r0)
 			opset(AVNORB, r0)
+			opset(AVSHUF4IB, r0)
+			opset(AVSHUF4IH, r0)
+			opset(AVSHUF4IW, r0)
+			opset(AVSHUF4IV, r0)
 
 		case AXVANDB:
 			opset(AXVORB, r0)
 			opset(AXVXORB, r0)
 			opset(AXVNORB, r0)
+			opset(AXVSHUF4IB, r0)
+			opset(AXVSHUF4IH, r0)
+			opset(AXVSHUF4IW, r0)
+			opset(AXVSHUF4IV, r0)
 
 		case AVANDV:
 			opset(AVORV, r0)
@@ -2155,6 +2163,12 @@ func (c *ctxt0) asmout(p *obj.Prog, o *Optab, out []uint32) {
 			r = int(p.To.Reg)
 		}
 
+		// the operand range available for instructions VSHUF4IV and XVSHUF4IV is [0, 15]
+		if p.As == AVSHUF4IV || p.As == AXVSHUF4IV {
+			operand := uint32(v)
+			c.checkoperand(p, operand, 15)
+		}
+
 		o1 = OP_8IRR(c.opirr(p.As), uint32(v), uint32(r), uint32(p.To.Reg))
 
 	case 24: // add $lcon,r1,r2
@@ -2695,6 +2709,13 @@ func (c *ctxt0) asmout(p *obj.Prog, o *Optab, out []uint32) {
 	out[4] = o5
 }
 
+// checkoperand checks if operand >= 0 && operand <= maxoperand
+func (c *ctxt0) checkoperand(p *obj.Prog, operand uint32, mask uint32) {
+	if (operand & ^mask) != 0 {
+		c.ctxt.Diag("operand out of range 0 to %d: %v", mask, p)
+	}
+}
+
 // checkindex checks if index >= 0 && index <= maxindex
 func (c *ctxt0) checkindex(p *obj.Prog, index uint32, mask uint32) {
 	if (index & ^mask) != 0 {
@@ -3813,6 +3834,22 @@ func (c *ctxt0) opirr(a obj.As) uint32 {
 		return 0xed1a << 15 // xvsubi.wu
 	case AXVSUBVU:
 		return 0xed1b << 15 // xvsubi.du
+	case AVSHUF4IB:
+		return 0x1ce4 << 18 // vshuf4i.b
+	case AVSHUF4IH:
+		return 0x1ce5 << 18 // vshuf4i.h
+	case AVSHUF4IW:
+		return 0x1ce6 << 18 // vshuf4i.w
+	case AVSHUF4IV:
+		return 0x1ce7 << 18 // vshuf4i.d
+	case AXVSHUF4IB:
+		return 0x1de4 << 18 // xvshuf4i.b
+	case AXVSHUF4IH:
+		return 0x1de5 << 18 // xvshuf4i.h
+	case AXVSHUF4IW:
+		return 0x1de6 << 18 // xvshuf4i.w
+	case AXVSHUF4IV:
+		return 0x1de7 << 18 // xvshuf4i.d
 	}
 
 	if a < 0 {