feat(arch): the rest of the SVE2 BFloat16 wall

Assisted-by: GLM 5.3 Flash

asm/arm64_ext_asm_test.go

@@ -822,6 +822,19 @@ func TestArm64AssembleExtensionStage4Golden(t *testing.T) {
 		{"ZBFMUL Z7.H, Z6.H, Z23.H", 0x650708d7},
 		{"ZBFSUB Z23.H, Z13.H, P1.M, Z13.H", 0x650186ed},
 		{"ZBFSUB Z7.H, Z6.H, Z23.H", 0x650704d7},
+		// The rest of the BFloat16 wall.
+		{"ZBFCVT Z13.S, P1.M, Z22.H", 0x658aa5b6},
+		{"ZBFCVT Z13.S, P1.Z, Z22.H", 0x649ac5b6},
+		{"ZBFCVTNT Z13.S, P1.M, Z22.H", 0x648aa5b6},
+		{"ZBFDOT Z7.H, Z6.H, Z23.S", 0x646780d7},
+		{"ZBFMLA Z23.H, Z13.H, P1.M, Z22.H", 0x653705b6},
+		{"ZBFMLALB Z7.H, Z6.H, Z23.S", 0x64e780d7},
+		{"ZBFMLALT Z7.H, Z6.H, Z23.S", 0x64e784d7},
+		{"ZBFMLS Z23.H, Z13.H, P1.M, Z22.H", 0x653725b6},
+		{"ZBFMLSLB Z7.H, Z6.H, Z23.S", 0x64e7a0d7},
+		{"ZBFMLSLT Z7.H, Z6.H, Z23.S", 0x64e7a4d7},
+		{"ZBFMMLA Z7.H, Z6.H, Z23.S", 0x6467e4d7},
+		{"ZBFSCALE Z23.H, Z13.H, P1.M, Z13.H", 0x650986ed},
 		// The multiple-structure loads and stores.
 		{"ZLD2B (R6)(R14), P4.Z, [Z13.B, Z14.B]", 0xa426d1cd},
 		{"ZLD2H (R6<<1)(R14), P4.Z, [Z13.H, Z14.H]", 0xa4a6d1cd},

asm/extension_test.go

@@ -327,6 +327,17 @@ func TestExtensionNamesARM64(t *testing.T) {
 		"ZBFMINNM",
 		"ZBFMUL",
 		"ZBFSUB",
+		"ZBFCVT",
+		"ZBFCVTNT",
+		"ZBFDOT",
+		"ZBFMLA",
+		"ZBFMLALB",
+		"ZBFMLALT",
+		"ZBFMLS",
+		"ZBFMLSLB",
+		"ZBFMLSLT",
+		"ZBFMMLA",
+		"ZBFSCALE",
 		"ZLD2B",
 		"ZLD2D",
 		"ZLD2H",
@@ -469,7 +480,7 @@ func TestExtensionNamesARM64(t *testing.T) {
 	if strings.Join(got, ",") != strings.Join(want, ",") {
 		t.Errorf("ExtensionNames(ARM64) = %v, want %v", got, want)
 	}
-	if n := len(arch.Extensions(arch.ARM64)); n != 470 {
-		t.Errorf("the family registers %d instructions, want 470", n)
+	if n := len(arch.Extensions(arch.ARM64)); n != 483 {
+		t.Errorf("the family registers %d instructions, want 483", n)
 	}
 }