[llvm] MC: X86 intel syntax: Support data32 and data16 better (PR #156287)

[Phoebe Wang via llvm-commits]

================
@@ -11,20 +11,12 @@
 // X86-16-SAME: encoding: [0x66,0x0f,0x01,0x16,0x00,0x00]
 data32 lgdt 0
 
-// X86-64: data16
-// X86-64: encoding: [0x66]
-// X86-64: lgdtq 0
-// X86-64: encoding: [0x0f,0x01,0x14,0x25,0x00,0x00,0x00,0x00]
-// X86-32: data16
-// X86-32: encoding: [0x66]
-// X86-32: lgdtl 0
-// X86-32: encoding: [0x0f,0x01,0x15,0x00,0x00,0x00,0x00]
+// X86-64: encoding: [0x66,0x0f,0x01,0x14,0x25,0x00,0x00,0x00,0x00]
+// X86-32: encoding: [0x66,0x0f,0x01,0x15,0x00,0x00,0x00,0x00]
----------------
phoebewang wrote:

Add the instruction check like `// X86-64: lgdtq 0`. The same below.

https://github.com/llvm/llvm-project/pull/156287