[llvm] [AMDGPU][True16][MC] 16bit vsrc and vdst support in MC (PR #104510)

[Brox Chen via llvm-commits]

================
@@ -583,11 +587,17 @@ v_cvt_f16_u16 v5, src_scc
 v_cvt_f16_u16 v127, 0xfe0b
 // GFX12: v_cvt_f16_u16_e32 v127, 0xfe0b ; encoding: [0xff,0xa0,0xfe,0x7e,0x0b,0xfe,0x00,0x00]
 
-v_cvt_f32_f16 v5, v1
-// GFX12: v_cvt_f32_f16_e32 v5, v1 ; encoding: [0x01,0x17,0x0a,0x7e]
+v_cvt_f32_f16 v5, v1.l
+// GFX12: v_cvt_f32_f16_e32 v5, v1.l ; encoding: [0x01,0x17,0x0a,0x7e]
+
+v_cvt_f32_f16 v5, v127.l
+// GFX12: v_cvt_f32_f16_e32 v5, v127.l ; encoding: [0x7f,0x17,0x0a,0x7e]
+
+v_cvt_f32_f16 v5, v1.h
+// GFX12: v_cvt_f32_f16_e32 v5, v1.h ; encoding: [0x81,0x17,0x0a,0x7e]
 
-v_cvt_f32_f16 v5, v127
-// GFX12: v_cvt_f32_f16_e32 v5, v127 ; encoding: [0x7f,0x17,0x0a,0x7e]
+v_cvt_f32_f16 v5, v127.h
+// GFX12: v_cvt_f32_f16_e32 v5, v127.h ; encoding: [0xff,0x17,0x0a,0x7e]
 
 v_cvt_f32_f16 v5, s1
----------------
broxigarchen wrote:

added

https://github.com/llvm/llvm-project/pull/104510