[llvm] [AMDGPU][MC] Fix disassembler problem for image_atomic with TFE (PR #112622)

[Mirko Brkušanin via llvm-commits]

================
@@ -192,10 +271,10 @@
 # VI: image_atomic_add v5, v1, s[8:15] dmask:0x2 unorm ; encoding: [0x00,0x12,0x48,0xf0,0x01,0x05,0x02,0x00]
 0x00,0x12,0x48,0xf0,0x01,0x05,0x02,0x00
 
-# VI: image_atomic_add v5, v1, s[8:15] dmask:0x7 unorm ; encoding: [0x00,0x17,0x48,0xf0,0x01,0x05,0x02,0x00]
+# VI: image_atomic_add v[5:7], v1, s[8:15] dmask:0x7 unorm ; encoding: [0x00,0x17,0x48,0xf0,0x01,0x05,0x02,0x00]
 0x00,0x17,0x48,0xf0,0x01,0x05,0x02,0x00
 
-# VI: image_atomic_add v[5:9], v1, s[8:15] dmask:0xf unorm ; encoding: [0x00,0x1f,0x48,0xf0,0x01,0x05,0x02,0x00]
+# VI: image_atomic_add v5, v1, s[8:15] dmask:0xf unorm ; encoding: [0x00,0x1f,0x48,0xf0,0x01,0x05,0x02,0x00]
----------------
mbrkusanin wrote:

Disassembler defaults to V1 variants. That would have to be `IMAGE_ATOMIC_ADD_V4_V*` but there are not `V4_V* `variants for `IMAGE_ATOMIC`s, only `V1`, `V2` (64 or 32+tfe) and `V3` (64+tfe). 

Only cmpswap needs forth variant and this patch deletes `V4`s for non cmpswap atomics.

https://github.com/llvm/llvm-project/pull/112622