[llvm] [X86] Support APX CMOV/CFCMOV instructions (PR #82592)

[Shengchen Kan via llvm-commits]

================
@@ -412,3 +412,9 @@
 # ATT:   movbeq	 %r16, %r17
 # INTEL: movbe	r17, r16
 0x62,0xec,0xfc,0x08,0x60,0xc8
+
+## cfcmov
+
+# ATT:   cfcmovbeq	%r15, %r9
----------------
KanRobert wrote:

Use r16 and r17 so that we can check that R2, B2 bits are decoded correctly for reverse encoding

https://github.com/llvm/llvm-project/pull/82592