[clang] [llvm] [AMDGPU][WIP] Extend readlane, writelane and readfirstlane intrinsic lowering for generic types (PR #89217)

[Matt Arsenault via llvm-commits]

================
@@ -243,11 +243,16 @@ def VOP_READFIRSTLANE : VOPProfile <[i32, i32, untyped, untyped]> {
 // FIXME: Specify SchedRW for READFIRSTLANE_B32
 // TODO: There is VOP3 encoding also
 def V_READFIRSTLANE_B32 : VOP1_Pseudo <"v_readfirstlane_b32", VOP_READFIRSTLANE,
-                                       getVOP1Pat<int_amdgcn_readfirstlane,
-                                                  VOP_READFIRSTLANE>.ret, 1> {
+                                       [], 1> {
   let isConvergent = 1;
 }
 
+foreach vt = Reg32Types.types in {
+  def : GCNPat<(vt (AMDGPUreadfirstlane (vt VRegOrLdsSrc_32:$src0))),
+        (V_READFIRSTLANE_B32 (vt VRegOrLdsSrc_32:$src0))
----------------
arsenm wrote:

I think GlobalISelEmitter is just ignoring PtrValueType

https://github.com/llvm/llvm-project/pull/89217