[llvm] AMDGPU/GlobalISel: RegBankLegalize rules for cluster_load_b32/b64/b128 (PR #196186)

[via llvm-commits]

================
@@ -1872,6 +1876,26 @@ RegBankLegalizeRules::RegBankLegalizeRules(const GCNSubtarget &_ST,
                     amdgcn_global_store_async_from_lds_b128})
       .Any({{}, {{}, {IntrId, VgprP1, VgprP3}}});
 
+  addRulesForIOpcs({amdgcn_cluster_load_b32})
+      .Any({{UniB32}, {{UniInVgprB32}, {IntrId, SgprP1, Imm, SgprB32_M0}}})
----------------
vangthao95 wrote:

For uniform cluster_load, doesn't selection use saddr form with p1 in sgpr?

https://github.com/llvm/llvm-project/pull/196186