merge changes for upstream/latest update

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_bwd_weight.hpp

@@ -791,8 +791,10 @@ struct GridwiseGemm_bk0mk1_bk0nk1_mn_xdlops_bwd_weight
             constexpr auto c_block_desc_mblock_mperblock_nblock_nperblock =
                 GetCBlockDescriptor_MBlock_MPerBlock_NBlock_NPerBlock();
 
+            void* p_shared = static_cast<void*>(p_shared_block);
+
             auto c_block_buf = make_dynamic_buffer<AddressSpaceEnum::Lds>(
-                static_cast<FloatC*>(p_shared_block),
+                static_cast<FloatC*>(p_shared),
                 c_block_desc_mblock_mperblock_nblock_nperblock.GetElementSpaceSize());
 
             static_assert(M1 == MWave, "");

include/ck/tensor_operation/gpu/grid/gridwise_gemm_xdlops_v2r4.hpp

@@ -249,7 +249,7 @@ struct GridwiseGemm_bk0mk1_bk0nk1_mn_xdlops_v2r4
         }();
 
         using BlockwiseGemm =
-            BlockwiseGemmXdlops_k0mk1_k0nk1_m0n0m1n1m2m3m4n2_v1<ThisThreadBlock,
+            BlockwiseGemmXdlops_k0mk1_k0nk1_m0n0m1n1m2m3m4n2_v1<BlockSize
                                                                 FloatAB,
                                                                 FloatAcc,
                                                                 decltype(a_k0_m_k1_block_desc),
@@ -453,7 +453,7 @@ struct GridwiseGemm_bk0mk1_bk0nk1_mn_xdlops_v2r4
         // sanity check
 
         auto blockwise_gemm =
-            BlockwiseGemmXdlops_k0mk1_k0nk1_m0n0m1n1m2m3m4n2_v1<ThisThreadBlock,
+            BlockwiseGemmXdlops_k0mk1_k0nk1_m0n0m1n1m2m3m4n2_v1<BlockSize,
                                                                 FloatAB,
                                                                 FloatAcc,
                                                                 decltype(a_k0_m_k1_block_desc),