[Cpu][Coverity] Cpu coverity fixes

src/plugins/intel_cpu/src/emitters/snippets/x64/utils.hpp

@@ -42,6 +42,12 @@ class jit_aux_gpr_holder {
 
     ~jit_aux_gpr_holder();
 
+    jit_aux_gpr_holder(const jit_aux_gpr_holder&) = delete;
+    jit_aux_gpr_holder& operator=(const jit_aux_gpr_holder&) = delete;
+
+    jit_aux_gpr_holder(jit_aux_gpr_holder&&) = delete;
+    jit_aux_gpr_holder& operator=(jit_aux_gpr_holder&&) = delete;
+
     [[nodiscard]] const Xbyak::Reg64& get_reg() const {
         return m_aux_gpr_idx;
     }

src/plugins/intel_cpu/src/graph_optimizer.cpp

@@ -867,27 +867,26 @@ void GraphOptimizer::FuseConvDeconvFCAndConvertOnWeights(Graph& graph) {
             parent = transpose->getParentEdgeAt(0)->getParent();
         }
 
-        const auto convert = parent;
-        if (!isSuitableConvert(convert)) {
+        if (!isSuitableConvert(parent)) {
             continue;
         }
 
-        const auto weights = convert->getParentEdgeAt(0)->getParent();
+        const auto weights = parent->getParentEdgeAt(0)->getParent();
         const auto weights_out_edge = weights->getChildEdges()[0].lock();
         const auto fc_weights_path_edge =
             transpose ? transpose->getParentEdgeAt(0) : fullyConnected->getParentEdgeAt(1);
         const auto inNum = weights_out_edge->getInputNum();
         const auto outNum = fc_weights_path_edge->getOutputNum();
-        const auto originalPrecision = convert->getOriginalInputPrecisionAtPort(0);
+        const auto originalPrecision = parent->getOriginalInputPrecisionAtPort(0);
         fullyConnected->setOriginalInputPrecisionAtPort(1, originalPrecision);
         if (transpose) {
             transpose->setOriginalInputPrecisionAtPort(0, originalPrecision);
             transpose->setOriginalOutputPrecisionAtPort(0, originalPrecision);
         }
         graph.RemoveEdge(fc_weights_path_edge);
         graph.CreateEdge(weights, transpose ? transpose : fullyConnected, inNum, outNum);
-        if (convert->getChildEdges().empty()) {
-            graph.DropNode(convert);
+        if (parent->getChildEdges().empty()) {
+            graph.DropNode(parent);
         }
     }
 }

src/plugins/intel_cpu/src/nodes/executors/dnnl/dnnl_matmul_primitive.cpp

@@ -372,7 +372,7 @@ static primitive_desc createPrimitiveDesc(const dnnl::memory::desc& inputDesc,
                 const size_t highestPriority = prim_desc_w_priority.priority;
                 if (priorityId < highestPriority) {
                     auto desc_copy = dnnl::primitive_desc(DnnlExtensionUtils::clone_primitive_desc(desc.get(true)));
-                    prim_desc_w_priority = {desc_copy, priorityId};
+                    prim_desc_w_priority = {std::move(desc_copy), priorityId};
                 }
             });
 

src/plugins/intel_cpu/src/nodes/interpolate.cpp

@@ -2509,7 +2509,7 @@ void Interpolate::prepareParams() {
     }
 
     if (canUseAclExecutor) {
-        interpAttrs.dataScales = scales5D;
+        interpAttrs.dataScales = std::move(scales5D);
 
         std::vector<MemoryDescPtr> srcMemoryDescs;
         for (size_t i = 0; i < getParentEdges().size(); i++) {

src/plugins/intel_cpu/src/transformations/cpu_opset/x64/pass/mlp_fusion.cpp

@@ -125,7 +125,7 @@ ov::intel_cpu::MLPFusionPass::MLPFusionPass() {
 
         // Determine gate_up_type based on pattern matching
         LLMMLPNode::GATE_UP_TYPE gate_up_type = LLMMLPNode::GATE_UP_TYPE::SEPARATE;
-        if (pattern_map.count(gate_up_proj_split)) {
+        if (pattern_map.find(gate_up_proj_split) != pattern_map.end()) {
             auto mlp_gated_up_node = pattern_map.at(mlp_gated_up).get_node_shared_ptr();
             auto input0 = mlp_gated_up_node->input_value(0);
             auto input1 = mlp_gated_up_node->input_value(1);