[CI][Lint] Retire `format.sh` and add `clang-tidy` to GHA workflow (#1044)

* [Lint] Retire `format.sh` and add `clang-tidy` to GHA workflow

* chore: update clang-tidy settings

* chore: upgrade clang-format and clang-tidy version

* lint: resolve clang-tidy errors

* [Maint] restore format.sh

* [CI] pre-commit autoupdate

* [Minor] fix `command -v` usage

src/transform/lower_thread_allreduce.cc

@@ -496,8 +496,9 @@ private:
       if (reduce_extent == 1) {
         // special case, no reduction is needed.
         std::vector<Stmt> stores;
+        stores.reserve(size);
         for (size_t i = 0; i < size; ++i) {
-          stores.push_back(BufferStore(buffers[i], values[i], {0}));
+          stores.emplace_back(BufferStore(buffers[i], values[i], {0}));
         }
         return SeqStmt::Flatten(stores);
       }
@@ -604,7 +605,7 @@ private:
       // Load reduction values, no synchronization needed.
       Array<PrimExpr> a, b;
       for (int i = 0; i < n_buffers; ++i) {
-        Buffer shared_buf = shared_bufs[i];
+        const Buffer &shared_buf = shared_bufs[i];
         BufferLoad val(shared_buf, zero_indices);
         ICHECK_EQ(val->dtype, dtypes[i]);
         a.push_back(val);
@@ -623,7 +624,7 @@ private:
         // branch with a warp sync call inside.
         PrimExpr other = WarpShuffle(builtin::tvm_warp_shuffle_down(),
                                      mask_buffer, val, offset);
-        Buffer local_buf = local_bufs[i];
+        const Buffer &local_buf = local_bufs[i];
         Stmt s = BufferStore(local_buf, other, zero_indices);
         seq->push_back(s);
 
@@ -639,7 +640,7 @@ private:
       std::vector<Stmt> stores;
       stores.reserve(n_buffers);
       for (int i = 0; i < n_buffers; ++i) {
-        Buffer buf = shared_bufs[i];
+        const Buffer &buf = shared_bufs[i];
         stores.push_back(BufferStore(buf, ret[i], zero_indices));
       }
 

src/transform/make_packed_api.cc

@@ -477,7 +477,7 @@ tvm::transform::Pass MakePackedAPI() {
     Map<GlobalVar, String> packed_func_methods;
     for (const auto &[gvar, base_func] : mod->functions) {
       if (auto opt = base_func.as<PrimFunc>()) {
-        auto prim_func = opt.value();
+        const auto &prim_func = opt.value();
         if (auto global_symbol = RequiresPackedAPI(prim_func)) {
           packed_func_methods.Set(gvar, global_symbol.value());
         }

src/transform/storage_access.cc

@@ -209,7 +209,7 @@ void TileLangStorageAccessVisitor::VisitStmt_(const ForNode *op) {
 bool IsThreadInvariant(const PrimExpr &cond) {
   if (auto call = cond.as<CallNode>()) {
     if (auto opt_call_op = call->op.as<Op>()) {
-      auto call_op = opt_call_op.value();
+      const auto &call_op = opt_call_op.value();
       if (call_op.same_as(builtin::tvm_thread_invariant())) {
         return true;
       }

src/transform/warp_specialized_rewriter.cc

@@ -530,10 +530,11 @@ private:
         block_stmt.push_back(block->body);
         cur_id++;
       }
-      new_body.push_back(MakeGroupBlock(block_stmt.size() == 1
-                                            ? block_stmt[0]
-                                            : SeqStmt(std::move(block_stmt)),
-                                        annotations));
+      new_body.push_back(MakeGroupBlock(
+          block_stmt.size() == 1 ? block_stmt[0]
+                                 // NOLINTNEXTLINE(performance-move-const-arg)
+                                 : SeqStmt(std::move(block_stmt)),
+          annotations));
     }
     Array<Integer> order_anno;
     Array<Integer> stage_anno;
@@ -697,10 +698,12 @@ private:
             continue;
           if (marker_.GetRole(op->seq[i]) == Role::kBoth) {
             block_stmt.push_back(seq_transformed[i]);
-            new_body.push_back(MakeGroupBlock(
-                block_stmt.size() == 1 ? block_stmt[0]
-                                       : SeqStmt(std::move(block_stmt)),
-                annotations));
+            new_body.push_back(
+                MakeGroupBlock(block_stmt.size() == 1
+                                   ? block_stmt[0]
+                                   // NOLINTNEXTLINE(performance-move-const-arg)
+                                   : SeqStmt(std::move(block_stmt)),
+                               annotations));
             continue;
           }
         }
@@ -734,10 +737,12 @@ private:
             }
           }
           collector.Clear();
-          new_body.push_back(MakeGroupBlock(
-              block_stmt.size() == 1 ? block_stmt[0]
-                                     : SeqStmt(std::move(block_stmt)),
-              annotations));
+          new_body.push_back(
+              MakeGroupBlock(block_stmt.size() == 1
+                                 ? block_stmt[0]
+                                 // NOLINTNEXTLINE(performance-move-const-arg)
+                                 : SeqStmt(std::move(block_stmt)),
+                             annotations));
         }
       }
     } else { // consumer case
@@ -766,10 +771,11 @@ private:
             }
           }
         }
-        new_body.push_back(MakeGroupBlock(block_stmt.size() == 1
-                                              ? block_stmt[0]
-                                              : SeqStmt(std::move(block_stmt)),
-                                          annotations));
+        new_body.push_back(MakeGroupBlock(
+            block_stmt.size() == 1 ? block_stmt[0]
+                                   // NOLINTNEXTLINE(performance-move-const-arg)
+                                   : SeqStmt(std::move(block_stmt)),
+            annotations));
       }
       // Filter out the producer stmts
       int cur_id = 0;