Add gil for offline debugger

mindspore/ccsrc/debug/debug_services.cc

@@ -1,5 +1,5 @@
 /**
- * Copyright 2019-2021 Huawei Technologies Co., Ltd
+ * Copyright 2019-2022 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -742,12 +742,15 @@ void DebugServices::ConvertToHostFormat(const std::map<std::string, std::vector<
     if (!files_to_convert_in_dir.empty()) {
       // Look for the installation path to the conver_async package. If not found, throw exception and terminate the
       // later task.
-      try {
-        auto pkg = pybind11::module::import("mindspore.offline_debug.convert_async");
-        auto convert_obj = pkg.attr("AsyncDumpConverter")(pybind11::cast(files_to_convert_in_dir), dump_key);
-        (void)convert_obj.attr("convert_files")();
-      } catch (pybind11::error_already_set &e) {
-        MS_LOG(EXCEPTION) << "Failed to convert async dump data: " << e.what();
+      {
+        pybind11::gil_scoped_acquire acquire;
+        try {
+          auto pkg = pybind11::module::import("mindspore.offline_debug.convert_async");
+          auto convert_obj = pkg.attr("AsyncDumpConverter")(pybind11::cast(files_to_convert_in_dir), dump_key);
+          (void)convert_obj.attr("convert_files")();
+        } catch (pybind11::error_already_set &e) {
+          MS_LOG(EXCEPTION) << "Failed to convert async dump data: " << e.what();
+        }
       }
       ProcessConvertToHostFormat(files_after_convert_in_dir, dump_key, result_list, file_format);
     }

mindspore/ccsrc/debug/debugger/debugger.cc

@@ -1,5 +1,5 @@
 /**
- * Copyright 2020-2021 Huawei Technologies Co., Ltd
+ * Copyright 2020-2022 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -621,7 +621,7 @@ void Debugger::CheckGraphPtr(const KernelGraphPtr &graph_ptr) {
       }
       if (debugger_enabled_) {
         LoadParametersAndConst();
-        // get graph proto and send to Mindinsight
+        // get graph proto and send to MindInsight
         auto graph_proto = graph_proto_list_.front();
         SendGraphAndSuspend(graph_proto);
         graph_proto_list_.clear();
@@ -700,7 +700,7 @@ void Debugger::SendGraphAndSuspend(const GraphProto &graph_proto) {
   if (!CheckSendMetadata()) {
     return;
   }
-  // send graph to Mindinsight server
+  // send graph to MindInsight server
   MS_EXCEPTION_IF_NULL(grpc_client_);
   EventReply reply = grpc_client_->SendGraph(graph_proto);
   if (reply.status() != reply.OK) {
@@ -874,11 +874,11 @@ void Debugger::CommandLoop() {
         ProcessKViewCMD(reply);
         break;
       case DebuggerCommand::kVersionMatchedCMD:
-        MS_LOG(ERROR) << "Received unexpected Version Matched CMD from Mindinsight.";
+        MS_LOG(ERROR) << "Received unexpected Version Matched CMD from MindInsight.";
         Exit();
         break;
       default:
-        MS_LOG(ERROR) << "Received unknown CMD from Mindinsight";
+        MS_LOG(ERROR) << "Received unknown CMD from MindInsight";
         Exit();
         break;
     }

mindspore/ccsrc/debug/debugger/debugger.h

@@ -1,5 +1,5 @@
 /**
- * Copyright 2020-2021 Huawei Technologies Co., Ltd
+ * Copyright 2020-2022 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -134,8 +134,8 @@ class Debugger : public std::enable_shared_from_this<Debugger> {
 
   void SetTrainingDone(bool training_done);
 
-  // returns true if reply received and mindspore version matched with mindinsight version
-  // version_check should be true if you want the function to do backend compatibility check with Mindinsight
+  // returns true if reply received and mindspore version matched with mindInsight version
+  // version_check should be true if you want the function to do backend compatibility check with MindInsight
   bool SendMetadata(bool version_check);
 
   bool CheckSendMetadata();

mindspore/ccsrc/debug/debugger/offline_debug/dbg_services.cc

@@ -1,5 +1,5 @@
 /**
- * Copyright 2021 Huawei Technologies Co., Ltd
+ * Copyright 2021-2022 Huawei Technologies Co., Ltd
  *
  * Licensed under the Apache License, Version 2.0 (the "License");
  * you may not use this file except in compliance with the License.
@@ -189,10 +189,9 @@ std::vector<watchpoint_hit_t> DbgServices::CheckWatchpoints(unsigned int iterati
   std::vector<std::string> file_paths;
 
   const bool init_dbg_suspend = (iteration == UINT_MAX);
-
-  tensor_list = debug_services_->ReadNeededDumpedTensors(iteration, &file_paths, error_on_no_value);
   {
-    py::gil_scoped_release release;
+    pybind11::gil_scoped_release release;
+    tensor_list = debug_services_->ReadNeededDumpedTensors(iteration, &file_paths, error_on_no_value);
     debug_services_->CheckWatchpoints(&name, &slot, &condition, &watchpoint_id, &parameters, &error_codes, overflow_ops,
                                       file_paths, &tensor_list, init_dbg_suspend, true, true, &rank_id, &root_graph_id,
                                       error_on_no_value);

mindspore/python/mindspore/train/callback/_summary_collector.py

@@ -1,4 +1,4 @@
-# Copyright 2020-2021 Huawei Technologies Co., Ltd
+# Copyright 2020-2022 Huawei Technologies Co., Ltd
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -102,9 +102,9 @@ class SummaryCollector(Callback):
             - collect_train_lineage (bool): Whether to collect lineage data for the training phase,
               this field will be displayed on the `lineage page \
               <https://www.mindspore.cn/mindinsight/docs/en/master/lineage_and_scalars_comparison.html>`_
-              of Mindinsight. Default: True.
+              of MindInsight. Default: True.
             - collect_eval_lineage (bool): Whether to collect lineage data for the evaluation phase,
-              this field will be displayed on the lineage page of Mindinsight. Default: True.
+              this field will be displayed on the lineage page of MindInsight. Default: True.
             - collect_input_data (bool): Whether to collect dataset for each training.
               Currently only image data is supported.
               If there are multiple columns of data in the dataset, the first column should be image data.