每日归档： 2022年2月8日

使用VSCode 调试Apollo无人车代码

原创文章，转载请注明： 转载自慢慢的回味

本文链接地址: 使用VSCode 调试Apollo无人车代码

深入研究Apollo的代码是学习自动驾驶的很好途径。很多前沿科技，比如图像识别，激光雷达，多传感器融合，路径规划都可以直接完整的学习。能够直接调试代码是比读代码更能加深理解。本文就介绍怎么去调试Apollo的代码。

Content:
  1. 下载代码
  2. 安装Docker
  3. 安装无线网卡驱动（可选）
  4. 安装Nvidia驱动
  5. 禁止自动更新内核来避免重新安装驱动
  6. 安装Nvidia Docker
  7. 启动Apollo项目
  8. 调试一个车道线识别程序

下载代码

系统使用Ubuntu 18.04版本。
https://gitee.com/ApolloAuto/apollo.git
比如目录为~/apollo。

安装Docker
curl -fsSL https://download.docker.com/linux/ubuntu/gpg | sudo gpg --dearmor -o /usr/share/keyrings/docker-archive-keyring.gpg
echo   "deb [arch=$(dpkg --print-architecture) signed-by=/usr/share/keyrings/docker-archive-keyring.gpg] https://download.docker.com/linux/ubuntu \
$(lsb_release -cs) stable" | sudo tee /etc/apt/sources.list.d/docker.list > /dev/null
 
sudo apt-get update
sudo apt-get install docker-ce docker-ce-cli containerd.io
systemctl daemon-reload
systemctl restart docker
安装无线网卡驱动（可选）

因为笔者的网卡Ubuntu没有自带，需要自行安装驱动。
下载驱动：https://codeload.github.com/gnab/rtl8812au/zip/refs/heads/master

make dkms_install
echo 8812au | sudo tee -a /etc/modules
insmod 8812au.ko
安装Nvidia驱动

本人的显卡为GTX1060。Apollo项目需要Nvidia显卡，否则大部分模块无法编译运行。
如下为安装显卡驱动的脚步程序。cuda安装后的路径可能有所不同。

sudo apt-get install linux-headers-$(uname -r)
distribution=$(. /etc/os-release;echo $ID$VERSION_ID | sed -e 's/\.//g')
wget https://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64/cuda-$distribution.pin
sudo mv cuda-$distribution.pin /etc/apt/preferences.d/cuda-repository-pin-600
#key的id可以会更新 https://developer.nvidia.com/blog/updating-the-cuda-linux-gpg-repository-key/
sudo apt-key adv --fetch-keys https://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64/7fa2af80.pub
echo "deb http://developer.download.nvidia.com/compute/cuda/repos/$distribution/x86_64 /" | sudo tee /etc/apt/sources.list.d/cuda.list
sudo apt-get update
sudo apt-get -y install cuda-drivers
export PATH=$PATH:/usr/local/cuda-11.2/bin
export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-11.2/lib64:/usr/local/cuda/cuda/lib64

安装完成后，如下命令可以查看显卡信息。请确保cuda已正确安装。

nvidia-smi
禁止自动更新内核来避免重新安装驱动

1 查看自己使用的内核

derek@ubuntu:~$ uname -a
Linux ubuntu 5.4.0-150-generic #167~18.04.1-Ubuntu SMP Wed May 24 00:51:42 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

2 查看目前安装的内核，并锁定它们

derek@ubuntu:~$ dpkg --get-selections|grep linux|grep 150
linux-headers-5.4.0-150-generic			install
linux-hwe-5.4-headers-5.4.0-150			install
linux-image-5.4.0-150-generic			install
linux-modules-5.4.0-150-generic			install
linux-modules-extra-5.4.0-150-generic		install
 
derek@ubuntu:~$ sudo apt-mark hold linux-headers-5.4.0-150-generic
linux-headers-5.4.0-150-generic set on hold.
derek@ubuntu:~$ sudo apt-mark hold linux-hwe-5.4-headers-5.4.0-150
linux-hwe-5.4-headers-5.4.0-150 set on hold.
derek@ubuntu:~$ sudo apt-mark hold linux-image-5.4.0-150-generic
linux-image-5.4.0-150-generic set on hold.
derek@ubuntu:~$ sudo apt-mark hold linux-modules-5.4.0-150-generic
linux-modules-5.4.0-150-generic set on hold.
derek@ubuntu:~$ sudo apt-mark hold linux-modules-extra-5.4.0-150-generic
linux-modules-extra-5.4.0-150-generic set on hold
 
derek@ubuntu:~$ dpkg --get-selections|grep linux|grep 150
linux-headers-5.4.0-150-generic			hold
linux-hwe-5.4-headers-5.4.0-150			hold
linux-image-5.4.0-150-generic			hold
linux-modules-5.4.0-150-generic			hold
linux-modules-extra-5.4.0-150-generic		hold

3 修改软件自动更新
修改如下2项

安装Nvidia Docker

Apollo需要运行在Docker预安装环境，以加快开发运行环境的统一性。

distribution=$(. /etc/os-release;echo $ID$VERSION_ID)    && curl -s -L https://nvidia.github.io/nvidia-docker/gpgkey | sudo apt-key add -    && curl -s -L https://nvidia.github.io/nvidia-docker/$distribution/nvidia-docker.list | sudo tee /etc/apt/sources.list.d/nvidia-docker.list
sudo apt-get update
sudo apt-get install -y nvidia-docker2
sudo systemctl restart docker

安装完成后，可启动示例Dockers程序查看。

sudo docker run --rm --gpus all nvidia/cuda:11.0-base nvidia-smi
启动Apollo项目

在源码目录~/apollo中，执行如下命令，进入Apollo Docker：
脚本会把源文件里面的目录以volumn的形式挂载进Docker环境。

./docker/scripts/dev_start.sh 
./docker/scripts/dev_into.sh

以后进入，直接启动已有的Docker即可：

docker ps -a
docker start 671567b64765
./docker/scripts/dev_into.sh
调试一个车道线识别程序

创建一个Bazel编译配置：
~/apollo/modules/perception/camera/test/BUILD
内容为：

load("@rules_cc//cc:defs.bzl", "cc_library", "cc_test")
load("//tools:cpplint.bzl", "cpplint")
 
package(default_visibility = ["//visibility:public"])
 
cc_test(
    name = "camera_lib_lane_detector_darkscnn_lane_detector_test",
    size = "medium",
    srcs = ["camera_lib_lane_detector_darkscnn_lane_detector_test.cc"],
    deps = [
        "//cyber",
        "//modules/perception/base",
        "//modules/perception/camera/lib/lane/detector/darkSCNN:darkSCNN_lane_detector",
        "//modules/perception/common/io:io_util",
        "@com_google_googletest//:gtest_main",
        "@opencv//:core",
    ],
)
 
cpplint()

现在可以编译Apollo程序了：

./apollo.sh build_dbg

启动GDB Server，供Docker外面的VS Code进行远程调试：

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/libtorch_gpu/lib/
apt update
apt install gdbserver
gdbserver 127.0.0.1:2222 bazel-bin/modules/perception/camera/test/camera_lib_lane_detector_darkscnn_lane_detector_test

启动VS Code，安装C++插件，然后点击Debug，添加如下配置即可调试Docker里面刚才启动的程序：(注意修改program的路径)

{
    "version": "0.2.0",
    "configurations": [
        {
            "name": "gdb Remote camera_lib_lane_postprocessor_darkscnn_lane_postprocessor_test",
            "type": "cppdbg",
            "request": "launch",
            "program": "~/apollo/.cache/bazel/540135163923dd7d5820f3ee4b306b32/execroot/apollo/bazel-out/k8-dbg/bin/modules/perception/camera/test/camera_lib_lane_detector_darkscnn_lane_detector_test",
            "args": ["myarg1", "myarg2", "myarg3"],
            "stopAtEntry": true,
            "environment": [],
            "externalConsole": false,
            "MIMode": "gdb",
            "miDebuggerPath": "gdb",
            "miDebuggerArgs": "gdb",
            "linux": {
                "MIMode": "gdb",
                "miDebuggerPath": "/usr/bin/gdb",
                "miDebuggerServerAddress": "127.0.0.1:2222",
            },
            "logging": {
                "moduleLoad": false,
                "engineLogging": false,
                "trace": false
            },
            "setupCommands": [
                {
                    "description": "Enable pretty-printing for gdb",
                    "text": "-enable-pretty-printing",
                    "ignoreFailures": true
                }
            ],
            "cwd": "${workspaceFolder}",
        }
    ]
}

本作品采用知识共享署名 4.0 国际许可协议进行许可。

使用VSCode 调试tensorflow c lib的简单方法

原创文章，转载请注明： 转载自慢慢的回味

本文链接地址: 使用VSCode 调试tensorflow c lib的简单方法

Content:
  1. 编写测试程序
  2. 添加Bazel build配置
  3. 编译测试程序
  4. VSCode Debug

回目录

编写测试程序

编写测试程序TensorflowTest.cpp，放到目录tensorflow/c下面。

//============================================================================
// Name        : TensorflowTest.cpp
// Author      : 
// Version     :
// Copyright   : Your copyright notice
// Description : Hello World in C++, Ansi-style
//============================================================================
 
#include <iostream>
#include "c_api.h"
#include "c_api_experimental.h"
#include "c_test_util.h"
 
#include <algorithm>
#include <cstddef>
#include <iterator>
#include <memory>
#include <vector>
#include <string.h>
 
using namespace std;
 
int main() {
	cout << "!!!Hello World!!!" << endl; // prints !!!Hello World!!!
	cout << "Hello from TensorFlow C library version" << TF_Version() << endl;
 
	TF_Status* s = TF_NewStatus();
	TF_Graph* graph = TF_NewGraph();
 
	// Construct the graph: A + 2 + B
	TF_Operation* a = Placeholder(graph, s, "A");
	cout << TF_Message(s);
 
	TF_Operation* b = Placeholder(graph, s, "B");
	cout << TF_Message(s);
 
	TF_Operation* one = ScalarConst(1, graph, s, "kone");
	cout << TF_Message(s);
 
	TF_Operation* two = ScalarConst(2, graph, s, "ktwo");
	cout << TF_Message(s);
 
	TF_Operation* three = Add(one, two, graph, s, "kthree");
	cout << TF_Message(s);
 
	TF_Operation* plus2 = Add(a, two, graph, s, "plus2");
	cout << TF_Message(s);
 
	TF_Operation* plusB = Add(plus2, b, graph, s, "plusB");
	cout << TF_Message(s);
 
	TF_Operation* plusC = Add(plusB, three, graph, s, "plusC");
	cout << TF_Message(s);
 
	// Setup a session and a partial run handle.  The partial run will allow
	// computation of A + 2 + B in two phases (calls to TF_SessionPRun):
	// 1. Feed A and get (A+2)
	// 2. Feed B and get (A+2)+B
	TF_SessionOptions* opts = TF_NewSessionOptions();
	TF_EnableXLACompilation(opts, true);
	TF_Session* sess = TF_NewSession(graph, opts, s);
	TF_DeleteSessionOptions(opts);
 
	TF_Output feeds[] = { TF_Output { a, 0 }, TF_Output { b, 0 } };
	TF_Output fetches[] = { TF_Output { plus2, 0 }, TF_Output { plusB, 0 }, TF_Output { plusC, 0 }  };
 
	const char* handle = nullptr;
	TF_SessionPRunSetup(sess, feeds, TF_ARRAYSIZE(feeds), fetches,
			TF_ARRAYSIZE(fetches), NULL, 0, &handle, s);
	cout << TF_Message(s);
 
	// Feed A and fetch A + 2.
	TF_Output feeds1[] = { TF_Output { a, 0 }, TF_Output { b, 0 } };
	TF_Output fetches1[] = { TF_Output { plus2, 0 }, TF_Output { plusB, 0 }, TF_Output { plusC, 0 } };
	TF_Tensor* feedValues1[] = { Int32Tensor(1), Int32Tensor(3) };
	TF_Tensor* fetchValues1[3];
	TF_SessionPRun(sess, handle, feeds1, feedValues1, 2, fetches1, fetchValues1,
			3, NULL, 0, s);
	cout << TF_Message(s);
	cout << *(static_cast<int*>(TF_TensorData(fetchValues1[0]))) << endl;
	cout << *(static_cast<int*>(TF_TensorData(fetchValues1[1]))) << endl;
	cout << *(static_cast<int*>(TF_TensorData(fetchValues1[2]))) << endl;
 
	// Clean up.
	TF_DeletePRunHandle(handle);
	TF_DeleteSession(sess, s);
	cout << TF_Message(s);
	TF_DeleteGraph(graph);
	TF_DeleteStatus(s);
	return 0;
}
添加Bazel build配置

修改tensorflow/c/BUILD文件：
替换

load("//tensorflow:tensorflow.bzl", "tf_cuda_cc_test")

为

load("//tensorflow:tensorflow.bzl", "tf_cuda_cc_test", "tf_cc_binary")

添加

tf_cc_binary(
    name = "TensorflowTest",
    testonly = 1,
    srcs = [
        "TensorflowTest.cpp",
    ],
    deps = [
        ":c_api_experimental",
        ":c_test_util",
        ":c_api",
    ],
)
编译测试程序
bazel build -s --config=dbg --config=noaws --config=nogcp --config=nohdfs --config=nonccl --config=xla //tensorflow/c:TensorflowTest --verbose_failures
VSCode Debug

启动VSCode，打开Tensorflow源代码根目录。
添加CPP插件

配置Debug，弹出菜单中选择gdb。
然后修改program参数，然后即可在代码上打断点调试
本作品采用知识共享署名 4.0 国际许可协议进行许可。