c++ - "gdb"调试器奇怪地跳过断点-6ren

c++ - "gdb"调试器奇怪地跳过断点

转载作者：太空宇宙更新时间：2023-11-04 02:32:00

这里有一小段我的代码:

 //l2tp_inspector.cc
 14   else if (f_info->key.proto == UDP_PROTO) {
 15     if (size >= 4) {
 16       uint32_t l2tp_part;
 17       l2tp_part = *((uint32_t*)(data));
 18 
 19       if ((l2tp_part & 0xFFFF0000) == 0xC802 &&                                                                                                                  
 20           (l2tp_part & 0x0000FFFF) == size) {
 21         f_info->protocol_id = Bina::Protocols::UDP_L2TP;
 22         f_info->application_id = Bina::Applications::UNKNOWN;
 23         return PROTOCOL_ACCEPTED_FINALIZED;
 24       }
 25     }
 26     return REJECTED;
 27   }

在编译代码并使用 gdb 调试器并在第 17、18 和 19 行设置断点后，我们得到以下输出:

test@ubuntu:~/prj$ gdb prj-loader 

GNU gdb (Ubuntu 7.11.90.20161005-0ubuntu1) 7.11.90.20161005-git
This GDB was configured as "x86_64-linux-gnu".

(gdb) b l2tp_inspector.cc:17
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 1 (l2tp_inspector.cc:17) pending.

(gdb) b l2tp_inspector.cc:18
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 2 (l2tp_inspector.cc:18) pending.

(gdb) b l2tp_inspector.cc:19
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 3 (l2tp_inspector.cc:19) pending.

(gdb) r
Starting program: ~/prj/prj-loader 
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[2017-02-02 08:44:52.267622] [0x00007ffff7fc9780] [info]    New id registered TCP with id 0
[2017-02-02 08:44:52.267899] [0x00007ffff7fc9780] [info]    New id registered TCP_PAYLOAD with id 1
[2017-02-02 08:44:52.267991] [0x00007ffff7fc9780] [info]    New id registered UDP with id 2
[2017-02-02 08:44:52.268025] [0x00007ffff7fc9780] [info]    New id registered UDP_PAYLOAD with id 3
[2017-02-02 08:44:52.268169] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:44:52.268201] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:44:52.268285] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:44:52.268380] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:44:52.268519] [0x00007ffff7fc9780] [info]    New id registered IPV4 with id 4
[2017-02-02 08:44:52.268553] [0x00007ffff7fc9780] [info]    New id registered IPV4FRAG with id 5
[2017-02-02 08:44:52.268632] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:44:52.268715] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:44:52.268795] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:44:52.268877] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:44:52.268972] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:44:52.269066] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:44:52.269149] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:44:52.269229] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:44:52.269370] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:44:52.269401] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:44:52.269481] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:44:52.269561] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:44:52.269643] [0x00007ffff7fc9780] [info]    Already registered id IPV4 with id 4
[2017-02-02 08:44:52.269723] [0x00007ffff7fc9780] [info]    Already registered id IPV4FRAG with id 5
[2017-02-02 08:44:52.269807] [0x00007ffff7fc9780] [info]    New id registered ETHERNET with id 6

Breakpoint 1, inspect (f_info=0x7fffffffce50, data=0x555555be439a "\310\002", size=106, s_module=@0x7fffffffcde0: 0x0)
    at ../src/modules/prj-lib/src/proto-inspectors/l2tp_inspector.cc:17
17        l2tp_part = *((uint32_t*)(data));

(gdb) s
Breakpoint 2, inspect (f_info=0x7fffffffce50, data=0x555555be439a "\310\002", size=106, s_module=@0x7fffffffcde0: 0x0)
    at ../src/modules/prj-lib/src/proto-inspectors/l2tp_inspector.cc:26
26      return REJECTED;

正如您在上面看到的，当我在第 19 行打断时，gdb 不会在那行停止并跳到第 26 行，为什么？

当我在第 19 行用 0xC8020000 替换 0xC802 时，它工作正常 gdb 不再错过第 19 行:

 //l2tp_inspector.cc
 14   else if (f_info->key.proto == UDP_PROTO) {
 15     if (size >= 4) {
 16       uint32_t l2tp_part;
 17       l2tp_part = *((uint32_t*)(data));
 18 
 19       if ((l2tp_part & 0xFFFF0000) == 0xC8020000 &&                                                                                                                  
 20           (l2tp_part & 0x0000FFFF) == size) {
 21         f_info->protocol_id = Bina::Protocols::UDP_L2TP;
 22         f_info->application_id = Bina::Applications::UNKNOWN;
 23         return PROTOCOL_ACCEPTED_FINALIZED;
 24       }
 25     }
 26     return REJECTED;
 27   }

国内生产总值产出:

test@ubuntu:~/prj$ gdb prj-loader 
GNU gdb (Ubuntu 7.11.90.20161005-0ubuntu1) 7.11.90.20161005-git
This GDB was configured as "x86_64-linux-gnu".

(gdb) b l2tp_inspector.cc:17
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 1 (l2tp_inspector.cc:17) pending.

(gdb) b l2tp_inspector.cc:18
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 2 (l2tp_inspector.cc:18) pending.

(gdb) b l2tp_inspector.cc:19
No source file named l2tp_inspector.cc.
Make breakpoint pending on future shared library load? (y or [n]) y
Breakpoint 3 (l2tp_inspector.cc:19) pending.

(gdb) r
Starting program: ~/prj/prj-loader 
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[2017-02-02 08:53:07.201600] [0x00007ffff7fc9780] [info]    New id registered TCP with id 0
[2017-02-02 08:53:07.201677] [0x00007ffff7fc9780] [info]    New id registered TCP_PAYLOAD with id 1
[2017-02-02 08:53:07.201711] [0x00007ffff7fc9780] [info]    New id registered UDP with id 2
[2017-02-02 08:53:07.201728] [0x00007ffff7fc9780] [info]    New id registered UDP_PAYLOAD with id 3
[2017-02-02 08:53:07.201757] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:53:07.201773] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:53:07.201817] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:53:07.201832] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:53:07.201853] [0x00007ffff7fc9780] [info]    New id registered IPV4 with id 4
[2017-02-02 08:53:07.201871] [0x00007ffff7fc9780] [info]    New id registered IPV4FRAG with id 5
[2017-02-02 08:53:07.201888] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:53:07.201903] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:53:07.201933] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:53:07.201951] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:53:07.201968] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:53:07.201983] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:53:07.201998] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:53:07.202023] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:53:07.202048] [0x00007ffff7fc9780] [info]    Already registered id TCP with id 0
[2017-02-02 08:53:07.202067] [0x00007ffff7fc9780] [info]    Already registered id TCP_PAYLOAD with id 1
[2017-02-02 08:53:07.202083] [0x00007ffff7fc9780] [info]    Already registered id UDP with id 2
[2017-02-02 08:53:07.202099] [0x00007ffff7fc9780] [info]    Already registered id UDP_PAYLOAD with id 3
[2017-02-02 08:53:07.202114] [0x00007ffff7fc9780] [info]    Already registered id IPV4 with id 4
[2017-02-02 08:53:07.202133] [0x00007ffff7fc9780] [info]    Already registered id IPV4FRAG with id 5
[2017-02-02 08:53:07.202150] [0x00007ffff7fc9780] [info]    New id registered ETHERNET with id 6

Breakpoint 1, inspect (f_info=0x7fffffffce50, data=0x555555be439a "\310\002", size=106, s_module=@0x7fffffffcde0: 0x0)
    at ../src/modules/prj-lib/src/proto-inspectors/l2tp_inspector.cc:17
17        l2tp_part = *((uint32_t*)(data));

(gdb) s
Breakpoint 2, inspect (f_info=0x7fffffffce50, data=0x555555be439a "\310\002", size=106, s_module=@0x7fffffffcde0: 0x0)
    at ../src/modules/prj-lib/src/proto-inspectors/l2tp_inspector.cc:19
19        if ((l2tp_part & 0xFFFF0000) == 0xC8020000 && 

(gdb) s
26      return REJECTED;

为什么 dbg 在第一个代码段中跳过了第 19 行，但在第二个代码段中停在了第 19 行？几乎所有东西都是一样的，唯一的区别就是值(value)。

最佳答案

我想我明白了。

简答:

编译器优化完全删除了 if 条件及其 block !

长答案:

(l2tp_part & 0xFFFF0000) == 0xC802 在 if 条件中的结果总是 false!因为我们正在比较一个 4 字节数字，其中 2 个最低有效字节等于 0x0000，与一个 2 字节非零数字。换句话说，我们将 0xXXXX0000 与 0x0000C802 进行比较，显然它们在任何情况下都不相等。所以整个条件总是假的，因为 if block 是多余的，编译器将其删除。所以调试器不能在那一行停止。

关于c++ - "gdb"调试器奇怪地跳过断点，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/41995063/

文章推荐： c - pipe2(...) vs pipe() + fcntl(...)，为什么不同？

文章推荐： javascript - 什么是 html 模板中的自上而下/自下而上包含

文章推荐： android - android 中的 Socket.io 和 node.js 示例

文章推荐： c - 这是一个普通的 Bug 还是一个聪明的预编译器 Trick？

javascript - 谷歌地图自动完成弹回已经清除的文本......奇怪......奇怪......奇怪
我有这种来自 Google map 自动完成的奇怪行为(或者我可能错过了某事)...想法？奇怪的: 您在输入中输入某物，例如“伦敦” 您按 [ENTER] 你按下 [CLEAR] 按钮你点击进入'输
Java意外类型做字符串比较，奇怪
这段代码与《Learning Java》(Oracle Press Books)一书中的代码完全一样，但它不起作用。我不明白为什么它不起作用，它应该起作用。我用 OpenJDK 和 Sun JDK 7
Powershell 对新行使用反引号 - 奇怪
示例 1 中究竟发生了什么？这是如何解析的？ # doesnt split on , [String]::Join(",",("aaaaa,aaaaa,aaaaa,aaaaa,aaaaa,aa
iphone - 指针类型不兼容？？奇怪
我需要获得方程式系统的解决方案。为此，我使用函数sgesv_()。一切都很好，它使我感到解决方案的正确结果。但是我得到一个奇怪的警告。警告:从不兼容的指针类型传递'sgesv_'的参数3 我正在
ios - 奇怪!动画完成后是否一直调用函数？
我目前在制作动画时遇到一个奇怪的问题: [UIView animateWithDuration:3 delay:0
jQuery 不工作 - 奇怪
alert('works'); $(window).load(function () { alert('does not work'); });
java - 静态内部类 - 奇怪
我的代码: public class MyTest { public class StringSorter implements Comparator { public
JavaScript 对象更新行为(奇怪？)
我正在学习 JavaScript。尝试理解代码， function foo (){ var a = b = {name: 'Hai'}; document.write(a.name +''
c++ - GetLastError() != 奇怪
这个问题不太可能帮助任何 future 的访问者；它只与一个小的地理区域、一个特定的时间点或一个非常狭窄的情况有关，这些情况并不普遍适用于互联网的全局受众。为了帮助使这个问题更广泛地适用，visit
Linux 环境 -i 奇怪
这按预期工作: [dgorur@ted ~]$ env -i env [dgorur@ted ~]$ 这样做: [dgorur@ted ~]$ env -i which date which: no
c++ - 指针增量 - 奇怪
struct BLA { int size_; int size()const{ return size_; } } int x; BLA b[ 2 ]; BLA * p = &b[
css - 图像垂直对齐与 css - 奇怪
我有以下代码: #test img {vertical-align: middle;} div#test { border: 1px solid green; height: 150px; li
gcc - (奇怪？)GCC 预处理器行为
我想大多数使用过 C/C++ 的人都对预处理器的工作原理有一定的直觉(或多或少)。直到今天我也是这么认为的，但事实证明我的直觉是错误的。故事是这样的: 今天我尝试了一些东西，但我无法解释结果。首先考虑
OCMock只生效一次，奇怪，为什么？或者我这边有什么问题？
我想为 TnSettings 做 mock，是的，如果通过以下方法编写代码，它就可以工作，问题是我们需要为每个案例编写 mock 代码，如果我们只 mock 一次然后执行多个案例，那么第二个将报告异常
c - 溢出？找不到来源，奇怪
我的项目中有以下两个结构 typedef volatile struct { unsigned char rx_buf[MAX_UART_BUF]; //Input buffer over U
c# - 奇怪，regex.split方法匹配一个null元素
Regex rx = new Regex(@"[+-]"); string[] substrings = rx.Split(expression); expression = "-9a3dcb
java - JDBC 连接被对等方关闭(奇怪)
我的两个应用程序遇到了一个奇怪的问题。这是设置: 两个 tomcat/java 应用程序，在同一个网络中运行，连接到相同的 MS-SQL-Server。一个应用程序，恰好按顺序位于 DMZ 中可从互联
Android OnLongClickListener 奇怪/不可靠的行为
我目前正在与 Android Api Lvl 8 上的 OnLongClickListener 作斗争。拿这段代码: this.webView.setOnLongClickListener(new
java - JUnit 奇怪 - 我未指定预期数量
这个问题不太可能帮助任何 future 的访问者；它只与一个小的地理区域、一个特定的时间点或一个非常狭窄的情况相关，这些情况并不普遍适用于互联网的全局受众。为了帮助使这个问题更广泛地适用，visit
objective-c - performSelectorOnMainThread 奇怪
只是遇到了奇怪的事情。我有以下代码: -(void)ImageDownloadCompleat { [self performSelectorOnMainThread:@selector(up

太空宇宙

个人简介

我是一名优秀的程序员,十分优秀！

作者热门文章

滴滴打车优惠券免费领取

全站热门文章

首页

博学

6Ren·AI

商城

c++ - "gdb"调试器奇怪地跳过断点