远程gdb调试不会停在断点处 [英] Remote gdb debugging does not stop at breakpoints

问题描述

我在调试修改后的Linux-3.11.0内核时遇到问题。
为了调试代码，我使用：

1. GDB 7.6


2. QEMU 1.6。 5


3. Linux内核3.11.0

我的系统具有Intel（R）Core ）运行Ubuntu 12.10的i7-2760QM CPU @ 2.40GHz我使用'make menuconfig'创建了一个配置文件来编译修改后的内核。 （在内核黑客类别中进行了修改）

一些.config文件行（我认为相关）在这里引用：

  CONFIG_DEBUG_INFO = y 
 CONFIG_DEBUG_KERNEL = y 
 CONFIG_MAGIC_SYSRQ = y 
 CONFIG_DEBUG_FS = y 
 CONFIG_HAVE_HW_BREAKPOINT = y 
 CONFIG_HAVE_MIXED_BREAKPOINTS_REGS = y 
  

我使用带-S选项的qemu来启动内核，使用gdb我连接到qemu并恢复执行：

 （gdb）目标远程：1234 
（gdb）继续
  

设置断点时，例如：
break schedule 我得到一个确认信息，例如：

 断点1在0xffffffff81736400：file kernel / sched / core。 c，2509行。
  

所以在这里一切似乎都好！

问题是gdb和qemu不会停在断点处！
他们只是继续运行......

我尝试按照某些地方的建议设置硬件断点，但是我收到消息：

 目标中没有硬件断点支持。 
  

这很奇怪，因为我设置了CONFIG_HAVE_HW_BREAKPOINT。

我也取消了 CONFIG_DEBUG_RODATA 的建议，但仍然没有结果...

试图用这里建议的.config文件搞乱： gdbserver里面qemu不停止在断点上，但我无法弄清楚.config文件和我的.config文件之间的相关区别......（有太多的不同！）

任何人都可以帮我吗？

谢谢！

解决方案

好的，我发现了这个问题，所以我在这里发布给其他人：
显然这是一个非常微妙的过程...我的配置文件很好。我只需要在连接到qemu之后定义断点，而不是在加载vmlinux之后定义断点。这些断点必须是硬件断点......否则它们不会中断！

I'm having problems with debugging my modified Linux-3.11.0 kernel. For debugging the code, I use:

1. GDB 7.6
2. QEMU 1.6.5
3. Linux kernel 3.11.0

My system has Intel(R) Core(TM) i7-2760QM CPU @ 2.40GHz running Ubuntu 12.10

I compiled the modified kernel with a configuration file that I created using 'make menuconfig'. (with modifications in the "kernel hacking" category)

Some of the .config file lines (that I find relevant) are quoted here:

CONFIG_DEBUG_INFO=y 
CONFIG_DEBUG_KERNEL=y
CONFIG_MAGIC_SYSRQ=y
CONFIG_DEBUG_FS=y
CONFIG_HAVE_HW_BREAKPOINT=y
CONFIG_HAVE_MIXED_BREAKPOINTS_REGS=y

I use qemu with the "-S" option to boot my kernel, and using gdb I connect to qemu and resume it's execution:

(gdb) target remote :1234
(gdb) continue

When setting breakpoints, for an example: break schedule I get a confirmation message such as:

Breakpoint 1 at 0xffffffff81736400: file kernel/sched/core.c, line 2509.

So till here everything seems fine!

The problem is that gdb and qemu does not stop at the breakpoints! They both simply continue running...

I tried to set hardware breakpoints as suggested in some places, but I get the message:

No hardware breakpoint support in the target.

which is weird because I have the CONFIG_HAVE_HW_BREAKPOINT set.

I also unset CONFIG_DEBUG_RODATA as suggested in other places,but still no results...

I tried to mess around with the .config file as suggested here: gdbserver inside qemu does not stop on breakpoints but I couldn't figure out the relevant differences between the working .config file and my .config file... (There are too much differences!)

Can anyone please help me?

Thank You!!

解决方案

OK, I found the problem so I'm posting here for other people: Apparently this is a very delicate procedure... My config file was fine. All I needed to do is to define the breakpoints only after I connect to qemu, and not after loading the vmlinux. These breakpoints have to be HARDWARE breakpoints... otherwise they don't break!

这篇关于远程gdb调试不会停在断点处的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！