MDM和Microblaze出错

大家好
我们有一个Spartan 6 SOC设计与Microblaze，MDM和其他一些外围设备。
对于这个系统我写了SW（在C中），一切似乎都有效。
但有时整个软件崩溃并且没有响应。
我们的系统中有一个看门狗，如果发生错误并且sw将不再刷新wd，通常应该触发。
但也没有重置，系统不再响应，没有任何反应。
我现在发现的是，在崩溃之前，sw跳转到向量0x18，这是断开向量。
根据文档，有两种方法可以跳转到这个向量（sw跳转或来自mdm的信号）。
通常我不会使用mdm，只是开发我调试，但是当我们测试我们的系统时，我们不会使用mdm（但它与mdm包含相同的设计）。
所以我的问题是，如果没有配置mdm可能会引发其中一个信号，也可能会停止mb？
我将在一个设计中删除mdm并禁用调试，但它真的很难验证，如果这是解决方案导致错误有时两分钟发生在一分钟后来与sw相同的精灵文件和相同的fgpa设计在百小时内没有
更长的时间。
是否还有更多关于mdm的文档以及如何调试微型激光器，因此我可以理解可能出现的错误（例如，如果mdm可以阻止mb，那么可能是系统在崩溃后不再响应的原因）。
亲切的问候和提前感谢
马丁



以下为原文

Hi all

We have a Spartan 6 SOC design with and Microblaze, MDM and some other peripherie. For this system I wrote SW (in C), everything seems to work. But sometimes the whole SW crashes and do not respond. We have a watchdog in our system, which normally should trigger if an error occured and the sw will no longer refresh the wd. But also no reset occurs, the system respond no longer and nothing happens. What I now discover is that before a crash the sw jumps to the vector 0x18 which is the break vector. According to the documentation, there are two ways to jump to this vector (either a sw jump or a signal from the mdm).

Normally I will not use the mdm, only druing developping I debug, but when we test our system we will not use the mdm (but its the same design with the mdm included). So my question is, it is possible that the mdm raises one of the signals if not configured and maybe also stops the mb? I will remove in one design the mdm and disable debugging but its really hard to verifiy if that is the solution cause the error occured sometimes twice in one minute and later with the same elf file from the sw and the same fgpa design in hundred hours no longer. Is there also more documentation about the mdm and how a microblaze is debugged, so I can understand what could maybe the error (for example if the mdm could stop the mb, whats maybe the reason why the system do not respong after the crash).

Kind regards and thanks in advance
Martin