开发者

Memory failures when calling new

It might be a little hard to explain everything about this error but I'll try my best. Unfortunately the code is so large that it would be unpractical to try to include it.

I am writing an academic OS for homework (nachos if you must know) and the most recent thing I needed to do was implement mutual exclusion in a core map. To do this I added a Lock and 开发者_运维知识库a condition variable per page in emulated main memory. After doing this I run my code and get through one call to the exception (in a completely different directory from core map) handler just fine, but the second time the function gets called I get and error on the following call: r=new Lock("read"); and it reads like this:

*** glibc detected *** /home/some_other_directories/workspace/nachos3_repo/vm/nachos: malloc(): memory corruption (fast): 0x0805fe20 ***

Just to see how it behaves I changed the allocation of that lock to be an extern in my system file (there are a lot of global externs there) and after that I get a seg fault on the call fout.open("old.txt"); which I traced back through the stack to be within a call to malloc in a call to new.

My best guess is that my heap is getting to large but I am not sure if that is the case or how to handle it if it is. can anyone shed light on this problem?


malloc is just the victim here. In some completely different part of the code you have written outside the allocated boundaries and thus have corrupted chain links that are used by the heap manager to track free/allocated blocks. As you try to allocate here, the heap manager is chasing these linked lists and lands into a protected area (most likely not allocated). You should do a review of your changes first and see if you can spot where you scribble the memory.

0

上一篇:

下一篇:

精彩评论

暂无评论...
验证码 换一张
取 消

最新问答

问答排行榜