Re: Memory Fragmentation - Weekly Report

["Andrei Soare" <andrei soare gmail com>]

<-- Week 4 -->

Hello,

First, I am very happy to tell you that I passed the exam I was most worried about :D and now I have 2 more, but easy ones, I will have time for GSoC.

Last week I didn't do any work, as I said in my last report.

These days, I have been struggling with that bug in memview.py and realized that it might not be memview at fault. When I run free_size to determine the size of the deallocated memory blocks, I encounter many situations when a free or realloc is called to a zone which has never been allocated before. That's why I get a page overflow in memview.py, because not all memory that should be freed, is actualy freed. So I look at my_malloc_hook: it creates a log file for each PID and I noticed that after I run pidgin with the malloc hooks, a huge file is created with the allocations log, but also some small files, that belong to other pids. I'm thinking about threads now. I modified it to put all the data into one file, not taking pids into account, but with the same result. Maybe, if they belong to different threads, the data is somehow mixed and that's where it fails.

I would appreciate a professional opinion on this :)

These days, I will just try to fix this problem.


Regards,
Andrei Soare