hi, I have a C source code .I have created an elf image using the source code along with startup code and linkerscript. I want to view the memory partion ie, length of code segment , data segment stack and heap.Is there any methode to view the allocated memory (ie, segment based addresses)?. Is there any way to monitor the heap allocation , ( to avoid the stack heap overlapping)? regards thejus
The linker can output a map file to show memory organisation of your code with teh following option: -Map <mapfile> With respect to heap monitoring. The sbrk()/sbrk_r() syscall function is ultimately how the system aquired heap. You have te implement this yourself (or use one of the examples), you can add code here to perform monitoring operations, but it is normally written to be bounded to a defined memory region. Of greater concern would be stack overflow, since that is more difficult to guard against. A common approach is to pre-fill the stack segment with a known and repeated value or sequence of values, then to determine stack usage, scan the segment to test for the 'high warter mark'. Clifford
hi Clifford, Actually i am using so many pointers (some size up to 8mb). So if there is a way to view the heap allocation we can judge whether heap over writes the stack. am i right? What is the difference between hw stack and sw stack.? how can we know exactly how much space is required for stack (in hw). regards thejus Clifford Slocombe wrote: > The linker can output a map file to show memory organisation of your > code with teh following option: > > -Map <mapfile> > > With respect to heap monitoring. The sbrk()/sbrk_r() syscall function is > ultimately how the system aquired heap. You have te implement this > yourself (or use one of the examples), you can add code here to perform > monitoring operations, but it is normally written to be bounded to a > defined memory region. Of greater concern would be stack overflow, since > that is more difficult to guard against. A common approach is to > pre-fill the stack segment with a known and repeated value or sequence > of values, then to determine stack usage, scan the segment to test for > the 'high warter mark'. > > Clifford