In this project I wrote a dynamic storage allocator for C programs, i.e., my own version of the malloc , free , and realloc routines. You are encouraged to explore the design space creatively and implement an allocator that is correct, efficient, and fast.

Your dynamic storage allocator will consist of the following four functions, which are declared in mm.h and defined in mm.c .

• int mm_init(void);
• void *mm_malloc(size_t size);
• void mm_free(void *ptr);
• void *mm_realloc(void *ptr, size_t size); The mm.c file we have given you implements a simple memory allocator based on an implicit free list, first fit placement, and boundary tag coalescing, as described in the textbook. Using this as a starting place, modify these functions (and possibly define other private static functions), so that they obey the following semantics:
• mm_init:Before calling mm_malloc , mm_realloc , or mm_free , the application program (i.e., the trace-drive n driver program that you will use to evaluate your implementation) calls mm_init to perform any necessary initializations, such as allocating the initial heap area. The return value should be −1 if there was a problem in performing the initialization, 0 otherwise.

The driver will call mm_init before running each trace (and after resetting the brk pointer). Therefore, your mm_init function should be able to reinitialize all state in your allocator each time it is called. In other words, you should not assume that it will only be called once.

• mm_malloc: The mm_malloc routine returns a pointer to an allocated block payload of at least size bytes. The entire allocated block should lie within the heap region and should not overlap with any other allocated chunk. We will compare your implementation to the version of malloc supplied in the standard C library ( libc ). Since the libc malloc always returns payload pointers that are aligned to 8 bytes, your malloc implementation should do likewise and always return 8-byte aligned pointers.
• mm_free: The mm_free routine frees the block pointed to by ptr It returns nothing. This routine is only guaranteed to work when the passed pointer ( ptr ) was returned by an earlier call to mm_malloc or mm_realloc has not yet been freed.
• mm_realloc: The mm_realloc routine returns a point er to an allocated region of at least size bytes with the following constraints:
• if ptr is NULL, the call is equivalent to mm_malloc(size);
• if size is equal to zero, the call is equivalent to mm_free(ptr);
• if ptr is not NULL, it must have been returned by an earlier call to mm_malloc or mm_realloc . The call to mm_realloc changes the size of the memory block pointed to by ptr (the old block ) to size bytes and returns the address of the ne w block. Notice that the address of the new block might be the same as th e old block, or it might be different, depending on your implementation, the amount of internal fragmentation in the old block, and the size of the realloc request.
  The contents of the new block are the same as those of the old ptr block, up to the minimum of the old and new sizes. Everything else is uninitialized. For example, if the old block is 8 bytes and the new block is 12 bytes, then the first 8 bytes of the new block are identical to the first 8 bytes of the old block and the last 4 bytes are uninitialized. Similarly, if the old block is 8 bytes and the new block is 4 bytes, then the contents of the new block are identical to the first 4 bytes of the old block.
  These semantics must match the semantics of the corresponding libc malloc , realloc , and free routines. Type man malloc for complete documentation.

In this project I have created a dynamic storage allocator for C programs making our own versions of malloc, free and realloc. For this purpose, I have used the approach “Segregated explicit free list” that contains 10 different categories of block sizes varying from 'less than 2^6' to 'greater than 2^14' in the intervals of the powers of 2.

Segregated free list uses different linked lists for different sized blocks. Each of these linked list can be seen as an array of explicit free list. I am using this approach because it outputs performance better than explicit free lists and currently adopted implicit free lists. It performs adding and freeing operations in constant time. In addition to that, segregated best fit approach is faster than the regular/unsegregated best fit since it always searches in a free list for an appropriate index (index signifies our individual linked list for appropriate size). If it fails to fit in this index then it searches for larger indices. It also controls fragmentation of segregated storage by everytime splicing and coalescing the blocks after searching and allocating the block.

• Defining additional macros for segregated free list which are basically getter and setter methods for getting and setting next and previous block pointers in accordance with our block structure.

• Defined additional functions to perform operations on free list, i.e. adding a free block to the free list, removing a newly allocated block from the free list, coalescing between the blocks, etc. In the new created coalescing routine(main_coalesce), I have considered all the four cases in which our block can be coalesced. Then by calling the regular coalescing routine, I am adding the new coalesced block to the free list and removing the block from the free list, on which coalescing was being performed. Hence add and remove functions will be used.

• Changed a routine of find_fit for segregated free list, in which I am using the best fit method to find a fit for a block of required size. Initially I will fetch the list index from where the appropriate size of block could be achieved and if it fails to find any free block in that particular linked list of the list index (i.e. if free list is empty for that particular list index), it will search for the best fit in the linked list of next list index and so on.

• Updated the malloc and free routines that can call the segregated free list routines which I made initially. Now it calls find_fit and for allocating the blocks it calls an updated 'place' routine (which is now able to handle splicing of under-allocated blocks and placing the smaller splice into an appropriate free list). Finally it calls extend_heap routine if it fails to allocate a block of the required size. In the 'free' routine there is no change from what was given earlier apart from calling our new main_coalesce routine instead of regular coalesce. Updating malloc and free turned out to be a boost in the performance up to 38%.

• Finally I updated the realloc routine boosting up the performance up to 95%. I have added a few new things in addition to our current implementation. Realloc checks if I can coalesce on the right and creates a resultant block with appropriate size. This greatly increases the performance. Also if old ptr is NULL, then it just calls malloc. And if size is zero then it calls free, and returns NULL.

• Made an updated version of heap consistency checker which can now check for the following :
  1. Checks Prologue header
  2. Checks if the block is double word aligned or not.
  3. Checks if header matches footer or not.
  4. Checks if contiguous free blocks somehow escaped coalescing.
  5. Checks if free block is actually in free list. The number of free blocks in heap and segregated list match.
  6. Checks if pointers in free list points to valid free blocks.
  7. Checks epilogue header.