Subversion Repositories shark

/*

 * Doubly indexed dynamic memory allocator (DIDMA)

 * Version 0.61

 *

 * Written by Miguel Masmano Tello <mmasmano@disca.upv.es>

 * Copyright (C) Dec, 2002 OCERA Consortium

 * Release under the terms of the GNU General Public License Version 2

 * shark porting Trimarchi Michael

 */

#include <kernel/func.h>

#include "didma.h"

#include "queuem.h"

#include "i386/bitops.h"

#include "i386/bits.h"

#ifndef NULL

#define NULL ((void *) 0)

#endif // #ifndef NULL

#define INIT_THREAD_MUTEX()

#define THREAD_LOCK()   kern_cli()

#define THREAD_UNLOCK() kern_sti()

// u8, u16, u32 are only defined in kernel headers

#define u8 char 

#define u16 unsigned short int 

#define u32 unsigned int 

#define MAX_SL_LOG2_INDEX 5 // Real value is 2**MAX_SL_INDEX

#define MAX_SL_INDEX (1 << MAX_SL_LOG2_INDEX)

#define F_OK 0

#define F_ERROR -1

#define MIN_LOG2_SIZE 5 // 32 bytes

#define MAX_LOG2_SIZE MAX_FL_INDEX

#define MIN_SIZE (1 << MIN_LOG2_SIZE)

#define MAX_SIZE (1 << MAX_LOG2_SIZE)

#define DIDMA__set_bit(num, mem) mem |= (1 << num)

#define DIDMA__clear_bit(num, mem) mem &= ~(1 << num)

/*

 * First, all necessary TADs are defined

 */

/*

 * the follow TAD will be a double level indexed array, the most important

 * thing is that the time is limited, it is because this dynamic memory manger

 * is designed to run with real time programs.

 *

 *     First level       Second level

 *     it is indexed     it is indexed by 2**n+(2**n/m*index_number)

 *     by power of 2        

 *                            0             1     m-1        m

 *                            ----> NULL   --> NULL          ----> NUL

 *                            |            |                 |

 *       -------         ---------------------...---------------------

 *  2**n |  n  |  -----> |2**n+(2**n/m*0)|...|   |...|2**n+(2**n/m*m)|

 *       |-----|         ---------------------...---------------------

 * 2**n-1| n-1 |  -----> ....                      |            

 *       |-----|                                   --->NULL

 *        .....

 */

/* DON'T TOUCH THESE MACROS */

#define MAGIC_NUMBER 0xA5A5

#define MAGIC_NUMBER_NONE 0x0000

#define BLOCK_FREE 0

#define BLOCK_USED 1

struct free_ptr_struct {

  struct block_header_struct *prev;

  struct block_header_struct *next;

  /*

   * first_l_index and second_l_index are used to store

   * mapping_function results, that's how we get some extra

   * nanoseconds

   */

  u8 first_l_index;

  u8 second_l_index;

};

typedef struct block_header_struct {

  /* magic_number is the id of the block */

  u16 magic_number;

  /* size of the block */

  size_t size;

  /* state of the block can be free or used */

  u8 state;

  struct block_header_struct *prev;

  struct block_header_struct *next;

  union {

    struct free_ptr_struct free_ptr;

    u8 buffer[sizeof(struct free_ptr_struct)];

  } ptr;

} block_header_t;

/* first level index array */

typedef struct fl_array_struct {

  /* ptr is pointer to next level */

  block_header_t **sl_array;

  /* n_blocks is the number of the blocks which they are pointed by sl_array */

  u32 bitmapSL;

} fl_array_t;

/*

 * fl_array will be our first level array.

 */

static fl_array_t *fl_array;// [MAX_FL_INDEX - MIN_LOG2_SIZE];

u32 bitmapFL;

/*

 * first_bh will be our first memory block allocated by MALLOC function

 */

static block_header_t *first_bh, *free_block;

static int all_ok = 0;

/*

 * header_overhead has size of blocks_header - 2 * ptr to next free block

 */

static int header_overhead = 0, total_size = 0;

/*

 * log2size () returns cell of log2 (len)

 * it does a search between MIN_SIZE and MAX_SIZE values

 */

static int log2size(size_t size, size_t *new_size)

{

  int i;

  if (size <= 0) {

    PRINT_MSG ("ERROR: Chunk length must be > 0\n");

    return F_ERROR;

  }

  // MIN_SIZE and MAX_SIZE must be defined  

  if (size <= MIN_SIZE){

    *new_size = MIN_SIZE;

    return MIN_LOG2_SIZE;

  }

  for (i = MIN_LOG2_SIZE;

       (i <= MAX_LOG2_SIZE) && (size > (*new_size = (1 << i))) ; i++);

  if (i > MAX_LOG2_SIZE) {

    PRINT_MSG ("ERROR: Maximum chunk size exceeded\n");

    return F_ERROR;

  }

  return i;

}

/*

 * caculate_reserve () returns the size that has to be reservate for

 * avoid fragmentation

 */

#define MAX(a, b) (a>b)?a:b

/*

  May be calculate_reserve can be improved

*/

static inline int calculate_reserve (int size){

  //  int internal_frag, dummy;

  //  header_overhead = (int) first_bh -> ptr.buffer - (int) first_bh;  

  //internal_frag = (1 << log2size(size, &dummy)) / MAX_SL_INDEX;

  //dummy = MAX((header_overhead * size / MAX_SL_INDEX) + internal_frag, size);

  return size;//(size + dummy);

}

/*

 * mapping_function () returns first and second level index

 *

 * first level index function is:

 * fl = log2 (size)

 *

 * and second level index function is:

 * sl = (size - 2**(log2size (size))) / (log2 (size) - log2 (MAX_SL_INDEX))

 *

 */

static int mapping_function (size_t size, int *fl, int *sl){

  int aux;

  int new_len;

  // first level = log2 (size)

  *fl = log2size (size, &new_len);

  if (new_len == size) {

    //  if 2**fl == size, second level must be 0

    *sl = 0;

  } else {

    -- *fl;

    aux = *fl - MAX_SL_LOG2_INDEX;

    *sl = (int)((size - (new_len >> 1)) >> aux);

  }

  *fl -= MIN_LOG2_SIZE;

  return F_OK;

}

/*

 * init_TAD () initialize the structure fl_array to NULL

 * and free_block with the initial blocks pointed by ptr

 *

 * its cost is O (n x m)

 */

static int init_TAD (size_t size, u8 *ptr){

  int n, aux, dummy;

  if (!(size > 0)) {

    PRINT_MSG ("ERROR: size must be > 0\n");

    return F_ERROR;

  }

  if (MAX_SL_INDEX > MIN_SIZE) {

    PRINT_MSG ("ERROR: MAX_SL_INDEX must be < MIN_SIZE\n");

    return F_ERROR;

  }

  total_size = size;

  MAX_FL_INDEX = log2size(size, &dummy);

  if (MAX_FL_INDEX < 0) return -1;

  fl_array = (fl_array_t *) SYSTEM_MALLOC ((MAX_FL_INDEX - MIN_LOG2_SIZE)

                                           * sizeof(fl_array_t));

  if (fl_array == NULL) return -1;

  _init_malloc_ (20);

  for (n = 0; n < MAX_FL_INDEX - MIN_LOG2_SIZE; n++) {

    fl_array [n].bitmapSL = 0;

    aux = (1 << (n + MIN_LOG2_SIZE));

    fl_array[n].sl_array = NULL;

  }

  bitmapFL = 0;

  header_overhead = (int) first_bh -> ptr.buffer - (int) first_bh;

  first_bh = (block_header_t *) ptr;

  first_bh -> magic_number = MAGIC_NUMBER;

  first_bh -> size = size - header_overhead;

  first_bh -> state = BLOCK_FREE;

  first_bh -> prev = NULL;

  first_bh -> next = NULL;

  first_bh -> ptr.free_ptr.prev = NULL;

  first_bh -> ptr.free_ptr.next = NULL;

  free_block = first_bh;

  //mapping_function (first_bh->size,&n, &m);

  //fl_array[n].sl_array[m].ptr= first_bh;

  //DIDMA__set_bit (m, fl_array[n].bitmapSL);

  //DIDMA__set_bit (n, bitmapFL);

  all_ok = 1;

  return F_OK;

}

/*

 * destroy_TAD return a pointer to the initial block

 */

static void *destroy_TAD (void){

  all_ok = 0;

  _destroy_malloc_();

  return (void *) first_bh;

}

/*

 * insert_TAD merge a free block pointed by ptr

 * with its buddies and then

 * insert it into fl_array

 *

 * The cost of this operation is

 * always O (K) = (1)

 */

static int insert_TAD (u8 *ptr){

  int fl, sl;

  block_header_t *bh = NULL, *bh2;

  int for_free_block = 0;

  if (!all_ok) return F_ERROR;

  bh = (block_header_t *) (ptr - header_overhead);

  if (bh -> magic_number != MAGIC_NUMBER)

    return F_ERROR;

  THREAD_LOCK();

  /*

   * now bh is a free block

   */

  bh -> state = BLOCK_FREE;

  bh -> ptr.free_ptr.prev = NULL;

  bh -> ptr.free_ptr.next = NULL;

  /*

   * first bh will be merge with its free buddies and

   * then it will be inserted with the free blocks

   */

  /*

   * it is used to know if bh have to be inserted into free_block or

   * into fl_array

   */

  if (bh -> next == NULL) for_free_block = 1;

  /* is it free the follow block? */

  if (bh -> next != NULL) {

    bh2 = bh -> next;

    if (bh2 -> magic_number == MAGIC_NUMBER &&

        bh2 -> state == BLOCK_FREE) {

      /* we are lucky, we can merge bh with the next block */

      if (bh2 == free_block) for_free_block = 1;

      bh2 -> magic_number = MAGIC_NUMBER_NONE;

      if (bh2 -> next != NULL)

        bh2 -> next -> prev = bh;

      if (!for_free_block) {

        if (bh2 -> ptr.free_ptr.next != NULL)

          bh2 -> ptr.free_ptr.next -> ptr.free_ptr.prev =

            bh2  -> ptr.free_ptr.prev;

        if (bh2 -> ptr.free_ptr.prev != NULL)

          bh2 -> ptr.free_ptr.prev -> ptr.free_ptr.next =

            bh2 -> ptr.free_ptr.next;

        //mapping_function (bh2 -> size, &fl, &sl);

        fl = bh2 -> ptr.free_ptr.first_l_index;

        sl = bh2 -> ptr.free_ptr.second_l_index;

        /* bh2 must be erased from fl_array */

        if (fl_array [fl].sl_array [sl] == bh2)

          fl_array [fl].sl_array [sl] = bh2 -> ptr.free_ptr.next;

          //fl_array [fl].n_blocks --;

          if (fl_array [fl].sl_array [sl] == NULL){

            //fl_array[fl].bitmapSL &= ~(1 << sl);

            DIDMA__clear_bit (sl, fl_array[fl].bitmapSL);

            if (!fl_array[fl].bitmapSL) {

              _free_ ((void *) fl_array [fl].sl_array);

              DIDMA__clear_bit (fl, bitmapFL);

            }

          }

      }

      bh -> size += bh2 -> size + header_overhead;

      bh -> next = bh2 -> next;

    }

  }

  /* is it free the previous block? */

  if (bh -> prev != NULL) {

    bh2 = bh -> prev;

    if (bh2 -> magic_number == MAGIC_NUMBER &&

      bh2 -> state == BLOCK_FREE) {

      bh -> magic_number = MAGIC_NUMBER_NONE;

      if (bh -> next != NULL)

        bh -> next -> prev = bh2;

      if (bh2 -> ptr.free_ptr.next != NULL)

        bh2 -> ptr.free_ptr.next -> ptr.free_ptr.prev =

          bh2  -> ptr.free_ptr.prev;

      if (bh2 -> ptr.free_ptr.prev != NULL)

        bh2 -> ptr.free_ptr.prev -> ptr.free_ptr.next =

          bh2 -> ptr.free_ptr.next;

      //mapping_function (bh2 -> size, &fl, &sl);

      fl = bh2 -> ptr.free_ptr.first_l_index;

      sl = bh2 -> ptr.free_ptr.second_l_index;

      if (fl_array [fl].sl_array [sl] == bh2)

        fl_array [fl].sl_array [sl] = bh2 -> ptr.free_ptr.next;

      if (fl_array[fl].sl_array[sl] == NULL){

        //fl_array[fl].bitmapSL &= ~(1 << sl);

        DIDMA__clear_bit (sl, fl_array[fl].bitmapSL);

        if (!fl_array[fl].bitmapSL) {

          _free_ ((void *) fl_array [fl].sl_array);

          DIDMA__clear_bit (fl, bitmapFL);

        }

      }

      bh2 -> size += bh -> size + header_overhead;

      bh2 -> next = bh -> next;

      bh = bh2;

    }

  }

  /*

   * and then we merge the free block with the initial memory

   */

    if (for_free_block) {

      free_block = bh;

      bh -> ptr.free_ptr.next = NULL;

      bh -> ptr.free_ptr.prev = NULL;

    } else {

      /*

       * or we insert the free block in the TAD

       */

      mapping_function (bh -> size, &fl, &sl);

      if (fl_array[fl].bitmapSL == 0)

        fl_array[fl].sl_array = (block_header_t **) _malloc_ ();

      if (fl_array[fl].sl_array == NULL) {

        PRINT_MSG("CRITICAL ERROR: there isn't enought memory for free\n");

        THREAD_UNLOCK();

        return F_ERROR;

      }

      bh -> ptr.free_ptr.first_l_index = fl;

      bh -> ptr.free_ptr.second_l_index = sl;

      bh -> ptr.free_ptr.next = fl_array [fl].sl_array [sl];

      bh -> ptr.free_ptr.prev = NULL;

      if (fl_array [fl].sl_array [sl] != NULL)

        fl_array [fl].sl_array [sl] -> ptr.free_ptr.prev = bh;

      fl_array [fl].sl_array [sl] = bh;

      DIDMA__set_bit (sl, fl_array[fl].bitmapSL);

      DIDMA__set_bit (fl, bitmapFL);

    }

  THREAD_UNLOCK();

  return F_OK;

}

/*

 * search_TAD searchs a free block of size 'size'

 * then this block will be splitted in two new blocks,

 * one of these new blocks will be given to the user and the

 * other will be inserted into a free blocks structure

 *

 * The cost of this operation is

 *      best case: (K) = (1)  

 *      worst case: (MAX_FL_LOG2_INDEX - MIN_FL_LOG2_INDEX + MAX_SL_INDEX + K)

 *                   = (1)

 * where K is an integer constant

 */

static u8 *search_TAD (size_t size) {

  int for_free_block = 0;

  int fl, sl, n;

  int found;

  block_header_t *bh = NULL, *bh2;

  if (!(size > 0) || size >= MAX_SIZE || !all_ok) return NULL;

  THREAD_LOCK();

  if (size <= MIN_SIZE) {

    size = MIN_SIZE;

    fl = 0;

    sl = 0;

  } else {

    mapping_function (size, &fl, &sl);

    if (++sl == MAX_SL_INDEX) {

      fl ++;

      sl = 0;

    }

    /*

     * This is the reason of the internal fragmentation

     * The block given is greater that the size demanded

     */

    size = (1 << (fl + MIN_LOG2_SIZE));

    size = size + ((size >> MAX_SL_LOG2_INDEX) * sl);

  }

  /* the size given can be bigger than the size demanded */

  found = 0;

  // we take the first free block from fl_array or a buddy of him

  sl = fl_array[fl].bitmapSL & ((~0) << sl);

  if (sl != 0) {

    sl = DIDMA_fls(sl);

    bh = fl_array [fl].sl_array [sl];

    fl_array [fl].sl_array [sl] = bh -> ptr.free_ptr.next;

    if (fl_array [fl].sl_array [sl] != NULL)

      fl_array [fl].sl_array [sl] -> ptr.free_ptr.prev = NULL;

    else {

      DIDMA__clear_bit (sl, fl_array[fl].bitmapSL);

      if (!fl_array[fl].bitmapSL) {

        DIDMA__clear_bit (fl, bitmapFL);

        _free_ ((void *) fl_array [fl].sl_array);

      }

    }

    found = 1;

    goto out;

  }

  /*

   * if fl_array is empty we will take a free block

   * from free_block pointer

   */

  if (free_block != NULL && free_block -> size >= size) {

    bh = free_block;

    free_block = NULL;

    for_free_block = 1;

    found = 1;

    goto out;

  }

  /*

   * A free block is searched

   *

   * Using a bitmap

   */

  fl = DIDMA_fls(bitmapFL & ((~0) << (fl + 1)));

  if (fl > 0) {

    sl = DIDMA_fls(fl_array[fl].bitmapSL);

    bh = fl_array [fl].sl_array [sl];

    fl_array [fl].sl_array [sl] = bh -> ptr.free_ptr.next;

    if (fl_array [fl].sl_array [sl] != NULL){

      fl_array [fl].sl_array [sl] -> ptr.free_ptr.prev = NULL;

    } else {

      DIDMA__clear_bit (sl, fl_array[fl].bitmapSL);

      if (!fl_array[fl].bitmapSL) {

        _free_ ((void *) fl_array[fl].sl_array);

        DIDMA__clear_bit (fl, bitmapFL);

      }

    }

    found = 1;

    goto out;

  }

 out:

  /*

   * HUGGGG, NOT ENOUGHT MEMORY

   * I think that we have done all that we have been able

   */

  if (!found) {

    PRINT_MSG ("ERROR: Memory pool exhausted!!!\n");

    THREAD_UNLOCK();

    return NULL;

  }

  /*

   * we can say: YESSSSSSSSSSS, we have enought memory!!!!

   */

  bh -> state = BLOCK_USED;

  /* can bh be splitted? */

  n = (int)(bh -> size - size - header_overhead);

  if (n >= (int) MIN_SIZE) {

    /*

     * Yes, bh will be splitted

     */

    bh2 = (block_header_t *) ((u8 *)(bh -> ptr.buffer) + size);    

    bh2 -> magic_number = MAGIC_NUMBER;

    bh2 -> state = BLOCK_FREE;

    bh2 -> size = n;

    if (bh -> next != NULL)

      bh -> next -> prev = bh2;

    bh2 -> next = bh -> next;

    bh -> next = bh2;

    bh2 -> prev = bh;

    bh -> size = size;

    if (!for_free_block) {

      mapping_function (bh2 -> size, &fl, &sl);

      if (fl_array[fl].bitmapSL == 0)

        fl_array[fl].sl_array = (block_header_t **) _malloc_ ();

      if (fl_array[fl].sl_array == NULL) {

        PRINT_MSG("CRITICAL ERROR: there isn't enought memory for free\n");

        return NULL;

      }

      bh2 -> ptr.free_ptr.first_l_index = fl;

      bh2 -> ptr.free_ptr.second_l_index = sl;

      bh2 -> ptr.free_ptr.prev = NULL;

      bh2 -> ptr.free_ptr.next = fl_array [fl].sl_array [sl];

      if (fl_array [fl].sl_array [sl] != NULL)

        fl_array [fl].sl_array [sl] -> ptr.free_ptr.prev = bh2;

      fl_array [fl].sl_array [sl] = bh2;

      DIDMA__set_bit (sl, fl_array[fl].bitmapSL);

      DIDMA__set_bit (fl, bitmapFL);

        //fl_array [fl].n_blocks ++;

    } else

      free_block = bh2;

  }

  THREAD_UNLOCK();

  return (u8 *) bh -> ptr.buffer;

}

int init_memory_pool (int max_size){

  u8 *buffer;

  int real_max_size = 0;

  if (max_size < 0) return -1;

  //buffer = (u8 *) SYSTEM_MALLOC (real_max_size * sizeof (u8));

  real_max_size = calculate_reserve(max_size);

  buffer = (u8 *) SYSTEM_MALLOC (real_max_size * sizeof (u8));

  if (buffer == NULL) return F_ERROR;

  return init_TAD (real_max_size, buffer);

}

void destroy_memory_pool(void){

        SYSTEM_FREE (destroy_TAD ());

}

/* see 'man malloc' */

void *rt_malloc (size_t size){

  return (void *) search_TAD (size);

}

/* see 'man free' */

void rt_free (void *ptr){

  if (ptr != NULL) insert_TAD ((u8 *) ptr);

}

/* see 'man calloc' */

void *rt_calloc (size_t nelem, size_t elem_size)

{

  void *p;

  if (!all_ok || nelem <= 0 || elem_size <= 0) return NULL;

  if ((p = rt_malloc (nelem * elem_size)) == NULL)

    return NULL;

  memset (p, 0, nelem * elem_size);

  return p;

}

/*

 * see man realloc  

 * be careful, realloc () is an expensive operation because

 * it uses malloc () and free ()

 */

void *rt_realloc (void *p, size_t new_len)

{

  u8 *aux;

  void *addr;

  int old_len, min;

  block_header_t *b;

  if (!all_ok) return NULL;

  if (p == NULL)

    return rt_malloc (new_len);

  else if (new_len == 0) {

    rt_free(p);

    return NULL;

  } else {

    /*

     * Now we try to achieve old size because

     * then we need it to copy all data in the new allocated

     * memory

     */

    aux = (u8 *) p;

    aux -= header_overhead;

    b = (block_header_t *) aux;

    if (b -> magic_number == MAGIC_NUMBER) {

      old_len = b -> size;

    } else {

      PRINT_MSG ("CRITICAL ERROR: block corrupted\n");

      return NULL;

    }

    addr = rt_malloc (new_len);

    if (addr == NULL) return NULL;

    min = (old_len > new_len)? new_len : old_len;

    memcpy (addr, p, min);

    rt_free (p);

    return addr;

  }

}

#ifdef __DEBUG__

/*

 * dump_mem () is a simple operation,

 * it only does a dumped of memory

 */

void dump_mem (void){

  u8 *aux;

  int n;

  if (!all_ok) return;

  aux = (u8 *) first_bh;

  for (n = 0; n < total_size; n++) {

    PRINT_DBG_H (aux [n]);

    PRINT_DBG_C (" ");

  }

  PRINT_DBG_C ("\n");

}

/*

 * I haven't anything to say about print_block ()

 */

static void print_block (block_header_t *b){

  if (b == NULL) return;

  PRINT_DBG_C (">>>> MNum 0x");

  PRINT_DBG_H (b -> magic_number);

  PRINT_DBG_C (" Address 0x");

  //PRINT_DBG_H (b);

  if (b -> state == BLOCK_FREE)

    PRINT_DBG_C (" State FREE");

  else

    PRINT_DBG_C (" State USED");

  PRINT_DBG_C (" Size ");

  PRINT_DBG_D (b -> size);

  PRINT_DBG_C ("\n---- Prev 0x");

  //PRINT_DBG_H (b -> prev);

  PRINT_DBG_C (" Next 0x");

  //PRINT_DBG_H (b -> next);

  if (b -> state == BLOCK_FREE){

    PRINT_DBG_C ("\n---- Prev Free 0x");

    //PRINT_DBG_H (b -> ptr.free_ptr.prev);

    PRINT_DBG_C (" Next Free 0x");

    //PRINT_DBG_H (b -> ptr.free_ptr.next);

  }

  PRINT_DBG_C ("\n");

}

/*

 * structure_context () show the content of all blocks

 */

void structure_context (void) {

  block_header_t *b;

  if (!all_ok) return;

  b = first_bh;

  PRINT_DBG_C ("\nALL BLOCKS\n");

  while (b != NULL) {

    print_block (b);

    b = b -> next;

  }

}

/*

 * global_state () returns overhead, free and used space

 */

void global_state (int *free, int *used, int *overhead){

  block_header_t *b;

  *free = 0;

  *used = 0;

  *overhead = 0;

  if (!all_ok) return;

  b = first_bh;

  while (b != NULL) {

    if (b -> state == BLOCK_USED) {

      *used += b -> size;

    } else {

      *free += b -> size;

    }

    b = b -> next;

    *overhead += header_overhead;

  }

}

/*

 * free_blocks_context () show the content

 * of free blocks

 */

void free_blocks_context (void){

  int i, j;

  block_header_t *b;

  if (!all_ok) return;

  PRINT_DBG_C ("\nFREE BLOCKS\n");

  PRINT_DBG_C ("MAIN BLOCK\n");

  print_block (free_block);

  PRINT_DBG_C ("FRAGMENTED BLOCKS\n");

  for (i = MAX_FL_INDEX - 1 - MIN_LOG2_SIZE; i >= 0; i--) {

    if (fl_array [i].bitmapSL > 0)

      for (j = MAX_SL_INDEX - 1; j >= 0; j--) {

        if (fl_array [i].sl_array[j] != NULL) {

          b = fl_array [i].sl_array [j];

          PRINT_DBG_C ("[");

          PRINT_DBG_D (i);

          PRINT_DBG_C ("] ");

          PRINT_DBG_D (1 << (i + MIN_LOG2_SIZE));

          PRINT_DBG_C (" bytes -> Free blocks: 0x");