Re: malloc

ramu wrote:
santosh wrote:
ramu wrote:
Hi,
Will the memory allocated by malloc and realloc be contiguous?

regards

Do you mean across seperate calls? If so, then no. The standard makes
no such guarantee. If you want such a feature, (why?), then you'll have
to roll your own allocator.

If I use both the calls ie. malloc and realloc together, how the allocated memory will be
ie. contiguous or not?

Please fix your quoting. You're quoting your own text.

You need to be more specific. As I mentioned earlier, for any single
allocation of malloc() or calloc() or realloc(), the bytes of the
allocated memory can be regarded as contiguous. But this cant be
assumed for seperate blocks of allocated memory, unless they were
originally allocated as a single block and later on divided up by
pointers into logically seperate blocks.

1.
int *p = malloc(100 * sizeof *p);
Here the 100 int's pointed to by p can be regarded as contiguous. i.e.
pointer arithmetic will work as expected.

2.
double *d1 = malloc(100 * sizeof *d1);
double *d2 = realloc(d2, 100 * sizeof *d2);
Here the elements of d1 have no necessary relationship with elements of
d2, i.e. pointer arithmetic will not work _across_ them, though within
each block, elements can be regarded as contiguous.

3.
unsigned char *p = malloc(100);
p = realloc(p, 200);
Again, an old pointer value, valid for the first allocation of p, will
not be valid when p has been reallocated, (unless of course, realloc()
fails).

4.
long *b = malloc(1000 * sizeof *b);
long *b1 = b, *b2 = b+250, *b3 = b2+500;
Here, b1, b2 and b3 point to logically seperate blocks of memory, but
since you've partitioned them from a single call to malloc(), they can
all be regarded as contiguous with each other, in the proper order, _as
long as_ you don't reallocate or move any of the blocks.

HTH

.

Relevant Pages

  • Re: resizing an array, is there a better way?
    ... >> a large virtual memory space such that they can be trivially extended. ... > when realloc() is called with a size of 13; ... > allocation is done just after the existing block, ...
    (comp.lang.c)
  • Re: realloc, copy and VM
    ... > When you realloc for more size, it may be necessary to allocate a new block, ... > copy memory and deallocate the old, for example if there is not enough free ... That *is* extending the original block. ... via lazy allocation and copy-on-write.) ...
    (comp.lang.c)
  • Re: malloc
    ... If I use both the calls ie. malloc and realloc together, how the allocated memory will be ... Again, an old pointer value, valid for the first allocation of p, will ...
    (comp.lang.c)
  • Re: when can realloc fail?
    ... will the allocation of ptr be freed? ... about what the Standard requires/allows. ... memory pointed to by p, basing the claim on the deallocation ... Don't call realloc() with the second ...
    (comp.lang.c)
  • Re: appending elements to an ALLOCATABLE array
    ... > That's not a realloc. ... this) of addressing more than one pre-allocated memory locations using ... > As a proposed solution, I don't think what he asked for works. ... > For pointer allocation, you only need one copy today. ...
    (comp.lang.fortran)