1
1
openmpi/ompi/class/ompi_bitmap.h

237 строки
8.3 KiB
C
Исходник Обычный вид История

/*
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
* University Research and Technology
* Corporation. All rights reserved.
* Copyright (c) 2004-2005 The University of Tennessee and The University
* of Tennessee Research Foundation. All rights
* reserved.
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
* University of Stuttgart. All rights reserved.
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
* $COPYRIGHT$
*
* Additional copyrights may follow
*
* $HEADER$
*
*/
/** @file
*
* A bitmap implementation. The bits start off with 0, so this bitmap
* has bits numbered as bit 0, bit 1, bit 2 and so on. This bitmap
* has auto-expansion capabilities, that is once the size is set
* during init, it can be automatically expanded by setting the bit
* beyond the current size. But note, this is allowed just when the
* bit is set -- so the valid functions are set_bit and
* find_and_set_bit. Other functions like clear, if passed a bit
* outside the initialized range will result in an error.
*
* Since these bitmaps are only used to track fortran handles (which
* MPI defines to be int's), it is assumed that we can never have more
* than OMPI_FORTRAN_HANDLE_MAX (which is min(INT_MAX, fortran
* INTEGER max)).
*
* --------------------------------------------------------------------
*
* There are several classes that have forked between their OMPI and
* ORTE implementations. As of this writing:
*
* - ompi_bitmap and orte_bitmap
* - ompi_pointer_array and orte_pointer_array
* - opal_value_array and orte_value_array
*
* Short version:
*
* They were split to accomodate a few differences between
* requirements (e.g., "size" parameters being int vs. size_t). It
* would be nice to re-merge them someday; there's a few technical
* issues that would need to be solved, but nothing impossible. But
* there's no pressing *need* to re-merge these, so they have fallen
* somewhat low on the priority list of things to do.
*
* Longer version:
*
* Although these are generic functionality classes, the ORTE versions
* split from the OMPI (soon to be OPAL) versions because of
* restrictions imposed by the OMPI versions. Specifically, the OMPI
* versions specifically limit size arguments to "int" (in multiple
* different ways, e.g.: types of "size" parameters to functions,
* maximum allowable values of size parameters, etc.). The ORTE
* functions need most size parameters to be of type size_t, not int.
* This is the most fundamental difference. In C++, we could have
* templated these functions, but we unfortunately can't easily do
* that in C (in hindsight, perhaps some preprocessor macros might
* have been sufficient, but...).
*
* Another, more subtle, reason why these were split because of the
* scary word "FORTRAN" that appears in some of the upper value limit
* checks in the OMPI versions. That is, we always check for max size
* against OMPI_FORTRAN_HANDLE_MAX. This value is simply min(INT_MAX,
* max value of Fortran INTEGER) -- it's the minimum of the maximum
* values of integers in C and Fortran. Usually, it's the same value
* (2^32), but the macro is there to ensure that even if it's
* different, we end up with a value that can be represented in both
* languages.
*
* This is because the primary purpose of these classes is to serve as
* an interface to the Fortran language bindings -- fortran handles
* may be directly represented as indices into arrays or bitmaps.
* Hence, the size has to be representable in both C and Fortran.
*
* Regardless, we need to check for *some* max value for the size of
* these entites. Perhaps the name "FORTRAN" in the macro is
* scary/misleading -- it can certainly be changed in the future if it
* would be more clear. So it's ok to have a max -- but perhaps
* changing the name would make it more palatable to both layers
* (remebering that the max value is still going to be enormous --
* usually 2^32; even if the size is of type size_t, you're going to
* run out of memory long before you have 2^32 entries). Or perhaps
* the max value can be parameterized to depend on whether the size
* type is int or size_t -- I'm sure something can be worked out.
*
* As mentioned above, if these really are the only two differences
* (int vs. size_t and the max sentinel value) -- and I'm pretty sure
* that they are -- then these classes can be re-merged someday,
* resulting in less code to maintain. This would be good for
* long-term maintenance. However, it's kinda low on the priority
* list. So it hasn't been done [yet]. If someone wants to do this,
* please feel free! :-)
*
* Given the fact that we just had yet another round of discussions
* about the splitting / re-merging of these classes (26 May 2005), it
* was decided to put this big comment in the hopes of:
*
* - someday motivating someone to re-merge the classes
* - prevent yet-another round of these discussions by documenting the
* issues once and for all :-)
*/
#ifndef OMPI_BITMAP_H
#define OMPI_BITMAP_H
#include "ompi_config.h"
#include <string.h>
#include "include/types.h"
#include "opal/class/opal_object.h"
#if defined(c_plusplus) || defined(__cplusplus)
extern "C" {
#endif
struct ompi_bitmap_t {
opal_object_t super; /**< Subclass of opal_object_t */
unsigned char *bitmap; /**< The actual bitmap array of characters */
int array_size; /**< The actual array size that maintains the bitmap */
int legal_numbits; /**< The number of bits which are legal (the
actual bitmap may contain more bits, since
it needs to be rounded to the nearest
char */
};
typedef struct ompi_bitmap_t ompi_bitmap_t;
OMPI_DECLSPEC OBJ_CLASS_DECLARATION(ompi_bitmap_t);
/**
* Initializes the bitmap and sets its size. This must be called
* before the bitmap can be actually used
*
* @param bitmap The input bitmap (IN)
* @param size The initial size of the bitmap in terms of bits (IN)
* @return OMPI error code or success
*
*/
OMPI_DECLSPEC int ompi_bitmap_init (ompi_bitmap_t *bm, int size);
/**
* Set a bit of the bitmap. If the bit asked for is beyond the current
* size of the bitmap, then the bitmap is extended to accomodate the
* bit
*
* @param bitmap The input bitmap (IN)
* @param bit The bit which is to be set (IN)
* @return OMPI error code or success
*
*/
OMPI_DECLSPEC int ompi_bitmap_set_bit(ompi_bitmap_t *bm, int bit);
/**
* Clear/unset a bit of the bitmap. If the bit is beyond the current
* size of the bitmap, an error is returned
*
* @param bitmap The input bitmap (IN)
* @param bit The bit which is to be cleared (IN)
* @return OMPI error code if the bit is out of range, else success
*
*/
OMPI_DECLSPEC int ompi_bitmap_clear_bit(ompi_bitmap_t *bm, int bit);
/**
* Find out if a bit is set in the bitmap
*
* @param bitmap The input bitmap (IN)
* @param bit The bit which is to be checked (IN)
* @return OMPI error code if the bit is out of range
* 1 if the bit is set
* 0 if the bit is not set
*
*/
OMPI_DECLSPEC int ompi_bitmap_is_set_bit(ompi_bitmap_t *bm, int bit);
/**
* Find the first clear bit in the bitmap and set it
*
* @param bitmap The input bitmap (IN)
* @param position Position of the first clear bit (OUT)
* @return err OMPI_SUCCESS on success
*/
OMPI_DECLSPEC int ompi_bitmap_find_and_set_first_unset_bit(ompi_bitmap_t *bm,
int *position);
/**
* Clear all bits in the bitmap
*
* @param bitmap The input bitmap (IN)
* @return OMPI error code if bm is NULL
*
*/
OMPI_DECLSPEC int ompi_bitmap_clear_all_bits(ompi_bitmap_t *bm);
/**
* Set all bits in the bitmap
* @param bitmap The input bitmap (IN)
* @return OMPI error code if bm is NULL
*
*/
OMPI_DECLSPEC int ompi_bitmap_set_all_bits(ompi_bitmap_t *bm);
/**
* Gives the current size (number of bits) in the bitmap. This is the
* legal (accessible) number of bits
*
* @param bitmap The input bitmap (IN)
* @return OMPI error code if bm is NULL
*
*/
static inline int ompi_bitmap_size(ompi_bitmap_t *bm)
{
return (NULL == bm) ? 0 : bm->legal_numbits;
}
#if defined(c_plusplus) || defined(__cplusplus)
}
#endif
#endif