Imagelibrary/rtems
1
0
Fork 1
mirror of https://gitlab.rtems.org/rtems/rtos/rtems.git synced 2025-12-27 06:58:19 +00:00
Code Issues Packages Projects Releases Wiki Activity

2009-09-25 Sebastian Huber <Sebastian.Huber@embedded-brains.de>

Browse Source 
* score/src/heap.c, score/include/rtems/score/heap.h: Reduced alignment
 	requirement for CPU_ALIGNMENT from four to two.
This commit is contained in:
Joel Sherrill
2009-09-25 17:49:32 +00:00
parent 0feb80855d
commit e89faf3ef4
3 changed files with 13 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
cpukit/ChangeLog
View File

@@ -1,3 +1,8 @@
2009-09-25 Sebastian Huber <Sebastian.Huber@embedded-brains.de>
* score/src/heap.c, score/include/rtems/score/heap.h: Reduced alignment
requirement for CPU_ALIGNMENT from four to two.
2009-09-25 Joel Sherrill <joel.sherrill@OARcorp.com>
* posix/include/rtems/posix/threadsup.h: Add no return atrribute to

16
cpukit/score/include/rtems/score/heap.h
View File

@@ -38,12 +38,10 @@ extern "C" {
* area. A heap control structure contains control information for the heap.
*
* The alignment routines could be made faster should we require only powers of
* two to be supported both for page size, alignment and boundary arguments.
* However, both workspace and malloc heaps are initialized with
* CPU_HEAP_ALIGNMENT as page size, and while all the BSPs seem to use
* CPU_ALIGNMENT (that is power of two) as CPU_HEAP_ALIGNMENT, for whatever
* reason CPU_HEAP_ALIGNMENT is only required to be multiple of CPU_ALIGNMENT
* and explicitly not required to be a power of two.
* two to be supported for page size, alignment and boundary arguments. The
* minimum alignment requirement for pages is currently CPU_ALIGNMENT and this
* value is only required to be multiple of two and explicitly not required to
* be a power of two.
*
* There are two kinds of blocks. One sort describes a free block from which
* we can allocate memory. The other blocks are used and provide an allocated
@@ -167,9 +165,9 @@ typedef struct Heap_Block {
* If the flag @c HEAP_PREV_BLOCK_USED is set, then the previous block is
* used, otherwise the previous block is free. A used previous block may
* claim the @a prev_size field for allocation. This trick allows to
* decrease the overhead in the used blocks by the size of the
* @a prev_size field. As sizes are always multiples of four, the two least
* significant bits are always zero. We use one of them to store the flag.
* decrease the overhead in the used blocks by the size of the @a prev_size
* field. As sizes are required to be multiples of two, the least
* significant bits would be always zero. We use this bit to store the flag.
*
* This field is always valid.
*/

3
cpukit/score/src/heap.c
View File

@@ -26,7 +26,7 @@
#include <rtems/system.h>
#include <rtems/score/heap.h>
#if CPU_ALIGNMENT == 0 || CPU_ALIGNMENT % 4 != 0
#if CPU_ALIGNMENT == 0 || CPU_ALIGNMENT % 2 != 0
#error "invalid CPU_ALIGNMENT value"
#endif
@@ -213,7 +213,6 @@ uintptr_t _Heap_Initialize(
stats->resizes = 0;
stats->instance = instance++;
_HAssert( _Heap_Is_aligned( CPU_ALIGNMENT, 4 ) );
_HAssert( _Heap_Is_aligned( heap->page_size, CPU_ALIGNMENT ) );
_HAssert( _Heap_Is_aligned( heap->min_block_size, page_size ) );
_HAssert(