/*************************************************************************
* *
* Open Dynamics Engine, Copyright (C) 2001,2002 Russell L. Smith. *
* All rights reserved. Email: russ@q12.org Web: www.q12.org *
* *
* This library is free software; you can redistribute it and/or *
* modify it under the terms of EITHER: *
* (1) The GNU Lesser General Public License as published by the Free *
* Software Foundation; either version 2.1 of the License, or (at *
* your option) any later version. The text of the GNU Lesser *
* General Public License is included with this library in the *
* file LICENSE.TXT. *
* (2) The BSD-style license that is included with this library in *
* the file LICENSE-BSD.TXT. *
* *
* This library is distributed in the hope that it will be useful, *
* but WITHOUT ANY WARRANTY; without even the implied warranty of *
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the files *
* LICENSE.TXT and LICENSE-BSD.TXT for more details. *
* *
*************************************************************************/
/* this comes from the `reuse' library. copy any changes back to the source.
these stack allocation functions are a replacement for alloca(), except that
they allocate memory from a separate pool.
advantages over alloca():
- consecutive allocations are guaranteed to be contiguous with increasing
address.
- functions can allocate stack memory that is returned to the caller,
in other words pushing and popping stack frames is optional.
disadvantages compared to alloca():
- less portable
- slightly slower, although still orders of magnitude faster than malloc().
- longjmp() and exceptions do not deallocate stack memory (but who cares?).
just like alloca():
- using too much stack memory does not fail gracefully, it fails with a
segfault.
*/
#ifndef _ODE_STACK_H_
#define _ODE_STACK_H_
#ifdef WIN32
#include "windows.h"
#endif
struct dStack {
char *base; // bottom of the stack
int size; // maximum size of the stack
char *pointer; // current top of the stack
char *frame; // linked list of stack frame ptrs
# ifdef WIN32 // stuff for windows:
int pagesize; // - page size - this is ASSUMED to be a power of 2
int committed; // - bytes committed in allocated region
#endif
// initialize the stack. `max_size' is the maximum size that the stack can
// reach. on unix and windows a `virtual' memory block of this size is
// mapped into the address space but does not actually consume physical
// memory until it is referenced - so it is safe to set this to a high value.
void init (int max_size);
// destroy the stack. this unmaps any virtual memory that was allocated.
void destroy();
// allocate `size' bytes from the stack and return a pointer to the allocated
// memory. `size' must be >= 0. the returned pointer will be aligned to the
// size of a long int.
char * alloc (int size)
{
char *ret = pointer;
pointer += ((size-1) | (sizeof(long int)-1) )+1;
# ifdef WIN32
// for windows we need to commit pages as they are required
if ((pointer-base) > committed) {
committed = ((pointer-base-1) | (pagesize-1))+1; // round up to pgsize
VirtualAlloc (base,committed,MEM_COMMIT,PAGE_READWRITE);
}
# endif
return ret;
}
// return the address that will be returned by the next call to alloc()
char *nextAlloc()
{
return pointer;
}
// push and pop the current size of the stack. pushFrame() saves the current
// frame pointer on the stack, and popFrame() retrieves it. a typical
// stack-using function will bracket alloc() calls with pushFrame() and
// popFrame(). both functions return the current stack pointer - this should
// be the same value for the two bracketing calls. calling popFrame() too
// many times will result in a segfault.
char * pushFrame()
{
char *newframe = pointer;
char **addr = (char**) alloc (sizeof(char*));
*addr = frame;
frame = newframe;
return newframe;
/* OLD CODE
*((char**)pointer) = frame;
frame = pointer;
char *ret = pointer;
pointer += sizeof(char*);
return ret;
*/
}
char * popFrame()
{
pointer = frame;
frame = *((char**)pointer);
return pointer;
}
};
#endif
syntax highlighted by Code2HTML, v. 0.9.1