nasm/nasmlib/raa.c
H. Peter Anvin 903ea54857 RAA: clean up the RAA infrastructure, make it support larger indicies
Make the RAA infrastructure a bit cleaner, make it support 64-bit
indicies, and reduce the memory overhead of a sparse or small RAA --
the old code would allocate a *minimum* of 256K for each RAA.  The new
code reduces that to 16K, and will not mandatorily allocate an entry
in the zero position.

The new shift, 11, was chosen so that a 32-bit RAA value will need 3
layers and a 64-bit value 6 layers, without excessive waste.

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2018-06-14 20:23:49 -07:00

204 lines
5.6 KiB
C

/* ----------------------------------------------------------------------- *
*
* Copyright 1996-2018 The NASM Authors - All Rights Reserved
* See the file AUTHORS included with the NASM distribution for
* the specific copyright holders.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following
* conditions are met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* * Redistributions in binary form must reproduce the above
* copyright notice, this list of conditions and the following
* disclaimer in the documentation and/or other materials provided
* with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND
* CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES,
* INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
* DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
* CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR
* OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
* EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* ----------------------------------------------------------------------- */
#include "nasmlib.h"
#include "raa.h"
#include "ilog2.h"
/*
* Routines to manage a dynamic random access array of int64_ts which
* may grow in size to be more than the largest single malloc'able
* chunk.
*/
#define RAA_LAYERSHIFT 11 /* 2^this many items per layer */
#define RAA_LAYERSIZE ((size_t)1 << RAA_LAYERSHIFT)
#define RAA_LAYERMASK (RAA_LAYERSIZE-1)
typedef struct RAA RAA;
typedef union RAA_UNION RAA_UNION;
typedef struct RAA_LEAF RAA_LEAF;
typedef struct RAA_BRANCH RAA_BRANCH;
union intorptr {
int64_t i;
void *p;
};
struct RAA {
/* Last position in this RAA */
raaindex endposn;
/*
* Number of layers below this one to get to the real data. 0
* means this structure is a leaf, holding RAA_LAYERSIZE real
* data items; 1 and above mean it's a branch, holding
* RAA_LAYERSIZE pointers to the next level branch or leaf
* structures.
*/
unsigned int layers;
/*
* Number of real data items spanned by one position in the
* `data' array at this level. This number is 0 trivially, for
* a leaf (level 0): for a level n branch it should be
* n*RAA_LAYERSHIFT.
*/
unsigned int shift;
/*
* The actual data
*/
union RAA_UNION {
struct RAA_LEAF {
union intorptr data[RAA_LAYERSIZE];
} l;
struct RAA_BRANCH {
struct RAA *data[RAA_LAYERSIZE];
} b;
} u;
};
#define LEAFSIZ (sizeof(RAA)-sizeof(RAA_UNION)+sizeof(RAA_LEAF))
#define BRANCHSIZ (sizeof(RAA)-sizeof(RAA_UNION)+sizeof(RAA_BRANCH))
static struct RAA *raa_init_layer(raaindex posn, unsigned int layers)
{
struct RAA *r;
raaindex posmask;
r = nasm_zalloc((layers == 0) ? LEAFSIZ : BRANCHSIZ);
r->shift = layers * RAA_LAYERSHIFT;
r->layers = layers;
posmask = ((raaindex)RAA_LAYERSIZE << r->shift) - 1;
r->endposn = posn | posmask;
return r;
}
void raa_free(struct RAA *r)
{
if (!r)
return;
if (r->layers) {
struct RAA **p = r->u.b.data;
size_t i;
for (i = 0; i < RAA_LAYERSIZE; i++)
raa_free(*p++);
}
nasm_free(r);
}
static const union intorptr *real_raa_read(struct RAA *r, raaindex posn)
{
nasm_assert(posn <= (~(raaindex)0 >> 1));
if (unlikely(!r || posn > r->endposn))
return NULL; /* Beyond the end */
while (r->layers) {
size_t l = (posn >> r->shift) & RAA_LAYERMASK;
r = r->u.b.data[l];
if (!r)
return NULL; /* Not present */
}
return &r->u.l.data[posn & RAA_LAYERMASK];
}
int64_t raa_read(struct RAA *r, raaindex pos)
{
const union intorptr *ip;
ip = real_raa_read(r, pos);
return ip ? ip->i : 0;
}
void *raa_read_ptr(struct RAA *r, raaindex pos)
{
const union intorptr *ip;
ip = real_raa_read(r, pos);
return ip ? ip->p : NULL;
}
static struct RAA *
real_raa_write(struct RAA *r, raaindex posn, union intorptr value)
{
struct RAA *result;
nasm_assert(posn <= (~(raaindex)0 >> 1));
if (unlikely(!r)) {
/* Create a new top-level RAA */
r = raa_init_layer(posn, ilog2_64(posn)/RAA_LAYERSHIFT);
} else {
while (unlikely(r->endposn < posn)) {
/* We need to add layers to an existing RAA */
struct RAA *s = raa_init_layer(r->endposn, r->layers + 1);
s->u.b.data[0] = r;
r = s;
}
}
result = r;
while (r->layers) {
struct RAA **s;
size_t l = (posn >> r->shift) & RAA_LAYERMASK;
s = &r->u.b.data[l];
if (unlikely(!*s))
*s = raa_init_layer(posn, r->layers - 1);
r = *s;
}
r->u.l.data[posn & RAA_LAYERMASK] = value;
return result;
}
struct RAA *raa_write(struct RAA *r, raaindex posn, int64_t value)
{
union intorptr ip;
ip.i = value;
return real_raa_write(r, posn, ip);
}
struct RAA *raa_write_ptr(struct RAA *r, raaindex posn, void *value)
{
union intorptr ip;
ip.p = value;
return real_raa_write(r, posn, ip);
}