#define _GNU_SOURCE
#include <stdlib.h>
#include <string.h>
#include <limits.h>
#include <stdint.h>
#include <errno.h>
#include <sys/mman.h>
#include "libc.h"
#include "atomic.h"
#include "pthread_impl.h"
#if defined(__GNUC__) && defined(__PIC__)
#define inline inline __attribute__((always_inline))
#endif
void *__mmap(void *, size_t, int, int, int, off_t);
int __munmap(void *, size_t);
void *__mremap(void *, size_t, size_t, int, ...);
int __madvise(void *, size_t, int);
struct chunk {
size_t psize, csize;
struct chunk *next, *prev;
};
struct bin {
volatile int lock[2];
struct chunk *head;
struct chunk *tail;
};
static struct {
volatile uint64_t binmap;
struct bin bins[64];
volatile int free_lock[2];
} mal;
#define SIZE_ALIGN (4*sizeof(size_t))
#define SIZE_MASK (-SIZE_ALIGN)
#define OVERHEAD (2*sizeof(size_t))
#define MMAP_THRESHOLD (0x1c00*SIZE_ALIGN)
#define DONTCARE 16
#define RECLAIM 163840
#define CHUNK_SIZE(c) ((c)->csize & -2)
#define CHUNK_PSIZE(c) ((c)->psize & -2)
#define PREV_CHUNK(c) ((struct chunk *)((char *)(c) - CHUNK_PSIZE(c)))
#define NEXT_CHUNK(c) ((struct chunk *)((char *)(c) + CHUNK_SIZE(c)))
#define MEM_TO_CHUNK(p) (struct chunk *)((char *)(p) - OVERHEAD)
#define CHUNK_TO_MEM(c) (void *)((char *)(c) + OVERHEAD)
#define BIN_TO_CHUNK(i) (MEM_TO_CHUNK(&mal.bins[i].head))
#define C_INUSE ((size_t)1)
#define IS_MMAPPED(c) !((c)->csize & (C_INUSE))
static inline void lock(volatile int *lk)
{
if (libc.threads_minus_1)
while(a_swap(lk, 1)) __wait(lk, lk+1, 1, 1);
}
static inline void unlock(volatile int *lk)
{
if (lk[0]) {
a_store(lk, 0);
if (lk[1]) __wake(lk, 1, 1);
}
}
static inline void lock_bin(int i)
{
lock(mal.bins[i].lock);
if (!mal.bins[i].head)
mal.bins[i].head = mal.bins[i].tail = BIN_TO_CHUNK(i);
}
static inline void unlock_bin(int i)
{
unlock(mal.bins[i].lock);
}
static int first_set(uint64_t x)
{
#if 1
return a_ctz_64(x);
#else#endif
}
static const unsigned char bin_tab[60] = {
32,33,34,35,36,36,37,37,38,38,39,39,
40,40,40,40,41,41,41,41,42,42,42,42,43,43,43,43,
44,44,44,44,44,44,44,44,45,45,45,45,45,45,45,45,
46,46,46,46,46,46,46,46,47,47,47,47,47,47,47,47,
};
static int bin_index(size_t x)
{
x = x / SIZE_ALIGN - 1;
if (x <= 32) return x;
if (x < 512) return bin_tab[x/8-4];
if (x > 0x1c00) return 63;
return bin_tab[x/128-4] + 16;
}
static int bin_index_up(size_t x)
{
x = x / SIZE_ALIGN - 1;
if (x <= 32) return x;
x--;
if (x < 512) return bin_tab[x/8-4] + 1;
return bin_tab[x/128-4] + 17;
}
#if 0#endif
void *__expand_heap(size_t *);
static struct chunk *expand_heap(size_t n)
{
static int heap_lock[2];
static void *end;
void *p;
struct chunk *w;
n += SIZE_ALIGN;
lock(heap_lock);
p = __expand_heap(&n);
if (!p) {
unlock(heap_lock);
return 0;
}
if (p != end) {
n -= SIZE_ALIGN;
p = (char *)p + SIZE_ALIGN;
w = MEM_TO_CHUNK(p);
w->psize = 0 | C_INUSE;
}
end = (char *)p + n;
w = MEM_TO_CHUNK(end);
w->psize = n | C_INUSE;
w->csize = 0 | C_INUSE;
w = MEM_TO_CHUNK(p);
w->csize = n | C_INUSE;
unlock(heap_lock);
return w;
}
static int adjust_size(size_t *n)
{
if (*n-1 > PTRDIFF_MAX - SIZE_ALIGN - PAGE_SIZE) {
if (*n) {
errno = ENOMEM;
return -1;
} else {
*n = SIZE_ALIGN;
return 0;
}
}
*n = (*n + OVERHEAD + SIZE_ALIGN - 1) & SIZE_MASK;
return 0;
}
static void unbin(struct chunk *c, int i)
{
if (c->prev == c->next)
a_and_64(&mal.binmap, ~(1ULL<<i));
c->prev->next = c->next;
c->next->prev = c->prev;
c->csize |= C_INUSE;
NEXT_CHUNK(c)->psize |= C_INUSE;
}
static int alloc_fwd(struct chunk *c)
{
int i;
size_t k;
while (!((k=c->csize) & C_INUSE)) {
i = bin_index(k);
lock_bin(i);
if (c->csize == k) {
unbin(c, i);
unlock_bin(i);
return 1;
}
unlock_bin(i);
}
return 0;
}
static int alloc_rev(struct chunk *c)
{
int i;
size_t k;
while (!((k=c->psize) & C_INUSE)) {
i = bin_index(k);
lock_bin(i);
if (c->psize == k) {
unbin(PREV_CHUNK(c), i);
unlock_bin(i);
return 1;
}
unlock_bin(i);
}
return 0;
}
static int pretrim(struct chunk *self, size_t n, int i, int j)
{
size_t n1;
struct chunk *next, *split;
if (j < 40) return 0;
if (j < i+3) {
if (j != 63) return 0;
n1 = CHUNK_SIZE(self);
if (n1-n <= MMAP_THRESHOLD) return 0;
} else {
n1 = CHUNK_SIZE(self);
}
if (bin_index(n1-n) != j) return 0;
next = NEXT_CHUNK(self);
split = (void *)((char *)self + n);
split->prev = self->prev;
split->next = self->next;
split->prev->next = split;
split->next->prev = split;
split->psize = n | C_INUSE;
split->csize = n1-n;
next->psize = n1-n;
self->csize = n | C_INUSE;
return 1;
}
static void trim(struct chunk *self, size_t n)
{
size_t n1 = CHUNK_SIZE(self);
struct chunk *next, *split;
if (n >= n1 - DONTCARE) return;
next = NEXT_CHUNK(self);
split = (void *)((char *)self + n);
split->psize = n | C_INUSE;
split->csize = n1-n | C_INUSE;
next->psize = n1-n | C_INUSE;
self->csize = n | C_INUSE;
free(CHUNK_TO_MEM(split));
}
void *malloc(size_t n)
{
struct chunk *c;
int i, j;
if (adjust_size(&n) < 0) return 0;
if (n > MMAP_THRESHOLD) {
size_t len = n + OVERHEAD + PAGE_SIZE - 1 & -PAGE_SIZE;
char *base = __mmap(0, len, PROT_READ|PROT_WRITE,
MAP_PRIVATE|MAP_ANONYMOUS, -1, 0);
if (base == (void *)-1) return 0;
c = (void *)(base + SIZE_ALIGN - OVERHEAD);
c->csize = len - (SIZE_ALIGN - OVERHEAD);
c->psize = SIZE_ALIGN - OVERHEAD;
return CHUNK_TO_MEM(c);
}
i = bin_index_up(n);
for (;;) {
uint64_t mask = mal.binmap & -(1ULL<<i);
if (!mask) {
c = expand_heap(n);
if (!c) return 0;
if (alloc_rev(c)) {
struct chunk *x = c;
c = PREV_CHUNK(c);
NEXT_CHUNK(x)->psize = c->csize =
x->csize + CHUNK_SIZE(c);
}
break;
}
j = first_set(mask);
lock_bin(j);
c = mal.bins[j].head;
if (c != BIN_TO_CHUNK(j)) {
if (!pretrim(c, n, i, j)) unbin(c, j);
unlock_bin(j);
break;
}
unlock_bin(j);
}
trim(c, n);
return CHUNK_TO_MEM(c);
}
void *__malloc0(size_t n)
{
void *p = malloc(n);
if (p && !IS_MMAPPED(MEM_TO_CHUNK(p))) {
size_t *z;
n = (n + sizeof *z - 1)/sizeof *z;
for (z=p; n; n--, z++) if (*z) *z=0;
}
return p;
}
void *realloc(void *p, size_t n)
{
struct chunk *self, *next;
size_t n0, n1;
void *new;
if (!p) return malloc(n);
if (adjust_size(&n) < 0) return 0;
self = MEM_TO_CHUNK(p);
n1 = n0 = CHUNK_SIZE(self);
if (IS_MMAPPED(self)) {
size_t extra = self->psize;
char *base = (char *)self - extra;
size_t oldlen = n0 + extra;
size_t newlen = n + extra;
if (extra & 1) a_crash();
if (newlen < PAGE_SIZE && (new = malloc(n))) {
memcpy(new, p, n-OVERHEAD);
free(p);
return new;
}
newlen = (newlen + PAGE_SIZE-1) & -PAGE_SIZE;
if (oldlen == newlen) return p;
base = __mremap(base, oldlen, newlen, MREMAP_MAYMOVE);
if (base == (void *)-1)
goto copy_realloc;
self = (void *)(base + extra);
self->csize = newlen - extra;
return CHUNK_TO_MEM(self);
}
next = NEXT_CHUNK(self);
if (next->psize != self->csize) a_crash();
if (n > n1 && alloc_fwd(next)) {
n1 += CHUNK_SIZE(next);
next = NEXT_CHUNK(next);
}
if (0 && n > n1 && alloc_rev(self)) {
self = PREV_CHUNK(self);
n1 += CHUNK_SIZE(self);
}
self->csize = n1 | C_INUSE;
next->psize = n1 | C_INUSE;
if (n <= n1) {
trim(self, n);
return CHUNK_TO_MEM(self);
}
copy_realloc:
new = malloc(n-OVERHEAD);
if (!new) return 0;
memcpy(new, p, n0-OVERHEAD);
free(CHUNK_TO_MEM(self));
return new;
}
void free(void *p)
{
struct chunk *self = MEM_TO_CHUNK(p);
struct chunk *next;
size_t final_size, new_size, size;
int reclaim=0;
int i;
if (!p) return;
if (IS_MMAPPED(self)) {
size_t extra = self->psize;
char *base = (char *)self - extra;
size_t len = CHUNK_SIZE(self) + extra;
if (extra & 1) a_crash();
__munmap(base, len);
return;
}
final_size = new_size = CHUNK_SIZE(self);
next = NEXT_CHUNK(self);
if (next->psize != self->csize) a_crash();
for (;;) {
if (self->psize & next->csize & C_INUSE) {
self->csize = final_size | C_INUSE;
next->psize = final_size | C_INUSE;
i = bin_index(final_size);
lock_bin(i);
lock(mal.free_lock);
if (self->psize & next->csize & C_INUSE)
break;
unlock(mal.free_lock);
unlock_bin(i);
}
if (alloc_rev(self)) {
self = PREV_CHUNK(self);
size = CHUNK_SIZE(self);
final_size += size;
if (new_size+size > RECLAIM && (new_size+size^size) > size)
reclaim = 1;
}
if (alloc_fwd(next)) {
size = CHUNK_SIZE(next);
final_size += size;
if (new_size+size > RECLAIM && (new_size+size^size) > size)
reclaim = 1;
next = NEXT_CHUNK(next);
}
}
if (!(mal.binmap & 1ULL<<i))
a_or_64(&mal.binmap, 1ULL<<i);
self->csize = final_size;
next->psize = final_size;
unlock(mal.free_lock);
self->next = BIN_TO_CHUNK(i);
self->prev = mal.bins[i].tail;
self->next->prev = self;
self->prev->next = self;
if (reclaim) {
uintptr_t a = (uintptr_t)self + SIZE_ALIGN+PAGE_SIZE-1 & -PAGE_SIZE;
uintptr_t b = (uintptr_t)next - SIZE_ALIGN & -PAGE_SIZE;
#if 1
__madvise((void *)a, b-a, MADV_DONTNEED);
#else#endif
}
unlock_bin(i);
}