// Copyright 2009 The Go Authors. All rights reserved.
|
// Copyright 2009 The Go Authors. All rights reserved.
|
// Use of this source code is governed by a BSD-style
|
// Use of this source code is governed by a BSD-style
|
// license that can be found in the LICENSE file.
|
// license that can be found in the LICENSE file.
|
|
|
// Garbage collector.
|
// Garbage collector.
|
|
|
#include "runtime.h"
|
#include "runtime.h"
|
#include "arch.h"
|
#include "arch.h"
|
#include "malloc.h"
|
#include "malloc.h"
|
|
|
#ifdef USING_SPLIT_STACK
|
#ifdef USING_SPLIT_STACK
|
|
|
extern void * __splitstack_find (void *, void *, size_t *, void **, void **,
|
extern void * __splitstack_find (void *, void *, size_t *, void **, void **,
|
void **);
|
void **);
|
|
|
extern void * __splitstack_find_context (void *context[10], size_t *, void **,
|
extern void * __splitstack_find_context (void *context[10], size_t *, void **,
|
void **, void **);
|
void **, void **);
|
|
|
#endif
|
#endif
|
|
|
enum {
|
enum {
|
Debug = 0,
|
Debug = 0,
|
PtrSize = sizeof(void*),
|
PtrSize = sizeof(void*),
|
DebugMark = 0, // run second pass to check mark
|
DebugMark = 0, // run second pass to check mark
|
|
|
// Four bits per word (see #defines below).
|
// Four bits per word (see #defines below).
|
wordsPerBitmapWord = sizeof(void*)*8/4,
|
wordsPerBitmapWord = sizeof(void*)*8/4,
|
bitShift = sizeof(void*)*8/4,
|
bitShift = sizeof(void*)*8/4,
|
};
|
};
|
|
|
// Bits in per-word bitmap.
|
// Bits in per-word bitmap.
|
// #defines because enum might not be able to hold the values.
|
// #defines because enum might not be able to hold the values.
|
//
|
//
|
// Each word in the bitmap describes wordsPerBitmapWord words
|
// Each word in the bitmap describes wordsPerBitmapWord words
|
// of heap memory. There are 4 bitmap bits dedicated to each heap word,
|
// of heap memory. There are 4 bitmap bits dedicated to each heap word,
|
// so on a 64-bit system there is one bitmap word per 16 heap words.
|
// so on a 64-bit system there is one bitmap word per 16 heap words.
|
// The bits in the word are packed together by type first, then by
|
// The bits in the word are packed together by type first, then by
|
// heap location, so each 64-bit bitmap word consists of, from top to bottom,
|
// heap location, so each 64-bit bitmap word consists of, from top to bottom,
|
// the 16 bitSpecial bits for the corresponding heap words, then the 16 bitMarked bits,
|
// the 16 bitSpecial bits for the corresponding heap words, then the 16 bitMarked bits,
|
// then the 16 bitNoPointers/bitBlockBoundary bits, then the 16 bitAllocated bits.
|
// then the 16 bitNoPointers/bitBlockBoundary bits, then the 16 bitAllocated bits.
|
// This layout makes it easier to iterate over the bits of a given type.
|
// This layout makes it easier to iterate over the bits of a given type.
|
//
|
//
|
// The bitmap starts at mheap.arena_start and extends *backward* from
|
// The bitmap starts at mheap.arena_start and extends *backward* from
|
// there. On a 64-bit system the off'th word in the arena is tracked by
|
// there. On a 64-bit system the off'th word in the arena is tracked by
|
// the off/16+1'th word before mheap.arena_start. (On a 32-bit system,
|
// the off/16+1'th word before mheap.arena_start. (On a 32-bit system,
|
// the only difference is that the divisor is 8.)
|
// the only difference is that the divisor is 8.)
|
//
|
//
|
// To pull out the bits corresponding to a given pointer p, we use:
|
// To pull out the bits corresponding to a given pointer p, we use:
|
//
|
//
|
// off = p - (uintptr*)mheap.arena_start; // word offset
|
// off = p - (uintptr*)mheap.arena_start; // word offset
|
// b = (uintptr*)mheap.arena_start - off/wordsPerBitmapWord - 1;
|
// b = (uintptr*)mheap.arena_start - off/wordsPerBitmapWord - 1;
|
// shift = off % wordsPerBitmapWord
|
// shift = off % wordsPerBitmapWord
|
// bits = *b >> shift;
|
// bits = *b >> shift;
|
// /* then test bits & bitAllocated, bits & bitMarked, etc. */
|
// /* then test bits & bitAllocated, bits & bitMarked, etc. */
|
//
|
//
|
#define bitAllocated ((uintptr)1<<(bitShift*0))
|
#define bitAllocated ((uintptr)1<<(bitShift*0))
|
#define bitNoPointers ((uintptr)1<<(bitShift*1)) /* when bitAllocated is set */
|
#define bitNoPointers ((uintptr)1<<(bitShift*1)) /* when bitAllocated is set */
|
#define bitMarked ((uintptr)1<<(bitShift*2)) /* when bitAllocated is set */
|
#define bitMarked ((uintptr)1<<(bitShift*2)) /* when bitAllocated is set */
|
#define bitSpecial ((uintptr)1<<(bitShift*3)) /* when bitAllocated is set - has finalizer or being profiled */
|
#define bitSpecial ((uintptr)1<<(bitShift*3)) /* when bitAllocated is set - has finalizer or being profiled */
|
#define bitBlockBoundary ((uintptr)1<<(bitShift*1)) /* when bitAllocated is NOT set */
|
#define bitBlockBoundary ((uintptr)1<<(bitShift*1)) /* when bitAllocated is NOT set */
|
|
|
#define bitMask (bitBlockBoundary | bitAllocated | bitMarked | bitSpecial)
|
#define bitMask (bitBlockBoundary | bitAllocated | bitMarked | bitSpecial)
|
|
|
// TODO: Make these per-M.
|
// TODO: Make these per-M.
|
static uint64 nhandoff;
|
static uint64 nhandoff;
|
|
|
static int32 gctrace;
|
static int32 gctrace;
|
|
|
typedef struct Workbuf Workbuf;
|
typedef struct Workbuf Workbuf;
|
struct Workbuf
|
struct Workbuf
|
{
|
{
|
Workbuf *next;
|
Workbuf *next;
|
uintptr nobj;
|
uintptr nobj;
|
byte *obj[512-2];
|
byte *obj[512-2];
|
};
|
};
|
|
|
typedef struct Finalizer Finalizer;
|
typedef struct Finalizer Finalizer;
|
struct Finalizer
|
struct Finalizer
|
{
|
{
|
void (*fn)(void*);
|
void (*fn)(void*);
|
void *arg;
|
void *arg;
|
const struct __go_func_type *ft;
|
const struct __go_func_type *ft;
|
};
|
};
|
|
|
typedef struct FinBlock FinBlock;
|
typedef struct FinBlock FinBlock;
|
struct FinBlock
|
struct FinBlock
|
{
|
{
|
FinBlock *alllink;
|
FinBlock *alllink;
|
FinBlock *next;
|
FinBlock *next;
|
int32 cnt;
|
int32 cnt;
|
int32 cap;
|
int32 cap;
|
Finalizer fin[1];
|
Finalizer fin[1];
|
};
|
};
|
|
|
|
|
static G *fing;
|
static G *fing;
|
static FinBlock *finq; // list of finalizers that are to be executed
|
static FinBlock *finq; // list of finalizers that are to be executed
|
static FinBlock *finc; // cache of free blocks
|
static FinBlock *finc; // cache of free blocks
|
static FinBlock *allfin; // list of all blocks
|
static FinBlock *allfin; // list of all blocks
|
static Lock finlock;
|
static Lock finlock;
|
static int32 fingwait;
|
static int32 fingwait;
|
|
|
static void runfinq(void*);
|
static void runfinq(void*);
|
static Workbuf* getempty(Workbuf*);
|
static Workbuf* getempty(Workbuf*);
|
static Workbuf* getfull(Workbuf*);
|
static Workbuf* getfull(Workbuf*);
|
static void putempty(Workbuf*);
|
static void putempty(Workbuf*);
|
static Workbuf* handoff(Workbuf*);
|
static Workbuf* handoff(Workbuf*);
|
|
|
static struct {
|
static struct {
|
Lock fmu;
|
Lock fmu;
|
Workbuf *full;
|
Workbuf *full;
|
Lock emu;
|
Lock emu;
|
Workbuf *empty;
|
Workbuf *empty;
|
uint32 nproc;
|
uint32 nproc;
|
volatile uint32 nwait;
|
volatile uint32 nwait;
|
volatile uint32 ndone;
|
volatile uint32 ndone;
|
Note alldone;
|
Note alldone;
|
Lock markgate;
|
Lock markgate;
|
Lock sweepgate;
|
Lock sweepgate;
|
MSpan *spans;
|
MSpan *spans;
|
|
|
Lock;
|
Lock;
|
byte *chunk;
|
byte *chunk;
|
uintptr nchunk;
|
uintptr nchunk;
|
} work;
|
} work;
|
|
|
// scanblock scans a block of n bytes starting at pointer b for references
|
// scanblock scans a block of n bytes starting at pointer b for references
|
// to other objects, scanning any it finds recursively until there are no
|
// to other objects, scanning any it finds recursively until there are no
|
// unscanned objects left. Instead of using an explicit recursion, it keeps
|
// unscanned objects left. Instead of using an explicit recursion, it keeps
|
// a work list in the Workbuf* structures and loops in the main function
|
// a work list in the Workbuf* structures and loops in the main function
|
// body. Keeping an explicit work list is easier on the stack allocator and
|
// body. Keeping an explicit work list is easier on the stack allocator and
|
// more efficient.
|
// more efficient.
|
static void
|
static void
|
scanblock(byte *b, int64 n)
|
scanblock(byte *b, int64 n)
|
{
|
{
|
byte *obj, *arena_start, *arena_used, *p;
|
byte *obj, *arena_start, *arena_used, *p;
|
void **vp;
|
void **vp;
|
uintptr size, *bitp, bits, shift, i, j, x, xbits, off, nobj, nproc;
|
uintptr size, *bitp, bits, shift, i, j, x, xbits, off, nobj, nproc;
|
MSpan *s;
|
MSpan *s;
|
PageID k;
|
PageID k;
|
void **wp;
|
void **wp;
|
Workbuf *wbuf;
|
Workbuf *wbuf;
|
bool keepworking;
|
bool keepworking;
|
|
|
if((int64)(uintptr)n != n || n < 0) {
|
if((int64)(uintptr)n != n || n < 0) {
|
// runtime_printf("scanblock %p %lld\n", b, (long long)n);
|
// runtime_printf("scanblock %p %lld\n", b, (long long)n);
|
runtime_throw("scanblock");
|
runtime_throw("scanblock");
|
}
|
}
|
|
|
// Memory arena parameters.
|
// Memory arena parameters.
|
arena_start = runtime_mheap.arena_start;
|
arena_start = runtime_mheap.arena_start;
|
arena_used = runtime_mheap.arena_used;
|
arena_used = runtime_mheap.arena_used;
|
nproc = work.nproc;
|
nproc = work.nproc;
|
|
|
wbuf = nil; // current work buffer
|
wbuf = nil; // current work buffer
|
wp = nil; // storage for next queued pointer (write pointer)
|
wp = nil; // storage for next queued pointer (write pointer)
|
nobj = 0; // number of queued objects
|
nobj = 0; // number of queued objects
|
|
|
// Scanblock helpers pass b==nil.
|
// Scanblock helpers pass b==nil.
|
// The main proc needs to return to make more
|
// The main proc needs to return to make more
|
// calls to scanblock. But if work.nproc==1 then
|
// calls to scanblock. But if work.nproc==1 then
|
// might as well process blocks as soon as we
|
// might as well process blocks as soon as we
|
// have them.
|
// have them.
|
keepworking = b == nil || work.nproc == 1;
|
keepworking = b == nil || work.nproc == 1;
|
|
|
// Align b to a word boundary.
|
// Align b to a word boundary.
|
off = (uintptr)b & (PtrSize-1);
|
off = (uintptr)b & (PtrSize-1);
|
if(off != 0) {
|
if(off != 0) {
|
b += PtrSize - off;
|
b += PtrSize - off;
|
n -= PtrSize - off;
|
n -= PtrSize - off;
|
}
|
}
|
|
|
for(;;) {
|
for(;;) {
|
// Each iteration scans the block b of length n, queueing pointers in
|
// Each iteration scans the block b of length n, queueing pointers in
|
// the work buffer.
|
// the work buffer.
|
if(Debug > 1)
|
if(Debug > 1)
|
runtime_printf("scanblock %p %lld\n", b, (long long) n);
|
runtime_printf("scanblock %p %lld\n", b, (long long) n);
|
|
|
vp = (void**)b;
|
vp = (void**)b;
|
n >>= (2+PtrSize/8); /* n /= PtrSize (4 or 8) */
|
n >>= (2+PtrSize/8); /* n /= PtrSize (4 or 8) */
|
for(i=0; i<(uintptr)n; i++) {
|
for(i=0; i<(uintptr)n; i++) {
|
obj = (byte*)vp[i];
|
obj = (byte*)vp[i];
|
|
|
// Words outside the arena cannot be pointers.
|
// Words outside the arena cannot be pointers.
|
if((byte*)obj < arena_start || (byte*)obj >= arena_used)
|
if((byte*)obj < arena_start || (byte*)obj >= arena_used)
|
continue;
|
continue;
|
|
|
// obj may be a pointer to a live object.
|
// obj may be a pointer to a live object.
|
// Try to find the beginning of the object.
|
// Try to find the beginning of the object.
|
|
|
// Round down to word boundary.
|
// Round down to word boundary.
|
obj = (void*)((uintptr)obj & ~((uintptr)PtrSize-1));
|
obj = (void*)((uintptr)obj & ~((uintptr)PtrSize-1));
|
|
|
// Find bits for this word.
|
// Find bits for this word.
|
off = (uintptr*)obj - (uintptr*)arena_start;
|
off = (uintptr*)obj - (uintptr*)arena_start;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
xbits = *bitp;
|
xbits = *bitp;
|
bits = xbits >> shift;
|
bits = xbits >> shift;
|
|
|
// Pointing at the beginning of a block?
|
// Pointing at the beginning of a block?
|
if((bits & (bitAllocated|bitBlockBoundary)) != 0)
|
if((bits & (bitAllocated|bitBlockBoundary)) != 0)
|
goto found;
|
goto found;
|
|
|
// Pointing just past the beginning?
|
// Pointing just past the beginning?
|
// Scan backward a little to find a block boundary.
|
// Scan backward a little to find a block boundary.
|
for(j=shift; j-->0; ) {
|
for(j=shift; j-->0; ) {
|
if(((xbits>>j) & (bitAllocated|bitBlockBoundary)) != 0) {
|
if(((xbits>>j) & (bitAllocated|bitBlockBoundary)) != 0) {
|
obj = (byte*)obj - (shift-j)*PtrSize;
|
obj = (byte*)obj - (shift-j)*PtrSize;
|
shift = j;
|
shift = j;
|
bits = xbits>>shift;
|
bits = xbits>>shift;
|
goto found;
|
goto found;
|
}
|
}
|
}
|
}
|
|
|
// Otherwise consult span table to find beginning.
|
// Otherwise consult span table to find beginning.
|
// (Manually inlined copy of MHeap_LookupMaybe.)
|
// (Manually inlined copy of MHeap_LookupMaybe.)
|
k = (uintptr)obj>>PageShift;
|
k = (uintptr)obj>>PageShift;
|
x = k;
|
x = k;
|
if(sizeof(void*) == 8)
|
if(sizeof(void*) == 8)
|
x -= (uintptr)arena_start>>PageShift;
|
x -= (uintptr)arena_start>>PageShift;
|
s = runtime_mheap.map[x];
|
s = runtime_mheap.map[x];
|
if(s == nil || k < s->start || k - s->start >= s->npages || s->state != MSpanInUse)
|
if(s == nil || k < s->start || k - s->start >= s->npages || s->state != MSpanInUse)
|
continue;
|
continue;
|
p = (byte*)((uintptr)s->start<<PageShift);
|
p = (byte*)((uintptr)s->start<<PageShift);
|
if(s->sizeclass == 0) {
|
if(s->sizeclass == 0) {
|
obj = p;
|
obj = p;
|
} else {
|
} else {
|
if((byte*)obj >= (byte*)s->limit)
|
if((byte*)obj >= (byte*)s->limit)
|
continue;
|
continue;
|
size = runtime_class_to_size[s->sizeclass];
|
size = runtime_class_to_size[s->sizeclass];
|
int32 i = ((byte*)obj - p)/size;
|
int32 i = ((byte*)obj - p)/size;
|
obj = p+i*size;
|
obj = p+i*size;
|
}
|
}
|
|
|
// Now that we know the object header, reload bits.
|
// Now that we know the object header, reload bits.
|
off = (uintptr*)obj - (uintptr*)arena_start;
|
off = (uintptr*)obj - (uintptr*)arena_start;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
xbits = *bitp;
|
xbits = *bitp;
|
bits = xbits >> shift;
|
bits = xbits >> shift;
|
|
|
found:
|
found:
|
// Now we have bits, bitp, and shift correct for
|
// Now we have bits, bitp, and shift correct for
|
// obj pointing at the base of the object.
|
// obj pointing at the base of the object.
|
// Only care about allocated and not marked.
|
// Only care about allocated and not marked.
|
if((bits & (bitAllocated|bitMarked)) != bitAllocated)
|
if((bits & (bitAllocated|bitMarked)) != bitAllocated)
|
continue;
|
continue;
|
if(nproc == 1)
|
if(nproc == 1)
|
*bitp |= bitMarked<<shift;
|
*bitp |= bitMarked<<shift;
|
else {
|
else {
|
for(;;) {
|
for(;;) {
|
x = *bitp;
|
x = *bitp;
|
if(x & (bitMarked<<shift))
|
if(x & (bitMarked<<shift))
|
goto continue_obj;
|
goto continue_obj;
|
if(runtime_casp((void**)bitp, (void*)x, (void*)(x|(bitMarked<<shift))))
|
if(runtime_casp((void**)bitp, (void*)x, (void*)(x|(bitMarked<<shift))))
|
break;
|
break;
|
}
|
}
|
}
|
}
|
|
|
// If object has no pointers, don't need to scan further.
|
// If object has no pointers, don't need to scan further.
|
if((bits & bitNoPointers) != 0)
|
if((bits & bitNoPointers) != 0)
|
continue;
|
continue;
|
|
|
// If another proc wants a pointer, give it some.
|
// If another proc wants a pointer, give it some.
|
if(nobj > 4 && work.nwait > 0 && work.full == nil) {
|
if(nobj > 4 && work.nwait > 0 && work.full == nil) {
|
wbuf->nobj = nobj;
|
wbuf->nobj = nobj;
|
wbuf = handoff(wbuf);
|
wbuf = handoff(wbuf);
|
nobj = wbuf->nobj;
|
nobj = wbuf->nobj;
|
wp = (void**)(wbuf->obj + nobj);
|
wp = (void**)(wbuf->obj + nobj);
|
}
|
}
|
|
|
// If buffer is full, get a new one.
|
// If buffer is full, get a new one.
|
if(wbuf == nil || nobj >= nelem(wbuf->obj)) {
|
if(wbuf == nil || nobj >= nelem(wbuf->obj)) {
|
if(wbuf != nil)
|
if(wbuf != nil)
|
wbuf->nobj = nobj;
|
wbuf->nobj = nobj;
|
wbuf = getempty(wbuf);
|
wbuf = getempty(wbuf);
|
wp = (void**)(wbuf->obj);
|
wp = (void**)(wbuf->obj);
|
nobj = 0;
|
nobj = 0;
|
}
|
}
|
*wp++ = obj;
|
*wp++ = obj;
|
nobj++;
|
nobj++;
|
continue_obj:;
|
continue_obj:;
|
}
|
}
|
|
|
// Done scanning [b, b+n). Prepare for the next iteration of
|
// Done scanning [b, b+n). Prepare for the next iteration of
|
// the loop by setting b and n to the parameters for the next block.
|
// the loop by setting b and n to the parameters for the next block.
|
|
|
// Fetch b from the work buffer.
|
// Fetch b from the work buffer.
|
if(nobj == 0) {
|
if(nobj == 0) {
|
if(!keepworking) {
|
if(!keepworking) {
|
putempty(wbuf);
|
putempty(wbuf);
|
return;
|
return;
|
}
|
}
|
// Emptied our buffer: refill.
|
// Emptied our buffer: refill.
|
wbuf = getfull(wbuf);
|
wbuf = getfull(wbuf);
|
if(wbuf == nil)
|
if(wbuf == nil)
|
return;
|
return;
|
nobj = wbuf->nobj;
|
nobj = wbuf->nobj;
|
wp = (void**)(wbuf->obj + wbuf->nobj);
|
wp = (void**)(wbuf->obj + wbuf->nobj);
|
}
|
}
|
b = *--wp;
|
b = *--wp;
|
nobj--;
|
nobj--;
|
|
|
// Ask span about size class.
|
// Ask span about size class.
|
// (Manually inlined copy of MHeap_Lookup.)
|
// (Manually inlined copy of MHeap_Lookup.)
|
x = (uintptr)b>>PageShift;
|
x = (uintptr)b>>PageShift;
|
if(sizeof(void*) == 8)
|
if(sizeof(void*) == 8)
|
x -= (uintptr)arena_start>>PageShift;
|
x -= (uintptr)arena_start>>PageShift;
|
s = runtime_mheap.map[x];
|
s = runtime_mheap.map[x];
|
if(s->sizeclass == 0)
|
if(s->sizeclass == 0)
|
n = s->npages<<PageShift;
|
n = s->npages<<PageShift;
|
else
|
else
|
n = runtime_class_to_size[s->sizeclass];
|
n = runtime_class_to_size[s->sizeclass];
|
}
|
}
|
}
|
}
|
|
|
// debug_scanblock is the debug copy of scanblock.
|
// debug_scanblock is the debug copy of scanblock.
|
// it is simpler, slower, single-threaded, recursive,
|
// it is simpler, slower, single-threaded, recursive,
|
// and uses bitSpecial as the mark bit.
|
// and uses bitSpecial as the mark bit.
|
static void
|
static void
|
debug_scanblock(byte *b, int64 n)
|
debug_scanblock(byte *b, int64 n)
|
{
|
{
|
byte *obj, *p;
|
byte *obj, *p;
|
void **vp;
|
void **vp;
|
uintptr size, *bitp, bits, shift, i, xbits, off;
|
uintptr size, *bitp, bits, shift, i, xbits, off;
|
MSpan *s;
|
MSpan *s;
|
|
|
if(!DebugMark)
|
if(!DebugMark)
|
runtime_throw("debug_scanblock without DebugMark");
|
runtime_throw("debug_scanblock without DebugMark");
|
|
|
if((int64)(uintptr)n != n || n < 0) {
|
if((int64)(uintptr)n != n || n < 0) {
|
//runtime_printf("debug_scanblock %p %D\n", b, n);
|
//runtime_printf("debug_scanblock %p %D\n", b, n);
|
runtime_throw("debug_scanblock");
|
runtime_throw("debug_scanblock");
|
}
|
}
|
|
|
// Align b to a word boundary.
|
// Align b to a word boundary.
|
off = (uintptr)b & (PtrSize-1);
|
off = (uintptr)b & (PtrSize-1);
|
if(off != 0) {
|
if(off != 0) {
|
b += PtrSize - off;
|
b += PtrSize - off;
|
n -= PtrSize - off;
|
n -= PtrSize - off;
|
}
|
}
|
|
|
vp = (void**)b;
|
vp = (void**)b;
|
n /= PtrSize;
|
n /= PtrSize;
|
for(i=0; i<(uintptr)n; i++) {
|
for(i=0; i<(uintptr)n; i++) {
|
obj = (byte*)vp[i];
|
obj = (byte*)vp[i];
|
|
|
// Words outside the arena cannot be pointers.
|
// Words outside the arena cannot be pointers.
|
if((byte*)obj < runtime_mheap.arena_start || (byte*)obj >= runtime_mheap.arena_used)
|
if((byte*)obj < runtime_mheap.arena_start || (byte*)obj >= runtime_mheap.arena_used)
|
continue;
|
continue;
|
|
|
// Round down to word boundary.
|
// Round down to word boundary.
|
obj = (void*)((uintptr)obj & ~((uintptr)PtrSize-1));
|
obj = (void*)((uintptr)obj & ~((uintptr)PtrSize-1));
|
|
|
// Consult span table to find beginning.
|
// Consult span table to find beginning.
|
s = runtime_MHeap_LookupMaybe(&runtime_mheap, obj);
|
s = runtime_MHeap_LookupMaybe(&runtime_mheap, obj);
|
if(s == nil)
|
if(s == nil)
|
continue;
|
continue;
|
|
|
|
|
p = (byte*)((uintptr)s->start<<PageShift);
|
p = (byte*)((uintptr)s->start<<PageShift);
|
if(s->sizeclass == 0) {
|
if(s->sizeclass == 0) {
|
obj = p;
|
obj = p;
|
size = (uintptr)s->npages<<PageShift;
|
size = (uintptr)s->npages<<PageShift;
|
} else {
|
} else {
|
if((byte*)obj >= (byte*)s->limit)
|
if((byte*)obj >= (byte*)s->limit)
|
continue;
|
continue;
|
size = runtime_class_to_size[s->sizeclass];
|
size = runtime_class_to_size[s->sizeclass];
|
int32 i = ((byte*)obj - p)/size;
|
int32 i = ((byte*)obj - p)/size;
|
obj = p+i*size;
|
obj = p+i*size;
|
}
|
}
|
|
|
// Now that we know the object header, reload bits.
|
// Now that we know the object header, reload bits.
|
off = (uintptr*)obj - (uintptr*)runtime_mheap.arena_start;
|
off = (uintptr*)obj - (uintptr*)runtime_mheap.arena_start;
|
bitp = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
bitp = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
xbits = *bitp;
|
xbits = *bitp;
|
bits = xbits >> shift;
|
bits = xbits >> shift;
|
|
|
// Now we have bits, bitp, and shift correct for
|
// Now we have bits, bitp, and shift correct for
|
// obj pointing at the base of the object.
|
// obj pointing at the base of the object.
|
// If not allocated or already marked, done.
|
// If not allocated or already marked, done.
|
if((bits & bitAllocated) == 0 || (bits & bitSpecial) != 0) // NOTE: bitSpecial not bitMarked
|
if((bits & bitAllocated) == 0 || (bits & bitSpecial) != 0) // NOTE: bitSpecial not bitMarked
|
continue;
|
continue;
|
*bitp |= bitSpecial<<shift;
|
*bitp |= bitSpecial<<shift;
|
if(!(bits & bitMarked))
|
if(!(bits & bitMarked))
|
runtime_printf("found unmarked block %p in %p\n", obj, vp+i);
|
runtime_printf("found unmarked block %p in %p\n", obj, vp+i);
|
|
|
// If object has no pointers, don't need to scan further.
|
// If object has no pointers, don't need to scan further.
|
if((bits & bitNoPointers) != 0)
|
if((bits & bitNoPointers) != 0)
|
continue;
|
continue;
|
|
|
debug_scanblock(obj, size);
|
debug_scanblock(obj, size);
|
}
|
}
|
}
|
}
|
|
|
// Get an empty work buffer off the work.empty list,
|
// Get an empty work buffer off the work.empty list,
|
// allocating new buffers as needed.
|
// allocating new buffers as needed.
|
static Workbuf*
|
static Workbuf*
|
getempty(Workbuf *b)
|
getempty(Workbuf *b)
|
{
|
{
|
if(work.nproc == 1) {
|
if(work.nproc == 1) {
|
// Put b on full list.
|
// Put b on full list.
|
if(b != nil) {
|
if(b != nil) {
|
b->next = work.full;
|
b->next = work.full;
|
work.full = b;
|
work.full = b;
|
}
|
}
|
// Grab from empty list if possible.
|
// Grab from empty list if possible.
|
b = work.empty;
|
b = work.empty;
|
if(b != nil) {
|
if(b != nil) {
|
work.empty = b->next;
|
work.empty = b->next;
|
goto haveb;
|
goto haveb;
|
}
|
}
|
} else {
|
} else {
|
// Put b on full list.
|
// Put b on full list.
|
if(b != nil) {
|
if(b != nil) {
|
runtime_lock(&work.fmu);
|
runtime_lock(&work.fmu);
|
b->next = work.full;
|
b->next = work.full;
|
work.full = b;
|
work.full = b;
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
}
|
}
|
// Grab from empty list if possible.
|
// Grab from empty list if possible.
|
runtime_lock(&work.emu);
|
runtime_lock(&work.emu);
|
b = work.empty;
|
b = work.empty;
|
if(b != nil)
|
if(b != nil)
|
work.empty = b->next;
|
work.empty = b->next;
|
runtime_unlock(&work.emu);
|
runtime_unlock(&work.emu);
|
if(b != nil)
|
if(b != nil)
|
goto haveb;
|
goto haveb;
|
}
|
}
|
|
|
// Need to allocate.
|
// Need to allocate.
|
runtime_lock(&work);
|
runtime_lock(&work);
|
if(work.nchunk < sizeof *b) {
|
if(work.nchunk < sizeof *b) {
|
work.nchunk = 1<<20;
|
work.nchunk = 1<<20;
|
work.chunk = runtime_SysAlloc(work.nchunk);
|
work.chunk = runtime_SysAlloc(work.nchunk);
|
}
|
}
|
b = (Workbuf*)work.chunk;
|
b = (Workbuf*)work.chunk;
|
work.chunk += sizeof *b;
|
work.chunk += sizeof *b;
|
work.nchunk -= sizeof *b;
|
work.nchunk -= sizeof *b;
|
runtime_unlock(&work);
|
runtime_unlock(&work);
|
|
|
haveb:
|
haveb:
|
b->nobj = 0;
|
b->nobj = 0;
|
return b;
|
return b;
|
}
|
}
|
|
|
static void
|
static void
|
putempty(Workbuf *b)
|
putempty(Workbuf *b)
|
{
|
{
|
if(b == nil)
|
if(b == nil)
|
return;
|
return;
|
|
|
if(work.nproc == 1) {
|
if(work.nproc == 1) {
|
b->next = work.empty;
|
b->next = work.empty;
|
work.empty = b;
|
work.empty = b;
|
return;
|
return;
|
}
|
}
|
|
|
runtime_lock(&work.emu);
|
runtime_lock(&work.emu);
|
b->next = work.empty;
|
b->next = work.empty;
|
work.empty = b;
|
work.empty = b;
|
runtime_unlock(&work.emu);
|
runtime_unlock(&work.emu);
|
}
|
}
|
|
|
// Get a full work buffer off the work.full list, or return nil.
|
// Get a full work buffer off the work.full list, or return nil.
|
static Workbuf*
|
static Workbuf*
|
getfull(Workbuf *b)
|
getfull(Workbuf *b)
|
{
|
{
|
int32 i;
|
int32 i;
|
Workbuf *b1;
|
Workbuf *b1;
|
|
|
if(work.nproc == 1) {
|
if(work.nproc == 1) {
|
// Put b on empty list.
|
// Put b on empty list.
|
if(b != nil) {
|
if(b != nil) {
|
b->next = work.empty;
|
b->next = work.empty;
|
work.empty = b;
|
work.empty = b;
|
}
|
}
|
// Grab from full list if possible.
|
// Grab from full list if possible.
|
// Since work.nproc==1, no one else is
|
// Since work.nproc==1, no one else is
|
// going to give us work.
|
// going to give us work.
|
b = work.full;
|
b = work.full;
|
if(b != nil)
|
if(b != nil)
|
work.full = b->next;
|
work.full = b->next;
|
return b;
|
return b;
|
}
|
}
|
|
|
putempty(b);
|
putempty(b);
|
|
|
// Grab buffer from full list if possible.
|
// Grab buffer from full list if possible.
|
for(;;) {
|
for(;;) {
|
b1 = work.full;
|
b1 = work.full;
|
if(b1 == nil)
|
if(b1 == nil)
|
break;
|
break;
|
runtime_lock(&work.fmu);
|
runtime_lock(&work.fmu);
|
if(work.full != nil) {
|
if(work.full != nil) {
|
b1 = work.full;
|
b1 = work.full;
|
work.full = b1->next;
|
work.full = b1->next;
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
return b1;
|
return b1;
|
}
|
}
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
}
|
}
|
|
|
runtime_xadd(&work.nwait, +1);
|
runtime_xadd(&work.nwait, +1);
|
for(i=0;; i++) {
|
for(i=0;; i++) {
|
b1 = work.full;
|
b1 = work.full;
|
if(b1 != nil) {
|
if(b1 != nil) {
|
runtime_lock(&work.fmu);
|
runtime_lock(&work.fmu);
|
if(work.full != nil) {
|
if(work.full != nil) {
|
runtime_xadd(&work.nwait, -1);
|
runtime_xadd(&work.nwait, -1);
|
b1 = work.full;
|
b1 = work.full;
|
work.full = b1->next;
|
work.full = b1->next;
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
return b1;
|
return b1;
|
}
|
}
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
continue;
|
continue;
|
}
|
}
|
if(work.nwait == work.nproc)
|
if(work.nwait == work.nproc)
|
return nil;
|
return nil;
|
if(i < 10)
|
if(i < 10)
|
runtime_procyield(20);
|
runtime_procyield(20);
|
else if(i < 20)
|
else if(i < 20)
|
runtime_osyield();
|
runtime_osyield();
|
else
|
else
|
runtime_usleep(100);
|
runtime_usleep(100);
|
}
|
}
|
}
|
}
|
|
|
static Workbuf*
|
static Workbuf*
|
handoff(Workbuf *b)
|
handoff(Workbuf *b)
|
{
|
{
|
int32 n;
|
int32 n;
|
Workbuf *b1;
|
Workbuf *b1;
|
|
|
// Make new buffer with half of b's pointers.
|
// Make new buffer with half of b's pointers.
|
b1 = getempty(nil);
|
b1 = getempty(nil);
|
n = b->nobj/2;
|
n = b->nobj/2;
|
b->nobj -= n;
|
b->nobj -= n;
|
b1->nobj = n;
|
b1->nobj = n;
|
runtime_memmove(b1->obj, b->obj+b->nobj, n*sizeof b1->obj[0]);
|
runtime_memmove(b1->obj, b->obj+b->nobj, n*sizeof b1->obj[0]);
|
nhandoff += n;
|
nhandoff += n;
|
|
|
// Put b on full list - let first half of b get stolen.
|
// Put b on full list - let first half of b get stolen.
|
runtime_lock(&work.fmu);
|
runtime_lock(&work.fmu);
|
b->next = work.full;
|
b->next = work.full;
|
work.full = b;
|
work.full = b;
|
runtime_unlock(&work.fmu);
|
runtime_unlock(&work.fmu);
|
|
|
return b1;
|
return b1;
|
}
|
}
|
|
|
// Scanstack calls scanblock on each of gp's stack segments.
|
// Scanstack calls scanblock on each of gp's stack segments.
|
static void
|
static void
|
scanstack(void (*scanblock)(byte*, int64), G *gp)
|
scanstack(void (*scanblock)(byte*, int64), G *gp)
|
{
|
{
|
#ifdef USING_SPLIT_STACK
|
#ifdef USING_SPLIT_STACK
|
M *mp;
|
M *mp;
|
void* sp;
|
void* sp;
|
size_t spsize;
|
size_t spsize;
|
void* next_segment;
|
void* next_segment;
|
void* next_sp;
|
void* next_sp;
|
void* initial_sp;
|
void* initial_sp;
|
|
|
if(gp == runtime_g()) {
|
if(gp == runtime_g()) {
|
// Scanning our own stack.
|
// Scanning our own stack.
|
sp = __splitstack_find(nil, nil, &spsize, &next_segment,
|
sp = __splitstack_find(nil, nil, &spsize, &next_segment,
|
&next_sp, &initial_sp);
|
&next_sp, &initial_sp);
|
} else if((mp = gp->m) != nil && mp->helpgc) {
|
} else if((mp = gp->m) != nil && mp->helpgc) {
|
// gchelper's stack is in active use and has no interesting pointers.
|
// gchelper's stack is in active use and has no interesting pointers.
|
return;
|
return;
|
} else {
|
} else {
|
// Scanning another goroutine's stack.
|
// Scanning another goroutine's stack.
|
// The goroutine is usually asleep (the world is stopped).
|
// The goroutine is usually asleep (the world is stopped).
|
|
|
// The exception is that if the goroutine is about to enter or might
|
// The exception is that if the goroutine is about to enter or might
|
// have just exited a system call, it may be executing code such
|
// have just exited a system call, it may be executing code such
|
// as schedlock and may have needed to start a new stack segment.
|
// as schedlock and may have needed to start a new stack segment.
|
// Use the stack segment and stack pointer at the time of
|
// Use the stack segment and stack pointer at the time of
|
// the system call instead, since that won't change underfoot.
|
// the system call instead, since that won't change underfoot.
|
if(gp->gcstack != nil) {
|
if(gp->gcstack != nil) {
|
sp = gp->gcstack;
|
sp = gp->gcstack;
|
spsize = gp->gcstack_size;
|
spsize = gp->gcstack_size;
|
next_segment = gp->gcnext_segment;
|
next_segment = gp->gcnext_segment;
|
next_sp = gp->gcnext_sp;
|
next_sp = gp->gcnext_sp;
|
initial_sp = gp->gcinitial_sp;
|
initial_sp = gp->gcinitial_sp;
|
} else {
|
} else {
|
sp = __splitstack_find_context(&gp->stack_context[0],
|
sp = __splitstack_find_context(&gp->stack_context[0],
|
&spsize, &next_segment,
|
&spsize, &next_segment,
|
&next_sp, &initial_sp);
|
&next_sp, &initial_sp);
|
}
|
}
|
}
|
}
|
if(sp != nil) {
|
if(sp != nil) {
|
scanblock(sp, spsize);
|
scanblock(sp, spsize);
|
while((sp = __splitstack_find(next_segment, next_sp,
|
while((sp = __splitstack_find(next_segment, next_sp,
|
&spsize, &next_segment,
|
&spsize, &next_segment,
|
&next_sp, &initial_sp)) != nil)
|
&next_sp, &initial_sp)) != nil)
|
scanblock(sp, spsize);
|
scanblock(sp, spsize);
|
}
|
}
|
#else
|
#else
|
M *mp;
|
M *mp;
|
byte* bottom;
|
byte* bottom;
|
byte* top;
|
byte* top;
|
|
|
if(gp == runtime_g()) {
|
if(gp == runtime_g()) {
|
// Scanning our own stack.
|
// Scanning our own stack.
|
bottom = (byte*)&gp;
|
bottom = (byte*)&gp;
|
} else if((mp = gp->m) != nil && mp->helpgc) {
|
} else if((mp = gp->m) != nil && mp->helpgc) {
|
// gchelper's stack is in active use and has no interesting pointers.
|
// gchelper's stack is in active use and has no interesting pointers.
|
return;
|
return;
|
} else {
|
} else {
|
// Scanning another goroutine's stack.
|
// Scanning another goroutine's stack.
|
// The goroutine is usually asleep (the world is stopped).
|
// The goroutine is usually asleep (the world is stopped).
|
bottom = (byte*)gp->gcnext_sp;
|
bottom = (byte*)gp->gcnext_sp;
|
if(bottom == nil)
|
if(bottom == nil)
|
return;
|
return;
|
}
|
}
|
top = (byte*)gp->gcinitial_sp + gp->gcstack_size;
|
top = (byte*)gp->gcinitial_sp + gp->gcstack_size;
|
if(top > bottom)
|
if(top > bottom)
|
scanblock(bottom, top - bottom);
|
scanblock(bottom, top - bottom);
|
else
|
else
|
scanblock(top, bottom - top);
|
scanblock(top, bottom - top);
|
#endif
|
#endif
|
}
|
}
|
|
|
// Markfin calls scanblock on the blocks that have finalizers:
|
// Markfin calls scanblock on the blocks that have finalizers:
|
// the things pointed at cannot be freed until the finalizers have run.
|
// the things pointed at cannot be freed until the finalizers have run.
|
static void
|
static void
|
markfin(void *v)
|
markfin(void *v)
|
{
|
{
|
uintptr size;
|
uintptr size;
|
|
|
size = 0;
|
size = 0;
|
if(!runtime_mlookup(v, (byte**)&v, &size, nil) || !runtime_blockspecial(v))
|
if(!runtime_mlookup(v, (byte**)&v, &size, nil) || !runtime_blockspecial(v))
|
runtime_throw("mark - finalizer inconsistency");
|
runtime_throw("mark - finalizer inconsistency");
|
|
|
// do not mark the finalizer block itself. just mark the things it points at.
|
// do not mark the finalizer block itself. just mark the things it points at.
|
scanblock(v, size);
|
scanblock(v, size);
|
}
|
}
|
|
|
struct root_list {
|
struct root_list {
|
struct root_list *next;
|
struct root_list *next;
|
struct root {
|
struct root {
|
void *decl;
|
void *decl;
|
size_t size;
|
size_t size;
|
} roots[];
|
} roots[];
|
};
|
};
|
|
|
static struct root_list* roots;
|
static struct root_list* roots;
|
|
|
void
|
void
|
__go_register_gc_roots (struct root_list* r)
|
__go_register_gc_roots (struct root_list* r)
|
{
|
{
|
// FIXME: This needs locking if multiple goroutines can call
|
// FIXME: This needs locking if multiple goroutines can call
|
// dlopen simultaneously.
|
// dlopen simultaneously.
|
r->next = roots;
|
r->next = roots;
|
roots = r;
|
roots = r;
|
}
|
}
|
|
|
static void
|
static void
|
debug_markfin(void *v)
|
debug_markfin(void *v)
|
{
|
{
|
uintptr size;
|
uintptr size;
|
|
|
if(!runtime_mlookup(v, (byte**)&v, &size, nil))
|
if(!runtime_mlookup(v, (byte**)&v, &size, nil))
|
runtime_throw("debug_mark - finalizer inconsistency");
|
runtime_throw("debug_mark - finalizer inconsistency");
|
debug_scanblock(v, size);
|
debug_scanblock(v, size);
|
}
|
}
|
|
|
// Mark
|
// Mark
|
static void
|
static void
|
mark(void (*scan)(byte*, int64))
|
mark(void (*scan)(byte*, int64))
|
{
|
{
|
struct root_list *pl;
|
struct root_list *pl;
|
G *gp;
|
G *gp;
|
FinBlock *fb;
|
FinBlock *fb;
|
|
|
// mark data+bss.
|
// mark data+bss.
|
for(pl = roots; pl != nil; pl = pl->next) {
|
for(pl = roots; pl != nil; pl = pl->next) {
|
struct root* pr = &pl->roots[0];
|
struct root* pr = &pl->roots[0];
|
while(1) {
|
while(1) {
|
void *decl = pr->decl;
|
void *decl = pr->decl;
|
if(decl == nil)
|
if(decl == nil)
|
break;
|
break;
|
scanblock(decl, pr->size);
|
scanblock(decl, pr->size);
|
pr++;
|
pr++;
|
}
|
}
|
}
|
}
|
|
|
scan((byte*)&runtime_m0, sizeof runtime_m0);
|
scan((byte*)&runtime_m0, sizeof runtime_m0);
|
scan((byte*)&runtime_g0, sizeof runtime_g0);
|
scan((byte*)&runtime_g0, sizeof runtime_g0);
|
scan((byte*)&runtime_allg, sizeof runtime_allg);
|
scan((byte*)&runtime_allg, sizeof runtime_allg);
|
scan((byte*)&runtime_allm, sizeof runtime_allm);
|
scan((byte*)&runtime_allm, sizeof runtime_allm);
|
runtime_MProf_Mark(scan);
|
runtime_MProf_Mark(scan);
|
runtime_time_scan(scan);
|
runtime_time_scan(scan);
|
|
|
// mark stacks
|
// mark stacks
|
for(gp=runtime_allg; gp!=nil; gp=gp->alllink) {
|
for(gp=runtime_allg; gp!=nil; gp=gp->alllink) {
|
switch(gp->status){
|
switch(gp->status){
|
default:
|
default:
|
runtime_printf("unexpected G.status %d\n", gp->status);
|
runtime_printf("unexpected G.status %d\n", gp->status);
|
runtime_throw("mark - bad status");
|
runtime_throw("mark - bad status");
|
case Gdead:
|
case Gdead:
|
break;
|
break;
|
case Grunning:
|
case Grunning:
|
if(gp != runtime_g())
|
if(gp != runtime_g())
|
runtime_throw("mark - world not stopped");
|
runtime_throw("mark - world not stopped");
|
scanstack(scan, gp);
|
scanstack(scan, gp);
|
break;
|
break;
|
case Grunnable:
|
case Grunnable:
|
case Gsyscall:
|
case Gsyscall:
|
case Gwaiting:
|
case Gwaiting:
|
scanstack(scan, gp);
|
scanstack(scan, gp);
|
break;
|
break;
|
}
|
}
|
}
|
}
|
|
|
// mark things pointed at by objects with finalizers
|
// mark things pointed at by objects with finalizers
|
if(scan == debug_scanblock)
|
if(scan == debug_scanblock)
|
runtime_walkfintab(debug_markfin, scan);
|
runtime_walkfintab(debug_markfin, scan);
|
else
|
else
|
runtime_walkfintab(markfin, scan);
|
runtime_walkfintab(markfin, scan);
|
|
|
for(fb=allfin; fb; fb=fb->alllink)
|
for(fb=allfin; fb; fb=fb->alllink)
|
scanblock((byte*)fb->fin, fb->cnt*sizeof(fb->fin[0]));
|
scanblock((byte*)fb->fin, fb->cnt*sizeof(fb->fin[0]));
|
|
|
// in multiproc mode, join in the queued work.
|
// in multiproc mode, join in the queued work.
|
scan(nil, 0);
|
scan(nil, 0);
|
}
|
}
|
|
|
static bool
|
static bool
|
handlespecial(byte *p, uintptr size)
|
handlespecial(byte *p, uintptr size)
|
{
|
{
|
void (*fn)(void*);
|
void (*fn)(void*);
|
const struct __go_func_type *ft;
|
const struct __go_func_type *ft;
|
FinBlock *block;
|
FinBlock *block;
|
Finalizer *f;
|
Finalizer *f;
|
|
|
if(!runtime_getfinalizer(p, true, &fn, &ft)) {
|
if(!runtime_getfinalizer(p, true, &fn, &ft)) {
|
runtime_setblockspecial(p, false);
|
runtime_setblockspecial(p, false);
|
runtime_MProf_Free(p, size);
|
runtime_MProf_Free(p, size);
|
return false;
|
return false;
|
}
|
}
|
|
|
runtime_lock(&finlock);
|
runtime_lock(&finlock);
|
if(finq == nil || finq->cnt == finq->cap) {
|
if(finq == nil || finq->cnt == finq->cap) {
|
if(finc == nil) {
|
if(finc == nil) {
|
finc = runtime_SysAlloc(PageSize);
|
finc = runtime_SysAlloc(PageSize);
|
finc->cap = (PageSize - sizeof(FinBlock)) / sizeof(Finalizer) + 1;
|
finc->cap = (PageSize - sizeof(FinBlock)) / sizeof(Finalizer) + 1;
|
finc->alllink = allfin;
|
finc->alllink = allfin;
|
allfin = finc;
|
allfin = finc;
|
}
|
}
|
block = finc;
|
block = finc;
|
finc = block->next;
|
finc = block->next;
|
block->next = finq;
|
block->next = finq;
|
finq = block;
|
finq = block;
|
}
|
}
|
f = &finq->fin[finq->cnt];
|
f = &finq->fin[finq->cnt];
|
finq->cnt++;
|
finq->cnt++;
|
f->fn = fn;
|
f->fn = fn;
|
f->ft = ft;
|
f->ft = ft;
|
f->arg = p;
|
f->arg = p;
|
runtime_unlock(&finlock);
|
runtime_unlock(&finlock);
|
return true;
|
return true;
|
}
|
}
|
|
|
// Sweep frees or collects finalizers for blocks not marked in the mark phase.
|
// Sweep frees or collects finalizers for blocks not marked in the mark phase.
|
// It clears the mark bits in preparation for the next GC round.
|
// It clears the mark bits in preparation for the next GC round.
|
static void
|
static void
|
sweep(void)
|
sweep(void)
|
{
|
{
|
M *m;
|
M *m;
|
MSpan *s;
|
MSpan *s;
|
int32 cl, n, npages;
|
int32 cl, n, npages;
|
uintptr size;
|
uintptr size;
|
byte *p;
|
byte *p;
|
MCache *c;
|
MCache *c;
|
byte *arena_start;
|
byte *arena_start;
|
|
|
m = runtime_m();
|
m = runtime_m();
|
arena_start = runtime_mheap.arena_start;
|
arena_start = runtime_mheap.arena_start;
|
|
|
for(;;) {
|
for(;;) {
|
s = work.spans;
|
s = work.spans;
|
if(s == nil)
|
if(s == nil)
|
break;
|
break;
|
if(!runtime_casp(&work.spans, s, s->allnext))
|
if(!runtime_casp(&work.spans, s, s->allnext))
|
continue;
|
continue;
|
|
|
if(s->state != MSpanInUse)
|
if(s->state != MSpanInUse)
|
continue;
|
continue;
|
|
|
p = (byte*)(s->start << PageShift);
|
p = (byte*)(s->start << PageShift);
|
cl = s->sizeclass;
|
cl = s->sizeclass;
|
if(cl == 0) {
|
if(cl == 0) {
|
size = s->npages<<PageShift;
|
size = s->npages<<PageShift;
|
n = 1;
|
n = 1;
|
} else {
|
} else {
|
// Chunk full of small blocks.
|
// Chunk full of small blocks.
|
size = runtime_class_to_size[cl];
|
size = runtime_class_to_size[cl];
|
npages = runtime_class_to_allocnpages[cl];
|
npages = runtime_class_to_allocnpages[cl];
|
n = (npages << PageShift) / size;
|
n = (npages << PageShift) / size;
|
}
|
}
|
|
|
// Sweep through n objects of given size starting at p.
|
// Sweep through n objects of given size starting at p.
|
// This thread owns the span now, so it can manipulate
|
// This thread owns the span now, so it can manipulate
|
// the block bitmap without atomic operations.
|
// the block bitmap without atomic operations.
|
for(; n > 0; n--, p += size) {
|
for(; n > 0; n--, p += size) {
|
uintptr off, *bitp, shift, bits;
|
uintptr off, *bitp, shift, bits;
|
|
|
off = (uintptr*)p - (uintptr*)arena_start;
|
off = (uintptr*)p - (uintptr*)arena_start;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
bitp = (uintptr*)arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
bits = *bitp>>shift;
|
bits = *bitp>>shift;
|
|
|
if((bits & bitAllocated) == 0)
|
if((bits & bitAllocated) == 0)
|
continue;
|
continue;
|
|
|
if((bits & bitMarked) != 0) {
|
if((bits & bitMarked) != 0) {
|
if(DebugMark) {
|
if(DebugMark) {
|
if(!(bits & bitSpecial))
|
if(!(bits & bitSpecial))
|
runtime_printf("found spurious mark on %p\n", p);
|
runtime_printf("found spurious mark on %p\n", p);
|
*bitp &= ~(bitSpecial<<shift);
|
*bitp &= ~(bitSpecial<<shift);
|
}
|
}
|
*bitp &= ~(bitMarked<<shift);
|
*bitp &= ~(bitMarked<<shift);
|
continue;
|
continue;
|
}
|
}
|
|
|
// Special means it has a finalizer or is being profiled.
|
// Special means it has a finalizer or is being profiled.
|
// In DebugMark mode, the bit has been coopted so
|
// In DebugMark mode, the bit has been coopted so
|
// we have to assume all blocks are special.
|
// we have to assume all blocks are special.
|
if(DebugMark || (bits & bitSpecial) != 0) {
|
if(DebugMark || (bits & bitSpecial) != 0) {
|
if(handlespecial(p, size))
|
if(handlespecial(p, size))
|
continue;
|
continue;
|
}
|
}
|
|
|
// Mark freed; restore block boundary bit.
|
// Mark freed; restore block boundary bit.
|
*bitp = (*bitp & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
*bitp = (*bitp & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
|
|
c = m->mcache;
|
c = m->mcache;
|
if(s->sizeclass == 0) {
|
if(s->sizeclass == 0) {
|
// Free large span.
|
// Free large span.
|
runtime_unmarkspan(p, 1<<PageShift);
|
runtime_unmarkspan(p, 1<<PageShift);
|
*(uintptr*)p = 1; // needs zeroing
|
*(uintptr*)p = 1; // needs zeroing
|
runtime_MHeap_Free(&runtime_mheap, s, 1);
|
runtime_MHeap_Free(&runtime_mheap, s, 1);
|
} else {
|
} else {
|
// Free small object.
|
// Free small object.
|
if(size > sizeof(uintptr))
|
if(size > sizeof(uintptr))
|
((uintptr*)p)[1] = 1; // mark as "needs to be zeroed"
|
((uintptr*)p)[1] = 1; // mark as "needs to be zeroed"
|
c->local_by_size[s->sizeclass].nfree++;
|
c->local_by_size[s->sizeclass].nfree++;
|
runtime_MCache_Free(c, p, s->sizeclass, size);
|
runtime_MCache_Free(c, p, s->sizeclass, size);
|
}
|
}
|
c->local_alloc -= size;
|
c->local_alloc -= size;
|
c->local_nfree++;
|
c->local_nfree++;
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
void
|
void
|
runtime_gchelper(void)
|
runtime_gchelper(void)
|
{
|
{
|
// Wait until main proc is ready for mark help.
|
// Wait until main proc is ready for mark help.
|
runtime_lock(&work.markgate);
|
runtime_lock(&work.markgate);
|
runtime_unlock(&work.markgate);
|
runtime_unlock(&work.markgate);
|
scanblock(nil, 0);
|
scanblock(nil, 0);
|
|
|
// Wait until main proc is ready for sweep help.
|
// Wait until main proc is ready for sweep help.
|
runtime_lock(&work.sweepgate);
|
runtime_lock(&work.sweepgate);
|
runtime_unlock(&work.sweepgate);
|
runtime_unlock(&work.sweepgate);
|
sweep();
|
sweep();
|
|
|
if(runtime_xadd(&work.ndone, +1) == work.nproc-1)
|
if(runtime_xadd(&work.ndone, +1) == work.nproc-1)
|
runtime_notewakeup(&work.alldone);
|
runtime_notewakeup(&work.alldone);
|
}
|
}
|
|
|
// Semaphore, not Lock, so that the goroutine
|
// Semaphore, not Lock, so that the goroutine
|
// reschedules when there is contention rather
|
// reschedules when there is contention rather
|
// than spinning.
|
// than spinning.
|
static uint32 gcsema = 1;
|
static uint32 gcsema = 1;
|
|
|
// Initialized from $GOGC. GOGC=off means no gc.
|
// Initialized from $GOGC. GOGC=off means no gc.
|
//
|
//
|
// Next gc is after we've allocated an extra amount of
|
// Next gc is after we've allocated an extra amount of
|
// memory proportional to the amount already in use.
|
// memory proportional to the amount already in use.
|
// If gcpercent=100 and we're using 4M, we'll gc again
|
// If gcpercent=100 and we're using 4M, we'll gc again
|
// when we get to 8M. This keeps the gc cost in linear
|
// when we get to 8M. This keeps the gc cost in linear
|
// proportion to the allocation cost. Adjusting gcpercent
|
// proportion to the allocation cost. Adjusting gcpercent
|
// just changes the linear constant (and also the amount of
|
// just changes the linear constant (and also the amount of
|
// extra memory used).
|
// extra memory used).
|
static int32 gcpercent = -2;
|
static int32 gcpercent = -2;
|
|
|
static void
|
static void
|
stealcache(void)
|
stealcache(void)
|
{
|
{
|
M *m;
|
M *m;
|
|
|
for(m=runtime_allm; m; m=m->alllink)
|
for(m=runtime_allm; m; m=m->alllink)
|
runtime_MCache_ReleaseAll(m->mcache);
|
runtime_MCache_ReleaseAll(m->mcache);
|
}
|
}
|
|
|
static void
|
static void
|
cachestats(void)
|
cachestats(void)
|
{
|
{
|
M *m;
|
M *m;
|
MCache *c;
|
MCache *c;
|
uint32 i;
|
uint32 i;
|
uint64 stacks_inuse;
|
uint64 stacks_inuse;
|
uint64 stacks_sys;
|
uint64 stacks_sys;
|
|
|
stacks_inuse = 0;
|
stacks_inuse = 0;
|
stacks_sys = 0;
|
stacks_sys = 0;
|
for(m=runtime_allm; m; m=m->alllink) {
|
for(m=runtime_allm; m; m=m->alllink) {
|
runtime_purgecachedstats(m);
|
runtime_purgecachedstats(m);
|
// stacks_inuse += m->stackalloc->inuse;
|
// stacks_inuse += m->stackalloc->inuse;
|
// stacks_sys += m->stackalloc->sys;
|
// stacks_sys += m->stackalloc->sys;
|
c = m->mcache;
|
c = m->mcache;
|
for(i=0; i<nelem(c->local_by_size); i++) {
|
for(i=0; i<nelem(c->local_by_size); i++) {
|
mstats.by_size[i].nmalloc += c->local_by_size[i].nmalloc;
|
mstats.by_size[i].nmalloc += c->local_by_size[i].nmalloc;
|
c->local_by_size[i].nmalloc = 0;
|
c->local_by_size[i].nmalloc = 0;
|
mstats.by_size[i].nfree += c->local_by_size[i].nfree;
|
mstats.by_size[i].nfree += c->local_by_size[i].nfree;
|
c->local_by_size[i].nfree = 0;
|
c->local_by_size[i].nfree = 0;
|
}
|
}
|
}
|
}
|
mstats.stacks_inuse = stacks_inuse;
|
mstats.stacks_inuse = stacks_inuse;
|
mstats.stacks_sys = stacks_sys;
|
mstats.stacks_sys = stacks_sys;
|
}
|
}
|
|
|
void
|
void
|
runtime_gc(int32 force)
|
runtime_gc(int32 force)
|
{
|
{
|
M *m;
|
M *m;
|
int64 t0, t1, t2, t3;
|
int64 t0, t1, t2, t3;
|
uint64 heap0, heap1, obj0, obj1;
|
uint64 heap0, heap1, obj0, obj1;
|
const byte *p;
|
const byte *p;
|
bool extra;
|
bool extra;
|
|
|
// Make sure all registers are saved on stack so that
|
// Make sure all registers are saved on stack so that
|
// scanstack sees them.
|
// scanstack sees them.
|
__builtin_unwind_init();
|
__builtin_unwind_init();
|
|
|
// The gc is turned off (via enablegc) until
|
// The gc is turned off (via enablegc) until
|
// the bootstrap has completed.
|
// the bootstrap has completed.
|
// Also, malloc gets called in the guts
|
// Also, malloc gets called in the guts
|
// of a number of libraries that might be
|
// of a number of libraries that might be
|
// holding locks. To avoid priority inversion
|
// holding locks. To avoid priority inversion
|
// problems, don't bother trying to run gc
|
// problems, don't bother trying to run gc
|
// while holding a lock. The next mallocgc
|
// while holding a lock. The next mallocgc
|
// without a lock will do the gc instead.
|
// without a lock will do the gc instead.
|
m = runtime_m();
|
m = runtime_m();
|
if(!mstats.enablegc || m->locks > 0 || runtime_panicking)
|
if(!mstats.enablegc || m->locks > 0 || runtime_panicking)
|
return;
|
return;
|
|
|
if(gcpercent == -2) { // first time through
|
if(gcpercent == -2) { // first time through
|
p = runtime_getenv("GOGC");
|
p = runtime_getenv("GOGC");
|
if(p == nil || p[0] == '\0')
|
if(p == nil || p[0] == '\0')
|
gcpercent = 100;
|
gcpercent = 100;
|
else if(runtime_strcmp((const char*)p, "off") == 0)
|
else if(runtime_strcmp((const char*)p, "off") == 0)
|
gcpercent = -1;
|
gcpercent = -1;
|
else
|
else
|
gcpercent = runtime_atoi(p);
|
gcpercent = runtime_atoi(p);
|
|
|
p = runtime_getenv("GOGCTRACE");
|
p = runtime_getenv("GOGCTRACE");
|
if(p != nil)
|
if(p != nil)
|
gctrace = runtime_atoi(p);
|
gctrace = runtime_atoi(p);
|
}
|
}
|
if(gcpercent < 0)
|
if(gcpercent < 0)
|
return;
|
return;
|
|
|
runtime_semacquire(&gcsema);
|
runtime_semacquire(&gcsema);
|
if(!force && mstats.heap_alloc < mstats.next_gc) {
|
if(!force && mstats.heap_alloc < mstats.next_gc) {
|
runtime_semrelease(&gcsema);
|
runtime_semrelease(&gcsema);
|
return;
|
return;
|
}
|
}
|
|
|
t0 = runtime_nanotime();
|
t0 = runtime_nanotime();
|
nhandoff = 0;
|
nhandoff = 0;
|
|
|
m->gcing = 1;
|
m->gcing = 1;
|
runtime_stoptheworld();
|
runtime_stoptheworld();
|
|
|
cachestats();
|
cachestats();
|
heap0 = mstats.heap_alloc;
|
heap0 = mstats.heap_alloc;
|
obj0 = mstats.nmalloc - mstats.nfree;
|
obj0 = mstats.nmalloc - mstats.nfree;
|
|
|
runtime_lock(&work.markgate);
|
runtime_lock(&work.markgate);
|
runtime_lock(&work.sweepgate);
|
runtime_lock(&work.sweepgate);
|
|
|
extra = false;
|
extra = false;
|
work.nproc = 1;
|
work.nproc = 1;
|
if(runtime_gomaxprocs > 1 && runtime_ncpu > 1) {
|
if(runtime_gomaxprocs > 1 && runtime_ncpu > 1) {
|
runtime_noteclear(&work.alldone);
|
runtime_noteclear(&work.alldone);
|
work.nproc += runtime_helpgc(&extra);
|
work.nproc += runtime_helpgc(&extra);
|
}
|
}
|
work.nwait = 0;
|
work.nwait = 0;
|
work.ndone = 0;
|
work.ndone = 0;
|
|
|
runtime_unlock(&work.markgate); // let the helpers in
|
runtime_unlock(&work.markgate); // let the helpers in
|
mark(scanblock);
|
mark(scanblock);
|
if(DebugMark)
|
if(DebugMark)
|
mark(debug_scanblock);
|
mark(debug_scanblock);
|
t1 = runtime_nanotime();
|
t1 = runtime_nanotime();
|
|
|
work.spans = runtime_mheap.allspans;
|
work.spans = runtime_mheap.allspans;
|
runtime_unlock(&work.sweepgate); // let the helpers in
|
runtime_unlock(&work.sweepgate); // let the helpers in
|
sweep();
|
sweep();
|
if(work.nproc > 1)
|
if(work.nproc > 1)
|
runtime_notesleep(&work.alldone);
|
runtime_notesleep(&work.alldone);
|
t2 = runtime_nanotime();
|
t2 = runtime_nanotime();
|
|
|
stealcache();
|
stealcache();
|
cachestats();
|
cachestats();
|
|
|
mstats.next_gc = mstats.heap_alloc+mstats.heap_alloc*gcpercent/100;
|
mstats.next_gc = mstats.heap_alloc+mstats.heap_alloc*gcpercent/100;
|
m->gcing = 0;
|
m->gcing = 0;
|
|
|
m->locks++; // disable gc during the mallocs in newproc
|
m->locks++; // disable gc during the mallocs in newproc
|
if(finq != nil) {
|
if(finq != nil) {
|
// kick off or wake up goroutine to run queued finalizers
|
// kick off or wake up goroutine to run queued finalizers
|
if(fing == nil)
|
if(fing == nil)
|
fing = __go_go(runfinq, nil);
|
fing = __go_go(runfinq, nil);
|
else if(fingwait) {
|
else if(fingwait) {
|
fingwait = 0;
|
fingwait = 0;
|
runtime_ready(fing);
|
runtime_ready(fing);
|
}
|
}
|
}
|
}
|
m->locks--;
|
m->locks--;
|
|
|
cachestats();
|
cachestats();
|
heap1 = mstats.heap_alloc;
|
heap1 = mstats.heap_alloc;
|
obj1 = mstats.nmalloc - mstats.nfree;
|
obj1 = mstats.nmalloc - mstats.nfree;
|
|
|
t3 = runtime_nanotime();
|
t3 = runtime_nanotime();
|
mstats.pause_ns[mstats.numgc%nelem(mstats.pause_ns)] = t3 - t0;
|
mstats.pause_ns[mstats.numgc%nelem(mstats.pause_ns)] = t3 - t0;
|
mstats.pause_total_ns += t3 - t0;
|
mstats.pause_total_ns += t3 - t0;
|
mstats.numgc++;
|
mstats.numgc++;
|
if(mstats.debuggc)
|
if(mstats.debuggc)
|
runtime_printf("pause %llu\n", (unsigned long long)t3-t0);
|
runtime_printf("pause %llu\n", (unsigned long long)t3-t0);
|
|
|
if(gctrace) {
|
if(gctrace) {
|
runtime_printf("gc%d(%d): %llu+%llu+%llu ms %llu -> %llu MB %llu -> %llu (%llu-%llu) objects %llu handoff\n",
|
runtime_printf("gc%d(%d): %llu+%llu+%llu ms %llu -> %llu MB %llu -> %llu (%llu-%llu) objects %llu handoff\n",
|
mstats.numgc, work.nproc, (unsigned long long)(t1-t0)/1000000, (unsigned long long)(t2-t1)/1000000, (unsigned long long)(t3-t2)/1000000,
|
mstats.numgc, work.nproc, (unsigned long long)(t1-t0)/1000000, (unsigned long long)(t2-t1)/1000000, (unsigned long long)(t3-t2)/1000000,
|
(unsigned long long)heap0>>20, (unsigned long long)heap1>>20, (unsigned long long)obj0, (unsigned long long)obj1,
|
(unsigned long long)heap0>>20, (unsigned long long)heap1>>20, (unsigned long long)obj0, (unsigned long long)obj1,
|
(unsigned long long) mstats.nmalloc, (unsigned long long)mstats.nfree,
|
(unsigned long long) mstats.nmalloc, (unsigned long long)mstats.nfree,
|
(unsigned long long) nhandoff);
|
(unsigned long long) nhandoff);
|
}
|
}
|
|
|
runtime_semrelease(&gcsema);
|
runtime_semrelease(&gcsema);
|
|
|
// If we could have used another helper proc, start one now,
|
// If we could have used another helper proc, start one now,
|
// in the hope that it will be available next time.
|
// in the hope that it will be available next time.
|
// It would have been even better to start it before the collection,
|
// It would have been even better to start it before the collection,
|
// but doing so requires allocating memory, so it's tricky to
|
// but doing so requires allocating memory, so it's tricky to
|
// coordinate. This lazy approach works out in practice:
|
// coordinate. This lazy approach works out in practice:
|
// we don't mind if the first couple gc rounds don't have quite
|
// we don't mind if the first couple gc rounds don't have quite
|
// the maximum number of procs.
|
// the maximum number of procs.
|
runtime_starttheworld(extra);
|
runtime_starttheworld(extra);
|
|
|
// give the queued finalizers, if any, a chance to run
|
// give the queued finalizers, if any, a chance to run
|
if(finq != nil)
|
if(finq != nil)
|
runtime_gosched();
|
runtime_gosched();
|
|
|
if(gctrace > 1 && !force)
|
if(gctrace > 1 && !force)
|
runtime_gc(1);
|
runtime_gc(1);
|
}
|
}
|
|
|
void runtime_ReadMemStats(MStats *)
|
void runtime_ReadMemStats(MStats *)
|
__asm__("libgo_runtime.runtime.ReadMemStats");
|
__asm__("libgo_runtime.runtime.ReadMemStats");
|
|
|
void
|
void
|
runtime_ReadMemStats(MStats *stats)
|
runtime_ReadMemStats(MStats *stats)
|
{
|
{
|
M *m;
|
M *m;
|
|
|
// Have to acquire gcsema to stop the world,
|
// Have to acquire gcsema to stop the world,
|
// because stoptheworld can only be used by
|
// because stoptheworld can only be used by
|
// one goroutine at a time, and there might be
|
// one goroutine at a time, and there might be
|
// a pending garbage collection already calling it.
|
// a pending garbage collection already calling it.
|
runtime_semacquire(&gcsema);
|
runtime_semacquire(&gcsema);
|
m = runtime_m();
|
m = runtime_m();
|
m->gcing = 1;
|
m->gcing = 1;
|
runtime_stoptheworld();
|
runtime_stoptheworld();
|
cachestats();
|
cachestats();
|
*stats = mstats;
|
*stats = mstats;
|
m->gcing = 0;
|
m->gcing = 0;
|
runtime_semrelease(&gcsema);
|
runtime_semrelease(&gcsema);
|
runtime_starttheworld(false);
|
runtime_starttheworld(false);
|
}
|
}
|
|
|
static void
|
static void
|
runfinq(void* dummy __attribute__ ((unused)))
|
runfinq(void* dummy __attribute__ ((unused)))
|
{
|
{
|
G* gp;
|
G* gp;
|
Finalizer *f;
|
Finalizer *f;
|
FinBlock *fb, *next;
|
FinBlock *fb, *next;
|
uint32 i;
|
uint32 i;
|
|
|
gp = runtime_g();
|
gp = runtime_g();
|
for(;;) {
|
for(;;) {
|
// There's no need for a lock in this section
|
// There's no need for a lock in this section
|
// because it only conflicts with the garbage
|
// because it only conflicts with the garbage
|
// collector, and the garbage collector only
|
// collector, and the garbage collector only
|
// runs when everyone else is stopped, and
|
// runs when everyone else is stopped, and
|
// runfinq only stops at the gosched() or
|
// runfinq only stops at the gosched() or
|
// during the calls in the for loop.
|
// during the calls in the for loop.
|
fb = finq;
|
fb = finq;
|
finq = nil;
|
finq = nil;
|
if(fb == nil) {
|
if(fb == nil) {
|
fingwait = 1;
|
fingwait = 1;
|
gp->status = Gwaiting;
|
gp->status = Gwaiting;
|
gp->waitreason = "finalizer wait";
|
gp->waitreason = "finalizer wait";
|
runtime_gosched();
|
runtime_gosched();
|
continue;
|
continue;
|
}
|
}
|
for(; fb; fb=next) {
|
for(; fb; fb=next) {
|
next = fb->next;
|
next = fb->next;
|
for(i=0; i<(uint32)fb->cnt; i++) {
|
for(i=0; i<(uint32)fb->cnt; i++) {
|
void *params[1];
|
void *params[1];
|
|
|
f = &fb->fin[i];
|
f = &fb->fin[i];
|
params[0] = &f->arg;
|
params[0] = &f->arg;
|
runtime_setblockspecial(f->arg, false);
|
runtime_setblockspecial(f->arg, false);
|
reflect_call(f->ft, (void*)f->fn, 0, 0, params, nil);
|
reflect_call(f->ft, (void*)f->fn, 0, 0, params, nil);
|
f->fn = nil;
|
f->fn = nil;
|
f->arg = nil;
|
f->arg = nil;
|
}
|
}
|
fb->cnt = 0;
|
fb->cnt = 0;
|
fb->next = finc;
|
fb->next = finc;
|
finc = fb;
|
finc = fb;
|
}
|
}
|
runtime_gc(1); // trigger another gc to clean up the finalized objects, if possible
|
runtime_gc(1); // trigger another gc to clean up the finalized objects, if possible
|
}
|
}
|
}
|
}
|
|
|
// mark the block at v of size n as allocated.
|
// mark the block at v of size n as allocated.
|
// If noptr is true, mark it as having no pointers.
|
// If noptr is true, mark it as having no pointers.
|
void
|
void
|
runtime_markallocated(void *v, uintptr n, bool noptr)
|
runtime_markallocated(void *v, uintptr n, bool noptr)
|
{
|
{
|
uintptr *b, obits, bits, off, shift;
|
uintptr *b, obits, bits, off, shift;
|
|
|
// if(0)
|
// if(0)
|
// runtime_printf("markallocated %p+%p\n", v, n);
|
// runtime_printf("markallocated %p+%p\n", v, n);
|
|
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
runtime_throw("markallocated: bad pointer");
|
runtime_throw("markallocated: bad pointer");
|
|
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
|
|
for(;;) {
|
for(;;) {
|
obits = *b;
|
obits = *b;
|
bits = (obits & ~(bitMask<<shift)) | (bitAllocated<<shift);
|
bits = (obits & ~(bitMask<<shift)) | (bitAllocated<<shift);
|
if(noptr)
|
if(noptr)
|
bits |= bitNoPointers<<shift;
|
bits |= bitNoPointers<<shift;
|
if(runtime_singleproc) {
|
if(runtime_singleproc) {
|
*b = bits;
|
*b = bits;
|
break;
|
break;
|
} else {
|
} else {
|
// more than one goroutine is potentially running: use atomic op
|
// more than one goroutine is potentially running: use atomic op
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
break;
|
break;
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
// mark the block at v of size n as freed.
|
// mark the block at v of size n as freed.
|
void
|
void
|
runtime_markfreed(void *v, uintptr n)
|
runtime_markfreed(void *v, uintptr n)
|
{
|
{
|
uintptr *b, obits, bits, off, shift;
|
uintptr *b, obits, bits, off, shift;
|
|
|
// if(0)
|
// if(0)
|
// runtime_printf("markallocated %p+%p\n", v, n);
|
// runtime_printf("markallocated %p+%p\n", v, n);
|
|
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
runtime_throw("markallocated: bad pointer");
|
runtime_throw("markallocated: bad pointer");
|
|
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
|
|
for(;;) {
|
for(;;) {
|
obits = *b;
|
obits = *b;
|
bits = (obits & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
bits = (obits & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
if(runtime_singleproc) {
|
if(runtime_singleproc) {
|
*b = bits;
|
*b = bits;
|
break;
|
break;
|
} else {
|
} else {
|
// more than one goroutine is potentially running: use atomic op
|
// more than one goroutine is potentially running: use atomic op
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
break;
|
break;
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
// check that the block at v of size n is marked freed.
|
// check that the block at v of size n is marked freed.
|
void
|
void
|
runtime_checkfreed(void *v, uintptr n)
|
runtime_checkfreed(void *v, uintptr n)
|
{
|
{
|
uintptr *b, bits, off, shift;
|
uintptr *b, bits, off, shift;
|
|
|
if(!runtime_checking)
|
if(!runtime_checking)
|
return;
|
return;
|
|
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
return; // not allocated, so okay
|
return; // not allocated, so okay
|
|
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start; // word offset
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
|
|
bits = *b>>shift;
|
bits = *b>>shift;
|
if((bits & bitAllocated) != 0) {
|
if((bits & bitAllocated) != 0) {
|
runtime_printf("checkfreed %p+%p: off=%p have=%p\n",
|
runtime_printf("checkfreed %p+%p: off=%p have=%p\n",
|
v, (void*)n, (void*)off, (void*)(bits & bitMask));
|
v, (void*)n, (void*)off, (void*)(bits & bitMask));
|
runtime_throw("checkfreed: not freed");
|
runtime_throw("checkfreed: not freed");
|
}
|
}
|
}
|
}
|
|
|
// mark the span of memory at v as having n blocks of the given size.
|
// mark the span of memory at v as having n blocks of the given size.
|
// if leftover is true, there is left over space at the end of the span.
|
// if leftover is true, there is left over space at the end of the span.
|
void
|
void
|
runtime_markspan(void *v, uintptr size, uintptr n, bool leftover)
|
runtime_markspan(void *v, uintptr size, uintptr n, bool leftover)
|
{
|
{
|
uintptr *b, off, shift;
|
uintptr *b, off, shift;
|
byte *p;
|
byte *p;
|
|
|
if((byte*)v+size*n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
if((byte*)v+size*n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
runtime_throw("markspan: bad pointer");
|
runtime_throw("markspan: bad pointer");
|
|
|
p = v;
|
p = v;
|
if(leftover) // mark a boundary just past end of last block too
|
if(leftover) // mark a boundary just past end of last block too
|
n++;
|
n++;
|
for(; n-- > 0; p += size) {
|
for(; n-- > 0; p += size) {
|
// Okay to use non-atomic ops here, because we control
|
// Okay to use non-atomic ops here, because we control
|
// the entire span, and each bitmap word has bits for only
|
// the entire span, and each bitmap word has bits for only
|
// one span, so no other goroutines are changing these
|
// one span, so no other goroutines are changing these
|
// bitmap words.
|
// bitmap words.
|
off = (uintptr*)p - (uintptr*)runtime_mheap.arena_start; // word offset
|
off = (uintptr*)p - (uintptr*)runtime_mheap.arena_start; // word offset
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
*b = (*b & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
*b = (*b & ~(bitMask<<shift)) | (bitBlockBoundary<<shift);
|
}
|
}
|
}
|
}
|
|
|
// unmark the span of memory at v of length n bytes.
|
// unmark the span of memory at v of length n bytes.
|
void
|
void
|
runtime_unmarkspan(void *v, uintptr n)
|
runtime_unmarkspan(void *v, uintptr n)
|
{
|
{
|
uintptr *p, *b, off;
|
uintptr *p, *b, off;
|
|
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
if((byte*)v+n > (byte*)runtime_mheap.arena_used || (byte*)v < runtime_mheap.arena_start)
|
runtime_throw("markspan: bad pointer");
|
runtime_throw("markspan: bad pointer");
|
|
|
p = v;
|
p = v;
|
off = p - (uintptr*)runtime_mheap.arena_start; // word offset
|
off = p - (uintptr*)runtime_mheap.arena_start; // word offset
|
if(off % wordsPerBitmapWord != 0)
|
if(off % wordsPerBitmapWord != 0)
|
runtime_throw("markspan: unaligned pointer");
|
runtime_throw("markspan: unaligned pointer");
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
n /= PtrSize;
|
n /= PtrSize;
|
if(n%wordsPerBitmapWord != 0)
|
if(n%wordsPerBitmapWord != 0)
|
runtime_throw("unmarkspan: unaligned length");
|
runtime_throw("unmarkspan: unaligned length");
|
// Okay to use non-atomic ops here, because we control
|
// Okay to use non-atomic ops here, because we control
|
// the entire span, and each bitmap word has bits for only
|
// the entire span, and each bitmap word has bits for only
|
// one span, so no other goroutines are changing these
|
// one span, so no other goroutines are changing these
|
// bitmap words.
|
// bitmap words.
|
n /= wordsPerBitmapWord;
|
n /= wordsPerBitmapWord;
|
while(n-- > 0)
|
while(n-- > 0)
|
*b-- = 0;
|
*b-- = 0;
|
}
|
}
|
|
|
bool
|
bool
|
runtime_blockspecial(void *v)
|
runtime_blockspecial(void *v)
|
{
|
{
|
uintptr *b, off, shift;
|
uintptr *b, off, shift;
|
|
|
if(DebugMark)
|
if(DebugMark)
|
return true;
|
return true;
|
|
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start;
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
|
|
return (*b & (bitSpecial<<shift)) != 0;
|
return (*b & (bitSpecial<<shift)) != 0;
|
}
|
}
|
|
|
void
|
void
|
runtime_setblockspecial(void *v, bool s)
|
runtime_setblockspecial(void *v, bool s)
|
{
|
{
|
uintptr *b, off, shift, bits, obits;
|
uintptr *b, off, shift, bits, obits;
|
|
|
if(DebugMark)
|
if(DebugMark)
|
return;
|
return;
|
|
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start;
|
off = (uintptr*)v - (uintptr*)runtime_mheap.arena_start;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
b = (uintptr*)runtime_mheap.arena_start - off/wordsPerBitmapWord - 1;
|
shift = off % wordsPerBitmapWord;
|
shift = off % wordsPerBitmapWord;
|
|
|
for(;;) {
|
for(;;) {
|
obits = *b;
|
obits = *b;
|
if(s)
|
if(s)
|
bits = obits | (bitSpecial<<shift);
|
bits = obits | (bitSpecial<<shift);
|
else
|
else
|
bits = obits & ~(bitSpecial<<shift);
|
bits = obits & ~(bitSpecial<<shift);
|
if(runtime_singleproc) {
|
if(runtime_singleproc) {
|
*b = bits;
|
*b = bits;
|
break;
|
break;
|
} else {
|
} else {
|
// more than one goroutine is potentially running: use atomic op
|
// more than one goroutine is potentially running: use atomic op
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
if(runtime_casp((void**)b, (void*)obits, (void*)bits))
|
break;
|
break;
|
}
|
}
|
}
|
}
|
}
|
}
|
|
|
void
|
void
|
runtime_MHeap_MapBits(MHeap *h)
|
runtime_MHeap_MapBits(MHeap *h)
|
{
|
{
|
// Caller has added extra mappings to the arena.
|
// Caller has added extra mappings to the arena.
|
// Add extra mappings of bitmap words as needed.
|
// Add extra mappings of bitmap words as needed.
|
// We allocate extra bitmap pieces in chunks of bitmapChunk.
|
// We allocate extra bitmap pieces in chunks of bitmapChunk.
|
enum {
|
enum {
|
bitmapChunk = 8192
|
bitmapChunk = 8192
|
};
|
};
|
uintptr n;
|
uintptr n;
|
|
|
n = (h->arena_used - h->arena_start) / wordsPerBitmapWord;
|
n = (h->arena_used - h->arena_start) / wordsPerBitmapWord;
|
n = (n+bitmapChunk-1) & ~(bitmapChunk-1);
|
n = (n+bitmapChunk-1) & ~(bitmapChunk-1);
|
if(h->bitmap_mapped >= n)
|
if(h->bitmap_mapped >= n)
|
return;
|
return;
|
|
|
runtime_SysMap(h->arena_start - n, n - h->bitmap_mapped);
|
runtime_SysMap(h->arena_start - n, n - h->bitmap_mapped);
|
h->bitmap_mapped = n;
|
h->bitmap_mapped = n;
|
}
|
}
|
|
|
