// reimburseSweepCredit records that unusableBytes bytes of a
// just-allocated span are not available for object allocation. This
// offsets the worst-case charge performed by deductSweepCredit.
func reimburseSweepCredit(unusableBytes uintptr) {
if mheap_.sweepPagesPerByte == 0 {
// Nobody cares about the credit. Avoid the atomic.
return
}
atomic.Xadd64(&mheap_.spanBytesAlloc, -int64(unusableBytes))
}
// If gcBlackenPromptly is true we are in the second mark phase phase so we allocate black.
//go:nowritebarrier
func gcmarknewobject_m(obj, size uintptr) {
if useCheckmark && !gcBlackenPromptly { // The world should be stopped so this should not happen.
throw("gcmarknewobject called while doing checkmark")
}
heapBitsForAddr(obj).setMarked()
atomic.Xadd64(&work.bytesMarked, int64(size))
}
// dispose returns any cached pointers to the global queue.
// The buffers are being put on the full queue so that the
// write barriers will not simply reacquire them before the
// GC can inspect them. This helps reduce the mutator's
// ability to hide pointers during the concurrent mark phase.
//
//go:nowritebarrier
func (w *gcWork) dispose() {
if wbuf := w.wbuf1.ptr(); wbuf != nil {
if wbuf.nobj == 0 {
putempty(wbuf, 212)
} else {
putfull(wbuf, 214)
}
w.wbuf1 = 0
wbuf = w.wbuf2.ptr()
if wbuf.nobj == 0 {
putempty(wbuf, 218)
} else {
putfull(wbuf, 220)
}
w.wbuf2 = 0
}
if w.bytesMarked != 0 {
// dispose happens relatively infrequently. If this
// atomic becomes a problem, we should first try to
// dispose less and if necessary aggregate in a per-P
// counter.
atomic.Xadd64(&work.bytesMarked, int64(w.bytesMarked))
w.bytesMarked = 0
}
if w.scanWork != 0 {
atomic.Xaddint64(&gcController.scanWork, w.scanWork)
w.scanWork = 0
}
}
// Variant of sync/atomic's TestUnaligned64:
func TestUnaligned64(t *testing.T) {
// Unaligned 64-bit atomics on 32-bit systems are
// a continual source of pain. Test that on 32-bit systems they crash
// instead of failing silently.
switch runtime.GOARCH {
default:
if unsafe.Sizeof(int(0)) != 4 {
t.Skip("test only runs on 32-bit systems")
}
case "amd64p32":
// amd64p32 can handle unaligned atomics.
t.Skipf("test not needed on %v", runtime.GOARCH)
}
x := make([]uint32, 4)
up64 := (*uint64)(unsafe.Pointer(&x[1])) // misaligned
p64 := (*int64)(unsafe.Pointer(&x[1])) // misaligned
shouldPanic(t, "Load64", func() { atomic.Load64(up64) })
shouldPanic(t, "Loadint64", func() { atomic.Loadint64(p64) })
shouldPanic(t, "Store64", func() { atomic.Store64(up64, 0) })
shouldPanic(t, "Xadd64", func() { atomic.Xadd64(up64, 1) })
shouldPanic(t, "Xchg64", func() { atomic.Xchg64(up64, 1) })
shouldPanic(t, "Cas64", func() { atomic.Cas64(up64, 1, 2) })
}
// Atomically decreases a given *system* memory stat. Same comments as
// mSysStatInc apply.
//go:nosplit
func mSysStatDec(sysStat *uint64, n uintptr) {
if sys.BigEndian != 0 {
atomic.Xadd64(sysStat, -int64(n))
return
}
if val := atomic.Xadduintptr((*uintptr)(unsafe.Pointer(sysStat)), uintptr(-int64(n))); val+n < n {
print("runtime: stat underflow: val ", val, ", n ", n, "\n")
exit(2)
}
}
// Atomically increases a given *system* memory stat. We are counting on this
// stat never overflowing a uintptr, so this function must only be used for
// system memory stats.
//
// The current implementation for little endian architectures is based on
// xadduintptr(), which is less than ideal: xadd64() should really be used.
// Using xadduintptr() is a stop-gap solution until arm supports xadd64() that
// doesn't use locks. (Locks are a problem as they require a valid G, which
// restricts their useability.)
//
// A side-effect of using xadduintptr() is that we need to check for
// overflow errors.
//go:nosplit
func mSysStatInc(sysStat *uint64, n uintptr) {
if _BigEndian != 0 {
atomic.Xadd64(sysStat, int64(n))
return
}
if val := atomic.Xadduintptr((*uintptr)(unsafe.Pointer(sysStat)), n); val < n {
print("runtime: stat overflow: val ", val, ", n ", n, "\n")
exit(2)
}
}
// deductSweepCredit deducts sweep credit for allocating a span of
// size spanBytes. This must be performed *before* the span is
// allocated to ensure the system has enough credit. If necessary, it
// performs sweeping to prevent going in to debt. If the caller will
// also sweep pages (e.g., for a large allocation), it can pass a
// non-zero callerSweepPages to leave that many pages unswept.
//
// deductSweepCredit makes a worst-case assumption that all spanBytes
// bytes of the ultimately allocated span will be available for object
// allocation. The caller should call reimburseSweepCredit if that
// turns out not to be the case once the span is allocated.
//
// deductSweepCredit is the core of the "proportional sweep" system.
// It uses statistics gathered by the garbage collector to perform
// enough sweeping so that all pages are swept during the concurrent
// sweep phase between GC cycles.
//
// mheap_ must NOT be locked.
func deductSweepCredit(spanBytes uintptr, callerSweepPages uintptr) {
if mheap_.sweepPagesPerByte == 0 {
// Proportional sweep is done or disabled.
return
}
// Account for this span allocation.
spanBytesAlloc := atomic.Xadd64(&mheap_.spanBytesAlloc, int64(spanBytes))
// Fix debt if necessary.
pagesOwed := int64(mheap_.sweepPagesPerByte * float64(spanBytesAlloc))
for pagesOwed-int64(atomic.Load64(&mheap_.pagesSwept)) > int64(callerSweepPages) {
if gosweepone() == ^uintptr(0) {
mheap_.sweepPagesPerByte = 0
break
}
}
}
// Return span from an MCache.
func (c *mcentral) uncacheSpan(s *mspan) {
lock(&c.lock)
s.incache = false
if s.allocCount == 0 {
throw("uncaching span but s.allocCount == 0")
}
cap := int32((s.npages << _PageShift) / s.elemsize)
n := cap - int32(s.allocCount)
if n > 0 {
c.empty.remove(s)
c.nonempty.insert(s)
// mCentral_CacheSpan conservatively counted
// unallocated slots in heap_live. Undo this.
atomic.Xadd64(&memstats.heap_live, -int64(n)*int64(s.elemsize))
}
unlock(&c.lock)
}
func testAtomic64() {
test_z64 = 42
test_x64 = 0
prefetcht0(uintptr(unsafe.Pointer(&test_z64)))
prefetcht1(uintptr(unsafe.Pointer(&test_z64)))
prefetcht2(uintptr(unsafe.Pointer(&test_z64)))
prefetchnta(uintptr(unsafe.Pointer(&test_z64)))
if atomic.Cas64(&test_z64, test_x64, 1) {
throw("cas64 failed")
}
if test_x64 != 0 {
throw("cas64 failed")
}
test_x64 = 42
if !atomic.Cas64(&test_z64, test_x64, 1) {
throw("cas64 failed")
}
if test_x64 != 42 || test_z64 != 1 {
throw("cas64 failed")
}
if atomic.Load64(&test_z64) != 1 {
throw("load64 failed")
}
atomic.Store64(&test_z64, (1<<40)+1)
if atomic.Load64(&test_z64) != (1<<40)+1 {
throw("store64 failed")
}
if atomic.Xadd64(&test_z64, (1<<40)+1) != (2<<40)+2 {
throw("xadd64 failed")
}
if atomic.Load64(&test_z64) != (2<<40)+2 {
throw("xadd64 failed")
}
if atomic.Xchg64(&test_z64, (3<<40)+3) != (2<<40)+2 {
throw("xchg64 failed")
}
if atomic.Load64(&test_z64) != (3<<40)+3 {
throw("xchg64 failed")
}
}
// oneNewExtraM allocates an m and puts it on the extra list.
func oneNewExtraM() {
// Create extra goroutine locked to extra m.
// The goroutine is the context in which the cgo callback will run.
// The sched.pc will never be returned to, but setting it to
// goexit makes clear to the traceback routines where
// the goroutine stack ends.
var g0SP unsafe.Pointer
var g0SPSize uintptr
mp := allocm(nil, true, &g0SP, &g0SPSize)
gp := malg(true, false, nil, nil)
gp.gcscanvalid = true // fresh G, so no dequeueRescan necessary
gp.gcRescan = -1
// malg returns status as Gidle, change to Gdead before adding to allg
// where GC will see it.
// gccgo uses Gdead here, not Gsyscall, because the split
// stack context is not initialized.
casgstatus(gp, _Gidle, _Gdead)
gp.m = mp
mp.curg = gp
mp.locked = _LockInternal
mp.lockedg = gp
gp.lockedm = mp
gp.goid = int64(atomic.Xadd64(&sched.goidgen, 1))
if raceenabled {
gp.racectx = racegostart(funcPC(newextram))
}
// put on allg for garbage collector
allgadd(gp)
// The context for gp will be set up in needm.
// Here we need to set the context for g0.
makeGContext(mp.g0, g0SP, g0SPSize)
// Add m to the extra list.
mnext := lockextra(true)
mp.schedlink.set(mnext)
unlockextra(mp)
}
func parfordo(desc *parfor) {
// Obtain 0-based thread index.
tid := atomic.Xadd(&desc.thrseq, 1) - 1
if tid >= desc.nthr {
print("tid=", tid, " nthr=", desc.nthr, "\n")
throw("parfor: invalid tid")
}
// If single-threaded, just execute the for serially.
body := desc.body
if desc.nthr == 1 {
for i := uint32(0); i < desc.cnt; i++ {
body(desc, i)
}
return
}
me := &desc.thr[tid]
mypos := &me.pos
for {
for {
// While there is local work,
// bump low index and execute the iteration.
pos := atomic.Xadd64(mypos, 1)
begin := uint32(pos) - 1
end := uint32(pos >> 32)
if begin < end {
body(desc, begin)
continue
}
break
}
// Out of work, need to steal something.
idle := false
for try := uint32(0); ; try++ {
// If we don't see any work for long enough,
// increment the done counter...
if try > desc.nthr*4 && !idle {
idle = true
atomic.Xadd(&desc.done, 1)
}
// ...if all threads have incremented the counter,
// we are done.
extra := uint32(0)
if !idle {
extra = 1
}
if desc.done+extra == desc.nthr {
if !idle {
atomic.Xadd(&desc.done, 1)
}
goto exit
}
// Choose a random victim for stealing.
var begin, end uint32
victim := fastrand1() % (desc.nthr - 1)
if victim >= tid {
victim++
}
victimpos := &desc.thr[victim].pos
for {
// See if it has any work.
pos := atomic.Load64(victimpos)
begin = uint32(pos)
end = uint32(pos >> 32)
if begin+1 >= end {
end = 0
begin = end
break
}
if idle {
atomic.Xadd(&desc.done, -1)
idle = false
}
begin2 := begin + (end-begin)/2
newpos := uint64(begin) | uint64(begin2)<<32
if atomic.Cas64(victimpos, pos, newpos) {
begin = begin2
break
}
}
if begin < end {
// Has successfully stolen some work.
if idle {
throw("parfor: should not be idle")
}
atomic.Store64(mypos, uint64(begin)|uint64(end)<<32)
me.nsteal++
me.nstealcnt += uint64(end) - uint64(begin)
break
}
// Backoff.
if try < desc.nthr {
// nothing
} else if try < 4*desc.nthr {
me.nprocyield++
procyield(20)
} else if !desc.wait {
// If a caller asked not to wait for the others, exit now
// (assume that most work is already done at this point).
if !idle {
atomic.Xadd(&desc.done, 1)
}
goto exit
} else if try < 6*desc.nthr {
me.nosyield++
osyield()
} else {
me.nsleep++
usleep(1)
}
}
}
exit:
atomic.Xadd64(&desc.nsteal, int64(me.nsteal))
atomic.Xadd64(&desc.nstealcnt, int64(me.nstealcnt))
atomic.Xadd64(&desc.nprocyield, int64(me.nprocyield))
atomic.Xadd64(&desc.nosyield, int64(me.nosyield))
atomic.Xadd64(&desc.nsleep, int64(me.nsleep))
me.nsteal = 0
me.nstealcnt = 0
me.nprocyield = 0
me.nosyield = 0
me.nsleep = 0
}
// 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.
// Returns true if the span was returned to heap.
// If preserve=true, don't return it to heap nor relink in MCentral lists;
// caller takes care of it.
//TODO go:nowritebarrier
func (s *mspan) sweep(preserve bool) bool {
// It's critical that we enter this function with preemption disabled,
// GC must not start while we are in the middle of this function.
_g_ := getg()
if _g_.m.locks == 0 && _g_.m.mallocing == 0 && _g_ != _g_.m.g0 {
throw("MSpan_Sweep: m is not locked")
}
sweepgen := mheap_.sweepgen
if s.state != mSpanInUse || s.sweepgen != sweepgen-1 {
print("MSpan_Sweep: state=", s.state, " sweepgen=", s.sweepgen, " mheap.sweepgen=", sweepgen, "\n")
throw("MSpan_Sweep: bad span state")
}
if trace.enabled {
traceGCSweepStart()
}
atomic.Xadd64(&mheap_.pagesSwept, int64(s.npages))
cl := s.sizeclass
size := s.elemsize
res := false
nfree := 0
c := _g_.m.mcache
freeToHeap := false
// The allocBits indicate which unmarked objects don't need to be
// processed since they were free at the end of the last GC cycle
// and were not allocated since then.
// If the allocBits index is >= s.freeindex and the bit
// is not marked then the object remains unallocated
// since the last GC.
// This situation is analogous to being on a freelist.
// Unlink & free special records for any objects we're about to free.
// Two complications here:
// 1. An object can have both finalizer and profile special records.
// In such case we need to queue finalizer for execution,
// mark the object as live and preserve the profile special.
// 2. A tiny object can have several finalizers setup for different offsets.
// If such object is not marked, we need to queue all finalizers at once.
// Both 1 and 2 are possible at the same time.
specialp := &s.specials
special := *specialp
for special != nil {
// A finalizer can be set for an inner byte of an object, find object beginning.
objIndex := uintptr(special.offset) / size
p := s.base() + objIndex*size
mbits := s.markBitsForIndex(objIndex)
if !mbits.isMarked() {
// This object is not marked and has at least one special record.
// Pass 1: see if it has at least one finalizer.
hasFin := false
endOffset := p - s.base() + size
for tmp := special; tmp != nil && uintptr(tmp.offset) < endOffset; tmp = tmp.next {
if tmp.kind == _KindSpecialFinalizer {
// Stop freeing of object if it has a finalizer.
mbits.setMarkedNonAtomic()
hasFin = true
break
}
}
// Pass 2: queue all finalizers _or_ handle profile record.
for special != nil && uintptr(special.offset) < endOffset {
// Find the exact byte for which the special was setup
// (as opposed to object beginning).
p := s.base() + uintptr(special.offset)
if special.kind == _KindSpecialFinalizer || !hasFin {
// Splice out special record.
y := special
special = special.next
*specialp = special
freespecial(y, unsafe.Pointer(p), size)
} else {
// This is profile record, but the object has finalizers (so kept alive).
// Keep special record.
specialp = &special.next
special = *specialp
}
}
} else {
// object is still live: keep special record
specialp = &special.next
special = *specialp
}
}
if debug.allocfreetrace != 0 || raceenabled || msanenabled {
// Find all newly freed objects. This doesn't have to
// efficient; allocfreetrace has massive overhead.
mbits := s.markBitsForBase()
abits := s.allocBitsForIndex(0)
for i := uintptr(0); i < s.nelems; i++ {
if !mbits.isMarked() && (abits.index < s.freeindex || abits.isMarked()) {
x := s.base() + i*s.elemsize
if debug.allocfreetrace != 0 {
tracefree(unsafe.Pointer(x), size)
}
if raceenabled {
racefree(unsafe.Pointer(x), size)
}
if msanenabled {
msanfree(unsafe.Pointer(x), size)
}
}
mbits.advance()
abits.advance()
}
}
// Count the number of free objects in this span.
nfree = s.countFree()
if cl == 0 && nfree != 0 {
s.needzero = 1
freeToHeap = true
}
nalloc := uint16(s.nelems) - uint16(nfree)
nfreed := s.allocCount - nalloc
if nalloc > s.allocCount {
print("runtime: nelems=", s.nelems, " nfree=", nfree, " nalloc=", nalloc, " previous allocCount=", s.allocCount, " nfreed=", nfreed, "\n")
throw("sweep increased allocation count")
}
s.allocCount = nalloc
wasempty := s.nextFreeIndex() == s.nelems
s.freeindex = 0 // reset allocation index to start of span.
// gcmarkBits becomes the allocBits.
// get a fresh cleared gcmarkBits in preparation for next GC
s.allocBits = s.gcmarkBits
s.gcmarkBits = newMarkBits(s.nelems)
// Initialize alloc bits cache.
s.refillAllocCache(0)
// We need to set s.sweepgen = h.sweepgen only when all blocks are swept,
// because of the potential for a concurrent free/SetFinalizer.
// But we need to set it before we make the span available for allocation
// (return it to heap or mcentral), because allocation code assumes that a
// span is already swept if available for allocation.
if freeToHeap || nfreed == 0 {
// The span must be in our exclusive ownership until we update sweepgen,
// check for potential races.
if s.state != mSpanInUse || s.sweepgen != sweepgen-1 {
print("MSpan_Sweep: state=", s.state, " sweepgen=", s.sweepgen, " mheap.sweepgen=", sweepgen, "\n")
throw("MSpan_Sweep: bad span state after sweep")
}
// Serialization point.
// At this point the mark bits are cleared and allocation ready
// to go so release the span.
atomic.Store(&s.sweepgen, sweepgen)
}
if nfreed > 0 && cl != 0 {
c.local_nsmallfree[cl] += uintptr(nfreed)
res = mheap_.central[cl].mcentral.freeSpan(s, preserve, wasempty)
// MCentral_FreeSpan updates sweepgen
} else if freeToHeap {
// Free large span to heap
// NOTE(rsc,dvyukov): The original implementation of efence
// in CL 22060046 used SysFree instead of SysFault, so that
// the operating system would eventually give the memory
// back to us again, so that an efence program could run
// longer without running out of memory. Unfortunately,
// calling SysFree here without any kind of adjustment of the
// heap data structures means that when the memory does
// come back to us, we have the wrong metadata for it, either in
// the MSpan structures or in the garbage collection bitmap.
// Using SysFault here means that the program will run out of
// memory fairly quickly in efence mode, but at least it won't
// have mysterious crashes due to confused memory reuse.
// It should be possible to switch back to SysFree if we also
// implement and then call some kind of MHeap_DeleteSpan.
if debug.efence > 0 {
s.limit = 0 // prevent mlookup from finding this span
sysFault(unsafe.Pointer(s.base()), size)
} else {
mheap_.freeSpan(s, 1)
}
c.local_nlargefree++
c.local_largefree += size
res = true
}
if !res {
// The span has been swept and is still in-use, so put
// it on the swept in-use list.
mheap_.sweepSpans[sweepgen/2%2].push(s)
}
if trace.enabled {
traceGCSweepDone()
}
return res
}
// Allocate a new span of npage pages from the heap for GC'd memory
// and record its size class in the HeapMap and HeapMapCache.
func (h *mheap) alloc_m(npage uintptr, sizeclass int32, large bool) *mspan {
_g_ := getg()
if _g_ != _g_.m.g0 {
throw("_mheap_alloc not on g0 stack")
}
lock(&h.lock)
// To prevent excessive heap growth, before allocating n pages
// we need to sweep and reclaim at least n pages.
if h.sweepdone == 0 {
// TODO(austin): This tends to sweep a large number of
// spans in order to find a few completely free spans
// (for example, in the garbage benchmark, this sweeps
// ~30x the number of pages its trying to allocate).
// If GC kept a bit for whether there were any marks
// in a span, we could release these free spans
// at the end of GC and eliminate this entirely.
h.reclaim(npage)
}
// transfer stats from cache to global
memstats.heap_scan += uint64(_g_.m.mcache.local_scan)
_g_.m.mcache.local_scan = 0
memstats.tinyallocs += uint64(_g_.m.mcache.local_tinyallocs)
_g_.m.mcache.local_tinyallocs = 0
s := h.allocSpanLocked(npage)
if s != nil {
// Record span info, because gc needs to be
// able to map interior pointer to containing span.
atomic.Store(&s.sweepgen, h.sweepgen)
s.state = _MSpanInUse
s.allocCount = 0
s.sizeclass = uint8(sizeclass)
if sizeclass == 0 {
s.elemsize = s.npages << _PageShift
s.divShift = 0
s.divMul = 0
s.divShift2 = 0
s.baseMask = 0
} else {
s.elemsize = uintptr(class_to_size[sizeclass])
m := &class_to_divmagic[sizeclass]
s.divShift = m.shift
s.divMul = m.mul
s.divShift2 = m.shift2
s.baseMask = m.baseMask
}
// update stats, sweep lists
h.pagesInUse += uint64(npage)
if large {
memstats.heap_objects++
atomic.Xadd64(&memstats.heap_live, int64(npage<<_PageShift))
// Swept spans are at the end of lists.
if s.npages < uintptr(len(h.free)) {
h.busy[s.npages].insertBack(s)
} else {
h.busylarge.insertBack(s)
}
}
}
// heap_scan and heap_live were updated.
if gcBlackenEnabled != 0 {
gcController.revise()
}
if trace.enabled {
traceHeapAlloc()
}
// h_spans is accessed concurrently without synchronization
// from other threads. Hence, there must be a store/store
// barrier here to ensure the writes to h_spans above happen
// before the caller can publish a pointer p to an object
// allocated from s. As soon as this happens, the garbage
// collector running on another processor could read p and
// look up s in h_spans. The unlock acts as the barrier to
// order these writes. On the read side, the data dependency
// between p and the index in h_spans orders the reads.
unlock(&h.lock)
return s
}
// 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.
// Returns true if the span was returned to heap.
// If preserve=true, don't return it to heap nor relink in MCentral lists;
// caller takes care of it.
//TODO go:nowritebarrier
func (s *mspan) sweep(preserve bool) bool {
// It's critical that we enter this function with preemption disabled,
// GC must not start while we are in the middle of this function.
_g_ := getg()
if _g_.m.locks == 0 && _g_.m.mallocing == 0 && _g_ != _g_.m.g0 {
throw("MSpan_Sweep: m is not locked")
}
sweepgen := mheap_.sweepgen
if s.state != mSpanInUse || s.sweepgen != sweepgen-1 {
print("MSpan_Sweep: state=", s.state, " sweepgen=", s.sweepgen, " mheap.sweepgen=", sweepgen, "\n")
throw("MSpan_Sweep: bad span state")
}
if trace.enabled {
traceGCSweepStart()
}
atomic.Xadd64(&mheap_.pagesSwept, int64(s.npages))
cl := s.sizeclass
size := s.elemsize
res := false
nfree := 0
var head, end gclinkptr
c := _g_.m.mcache
freeToHeap := false
// Mark any free objects in this span so we don't collect them.
sstart := uintptr(s.start << _PageShift)
for link := s.freelist; link.ptr() != nil; link = link.ptr().next {
if uintptr(link) < sstart || s.limit <= uintptr(link) {
// Free list is corrupted.
dumpFreeList(s)
throw("free list corrupted")
}
heapBitsForAddr(uintptr(link)).setMarkedNonAtomic()
}
// Unlink & free special records for any objects we're about to free.
// Two complications here:
// 1. An object can have both finalizer and profile special records.
// In such case we need to queue finalizer for execution,
// mark the object as live and preserve the profile special.
// 2. A tiny object can have several finalizers setup for different offsets.
// If such object is not marked, we need to queue all finalizers at once.
// Both 1 and 2 are possible at the same time.
specialp := &s.specials
special := *specialp
for special != nil {
// A finalizer can be set for an inner byte of an object, find object beginning.
p := uintptr(s.start<<_PageShift) + uintptr(special.offset)/size*size
hbits := heapBitsForAddr(p)
if !hbits.isMarked() {
// This object is not marked and has at least one special record.
// Pass 1: see if it has at least one finalizer.
hasFin := false
endOffset := p - uintptr(s.start<<_PageShift) + size
for tmp := special; tmp != nil && uintptr(tmp.offset) < endOffset; tmp = tmp.next {
if tmp.kind == _KindSpecialFinalizer {
// Stop freeing of object if it has a finalizer.
hbits.setMarkedNonAtomic()
hasFin = true
break
}
}
// Pass 2: queue all finalizers _or_ handle profile record.
for special != nil && uintptr(special.offset) < endOffset {
// Find the exact byte for which the special was setup
// (as opposed to object beginning).
p := uintptr(s.start<<_PageShift) + uintptr(special.offset)
if special.kind == _KindSpecialFinalizer || !hasFin {
// Splice out special record.
y := special
special = special.next
*specialp = special
freespecial(y, unsafe.Pointer(p), size)
} else {
// This is profile record, but the object has finalizers (so kept alive).
// Keep special record.
specialp = &special.next
special = *specialp
}
}
} else {
// object is still live: keep special record
specialp = &special.next
special = *specialp
}
}
// Sweep through n objects of given size starting at p.
// This thread owns the span now, so it can manipulate
// the block bitmap without atomic operations.
size, n, _ := s.layout()
heapBitsSweepSpan(s.base(), size, n, func(p uintptr) {
// At this point we know that we are looking at garbage object
// that needs to be collected.
if debug.allocfreetrace != 0 {
tracefree(unsafe.Pointer(p), size)
}
if msanenabled {
msanfree(unsafe.Pointer(p), size)
}
// Reset to allocated+noscan.
if cl == 0 {
// Free large span.
if preserve {
throw("can't preserve large span")
}
heapBitsForSpan(p).initSpan(s.layout())
s.needzero = 1
// Free the span after heapBitsSweepSpan
// returns, since it's not done with the span.
freeToHeap = true
} else {
// Free small object.
if size > 2*ptrSize {
*(*uintptr)(unsafe.Pointer(p + ptrSize)) = uintptrMask & 0xdeaddeaddeaddead // mark as "needs to be zeroed"
} else if size > ptrSize {
*(*uintptr)(unsafe.Pointer(p + ptrSize)) = 0
}
if head.ptr() == nil {
head = gclinkptr(p)
} else {
end.ptr().next = gclinkptr(p)
}
end = gclinkptr(p)
end.ptr().next = gclinkptr(0x0bade5)
nfree++
}
})
// We need to set s.sweepgen = h.sweepgen only when all blocks are swept,
// because of the potential for a concurrent free/SetFinalizer.
// But we need to set it before we make the span available for allocation
// (return it to heap or mcentral), because allocation code assumes that a
// span is already swept if available for allocation.
if freeToHeap || nfree == 0 {
// The span must be in our exclusive ownership until we update sweepgen,
// check for potential races.
if s.state != mSpanInUse || s.sweepgen != sweepgen-1 {
print("MSpan_Sweep: state=", s.state, " sweepgen=", s.sweepgen, " mheap.sweepgen=", sweepgen, "\n")
throw("MSpan_Sweep: bad span state after sweep")
}
atomic.Store(&s.sweepgen, sweepgen)
}
if nfree > 0 {
c.local_nsmallfree[cl] += uintptr(nfree)
res = mheap_.central[cl].mcentral.freeSpan(s, int32(nfree), head, end, preserve)
// MCentral_FreeSpan updates sweepgen
} else if freeToHeap {
// Free large span to heap
// NOTE(rsc,dvyukov): The original implementation of efence
// in CL 22060046 used SysFree instead of SysFault, so that
// the operating system would eventually give the memory
// back to us again, so that an efence program could run
// longer without running out of memory. Unfortunately,
// calling SysFree here without any kind of adjustment of the
// heap data structures means that when the memory does
// come back to us, we have the wrong metadata for it, either in
// the MSpan structures or in the garbage collection bitmap.
// Using SysFault here means that the program will run out of
// memory fairly quickly in efence mode, but at least it won't
// have mysterious crashes due to confused memory reuse.
// It should be possible to switch back to SysFree if we also
// implement and then call some kind of MHeap_DeleteSpan.
if debug.efence > 0 {
s.limit = 0 // prevent mlookup from finding this span
sysFault(unsafe.Pointer(uintptr(s.start<<_PageShift)), size)
} else {
mheap_.freeSpan(s, 1)
}
c.local_nlargefree++
c.local_largefree += size
res = true
}
if trace.enabled {
traceGCSweepDone()
}
return res
}
// Allocate a span to use in an MCache.
func (c *mcentral) cacheSpan() *mspan {
// Deduct credit for this span allocation and sweep if necessary.
spanBytes := uintptr(class_to_allocnpages[c.sizeclass]) * _PageSize
deductSweepCredit(spanBytes, 0)
lock(&c.lock)
sg := mheap_.sweepgen
retry:
var s *mspan
for s = c.nonempty.first; s != nil; s = s.next {
if s.sweepgen == sg-2 && atomic.Cas(&s.sweepgen, sg-2, sg-1) {
c.nonempty.remove(s)
c.empty.insertBack(s)
unlock(&c.lock)
s.sweep(true)
goto havespan
}
if s.sweepgen == sg-1 {
// the span is being swept by background sweeper, skip
continue
}
// we have a nonempty span that does not require sweeping, allocate from it
c.nonempty.remove(s)
c.empty.insertBack(s)
unlock(&c.lock)
goto havespan
}
for s = c.empty.first; s != nil; s = s.next {
if s.sweepgen == sg-2 && atomic.Cas(&s.sweepgen, sg-2, sg-1) {
// we have an empty span that requires sweeping,
// sweep it and see if we can free some space in it
c.empty.remove(s)
// swept spans are at the end of the list
c.empty.insertBack(s)
unlock(&c.lock)
s.sweep(true)
freeIndex := s.nextFreeIndex()
if freeIndex != s.nelems {
s.freeindex = freeIndex
goto havespan
}
lock(&c.lock)
// the span is still empty after sweep
// it is already in the empty list, so just retry
goto retry
}
if s.sweepgen == sg-1 {
// the span is being swept by background sweeper, skip
continue
}
// already swept empty span,
// all subsequent ones must also be either swept or in process of sweeping
break
}
unlock(&c.lock)
// Replenish central list if empty.
s = c.grow()
if s == nil {
return nil
}
lock(&c.lock)
c.empty.insertBack(s)
unlock(&c.lock)
// At this point s is a non-empty span, queued at the end of the empty list,
// c is unlocked.
havespan:
cap := int32((s.npages << _PageShift) / s.elemsize)
n := cap - int32(s.allocCount)
if n == 0 || s.freeindex == s.nelems || uintptr(s.allocCount) == s.nelems {
throw("span has no free objects")
}
usedBytes := uintptr(s.allocCount) * s.elemsize
if usedBytes > 0 {
reimburseSweepCredit(usedBytes)
}
atomic.Xadd64(&memstats.heap_live, int64(spanBytes)-int64(usedBytes))
if trace.enabled {
// heap_live changed.
traceHeapAlloc()
}
if gcBlackenEnabled != 0 {
// heap_live changed.
gcController.revise()
}
s.incache = true
freeByteBase := s.freeindex &^ (64 - 1)
whichByte := freeByteBase / 8
// Init alloc bits cache.
s.refillAllocCache(whichByte)
// Adjust the allocCache so that s.freeindex corresponds to the low bit in
// s.allocCache.
s.allocCache >>= s.freeindex % 64
return s
}
