Source file src/runtime/map.go

     1  // Copyright 2014 The Go Authors. All rights reserved.
     2  // Use of this source code is governed by a BSD-style
     3  // license that can be found in the LICENSE file.
     4  
     5  package runtime
     6  
     7  // This file contains the implementation of Go's map type.
     8  //
     9  // A map is just a hash table. The data is arranged
    10  // into an array of buckets. Each bucket contains up to
    11  // 8 key/elem pairs. The low-order bits of the hash are
    12  // used to select a bucket. Each bucket contains a few
    13  // high-order bits of each hash to distinguish the entries
    14  // within a single bucket.
    15  //
    16  // If more than 8 keys hash to a bucket, we chain on
    17  // extra buckets.
    18  //
    19  // When the hashtable grows, we allocate a new array
    20  // of buckets twice as big. Buckets are incrementally
    21  // copied from the old bucket array to the new bucket array.
    22  //
    23  // Map iterators walk through the array of buckets and
    24  // return the keys in walk order (bucket #, then overflow
    25  // chain order, then bucket index).  To maintain iteration
    26  // semantics, we never move keys within their bucket (if
    27  // we did, keys might be returned 0 or 2 times).  When
    28  // growing the table, iterators remain iterating through the
    29  // old table and must check the new table if the bucket
    30  // they are iterating through has been moved ("evacuated")
    31  // to the new table.
    32  
    33  // Picking loadFactor: too large and we have lots of overflow
    34  // buckets, too small and we waste a lot of space. I wrote
    35  // a simple program to check some stats for different loads:
    36  // (64-bit, 8 byte keys and elems)
    37  //  loadFactor    %overflow  bytes/entry     hitprobe    missprobe
    38  //        4.00         2.13        20.77         3.00         4.00
    39  //        4.50         4.05        17.30         3.25         4.50
    40  //        5.00         6.85        14.77         3.50         5.00
    41  //        5.50        10.55        12.94         3.75         5.50
    42  //        6.00        15.27        11.67         4.00         6.00
    43  //        6.50        20.90        10.79         4.25         6.50
    44  //        7.00        27.14        10.15         4.50         7.00
    45  //        7.50        34.03         9.73         4.75         7.50
    46  //        8.00        41.10         9.40         5.00         8.00
    47  //
    48  // %overflow   = percentage of buckets which have an overflow bucket
    49  // bytes/entry = overhead bytes used per key/elem pair
    50  // hitprobe    = # of entries to check when looking up a present key
    51  // missprobe   = # of entries to check when looking up an absent key
    52  //
    53  // Keep in mind this data is for maximally loaded tables, i.e. just
    54  // before the table grows. Typical tables will be somewhat less loaded.
    55  
    56  import (
    57  	"internal/abi"
    58  	"internal/goarch"
    59  	"runtime/internal/atomic"
    60  	"runtime/internal/math"
    61  	"unsafe"
    62  )
    63  
    64  const (
    65  	// Maximum number of key/elem pairs a bucket can hold.
    66  	bucketCntBits = 3
    67  	bucketCnt     = 1 << bucketCntBits
    68  
    69  	// Maximum average load of a bucket that triggers growth is 6.5.
    70  	// Represent as loadFactorNum/loadFactorDen, to allow integer math.
    71  	loadFactorNum = 13
    72  	loadFactorDen = 2
    73  
    74  	// Maximum key or elem size to keep inline (instead of mallocing per element).
    75  	// Must fit in a uint8.
    76  	// Fast versions cannot handle big elems - the cutoff size for
    77  	// fast versions in cmd/compile/internal/gc/walk.go must be at most this elem.
    78  	maxKeySize  = 128
    79  	maxElemSize = 128
    80  
    81  	// data offset should be the size of the bmap struct, but needs to be
    82  	// aligned correctly. For amd64p32 this means 64-bit alignment
    83  	// even though pointers are 32 bit.
    84  	dataOffset = unsafe.Offsetof(struct {
    85  		b bmap
    86  		v int64
    87  	}{}.v)
    88  
    89  	// Possible tophash values. We reserve a few possibilities for special marks.
    90  	// Each bucket (including its overflow buckets, if any) will have either all or none of its
    91  	// entries in the evacuated* states (except during the evacuate() method, which only happens
    92  	// during map writes and thus no one else can observe the map during that time).
    93  	emptyRest      = 0 // this cell is empty, and there are no more non-empty cells at higher indexes or overflows.
    94  	emptyOne       = 1 // this cell is empty
    95  	evacuatedX     = 2 // key/elem is valid.  Entry has been evacuated to first half of larger table.
    96  	evacuatedY     = 3 // same as above, but evacuated to second half of larger table.
    97  	evacuatedEmpty = 4 // cell is empty, bucket is evacuated.
    98  	minTopHash     = 5 // minimum tophash for a normal filled cell.
    99  
   100  	// flags
   101  	iterator     = 1 // there may be an iterator using buckets
   102  	oldIterator  = 2 // there may be an iterator using oldbuckets
   103  	hashWriting  = 4 // a goroutine is writing to the map
   104  	sameSizeGrow = 8 // the current map growth is to a new map of the same size
   105  
   106  	// sentinel bucket ID for iterator checks
   107  	noCheck = 1<<(8*goarch.PtrSize) - 1
   108  )
   109  
   110  // isEmpty reports whether the given tophash array entry represents an empty bucket entry.
   111  func isEmpty(x uint8) bool {
   112  	return x <= emptyOne
   113  }
   114  
   115  // A header for a Go map.
   116  type hmap struct {
   117  	// Note: the format of the hmap is also encoded in cmd/compile/internal/reflectdata/reflect.go.
   118  	// Make sure this stays in sync with the compiler's definition.
   119  	count     int // # live cells == size of map.  Must be first (used by len() builtin)
   120  	flags     uint8
   121  	B         uint8  // log_2 of # of buckets (can hold up to loadFactor * 2^B items)
   122  	noverflow uint16 // approximate number of overflow buckets; see incrnoverflow for details
   123  	hash0     uint32 // hash seed
   124  
   125  	buckets    unsafe.Pointer // array of 2^B Buckets. may be nil if count==0.
   126  	oldbuckets unsafe.Pointer // previous bucket array of half the size, non-nil only when growing
   127  	nevacuate  uintptr        // progress counter for evacuation (buckets less than this have been evacuated)
   128  
   129  	extra *mapextra // optional fields
   130  }
   131  
   132  // mapextra holds fields that are not present on all maps.
   133  type mapextra struct {
   134  	// If both key and elem do not contain pointers and are inline, then we mark bucket
   135  	// type as containing no pointers. This avoids scanning such maps.
   136  	// However, bmap.overflow is a pointer. In order to keep overflow buckets
   137  	// alive, we store pointers to all overflow buckets in hmap.extra.overflow and hmap.extra.oldoverflow.
   138  	// overflow and oldoverflow are only used if key and elem do not contain pointers.
   139  	// overflow contains overflow buckets for hmap.buckets.
   140  	// oldoverflow contains overflow buckets for hmap.oldbuckets.
   141  	// The indirection allows to store a pointer to the slice in hiter.
   142  	overflow    *[]*bmap
   143  	oldoverflow *[]*bmap
   144  
   145  	// nextOverflow holds a pointer to a free overflow bucket.
   146  	nextOverflow *bmap
   147  }
   148  
   149  // A bucket for a Go map.
   150  type bmap struct {
   151  	// tophash generally contains the top byte of the hash value
   152  	// for each key in this bucket. If tophash[0] < minTopHash,
   153  	// tophash[0] is a bucket evacuation state instead.
   154  	tophash [bucketCnt]uint8
   155  	// Followed by bucketCnt keys and then bucketCnt elems.
   156  	// NOTE: packing all the keys together and then all the elems together makes the
   157  	// code a bit more complicated than alternating key/elem/key/elem/... but it allows
   158  	// us to eliminate padding which would be needed for, e.g., map[int64]int8.
   159  	// Followed by an overflow pointer.
   160  }
   161  
   162  // A hash iteration structure.
   163  // If you modify hiter, also change cmd/compile/internal/reflectdata/reflect.go
   164  // and reflect/value.go to match the layout of this structure.
   165  type hiter struct {
   166  	key         unsafe.Pointer // Must be in first position.  Write nil to indicate iteration end (see cmd/compile/internal/walk/range.go).
   167  	elem        unsafe.Pointer // Must be in second position (see cmd/compile/internal/walk/range.go).
   168  	t           *maptype
   169  	h           *hmap
   170  	buckets     unsafe.Pointer // bucket ptr at hash_iter initialization time
   171  	bptr        *bmap          // current bucket
   172  	overflow    *[]*bmap       // keeps overflow buckets of hmap.buckets alive
   173  	oldoverflow *[]*bmap       // keeps overflow buckets of hmap.oldbuckets alive
   174  	startBucket uintptr        // bucket iteration started at
   175  	offset      uint8          // intra-bucket offset to start from during iteration (should be big enough to hold bucketCnt-1)
   176  	wrapped     bool           // already wrapped around from end of bucket array to beginning
   177  	B           uint8
   178  	i           uint8
   179  	bucket      uintptr
   180  	checkBucket uintptr
   181  }
   182  
   183  // bucketShift returns 1<<b, optimized for code generation.
   184  func bucketShift(b uint8) uintptr {
   185  	// Masking the shift amount allows overflow checks to be elided.
   186  	return uintptr(1) << (b & (goarch.PtrSize*8 - 1))
   187  }
   188  
   189  // bucketMask returns 1<<b - 1, optimized for code generation.
   190  func bucketMask(b uint8) uintptr {
   191  	return bucketShift(b) - 1
   192  }
   193  
   194  // tophash calculates the tophash value for hash.
   195  func tophash(hash uintptr) uint8 {
   196  	top := uint8(hash >> (goarch.PtrSize*8 - 8))
   197  	if top < minTopHash {
   198  		top += minTopHash
   199  	}
   200  	return top
   201  }
   202  
   203  func evacuated(b *bmap) bool {
   204  	h := b.tophash[0]
   205  	return h > emptyOne && h < minTopHash
   206  }
   207  
   208  func (b *bmap) overflow(t *maptype) *bmap {
   209  	return *(**bmap)(add(unsafe.Pointer(b), uintptr(t.bucketsize)-goarch.PtrSize))
   210  }
   211  
   212  func (b *bmap) setoverflow(t *maptype, ovf *bmap) {
   213  	*(**bmap)(add(unsafe.Pointer(b), uintptr(t.bucketsize)-goarch.PtrSize)) = ovf
   214  }
   215  
   216  func (b *bmap) keys() unsafe.Pointer {
   217  	return add(unsafe.Pointer(b), dataOffset)
   218  }
   219  
   220  // incrnoverflow increments h.noverflow.
   221  // noverflow counts the number of overflow buckets.
   222  // This is used to trigger same-size map growth.
   223  // See also tooManyOverflowBuckets.
   224  // To keep hmap small, noverflow is a uint16.
   225  // When there are few buckets, noverflow is an exact count.
   226  // When there are many buckets, noverflow is an approximate count.
   227  func (h *hmap) incrnoverflow() {
   228  	// We trigger same-size map growth if there are
   229  	// as many overflow buckets as buckets.
   230  	// We need to be able to count to 1<<h.B.
   231  	if h.B < 16 {
   232  		h.noverflow++
   233  		return
   234  	}
   235  	// Increment with probability 1/(1<<(h.B-15)).
   236  	// When we reach 1<<15 - 1, we will have approximately
   237  	// as many overflow buckets as buckets.
   238  	mask := uint32(1)<<(h.B-15) - 1
   239  	// Example: if h.B == 18, then mask == 7,
   240  	// and fastrand & 7 == 0 with probability 1/8.
   241  	if fastrand()&mask == 0 {
   242  		h.noverflow++
   243  	}
   244  }
   245  
   246  func (h *hmap) newoverflow(t *maptype, b *bmap) *bmap {
   247  	var ovf *bmap
   248  	if h.extra != nil && h.extra.nextOverflow != nil {
   249  		// We have preallocated overflow buckets available.
   250  		// See makeBucketArray for more details.
   251  		ovf = h.extra.nextOverflow
   252  		if ovf.overflow(t) == nil {
   253  			// We're not at the end of the preallocated overflow buckets. Bump the pointer.
   254  			h.extra.nextOverflow = (*bmap)(add(unsafe.Pointer(ovf), uintptr(t.bucketsize)))
   255  		} else {
   256  			// This is the last preallocated overflow bucket.
   257  			// Reset the overflow pointer on this bucket,
   258  			// which was set to a non-nil sentinel value.
   259  			ovf.setoverflow(t, nil)
   260  			h.extra.nextOverflow = nil
   261  		}
   262  	} else {
   263  		ovf = (*bmap)(newobject(t.bucket))
   264  	}
   265  	h.incrnoverflow()
   266  	if t.bucket.ptrdata == 0 {
   267  		h.createOverflow()
   268  		*h.extra.overflow = append(*h.extra.overflow, ovf)
   269  	}
   270  	b.setoverflow(t, ovf)
   271  	return ovf
   272  }
   273  
   274  func (h *hmap) createOverflow() {
   275  	if h.extra == nil {
   276  		h.extra = new(mapextra)
   277  	}
   278  	if h.extra.overflow == nil {
   279  		h.extra.overflow = new([]*bmap)
   280  	}
   281  }
   282  
   283  func makemap64(t *maptype, hint int64, h *hmap) *hmap {
   284  	if int64(int(hint)) != hint {
   285  		hint = 0
   286  	}
   287  	return makemap(t, int(hint), h)
   288  }
   289  
   290  // makemap_small implements Go map creation for make(map[k]v) and
   291  // make(map[k]v, hint) when hint is known to be at most bucketCnt
   292  // at compile time and the map needs to be allocated on the heap.
   293  func makemap_small() *hmap {
   294  	h := new(hmap)
   295  	h.hash0 = fastrand()
   296  	return h
   297  }
   298  
   299  // makemap implements Go map creation for make(map[k]v, hint).
   300  // If the compiler has determined that the map or the first bucket
   301  // can be created on the stack, h and/or bucket may be non-nil.
   302  // If h != nil, the map can be created directly in h.
   303  // If h.buckets != nil, bucket pointed to can be used as the first bucket.
   304  func makemap(t *maptype, hint int, h *hmap) *hmap {
   305  	mem, overflow := math.MulUintptr(uintptr(hint), t.bucket.size)
   306  	if overflow || mem > maxAlloc {
   307  		hint = 0
   308  	}
   309  
   310  	// initialize Hmap
   311  	if h == nil {
   312  		h = new(hmap)
   313  	}
   314  	h.hash0 = fastrand()
   315  
   316  	// Find the size parameter B which will hold the requested # of elements.
   317  	// For hint < 0 overLoadFactor returns false since hint < bucketCnt.
   318  	B := uint8(0)
   319  	for overLoadFactor(hint, B) {
   320  		B++
   321  	}
   322  	h.B = B
   323  
   324  	// allocate initial hash table
   325  	// if B == 0, the buckets field is allocated lazily later (in mapassign)
   326  	// If hint is large zeroing this memory could take a while.
   327  	if h.B != 0 {
   328  		var nextOverflow *bmap
   329  		h.buckets, nextOverflow = makeBucketArray(t, h.B, nil)
   330  		if nextOverflow != nil {
   331  			h.extra = new(mapextra)
   332  			h.extra.nextOverflow = nextOverflow
   333  		}
   334  	}
   335  
   336  	return h
   337  }
   338  
   339  // makeBucketArray initializes a backing array for map buckets.
   340  // 1<<b is the minimum number of buckets to allocate.
   341  // dirtyalloc should either be nil or a bucket array previously
   342  // allocated by makeBucketArray with the same t and b parameters.
   343  // If dirtyalloc is nil a new backing array will be alloced and
   344  // otherwise dirtyalloc will be cleared and reused as backing array.
   345  func makeBucketArray(t *maptype, b uint8, dirtyalloc unsafe.Pointer) (buckets unsafe.Pointer, nextOverflow *bmap) {
   346  	base := bucketShift(b)
   347  	nbuckets := base
   348  	// For small b, overflow buckets are unlikely.
   349  	// Avoid the overhead of the calculation.
   350  	if b >= 4 {
   351  		// Add on the estimated number of overflow buckets
   352  		// required to insert the median number of elements
   353  		// used with this value of b.
   354  		nbuckets += bucketShift(b - 4)
   355  		sz := t.bucket.size * nbuckets
   356  		up := roundupsize(sz)
   357  		if up != sz {
   358  			nbuckets = up / t.bucket.size
   359  		}
   360  	}
   361  
   362  	if dirtyalloc == nil {
   363  		buckets = newarray(t.bucket, int(nbuckets))
   364  	} else {
   365  		// dirtyalloc was previously generated by
   366  		// the above newarray(t.bucket, int(nbuckets))
   367  		// but may not be empty.
   368  		buckets = dirtyalloc
   369  		size := t.bucket.size * nbuckets
   370  		if t.bucket.ptrdata != 0 {
   371  			memclrHasPointers(buckets, size)
   372  		} else {
   373  			memclrNoHeapPointers(buckets, size)
   374  		}
   375  	}
   376  
   377  	if base != nbuckets {
   378  		// We preallocated some overflow buckets.
   379  		// To keep the overhead of tracking these overflow buckets to a minimum,
   380  		// we use the convention that if a preallocated overflow bucket's overflow
   381  		// pointer is nil, then there are more available by bumping the pointer.
   382  		// We need a safe non-nil pointer for the last overflow bucket; just use buckets.
   383  		nextOverflow = (*bmap)(add(buckets, base*uintptr(t.bucketsize)))
   384  		last := (*bmap)(add(buckets, (nbuckets-1)*uintptr(t.bucketsize)))
   385  		last.setoverflow(t, (*bmap)(buckets))
   386  	}
   387  	return buckets, nextOverflow
   388  }
   389  
   390  // mapaccess1 returns a pointer to h[key].  Never returns nil, instead
   391  // it will return a reference to the zero object for the elem type if
   392  // the key is not in the map.
   393  // NOTE: The returned pointer may keep the whole map live, so don't
   394  // hold onto it for very long.
   395  func mapaccess1(t *maptype, h *hmap, key unsafe.Pointer) unsafe.Pointer {
   396  	if raceenabled && h != nil {
   397  		callerpc := getcallerpc()
   398  		pc := abi.FuncPCABIInternal(mapaccess1)
   399  		racereadpc(unsafe.Pointer(h), callerpc, pc)
   400  		raceReadObjectPC(t.key, key, callerpc, pc)
   401  	}
   402  	if msanenabled && h != nil {
   403  		msanread(key, t.key.size)
   404  	}
   405  	if asanenabled && h != nil {
   406  		asanread(key, t.key.size)
   407  	}
   408  	if h == nil || h.count == 0 {
   409  		if t.hashMightPanic() {
   410  			t.hasher(key, 0) // see issue 23734
   411  		}
   412  		return unsafe.Pointer(&zeroVal[0])
   413  	}
   414  	if h.flags&hashWriting != 0 {
   415  		throw("concurrent map read and map write")
   416  	}
   417  	hash := t.hasher(key, uintptr(h.hash0))
   418  	m := bucketMask(h.B)
   419  	b := (*bmap)(add(h.buckets, (hash&m)*uintptr(t.bucketsize)))
   420  	if c := h.oldbuckets; c != nil {
   421  		if !h.sameSizeGrow() {
   422  			// There used to be half as many buckets; mask down one more power of two.
   423  			m >>= 1
   424  		}
   425  		oldb := (*bmap)(add(c, (hash&m)*uintptr(t.bucketsize)))
   426  		if !evacuated(oldb) {
   427  			b = oldb
   428  		}
   429  	}
   430  	top := tophash(hash)
   431  bucketloop:
   432  	for ; b != nil; b = b.overflow(t) {
   433  		for i := uintptr(0); i < bucketCnt; i++ {
   434  			if b.tophash[i] != top {
   435  				if b.tophash[i] == emptyRest {
   436  					break bucketloop
   437  				}
   438  				continue
   439  			}
   440  			k := add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   441  			if t.indirectkey() {
   442  				k = *((*unsafe.Pointer)(k))
   443  			}
   444  			if t.key.equal(key, k) {
   445  				e := add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   446  				if t.indirectelem() {
   447  					e = *((*unsafe.Pointer)(e))
   448  				}
   449  				return e
   450  			}
   451  		}
   452  	}
   453  	return unsafe.Pointer(&zeroVal[0])
   454  }
   455  
   456  func mapaccess2(t *maptype, h *hmap, key unsafe.Pointer) (unsafe.Pointer, bool) {
   457  	if raceenabled && h != nil {
   458  		callerpc := getcallerpc()
   459  		pc := abi.FuncPCABIInternal(mapaccess2)
   460  		racereadpc(unsafe.Pointer(h), callerpc, pc)
   461  		raceReadObjectPC(t.key, key, callerpc, pc)
   462  	}
   463  	if msanenabled && h != nil {
   464  		msanread(key, t.key.size)
   465  	}
   466  	if asanenabled && h != nil {
   467  		asanread(key, t.key.size)
   468  	}
   469  	if h == nil || h.count == 0 {
   470  		if t.hashMightPanic() {
   471  			t.hasher(key, 0) // see issue 23734
   472  		}
   473  		return unsafe.Pointer(&zeroVal[0]), false
   474  	}
   475  	if h.flags&hashWriting != 0 {
   476  		throw("concurrent map read and map write")
   477  	}
   478  	hash := t.hasher(key, uintptr(h.hash0))
   479  	m := bucketMask(h.B)
   480  	b := (*bmap)(add(h.buckets, (hash&m)*uintptr(t.bucketsize)))
   481  	if c := h.oldbuckets; c != nil {
   482  		if !h.sameSizeGrow() {
   483  			// There used to be half as many buckets; mask down one more power of two.
   484  			m >>= 1
   485  		}
   486  		oldb := (*bmap)(add(c, (hash&m)*uintptr(t.bucketsize)))
   487  		if !evacuated(oldb) {
   488  			b = oldb
   489  		}
   490  	}
   491  	top := tophash(hash)
   492  bucketloop:
   493  	for ; b != nil; b = b.overflow(t) {
   494  		for i := uintptr(0); i < bucketCnt; i++ {
   495  			if b.tophash[i] != top {
   496  				if b.tophash[i] == emptyRest {
   497  					break bucketloop
   498  				}
   499  				continue
   500  			}
   501  			k := add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   502  			if t.indirectkey() {
   503  				k = *((*unsafe.Pointer)(k))
   504  			}
   505  			if t.key.equal(key, k) {
   506  				e := add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   507  				if t.indirectelem() {
   508  					e = *((*unsafe.Pointer)(e))
   509  				}
   510  				return e, true
   511  			}
   512  		}
   513  	}
   514  	return unsafe.Pointer(&zeroVal[0]), false
   515  }
   516  
   517  // returns both key and elem. Used by map iterator
   518  func mapaccessK(t *maptype, h *hmap, key unsafe.Pointer) (unsafe.Pointer, unsafe.Pointer) {
   519  	if h == nil || h.count == 0 {
   520  		return nil, nil
   521  	}
   522  	hash := t.hasher(key, uintptr(h.hash0))
   523  	m := bucketMask(h.B)
   524  	b := (*bmap)(add(h.buckets, (hash&m)*uintptr(t.bucketsize)))
   525  	if c := h.oldbuckets; c != nil {
   526  		if !h.sameSizeGrow() {
   527  			// There used to be half as many buckets; mask down one more power of two.
   528  			m >>= 1
   529  		}
   530  		oldb := (*bmap)(add(c, (hash&m)*uintptr(t.bucketsize)))
   531  		if !evacuated(oldb) {
   532  			b = oldb
   533  		}
   534  	}
   535  	top := tophash(hash)
   536  bucketloop:
   537  	for ; b != nil; b = b.overflow(t) {
   538  		for i := uintptr(0); i < bucketCnt; i++ {
   539  			if b.tophash[i] != top {
   540  				if b.tophash[i] == emptyRest {
   541  					break bucketloop
   542  				}
   543  				continue
   544  			}
   545  			k := add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   546  			if t.indirectkey() {
   547  				k = *((*unsafe.Pointer)(k))
   548  			}
   549  			if t.key.equal(key, k) {
   550  				e := add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   551  				if t.indirectelem() {
   552  					e = *((*unsafe.Pointer)(e))
   553  				}
   554  				return k, e
   555  			}
   556  		}
   557  	}
   558  	return nil, nil
   559  }
   560  
   561  func mapaccess1_fat(t *maptype, h *hmap, key, zero unsafe.Pointer) unsafe.Pointer {
   562  	e := mapaccess1(t, h, key)
   563  	if e == unsafe.Pointer(&zeroVal[0]) {
   564  		return zero
   565  	}
   566  	return e
   567  }
   568  
   569  func mapaccess2_fat(t *maptype, h *hmap, key, zero unsafe.Pointer) (unsafe.Pointer, bool) {
   570  	e := mapaccess1(t, h, key)
   571  	if e == unsafe.Pointer(&zeroVal[0]) {
   572  		return zero, false
   573  	}
   574  	return e, true
   575  }
   576  
   577  // Like mapaccess, but allocates a slot for the key if it is not present in the map.
   578  func mapassign(t *maptype, h *hmap, key unsafe.Pointer) unsafe.Pointer {
   579  	if h == nil {
   580  		panic(plainError("assignment to entry in nil map"))
   581  	}
   582  	if raceenabled {
   583  		callerpc := getcallerpc()
   584  		pc := abi.FuncPCABIInternal(mapassign)
   585  		racewritepc(unsafe.Pointer(h), callerpc, pc)
   586  		raceReadObjectPC(t.key, key, callerpc, pc)
   587  	}
   588  	if msanenabled {
   589  		msanread(key, t.key.size)
   590  	}
   591  	if asanenabled {
   592  		asanread(key, t.key.size)
   593  	}
   594  	if h.flags&hashWriting != 0 {
   595  		throw("concurrent map writes")
   596  	}
   597  	hash := t.hasher(key, uintptr(h.hash0))
   598  
   599  	// Set hashWriting after calling t.hasher, since t.hasher may panic,
   600  	// in which case we have not actually done a write.
   601  	h.flags ^= hashWriting
   602  
   603  	if h.buckets == nil {
   604  		h.buckets = newobject(t.bucket) // newarray(t.bucket, 1)
   605  	}
   606  
   607  again:
   608  	bucket := hash & bucketMask(h.B)
   609  	if h.growing() {
   610  		growWork(t, h, bucket)
   611  	}
   612  	b := (*bmap)(add(h.buckets, bucket*uintptr(t.bucketsize)))
   613  	top := tophash(hash)
   614  
   615  	var inserti *uint8
   616  	var insertk unsafe.Pointer
   617  	var elem unsafe.Pointer
   618  bucketloop:
   619  	for {
   620  		for i := uintptr(0); i < bucketCnt; i++ {
   621  			if b.tophash[i] != top {
   622  				if isEmpty(b.tophash[i]) && inserti == nil {
   623  					inserti = &b.tophash[i]
   624  					insertk = add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   625  					elem = add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   626  				}
   627  				if b.tophash[i] == emptyRest {
   628  					break bucketloop
   629  				}
   630  				continue
   631  			}
   632  			k := add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   633  			if t.indirectkey() {
   634  				k = *((*unsafe.Pointer)(k))
   635  			}
   636  			if !t.key.equal(key, k) {
   637  				continue
   638  			}
   639  			// already have a mapping for key. Update it.
   640  			if t.needkeyupdate() {
   641  				typedmemmove(t.key, k, key)
   642  			}
   643  			elem = add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   644  			goto done
   645  		}
   646  		ovf := b.overflow(t)
   647  		if ovf == nil {
   648  			break
   649  		}
   650  		b = ovf
   651  	}
   652  
   653  	// Did not find mapping for key. Allocate new cell & add entry.
   654  
   655  	// If we hit the max load factor or we have too many overflow buckets,
   656  	// and we're not already in the middle of growing, start growing.
   657  	if !h.growing() && (overLoadFactor(h.count+1, h.B) || tooManyOverflowBuckets(h.noverflow, h.B)) {
   658  		hashGrow(t, h)
   659  		goto again // Growing the table invalidates everything, so try again
   660  	}
   661  
   662  	if inserti == nil {
   663  		// The current bucket and all the overflow buckets connected to it are full, allocate a new one.
   664  		newb := h.newoverflow(t, b)
   665  		inserti = &newb.tophash[0]
   666  		insertk = add(unsafe.Pointer(newb), dataOffset)
   667  		elem = add(insertk, bucketCnt*uintptr(t.keysize))
   668  	}
   669  
   670  	// store new key/elem at insert position
   671  	if t.indirectkey() {
   672  		kmem := newobject(t.key)
   673  		*(*unsafe.Pointer)(insertk) = kmem
   674  		insertk = kmem
   675  	}
   676  	if t.indirectelem() {
   677  		vmem := newobject(t.elem)
   678  		*(*unsafe.Pointer)(elem) = vmem
   679  	}
   680  	typedmemmove(t.key, insertk, key)
   681  	*inserti = top
   682  	h.count++
   683  
   684  done:
   685  	if h.flags&hashWriting == 0 {
   686  		throw("concurrent map writes")
   687  	}
   688  	h.flags &^= hashWriting
   689  	if t.indirectelem() {
   690  		elem = *((*unsafe.Pointer)(elem))
   691  	}
   692  	return elem
   693  }
   694  
   695  func mapdelete(t *maptype, h *hmap, key unsafe.Pointer) {
   696  	if raceenabled && h != nil {
   697  		callerpc := getcallerpc()
   698  		pc := abi.FuncPCABIInternal(mapdelete)
   699  		racewritepc(unsafe.Pointer(h), callerpc, pc)
   700  		raceReadObjectPC(t.key, key, callerpc, pc)
   701  	}
   702  	if msanenabled && h != nil {
   703  		msanread(key, t.key.size)
   704  	}
   705  	if asanenabled && h != nil {
   706  		asanread(key, t.key.size)
   707  	}
   708  	if h == nil || h.count == 0 {
   709  		if t.hashMightPanic() {
   710  			t.hasher(key, 0) // see issue 23734
   711  		}
   712  		return
   713  	}
   714  	if h.flags&hashWriting != 0 {
   715  		throw("concurrent map writes")
   716  	}
   717  
   718  	hash := t.hasher(key, uintptr(h.hash0))
   719  
   720  	// Set hashWriting after calling t.hasher, since t.hasher may panic,
   721  	// in which case we have not actually done a write (delete).
   722  	h.flags ^= hashWriting
   723  
   724  	bucket := hash & bucketMask(h.B)
   725  	if h.growing() {
   726  		growWork(t, h, bucket)
   727  	}
   728  	b := (*bmap)(add(h.buckets, bucket*uintptr(t.bucketsize)))
   729  	bOrig := b
   730  	top := tophash(hash)
   731  search:
   732  	for ; b != nil; b = b.overflow(t) {
   733  		for i := uintptr(0); i < bucketCnt; i++ {
   734  			if b.tophash[i] != top {
   735  				if b.tophash[i] == emptyRest {
   736  					break search
   737  				}
   738  				continue
   739  			}
   740  			k := add(unsafe.Pointer(b), dataOffset+i*uintptr(t.keysize))
   741  			k2 := k
   742  			if t.indirectkey() {
   743  				k2 = *((*unsafe.Pointer)(k2))
   744  			}
   745  			if !t.key.equal(key, k2) {
   746  				continue
   747  			}
   748  			// Only clear key if there are pointers in it.
   749  			if t.indirectkey() {
   750  				*(*unsafe.Pointer)(k) = nil
   751  			} else if t.key.ptrdata != 0 {
   752  				memclrHasPointers(k, t.key.size)
   753  			}
   754  			e := add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+i*uintptr(t.elemsize))
   755  			if t.indirectelem() {
   756  				*(*unsafe.Pointer)(e) = nil
   757  			} else if t.elem.ptrdata != 0 {
   758  				memclrHasPointers(e, t.elem.size)
   759  			} else {
   760  				memclrNoHeapPointers(e, t.elem.size)
   761  			}
   762  			b.tophash[i] = emptyOne
   763  			// If the bucket now ends in a bunch of emptyOne states,
   764  			// change those to emptyRest states.
   765  			// It would be nice to make this a separate function, but
   766  			// for loops are not currently inlineable.
   767  			if i == bucketCnt-1 {
   768  				if b.overflow(t) != nil && b.overflow(t).tophash[0] != emptyRest {
   769  					goto notLast
   770  				}
   771  			} else {
   772  				if b.tophash[i+1] != emptyRest {
   773  					goto notLast
   774  				}
   775  			}
   776  			for {
   777  				b.tophash[i] = emptyRest
   778  				if i == 0 {
   779  					if b == bOrig {
   780  						break // beginning of initial bucket, we're done.
   781  					}
   782  					// Find previous bucket, continue at its last entry.
   783  					c := b
   784  					for b = bOrig; b.overflow(t) != c; b = b.overflow(t) {
   785  					}
   786  					i = bucketCnt - 1
   787  				} else {
   788  					i--
   789  				}
   790  				if b.tophash[i] != emptyOne {
   791  					break
   792  				}
   793  			}
   794  		notLast:
   795  			h.count--
   796  			// Reset the hash seed to make it more difficult for attackers to
   797  			// repeatedly trigger hash collisions. See issue 25237.
   798  			if h.count == 0 {
   799  				h.hash0 = fastrand()
   800  			}
   801  			break search
   802  		}
   803  	}
   804  
   805  	if h.flags&hashWriting == 0 {
   806  		throw("concurrent map writes")
   807  	}
   808  	h.flags &^= hashWriting
   809  }
   810  
   811  // mapiterinit initializes the hiter struct used for ranging over maps.
   812  // The hiter struct pointed to by 'it' is allocated on the stack
   813  // by the compilers order pass or on the heap by reflect_mapiterinit.
   814  // Both need to have zeroed hiter since the struct contains pointers.
   815  func mapiterinit(t *maptype, h *hmap, it *hiter) {
   816  	if raceenabled && h != nil {
   817  		callerpc := getcallerpc()
   818  		racereadpc(unsafe.Pointer(h), callerpc, abi.FuncPCABIInternal(mapiterinit))
   819  	}
   820  
   821  	it.t = t
   822  	if h == nil || h.count == 0 {
   823  		return
   824  	}
   825  
   826  	if unsafe.Sizeof(hiter{})/goarch.PtrSize != 12 {
   827  		throw("hash_iter size incorrect") // see cmd/compile/internal/reflectdata/reflect.go
   828  	}
   829  	it.h = h
   830  
   831  	// grab snapshot of bucket state
   832  	it.B = h.B
   833  	it.buckets = h.buckets
   834  	if t.bucket.ptrdata == 0 {
   835  		// Allocate the current slice and remember pointers to both current and old.
   836  		// This preserves all relevant overflow buckets alive even if
   837  		// the table grows and/or overflow buckets are added to the table
   838  		// while we are iterating.
   839  		h.createOverflow()
   840  		it.overflow = h.extra.overflow
   841  		it.oldoverflow = h.extra.oldoverflow
   842  	}
   843  
   844  	// decide where to start
   845  	r := uintptr(fastrand())
   846  	if h.B > 31-bucketCntBits {
   847  		r += uintptr(fastrand()) << 31
   848  	}
   849  	it.startBucket = r & bucketMask(h.B)
   850  	it.offset = uint8(r >> h.B & (bucketCnt - 1))
   851  
   852  	// iterator state
   853  	it.bucket = it.startBucket
   854  
   855  	// Remember we have an iterator.
   856  	// Can run concurrently with another mapiterinit().
   857  	if old := h.flags; old&(iterator|oldIterator) != iterator|oldIterator {
   858  		atomic.Or8(&h.flags, iterator|oldIterator)
   859  	}
   860  
   861  	mapiternext(it)
   862  }
   863  
   864  func mapiternext(it *hiter) {
   865  	h := it.h
   866  	if raceenabled {
   867  		callerpc := getcallerpc()
   868  		racereadpc(unsafe.Pointer(h), callerpc, abi.FuncPCABIInternal(mapiternext))
   869  	}
   870  	if h.flags&hashWriting != 0 {
   871  		throw("concurrent map iteration and map write")
   872  	}
   873  	t := it.t
   874  	bucket := it.bucket
   875  	b := it.bptr
   876  	i := it.i
   877  	checkBucket := it.checkBucket
   878  
   879  next:
   880  	if b == nil {
   881  		if bucket == it.startBucket && it.wrapped {
   882  			// end of iteration
   883  			it.key = nil
   884  			it.elem = nil
   885  			return
   886  		}
   887  		if h.growing() && it.B == h.B {
   888  			// Iterator was started in the middle of a grow, and the grow isn't done yet.
   889  			// If the bucket we're looking at hasn't been filled in yet (i.e. the old
   890  			// bucket hasn't been evacuated) then we need to iterate through the old
   891  			// bucket and only return the ones that will be migrated to this bucket.
   892  			oldbucket := bucket & it.h.oldbucketmask()
   893  			b = (*bmap)(add(h.oldbuckets, oldbucket*uintptr(t.bucketsize)))
   894  			if !evacuated(b) {
   895  				checkBucket = bucket
   896  			} else {
   897  				b = (*bmap)(add(it.buckets, bucket*uintptr(t.bucketsize)))
   898  				checkBucket = noCheck
   899  			}
   900  		} else {
   901  			b = (*bmap)(add(it.buckets, bucket*uintptr(t.bucketsize)))
   902  			checkBucket = noCheck
   903  		}
   904  		bucket++
   905  		if bucket == bucketShift(it.B) {
   906  			bucket = 0
   907  			it.wrapped = true
   908  		}
   909  		i = 0
   910  	}
   911  	for ; i < bucketCnt; i++ {
   912  		offi := (i + it.offset) & (bucketCnt - 1)
   913  		if isEmpty(b.tophash[offi]) || b.tophash[offi] == evacuatedEmpty {
   914  			// TODO: emptyRest is hard to use here, as we start iterating
   915  			// in the middle of a bucket. It's feasible, just tricky.
   916  			continue
   917  		}
   918  		k := add(unsafe.Pointer(b), dataOffset+uintptr(offi)*uintptr(t.keysize))
   919  		if t.indirectkey() {
   920  			k = *((*unsafe.Pointer)(k))
   921  		}
   922  		e := add(unsafe.Pointer(b), dataOffset+bucketCnt*uintptr(t.keysize)+uintptr(offi)*uintptr(t.elemsize))
   923  		if checkBucket != noCheck && !h.sameSizeGrow() {
   924  			// Special case: iterator was started during a grow to a larger size
   925  			// and the grow is not done yet. We're working on a bucket whose
   926  			// oldbucket has not been evacuated yet. Or at least, it wasn't
   927  			// evacuated when we started the bucket. So we're iterating
   928  			// through the oldbucket, skipping any keys that will go
   929  			// to the other new bucket (each oldbucket expands to two
   930  			// buckets during a grow).
   931  			if t.reflexivekey() || t.key.equal(k, k) {
   932  				// If the item in the oldbucket is not destined for
   933  				// the current new bucket in the iteration, skip it.
   934  				hash := t.hasher(k, uintptr(h.hash0))
   935  				if hash&bucketMask(it.B) != checkBucket {
   936  					continue
   937  				}
   938  			} else {
   939  				// Hash isn't repeatable if k != k (NaNs).  We need a
   940  				// repeatable and randomish choice of which direction
   941  				// to send NaNs during evacuation. We'll use the low
   942  				// bit of tophash to decide which way NaNs go.
   943  				// NOTE: this case is why we need two evacuate tophash
   944  				// values, evacuatedX and evacuatedY, that differ in
   945  				// their low bit.
   946  				if checkBucket>>(it.B-1) != uintptr(b.tophash[offi]&1) {
   947  					continue
   948  				}
   949  			}
   950  		}
   951  		if (b.tophash[offi] != evacuatedX && b.tophash[offi] != evacuatedY) ||
   952  			!(t.reflexivekey() || t.key.equal(k, k)) {
   953  			// This is the golden data, we can return it.
   954  			// OR
   955  			// key!=key, so the entry can't be deleted or updated, so we can just return it.
   956  			// That's lucky for us because when key!=key we can't look it up successfully.
   957  			it.key = k
   958  			if t.indirectelem() {
   959  				e = *((*unsafe.Pointer)(e))
   960  			}
   961  			it.elem = e
   962  		} else {
   963  			// The hash table has grown since the iterator was started.
   964  			// The golden data for this key is now somewhere else.
   965  			// Check the current hash table for the data.
   966  			// This code handles the case where the key
   967  			// has been deleted, updated, or deleted and reinserted.
   968  			// NOTE: we need to regrab the key as it has potentially been
   969  			// updated to an equal() but not identical key (e.g. +0.0 vs -0.0).
   970  			rk, re := mapaccessK(t, h, k)
   971  			if rk == nil {
   972  				continue // key has been deleted
   973  			}
   974  			it.key = rk
   975  			it.elem = re
   976  		}
   977  		it.bucket = bucket
   978  		if it.bptr != b { // avoid unnecessary write barrier; see issue 14921
   979  			it.bptr = b
   980  		}
   981  		it.i = i + 1
   982  		it.checkBucket = checkBucket
   983  		return
   984  	}
   985  	b = b.overflow(t)
   986  	i = 0
   987  	goto next
   988  }
   989  
   990  // mapclear deletes all keys from a map.
   991  func mapclear(t *maptype, h *hmap) {
   992  	if raceenabled && h != nil {
   993  		callerpc := getcallerpc()
   994  		pc := abi.FuncPCABIInternal(mapclear)
   995  		racewritepc(unsafe.Pointer(h), callerpc, pc)
   996  	}
   997  
   998  	if h == nil || h.count == 0 {
   999  		return
  1000  	}
  1001  
  1002  	if h.flags&hashWriting != 0 {
  1003  		throw("concurrent map writes")
  1004  	}
  1005  
  1006  	h.flags ^= hashWriting
  1007  
  1008  	h.flags &^= sameSizeGrow
  1009  	h.oldbuckets = nil
  1010  	h.nevacuate = 0
  1011  	h.noverflow = 0
  1012  	h.count = 0
  1013  
  1014  	// Reset the hash seed to make it more difficult for attackers to
  1015  	// repeatedly trigger hash collisions. See issue 25237.
  1016  	h.hash0 = fastrand()
  1017  
  1018  	// Keep the mapextra allocation but clear any extra information.
  1019  	if h.extra != nil {
  1020  		*h.extra = mapextra{}
  1021  	}
  1022  
  1023  	// makeBucketArray clears the memory pointed to by h.buckets
  1024  	// and recovers any overflow buckets by generating them
  1025  	// as if h.buckets was newly alloced.
  1026  	_, nextOverflow := makeBucketArray(t, h.B, h.buckets)
  1027  	if nextOverflow != nil {
  1028  		// If overflow buckets are created then h.extra
  1029  		// will have been allocated during initial bucket creation.
  1030  		h.extra.nextOverflow = nextOverflow
  1031  	}
  1032  
  1033  	if h.flags&hashWriting == 0 {
  1034  		throw("concurrent map writes")
  1035  	}
  1036  	h.flags &^= hashWriting
  1037  }
  1038  
  1039  func hashGrow(t *maptype, h *hmap) {
  1040  	// If we've hit the load factor, get bigger.
  1041  	// Otherwise, there are too many overflow buckets,
  1042  	// so keep the same number of buckets and "grow" laterally.
  1043  	bigger := uint8(1)
  1044  	if !overLoadFactor(h.count+1, h.B) {
  1045  		bigger = 0
  1046  		h.flags |= sameSizeGrow
  1047  	}
  1048  	oldbuckets := h.buckets
  1049  	newbuckets, nextOverflow := makeBucketArray(t, h.B+bigger, nil)
  1050  
  1051  	flags := h.flags &^ (iterator | oldIterator)
  1052  	if h.flags&iterator != 0 {
  1053  		flags |= oldIterator
  1054  	}
  1055  	// commit the grow (atomic wrt gc)
  1056  	h.B += bigger
  1057  	h.flags = flags
  1058  	h.oldbuckets = oldbuckets
  1059  	h.buckets = newbuckets
  1060  	h.nevacuate = 0
  1061  	h.noverflow = 0
  1062  
  1063  	if h.extra != nil && h.extra.overflow != nil {
  1064  		// Promote current overflow buckets to the old generation.
  1065  		if h.extra.oldoverflow != nil {
  1066  			throw("oldoverflow is not nil")
  1067  		}
  1068  		h.extra.oldoverflow = h.extra.overflow
  1069  		h.extra.overflow = nil
  1070  	}
  1071  	if nextOverflow != nil {
  1072  		if h.extra == nil {
  1073  			h.extra = new(mapextra)
  1074  		}
  1075  		h.extra.nextOverflow = nextOverflow
  1076  	}
  1077  
  1078  	// the actual copying of the hash table data is done incrementally
  1079  	// by growWork() and evacuate().
  1080  }
  1081  
  1082  // overLoadFactor reports whether count items placed in 1<<B buckets is over loadFactor.
  1083  func overLoadFactor(count int, B uint8) bool {
  1084  	return count > bucketCnt && uintptr(count) > loadFactorNum*(bucketShift(B)/loadFactorDen)
  1085  }
  1086  
  1087  // tooManyOverflowBuckets reports whether noverflow buckets is too many for a map with 1<<B buckets.
  1088  // Note that most of these overflow buckets must be in sparse use;
  1089  // if use was dense, then we'd have already triggered regular map growth.
  1090  func tooManyOverflowBuckets(noverflow uint16, B uint8) bool {
  1091  	// If the threshold is too low, we do extraneous work.
  1092  	// If the threshold is too high, maps that grow and shrink can hold on to lots of unused memory.
  1093  	// "too many" means (approximately) as many overflow buckets as regular buckets.
  1094  	// See incrnoverflow for more details.
  1095  	if B > 15 {
  1096  		B = 15
  1097  	}
  1098  	// The compiler doesn't see here that B < 16; mask B to generate shorter shift code.
  1099  	return noverflow >= uint16(1)<<(B&15)
  1100  }
  1101  
  1102  // growing reports whether h is growing. The growth may be to the same size or bigger.
  1103  func (h *hmap) growing() bool {
  1104  	return h.oldbuckets != nil
  1105  }
  1106  
  1107  // sameSizeGrow reports whether the current growth is to a map of the same size.
  1108  func (h *hmap) sameSizeGrow() bool {
  1109  	return h.flags&sameSizeGrow != 0
  1110  }
  1111  
  1112  // noldbuckets calculates the number of buckets prior to the current map growth.
  1113  func (h *hmap) noldbuckets() uintptr {
  1114  	oldB := h.B
  1115  	if !h.sameSizeGrow() {
  1116  		oldB--
  1117  	}
  1118  	return bucketShift(oldB)
  1119  }
  1120  
  1121  // oldbucketmask provides a mask that can be applied to calculate n % noldbuckets().
  1122  func (h *hmap) oldbucketmask() uintptr {
  1123  	return h.noldbuckets() - 1
  1124  }
  1125  
  1126  func growWork(t *maptype, h *hmap, bucket uintptr) {
  1127  	// make sure we evacuate the oldbucket corresponding
  1128  	// to the bucket we're about to use
  1129  	evacuate(t, h, bucket&h.oldbucketmask())
  1130  
  1131  	// evacuate one more oldbucket to make progress on growing
  1132  	if h.growing() {
  1133  		evacuate(t, h, h.nevacuate)
  1134  	}
  1135  }
  1136  
  1137  func bucketEvacuated(t *maptype, h *hmap, bucket uintptr) bool {
  1138  	b := (*bmap)(add(h.oldbuckets, bucket*uintptr(t.bucketsize)))
  1139  	return evacuated(b)
  1140  }
  1141  
  1142  // evacDst is an evacuation destination.
  1143  type evacDst struct {
  1144  	b *bmap          // current destination bucket
  1145  	i int            // key/elem index into b
  1146  	k unsafe.Pointer // pointer to current key storage
  1147  	e unsafe.Pointer // pointer to current elem storage
  1148  }
  1149  
  1150  func evacuate(t *maptype, h *hmap, oldbucket uintptr) {
  1151  	b := (*bmap)(add(h.oldbuckets, oldbucket*uintptr(t.bucketsize)))
  1152  	newbit := h.noldbuckets()
  1153  	if !evacuated(b) {
  1154  		// TODO: reuse overflow buckets instead of using new ones, if there
  1155  		// is no iterator using the old buckets.  (If !oldIterator.)
  1156  
  1157  		// xy contains the x and y (low and high) evacuation destinations.
  1158  		var xy [2]evacDst
  1159  		x := &xy[0]
  1160  		x.b = (*bmap)(add(h.buckets, oldbucket*uintptr(t.bucketsize)))
  1161  		x.k = add(unsafe.Pointer(x.b), dataOffset)
  1162  		x.e = add(x.k, bucketCnt*uintptr(t.keysize))
  1163  
  1164  		if !h.sameSizeGrow() {
  1165  			// Only calculate y pointers if we're growing bigger.
  1166  			// Otherwise GC can see bad pointers.
  1167  			y := &xy[1]
  1168  			y.b = (*bmap)(add(h.buckets, (oldbucket+newbit)*uintptr(t.bucketsize)))
  1169  			y.k = add(unsafe.Pointer(y.b), dataOffset)
  1170  			y.e = add(y.k, bucketCnt*uintptr(t.keysize))
  1171  		}
  1172  
  1173  		for ; b != nil; b = b.overflow(t) {
  1174  			k := add(unsafe.Pointer(b), dataOffset)
  1175  			e := add(k, bucketCnt*uintptr(t.keysize))
  1176  			for i := 0; i < bucketCnt; i, k, e = i+1, add(k, uintptr(t.keysize)), add(e, uintptr(t.elemsize)) {
  1177  				top := b.tophash[i]
  1178  				if isEmpty(top) {
  1179  					b.tophash[i] = evacuatedEmpty
  1180  					continue
  1181  				}
  1182  				if top < minTopHash {
  1183  					throw("bad map state")
  1184  				}
  1185  				k2 := k
  1186  				if t.indirectkey() {
  1187  					k2 = *((*unsafe.Pointer)(k2))
  1188  				}
  1189  				var useY uint8
  1190  				if !h.sameSizeGrow() {
  1191  					// Compute hash to make our evacuation decision (whether we need
  1192  					// to send this key/elem to bucket x or bucket y).
  1193  					hash := t.hasher(k2, uintptr(h.hash0))
  1194  					if h.flags&iterator != 0 && !t.reflexivekey() && !t.key.equal(k2, k2) {
  1195  						// If key != key (NaNs), then the hash could be (and probably
  1196  						// will be) entirely different from the old hash. Moreover,
  1197  						// it isn't reproducible. Reproducibility is required in the
  1198  						// presence of iterators, as our evacuation decision must
  1199  						// match whatever decision the iterator made.
  1200  						// Fortunately, we have the freedom to send these keys either
  1201  						// way. Also, tophash is meaningless for these kinds of keys.
  1202  						// We let the low bit of tophash drive the evacuation decision.
  1203  						// We recompute a new random tophash for the next level so
  1204  						// these keys will get evenly distributed across all buckets
  1205  						// after multiple grows.
  1206  						useY = top & 1
  1207  						top = tophash(hash)
  1208  					} else {
  1209  						if hash&newbit != 0 {
  1210  							useY = 1
  1211  						}
  1212  					}
  1213  				}
  1214  
  1215  				if evacuatedX+1 != evacuatedY || evacuatedX^1 != evacuatedY {
  1216  					throw("bad evacuatedN")
  1217  				}
  1218  
  1219  				b.tophash[i] = evacuatedX + useY // evacuatedX + 1 == evacuatedY
  1220  				dst := &xy[useY]                 // evacuation destination
  1221  
  1222  				if dst.i == bucketCnt {
  1223  					dst.b = h.newoverflow(t, dst.b)
  1224  					dst.i = 0
  1225  					dst.k = add(unsafe.Pointer(dst.b), dataOffset)
  1226  					dst.e = add(dst.k, bucketCnt*uintptr(t.keysize))
  1227  				}
  1228  				dst.b.tophash[dst.i&(bucketCnt-1)] = top // mask dst.i as an optimization, to avoid a bounds check
  1229  				if t.indirectkey() {
  1230  					*(*unsafe.Pointer)(dst.k) = k2 // copy pointer
  1231  				} else {
  1232  					typedmemmove(t.key, dst.k, k) // copy elem
  1233  				}
  1234  				if t.indirectelem() {
  1235  					*(*unsafe.Pointer)(dst.e) = *(*unsafe.Pointer)(e)
  1236  				} else {
  1237  					typedmemmove(t.elem, dst.e, e)
  1238  				}
  1239  				dst.i++
  1240  				// These updates might push these pointers past the end of the
  1241  				// key or elem arrays.  That's ok, as we have the overflow pointer
  1242  				// at the end of the bucket to protect against pointing past the
  1243  				// end of the bucket.
  1244  				dst.k = add(dst.k, uintptr(t.keysize))
  1245  				dst.e = add(dst.e, uintptr(t.elemsize))
  1246  			}
  1247  		}
  1248  		// Unlink the overflow buckets & clear key/elem to help GC.
  1249  		if h.flags&oldIterator == 0 && t.bucket.ptrdata != 0 {
  1250  			b := add(h.oldbuckets, oldbucket*uintptr(t.bucketsize))
  1251  			// Preserve b.tophash because the evacuation
  1252  			// state is maintained there.
  1253  			ptr := add(b, dataOffset)
  1254  			n := uintptr(t.bucketsize) - dataOffset
  1255  			memclrHasPointers(ptr, n)
  1256  		}
  1257  	}
  1258  
  1259  	if oldbucket == h.nevacuate {
  1260  		advanceEvacuationMark(h, t, newbit)
  1261  	}
  1262  }
  1263  
  1264  func advanceEvacuationMark(h *hmap, t *maptype, newbit uintptr) {
  1265  	h.nevacuate++
  1266  	// Experiments suggest that 1024 is overkill by at least an order of magnitude.
  1267  	// Put it in there as a safeguard anyway, to ensure O(1) behavior.
  1268  	stop := h.nevacuate + 1024
  1269  	if stop > newbit {
  1270  		stop = newbit
  1271  	}
  1272  	for h.nevacuate != stop && bucketEvacuated(t, h, h.nevacuate) {
  1273  		h.nevacuate++
  1274  	}
  1275  	if h.nevacuate == newbit { // newbit == # of oldbuckets
  1276  		// Growing is all done. Free old main bucket array.
  1277  		h.oldbuckets = nil
  1278  		// Can discard old overflow buckets as well.
  1279  		// If they are still referenced by an iterator,
  1280  		// then the iterator holds a pointers to the slice.
  1281  		if h.extra != nil {
  1282  			h.extra.oldoverflow = nil
  1283  		}
  1284  		h.flags &^= sameSizeGrow
  1285  	}
  1286  }
  1287  
  1288  // Reflect stubs. Called from ../reflect/asm_*.s
  1289  
  1290  //go:linkname reflect_makemap reflect.makemap
  1291  func reflect_makemap(t *maptype, cap int) *hmap {
  1292  	// Check invariants and reflects math.
  1293  	if t.key.equal == nil {
  1294  		throw("runtime.reflect_makemap: unsupported map key type")
  1295  	}
  1296  	if t.key.size > maxKeySize && (!t.indirectkey() || t.keysize != uint8(goarch.PtrSize)) ||
  1297  		t.key.size <= maxKeySize && (t.indirectkey() || t.keysize != uint8(t.key.size)) {
  1298  		throw("key size wrong")
  1299  	}
  1300  	if t.elem.size > maxElemSize && (!t.indirectelem() || t.elemsize != uint8(goarch.PtrSize)) ||
  1301  		t.elem.size <= maxElemSize && (t.indirectelem() || t.elemsize != uint8(t.elem.size)) {
  1302  		throw("elem size wrong")
  1303  	}
  1304  	if t.key.align > bucketCnt {
  1305  		throw("key align too big")
  1306  	}
  1307  	if t.elem.align > bucketCnt {
  1308  		throw("elem align too big")
  1309  	}
  1310  	if t.key.size%uintptr(t.key.align) != 0 {
  1311  		throw("key size not a multiple of key align")
  1312  	}
  1313  	if t.elem.size%uintptr(t.elem.align) != 0 {
  1314  		throw("elem size not a multiple of elem align")
  1315  	}
  1316  	if bucketCnt < 8 {
  1317  		throw("bucketsize too small for proper alignment")
  1318  	}
  1319  	if dataOffset%uintptr(t.key.align) != 0 {
  1320  		throw("need padding in bucket (key)")
  1321  	}
  1322  	if dataOffset%uintptr(t.elem.align) != 0 {
  1323  		throw("need padding in bucket (elem)")
  1324  	}
  1325  
  1326  	return makemap(t, cap, nil)
  1327  }
  1328  
  1329  //go:linkname reflect_mapaccess reflect.mapaccess
  1330  func reflect_mapaccess(t *maptype, h *hmap, key unsafe.Pointer) unsafe.Pointer {
  1331  	elem, ok := mapaccess2(t, h, key)
  1332  	if !ok {
  1333  		// reflect wants nil for a missing element
  1334  		elem = nil
  1335  	}
  1336  	return elem
  1337  }
  1338  
  1339  //go:linkname reflect_mapaccess_faststr reflect.mapaccess_faststr
  1340  func reflect_mapaccess_faststr(t *maptype, h *hmap, key string) unsafe.Pointer {
  1341  	elem, ok := mapaccess2_faststr(t, h, key)
  1342  	if !ok {
  1343  		// reflect wants nil for a missing element
  1344  		elem = nil
  1345  	}
  1346  	return elem
  1347  }
  1348  
  1349  //go:linkname reflect_mapassign reflect.mapassign
  1350  func reflect_mapassign(t *maptype, h *hmap, key unsafe.Pointer, elem unsafe.Pointer) {
  1351  	p := mapassign(t, h, key)
  1352  	typedmemmove(t.elem, p, elem)
  1353  }
  1354  
  1355  //go:linkname reflect_mapassign_faststr reflect.mapassign_faststr
  1356  func reflect_mapassign_faststr(t *maptype, h *hmap, key string, elem unsafe.Pointer) {
  1357  	p := mapassign_faststr(t, h, key)
  1358  	typedmemmove(t.elem, p, elem)
  1359  }
  1360  
  1361  //go:linkname reflect_mapdelete reflect.mapdelete
  1362  func reflect_mapdelete(t *maptype, h *hmap, key unsafe.Pointer) {
  1363  	mapdelete(t, h, key)
  1364  }
  1365  
  1366  //go:linkname reflect_mapdelete_faststr reflect.mapdelete_faststr
  1367  func reflect_mapdelete_faststr(t *maptype, h *hmap, key string) {
  1368  	mapdelete_faststr(t, h, key)
  1369  }
  1370  
  1371  //go:linkname reflect_mapiterinit reflect.mapiterinit
  1372  func reflect_mapiterinit(t *maptype, h *hmap, it *hiter) {
  1373  	mapiterinit(t, h, it)
  1374  }
  1375  
  1376  //go:linkname reflect_mapiternext reflect.mapiternext
  1377  func reflect_mapiternext(it *hiter) {
  1378  	mapiternext(it)
  1379  }
  1380  
  1381  //go:linkname reflect_mapiterkey reflect.mapiterkey
  1382  func reflect_mapiterkey(it *hiter) unsafe.Pointer {
  1383  	return it.key
  1384  }
  1385  
  1386  //go:linkname reflect_mapiterelem reflect.mapiterelem
  1387  func reflect_mapiterelem(it *hiter) unsafe.Pointer {
  1388  	return it.elem
  1389  }
  1390  
  1391  //go:linkname reflect_maplen reflect.maplen
  1392  func reflect_maplen(h *hmap) int {
  1393  	if h == nil {
  1394  		return 0
  1395  	}
  1396  	if raceenabled {
  1397  		callerpc := getcallerpc()
  1398  		racereadpc(unsafe.Pointer(h), callerpc, abi.FuncPCABIInternal(reflect_maplen))
  1399  	}
  1400  	return h.count
  1401  }
  1402  
  1403  //go:linkname reflectlite_maplen internal/reflectlite.maplen
  1404  func reflectlite_maplen(h *hmap) int {
  1405  	if h == nil {
  1406  		return 0
  1407  	}
  1408  	if raceenabled {
  1409  		callerpc := getcallerpc()
  1410  		racereadpc(unsafe.Pointer(h), callerpc, abi.FuncPCABIInternal(reflect_maplen))
  1411  	}
  1412  	return h.count
  1413  }
  1414  
  1415  const maxZero = 1024 // must match value in reflect/value.go:maxZero cmd/compile/internal/gc/walk.go:zeroValSize
  1416  var zeroVal [maxZero]byte
  1417  

View as plain text