summaryrefslogtreecommitdiff
path: root/standalone/local_cache.h
blob: c97095d6be96d48a9dd7a702f00260387eef26e9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
//===-- local_cache.h -------------------------------------------*- C++ -*-===//
//
// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
// See https://llvm.org/LICENSE.txt for license information.
// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
//
//===----------------------------------------------------------------------===//

#ifndef SCUDO_LOCAL_CACHE_H_
#define SCUDO_LOCAL_CACHE_H_

#include "internal_defs.h"
#include "list.h"
#include "platform.h"
#include "report.h"
#include "stats.h"
#include "string_utils.h"

namespace scudo {

template <class SizeClassAllocator> struct SizeClassAllocatorLocalCache {
  typedef typename SizeClassAllocator::SizeClassMap SizeClassMap;
  typedef typename SizeClassAllocator::CompactPtrT CompactPtrT;

  struct TransferBatch {
    static const u16 MaxNumCached = SizeClassMap::MaxNumCachedHint;
    void setFromArray(CompactPtrT *Array, u16 N) {
      DCHECK_LE(N, MaxNumCached);
      Count = N;
      memcpy(Batch, Array, sizeof(Batch[0]) * Count);
    }
    void appendFromArray(CompactPtrT *Array, u16 N) {
      DCHECK_LE(N, MaxNumCached - Count);
      memcpy(Batch + Count, Array, sizeof(Batch[0]) * N);
      // u16 will be promoted to int by arithmetic type conversion.
      Count = static_cast<u16>(Count + N);
    }
    void clear() { Count = 0; }
    void add(CompactPtrT P) {
      DCHECK_LT(Count, MaxNumCached);
      Batch[Count++] = P;
    }
    void copyToArray(CompactPtrT *Array) const {
      memcpy(Array, Batch, sizeof(Batch[0]) * Count);
    }
    u16 getCount() const { return Count; }
    CompactPtrT get(u16 I) const {
      DCHECK_LE(I, Count);
      return Batch[I];
    }
    static u16 getMaxCached(uptr Size) {
      return Min(MaxNumCached, SizeClassMap::getMaxCachedHint(Size));
    }
    TransferBatch *Next;

  private:
    CompactPtrT Batch[MaxNumCached];
    u16 Count;
  };

  // A BatchGroup is used to collect blocks. Each group has a group id to
  // identify the group kind of contained blocks.
  struct BatchGroup {
    // `Next` is used by IntrusiveList.
    BatchGroup *Next;
    // The compact base address of each group
    uptr CompactPtrGroupBase;
    // Cache value of TransferBatch::getMaxCached()
    u16 MaxCachedPerBatch;
    // Number of blocks pushed into this group. This is an increment-only
    // counter.
    uptr PushedBlocks;
    // This is used to track how many bytes are not in-use since last time we
    // tried to release pages.
    uptr BytesInBGAtLastCheckpoint;
    // Blocks are managed by TransferBatch in a list.
    SinglyLinkedList<TransferBatch> Batches;
  };

  static_assert(sizeof(BatchGroup) <= sizeof(TransferBatch),
                "BatchGroup uses the same class size as TransferBatch");

  void init(GlobalStats *S, SizeClassAllocator *A) {
    DCHECK(isEmpty());
    Stats.init();
    if (LIKELY(S))
      S->link(&Stats);
    Allocator = A;
  }

  void destroy(GlobalStats *S) {
    drain();
    if (LIKELY(S))
      S->unlink(&Stats);
  }

  void *allocate(uptr ClassId) {
    DCHECK_LT(ClassId, NumClasses);
    PerClass *C = &PerClassArray[ClassId];
    if (C->Count == 0) {
      if (UNLIKELY(!refill(C, ClassId)))
        return nullptr;
      DCHECK_GT(C->Count, 0);
    }
    // We read ClassSize first before accessing Chunks because it's adjacent to
    // Count, while Chunks might be further off (depending on Count). That keeps
    // the memory accesses in close quarters.
    const uptr ClassSize = C->ClassSize;
    CompactPtrT CompactP = C->Chunks[--C->Count];
    Stats.add(StatAllocated, ClassSize);
    Stats.sub(StatFree, ClassSize);
    return Allocator->decompactPtr(ClassId, CompactP);
  }

  void deallocate(uptr ClassId, void *P) {
    CHECK_LT(ClassId, NumClasses);
    PerClass *C = &PerClassArray[ClassId];
    // We still have to initialize the cache in the event that the first heap
    // operation in a thread is a deallocation.
    initCacheMaybe(C);
    if (C->Count == C->MaxCount)
      drain(C, ClassId);
    // See comment in allocate() about memory accesses.
    const uptr ClassSize = C->ClassSize;
    C->Chunks[C->Count++] =
        Allocator->compactPtr(ClassId, reinterpret_cast<uptr>(P));
    Stats.sub(StatAllocated, ClassSize);
    Stats.add(StatFree, ClassSize);
  }

  bool isEmpty() const {
    for (uptr I = 0; I < NumClasses; ++I)
      if (PerClassArray[I].Count)
        return false;
    return true;
  }

  void drain() {
    // Drain BatchClassId last as createBatch can refill it.
    for (uptr I = 0; I < NumClasses; ++I) {
      if (I == BatchClassId)
        continue;
      while (PerClassArray[I].Count > 0)
        drain(&PerClassArray[I], I);
    }
    while (PerClassArray[BatchClassId].Count > 0)
      drain(&PerClassArray[BatchClassId], BatchClassId);
    DCHECK(isEmpty());
  }

  TransferBatch *createBatch(uptr ClassId, void *B) {
    if (ClassId != BatchClassId)
      B = allocate(BatchClassId);
    if (UNLIKELY(!B))
      reportOutOfMemory(SizeClassAllocator::getSizeByClassId(BatchClassId));
    return reinterpret_cast<TransferBatch *>(B);
  }

  BatchGroup *createGroup() {
    void *Ptr = allocate(BatchClassId);
    if (UNLIKELY(!Ptr))
      reportOutOfMemory(SizeClassAllocator::getSizeByClassId(BatchClassId));
    return reinterpret_cast<BatchGroup *>(Ptr);
  }

  LocalStats &getStats() { return Stats; }

  void getStats(ScopedString *Str) {
    bool EmptyCache = true;
    for (uptr I = 0; I < NumClasses; ++I) {
      if (PerClassArray[I].Count == 0)
        continue;

      EmptyCache = false;
      // The size of BatchClass is set to 0 intentionally. See the comment in
      // initCache() for more details.
      const uptr ClassSize = I == BatchClassId
                                 ? SizeClassAllocator::getSizeByClassId(I)
                                 : PerClassArray[I].ClassSize;
      // Note that the string utils don't support printing u16 thus we cast it
      // to a common use type uptr.
      Str->append("    %02zu (%6zu): cached: %4zu max: %4zu\n", I, ClassSize,
                  static_cast<uptr>(PerClassArray[I].Count),
                  static_cast<uptr>(PerClassArray[I].MaxCount));
    }

    if (EmptyCache)
      Str->append("    No block is cached.\n");
  }

private:
  static const uptr NumClasses = SizeClassMap::NumClasses;
  static const uptr BatchClassId = SizeClassMap::BatchClassId;
  struct alignas(SCUDO_CACHE_LINE_SIZE) PerClass {
    u16 Count;
    u16 MaxCount;
    // Note: ClassSize is zero for the transfer batch.
    uptr ClassSize;
    CompactPtrT Chunks[2 * TransferBatch::MaxNumCached];
  };
  PerClass PerClassArray[NumClasses] = {};
  LocalStats Stats;
  SizeClassAllocator *Allocator = nullptr;

  ALWAYS_INLINE void initCacheMaybe(PerClass *C) {
    if (LIKELY(C->MaxCount))
      return;
    initCache();
    DCHECK_NE(C->MaxCount, 0U);
  }

  NOINLINE void initCache() {
    for (uptr I = 0; I < NumClasses; I++) {
      PerClass *P = &PerClassArray[I];
      const uptr Size = SizeClassAllocator::getSizeByClassId(I);
      P->MaxCount = static_cast<u16>(2 * TransferBatch::getMaxCached(Size));
      if (I != BatchClassId) {
        P->ClassSize = Size;
      } else {
        // ClassSize in this struct is only used for malloc/free stats, which
        // should only track user allocations, not internal movements.
        P->ClassSize = 0;
      }
    }
  }

  void destroyBatch(uptr ClassId, void *B) {
    if (ClassId != BatchClassId)
      deallocate(BatchClassId, B);
  }

  NOINLINE bool refill(PerClass *C, uptr ClassId) {
    initCacheMaybe(C);
    TransferBatch *B = Allocator->popBatch(this, ClassId);
    if (UNLIKELY(!B))
      return false;
    DCHECK_GT(B->getCount(), 0);
    C->Count = B->getCount();
    B->copyToArray(C->Chunks);
    B->clear();
    destroyBatch(ClassId, B);
    return true;
  }

  NOINLINE void drain(PerClass *C, uptr ClassId) {
    const u16 Count = Min(static_cast<u16>(C->MaxCount / 2), C->Count);
    Allocator->pushBlocks(this, ClassId, &C->Chunks[0], Count);
    // u16 will be promoted to int by arithmetic type conversion.
    C->Count = static_cast<u16>(C->Count - Count);
    for (u16 I = 0; I < C->Count; I++)
      C->Chunks[I] = C->Chunks[I + Count];
  }
};

} // namespace scudo

#endif // SCUDO_LOCAL_CACHE_H_