[sanitizer] make LargeMmapAllocator::GetBlockBegin faster by not using a linked list

llvm-svn: 171035

compiler-rt/lib/sanitizer_common/sanitizer_allocator.h

@@ -714,11 +714,9 @@ class LargeMmapAllocator {
     h->map_size = map_size;
     {
       SpinMutexLock l(&mutex_);
-      h->next = list_;
+      uptr idx = n_chunks_++;
-      h->prev = 0;
+      h->chunk_idx = idx;
-      if (list_)
+      chunks_[idx] = h;
-        list_->prev = h;
-      list_ = h;
     }
     return reinterpret_cast<void*>(res);
   }
@@ -727,14 +725,12 @@ class LargeMmapAllocator {
     Header *h = GetHeader(p);
     {
       SpinMutexLock l(&mutex_);
-      Header *prev = h->prev;
+      uptr idx = h->chunk_idx;
-      Header *next = h->next;
+      CHECK_EQ(chunks_[idx], h);
-      if (prev)
+      CHECK_LT(idx, n_chunks_);
-        prev->next = next;
+      chunks_[idx] = chunks_[n_chunks_ - 1];
-      if (next)
+      chunks_[idx]->chunk_idx = idx;
-        next->prev = prev;
+      n_chunks_--;
-      if (h == list_)
-        list_ = next;
     }
     MapUnmapCallback().OnUnmap(h->map_beg, h->map_size);
     UnmapOrDie(reinterpret_cast<void*>(h->map_beg), h->map_size);
@@ -743,8 +739,10 @@ class LargeMmapAllocator {
   uptr TotalMemoryUsed() {
     SpinMutexLock l(&mutex_);
     uptr res = 0;
-    for (Header *l = list_; l; l = l->next) {
+    for (uptr i = 0; i < n_chunks_; i++) {
-      res += RoundUpMapSize(l->size);
+      Header *h = chunks_[i];
+      CHECK_EQ(h->chunk_idx, i);
+      res += RoundUpMapSize(h->size);
     }
     return res;
   }
@@ -765,20 +763,32 @@ class LargeMmapAllocator {
   void *GetBlockBegin(void *ptr) {
     uptr p = reinterpret_cast<uptr>(ptr);
     SpinMutexLock l(&mutex_);
-    for (Header *l = list_; l; l = l->next) {
+    uptr nearest_chunk = 0;
-      if (p >= l->map_beg && p < l->map_beg + l->map_size)
+    // Cache-friendly linear search.
-        return GetUser(l);
+    for (uptr i = 0; i < n_chunks_; i++) {
+      uptr ch = reinterpret_cast<uptr>(chunks_[i]);
+      if (p < ch) continue;  // p is at left to this chunk, skip it.
+      if (p - ch < p - nearest_chunk)
+        nearest_chunk = ch;
     }
+    if (!nearest_chunk)
       return 0;
+    Header *h = reinterpret_cast<Header *>(nearest_chunk);
+    CHECK_GE(nearest_chunk, h->map_beg);
+    CHECK_LT(nearest_chunk, h->map_beg + h->map_size);
+    CHECK_LE(nearest_chunk, p);
+    if (h->map_beg + h->map_size < p)
+      return 0;
+    return GetUser(h);
   }
 
  private:
+  static const int kMaxNumChunks = 1 << FIRST_32_SECOND_64(15, 18);
   struct Header {
     uptr map_beg;
     uptr map_size;
     uptr size;
-    Header *next;
+    uptr chunk_idx;
-    Header *prev;
   };
 
   Header *GetHeader(uptr p) {
@@ -797,7 +807,8 @@ class LargeMmapAllocator {
   }
 
   uptr page_size_;
-  Header *list_;
+  Header *chunks_[kMaxNumChunks];
+  uptr n_chunks_;
   SpinMutex mutex_;
 };
 

compiler-rt/lib/sanitizer_common/tests/sanitizer_allocator_test.cc

@@ -244,12 +244,13 @@ TEST(SanitizerCommon, LargeMmapAllocator) {
   LargeMmapAllocator<> a;
   a.Init();
 
-  static const int kNumAllocs = 100;
+  static const int kNumAllocs = 1000;
   char *allocated[kNumAllocs];
-  static const uptr size = 1000;
+  static const uptr size = 4000;
   // Allocate some.
   for (int i = 0; i < kNumAllocs; i++) {
     allocated[i] = (char *)a.Allocate(size, 1);
+    CHECK(a.PointerIsMine(allocated[i]));
   }
   // Deallocate all.
   CHECK_GT(a.TotalMemoryUsed(), size * kNumAllocs);
@@ -269,6 +270,11 @@ TEST(SanitizerCommon, LargeMmapAllocator) {
     *meta = i;
     allocated[i] = x;
   }
+  for (int i = 0; i < kNumAllocs * kNumAllocs; i++) {
+    char *p = allocated[i % kNumAllocs];
+    CHECK(a.PointerIsMine(p));
+    CHECK(a.PointerIsMine(p + 2000));
+  }
   CHECK_GT(a.TotalMemoryUsed(), size * kNumAllocs);
   // Deallocate all in reverse order.
   for (int i = 0; i < kNumAllocs; i++) {
@@ -280,9 +286,12 @@ TEST(SanitizerCommon, LargeMmapAllocator) {
     a.Deallocate(p);
   }
   CHECK_EQ(a.TotalMemoryUsed(), 0);
+
+  // Test alignments.
   uptr max_alignment = SANITIZER_WORDSIZE == 64 ? (1 << 28) : (1 << 24);
   for (uptr alignment = 8; alignment <= max_alignment; alignment *= 2) {
-    for (int i = 0; i < kNumAllocs; i++) {
+    const uptr kNumAlignedAllocs = 100;
+    for (int i = 0; i < kNumAlignedAllocs; i++) {
       uptr size = ((i % 10) + 1) * 4096;
       char *p = allocated[i] = (char *)a.Allocate(size, alignment);
       CHECK_EQ(p, a.GetBlockBegin(p));
@@ -291,7 +300,7 @@ TEST(SanitizerCommon, LargeMmapAllocator) {
       CHECK_EQ(0, (uptr)allocated[i] % alignment);
       p[0] = p[size - 1] = 0;
     }
-    for (int i = 0; i < kNumAllocs; i++) {
+    for (int i = 0; i < kNumAlignedAllocs; i++) {
       a.Deallocate(allocated[i]);
     }
   }
@@ -381,7 +390,7 @@ void TestSizeClassAllocatorLocalCache() {
   const int kNumIter = 100;
   uptr saved_total = 0;
   for (int class_id = 1; class_id <= 5; class_id++) {
-    for (int i = 0; i < kNumIter; i++) {
+    for (int it = 0; it < kNumIter; it++) {
       void *allocated[kNumAllocs];
       for (uptr i = 0; i < kNumAllocs; i++) {
         allocated[i] = cache.Allocate(a, class_id);
@@ -391,7 +400,7 @@ void TestSizeClassAllocatorLocalCache() {
       }
       cache.Drain(a);
       uptr total_allocated = a->TotalMemoryUsed();
-      if (i)
+      if (it)
         CHECK_EQ(saved_total, total_allocated);
       saved_total = total_allocated;
     }