SampleSet -> HistogramSamples which can be reused by SparseHistogram

BUG=139612

Review URL: https://chromiumcodereview.appspot.com/10829466

git-svn-id: svn://svn.chromium.org/chrome/trunk/src@158166 0039d316-1c4b-4281-b951-d872f2087c98

12 files changed, 899 insertions, 350 deletions

diff --git a/base/metrics/bucket_ranges.h b/base/metrics/bucket_ranges.h
index b77f52e..90b9047 100644
--- a/base/metrics/bucket_ranges.h
+++ b/base/metrics/bucket_ranges.h
@@ -26,7 +26,7 @@
 
 namespace base {
 
-class BASE_EXPORT_PRIVATE BucketRanges {
+class BASE_EXPORT BucketRanges {
  public:
   typedef std::vector<HistogramBase::Sample> Ranges;
 
diff --git a/base/metrics/histogram.cc b/base/metrics/histogram.cc
index dbc76ef..9e490a0 100644
--- a/base/metrics/histogram.cc
+++ b/base/metrics/histogram.cc
@@ -34,104 +34,6 @@ typedef HistogramBase::Sample Sample;
 // static
 const size_t Histogram::kBucketCount_MAX = 16384u;
 
-Histogram::SampleSet::SampleSet(size_t size)
-    : counts_(size, 0),
-      sum_(0),
-      redundant_count_(0) {}
-
-Histogram::SampleSet::SampleSet()
-    : counts_(),
-      sum_(0),
-      redundant_count_(0) {}
-
-Histogram::SampleSet::~SampleSet() {}
-
-void Histogram::SampleSet::Resize(size_t size) {
-  counts_.resize(size, 0);
-}
-
-void Histogram::SampleSet::Accumulate(Sample value,  Count count,
-                                      size_t index) {
-  DCHECK(count == 1 || count == -1);
-  counts_[index] += count;
-  sum_ += count * value;
-  redundant_count_ += count;
-  DCHECK_GE(counts_[index], 0);
-  DCHECK_GE(sum_, 0);
-  DCHECK_GE(redundant_count_, 0);
-}
-
-Count Histogram::SampleSet::TotalCount() const {
-  Count total = 0;
-  for (Counts::const_iterator it = counts_.begin();
-       it != counts_.end();
-       ++it) {
-    total += *it;
-  }
-  return total;
-}
-
-void Histogram::SampleSet::Add(const SampleSet& other) {
-  DCHECK_EQ(counts_.size(), other.counts_.size());
-  sum_ += other.sum_;
-  redundant_count_ += other.redundant_count_;
-  for (size_t index = 0; index < counts_.size(); ++index)
-    counts_[index] += other.counts_[index];
-}
-
-void Histogram::SampleSet::Subtract(const SampleSet& other) {
-  DCHECK_EQ(counts_.size(), other.counts_.size());
-  // Note: Race conditions in snapshotting a sum may lead to (temporary)
-  // negative values when snapshots are later combined (and deltas calculated).
-  // As a result, we don't currently CHCEK() for positive values.
-  sum_ -= other.sum_;
-  redundant_count_ -= other.redundant_count_;
-  for (size_t index = 0; index < counts_.size(); ++index) {
-    counts_[index] -= other.counts_[index];
-    DCHECK_GE(counts_[index], 0);
-  }
-}
-
-bool Histogram::SampleSet::Serialize(Pickle* pickle) const {
-  pickle->WriteInt64(sum_);
-  pickle->WriteInt64(redundant_count_);
-  pickle->WriteUInt64(counts_.size());
-
-  for (size_t index = 0; index < counts_.size(); ++index) {
-    pickle->WriteInt(counts_[index]);
-  }
-
-  return true;
-}
-
-bool Histogram::SampleSet::Deserialize(PickleIterator* iter) {
-  DCHECK_EQ(counts_.size(), 0u);
-  DCHECK_EQ(sum_, 0);
-  DCHECK_EQ(redundant_count_, 0);
-
-  uint64 counts_size;
-
-  if (!iter->ReadInt64(&sum_) ||
-      !iter->ReadInt64(&redundant_count_) ||
-      !iter->ReadUInt64(&counts_size)) {
-    return false;
-  }
-
-  if (counts_size == 0)
-    return false;
-
-  int count = 0;
-  for (uint64 index = 0; index < counts_size; ++index) {
-    int i;
-    if (!iter->ReadInt(&i))
-      return false;
-    counts_.push_back(i);
-    count += i;
-  }
-  DCHECK_EQ(count, redundant_count_);
-  return count == redundant_count_;
-}
-
 // TODO(rtenneti): delete this code after debugging.
 void CheckCorruption(const Histogram& histogram, bool new_histogram) {
   const std::string& histogram_name = histogram.histogram_name();
@@ -245,24 +147,27 @@ void Histogram::InitializeBucketRanges(Sample minimum,
   ranges->ResetChecksum();
 }
 
-// static
 void Histogram::Add(int value) {
+  DCHECK_EQ(0, ranges(0));
+  DCHECK_EQ(kSampleType_MAX, ranges(bucket_count_));
+
   if (value > kSampleType_MAX - 1)
     value = kSampleType_MAX - 1;
   if (value < 0)
     value = 0;
-  size_t index = BucketIndex(value);
-  DCHECK_GE(value, ranges(index));
-  DCHECK_LT(value, ranges(index + 1));
-  Accumulate(value, 1, index);
+  samples_->Accumulate(value, 1);
 }
 
 void Histogram::AddBoolean(bool value) {
   DCHECK(false);
 }
 
-void Histogram::AddSampleSet(const SampleSet& sample) {
-  sample_.Add(sample);
+void Histogram::AddSamples(const HistogramSamples& samples) {
+  samples_->Add(samples);
+}
+
+bool Histogram::AddSamplesFromPickle(PickleIterator* iter) {
+  return samples_->AddFromPickle(iter);
 }
 
 void Histogram::SetRangeDescriptions(const DescriptionPair descriptions[]) {
@@ -283,7 +188,7 @@ void Histogram::WriteAscii(string* output) const {
 
 // static
 string Histogram::SerializeHistogramInfo(const Histogram& histogram,
-                                         const SampleSet& snapshot) {
+                                         const HistogramSamples& snapshot) {
   DCHECK_NE(NOT_VALID_IN_RENDERER, histogram.histogram_type());
   DCHECK(histogram.bucket_ranges()->HasValidChecksum());
 
@@ -296,10 +201,10 @@ string Histogram::SerializeHistogramInfo(const Histogram& histogram,
   pickle.WriteInt(histogram.histogram_type());
   pickle.WriteInt(histogram.flags());
 
-  snapshot.Serialize(&pickle);
-
   histogram.SerializeRanges(&pickle);
 
+  snapshot.Serialize(&pickle);
+
   return string(static_cast<const char*>(pickle.data()), pickle.size());
 }
 
@@ -318,7 +223,6 @@ bool Histogram::DeserializeHistogramInfo(const string& histogram_info) {
   uint32 range_checksum;
   int histogram_type;
   int pickle_flags;
-  SampleSet sample;
 
   PickleIterator iter(pickle);
   if (!iter.ReadString(&histogram_name) ||
@@ -327,8 +231,7 @@ bool Histogram::DeserializeHistogramInfo(const string& histogram_info) {
       !iter.ReadUInt64(&bucket_count) ||
       !iter.ReadUInt32(&range_checksum) ||
       !iter.ReadInt(&histogram_type) ||
-      !iter.ReadInt(&pickle_flags) ||
-      !sample.Histogram::SampleSet::Deserialize(&iter)) {
+      !iter.ReadInt(&pickle_flags)) {
     DLOG(ERROR) << "Pickle error decoding Histogram: " << histogram_name;
     return false;
   }
@@ -382,23 +285,21 @@ bool Histogram::DeserializeHistogramInfo(const string& histogram_info) {
   if (render_histogram->flags() & kIPCSerializationSourceFlag) {
     DVLOG(1) << "Single process mode, histogram observed and not copied: "
              << histogram_name;
-  } else {
-    DCHECK_EQ(flags & render_histogram->flags(), flags);
-    render_histogram->AddSampleSet(sample);
+    return true;
   }
 
-  return true;
+  DCHECK_EQ(flags & render_histogram->flags(), flags);
+  return render_histogram->AddSamplesFromPickle(&iter);
 }
 
+// static
+const int Histogram::kCommonRaceBasedCountMismatch = 5;
 
-// Validate a sample and related histogram.
 Histogram::Inconsistencies Histogram::FindCorruption(
-    const SampleSet& snapshot) const {
+    const HistogramSamples& samples) const {
   int inconsistencies = NO_INCONSISTENCIES;
   Sample previous_range = -1;  // Bottom range is always 0.
-  int64 count = 0;
   for (size_t index = 0; index < bucket_count(); ++index) {
-    count += snapshot.counts(index);
     int new_range = ranges(index);
     if (previous_range >= new_range)
       inconsistencies |= BUCKET_ORDER_ERROR;
@@ -408,20 +309,11 @@ Histogram::Inconsistencies Histogram::FindCorruption(
   if (!bucket_ranges()->HasValidChecksum())
     inconsistencies |= RANGE_CHECKSUM_ERROR;
 
-  int64 delta64 = snapshot.redundant_count() - count;
+  int64 delta64 = samples.redundant_count() - samples.TotalCount();
   if (delta64 != 0) {
     int delta = static_cast<int>(delta64);
     if (delta != delta64)
       delta = INT_MAX;  // Flag all giant errors as INT_MAX.
-    // Since snapshots of histograms are taken asynchronously relative to
-    // sampling (and snapped from different threads), it is pretty likely that
-    // we'll catch a redundant count that doesn't match the sample count.  We
-    // allow for a certain amount of slop before flagging this as an
-    // inconsistency.  Even with an inconsistency, we'll snapshot it again (for
-    // UMA in about a half hour, so we'll eventually get the data, if it was
-    // not the result of a corruption.  If histograms show that 1 is "too tight"
-    // then we may try to use 2 or 3 for this slop value.
-    const int kCommonRaceBasedCountMismatch = 5;
     if (delta > 0) {
       UMA_HISTOGRAM_COUNTS("Histogram.InconsistentCountHigh", delta);
       if (delta > kCommonRaceBasedCountMismatch)
@@ -448,11 +340,10 @@ size_t Histogram::bucket_count() const {
   return bucket_count_;
 }
 
-// Do a safe atomic snapshot of sample data.
-// This implementation assumes we are on a safe single thread.
-void Histogram::SnapshotSample(SampleSet* sample) const {
-  // Note locking not done in this version!!!
-  *sample = sample_;
+scoped_ptr<SampleVector> Histogram::SnapshotSamples() const {
+  scoped_ptr<SampleVector> samples(new SampleVector(bucket_ranges()));
+  samples->Add(*samples_);
+  return samples.Pass();
 }
 
 bool Histogram::HasConstructionArguments(Sample minimum,
@@ -471,8 +362,10 @@ Histogram::Histogram(const string& name,
     bucket_ranges_(ranges),
     declared_min_(minimum),
     declared_max_(maximum),
-    bucket_count_(bucket_count),
-    sample_(bucket_count) {}
+    bucket_count_(bucket_count) {
+  if (ranges)
+    samples_.reset(new SampleVector(ranges));
+}
 
 Histogram::~Histogram() {
   if (StatisticsRecorder::dump_on_exit()) {
@@ -519,31 +412,6 @@ bool Histogram::PrintEmptyBucket(size_t index) const {
   return true;
 }
 
-size_t Histogram::BucketIndex(Sample value) const {
-  // Use simple binary search.  This is very general, but there are better
-  // approaches if we knew that the buckets were linearly distributed.
-  DCHECK_LE(ranges(0), value);
-  DCHECK_GT(ranges(bucket_count()), value);
-  size_t under = 0;
-  size_t over = bucket_count();
-  size_t mid;
-
-  do {
-    DCHECK_GE(over, under);
-    mid = under + (over - under)/2;
-    if (mid == under)
-      break;
-    if (ranges(mid) <= value)
-      under = mid;
-    else
-      over = mid;
-  } while (true);
-
-  DCHECK_LE(ranges(mid), value);
-  CHECK_GT(ranges(mid+1), value);
-  return mid;
-}
-
 // Use the actual bucket widths (like a linear histogram) until the widths get
 // over some transition value, and then use that transition width.  Exponentials
 // get so big so fast (and we don't expect to see a lot of entries in the large
@@ -567,12 +435,6 @@ const string Histogram::GetAsciiBucketRange(size_t i) const {
   return result;
 }
 
-// Update histogram data with new sample.
-void Histogram::Accumulate(Sample value, Count count, size_t index) {
-  // Note locking not done in this version!!!
-  sample_.Accumulate(value, count, index);
-}
-
 //------------------------------------------------------------------------------
 // Private methods
 
@@ -581,22 +443,21 @@ void Histogram::WriteAsciiImpl(bool graph_it,
                                string* output) const {
   // Get local (stack) copies of all effectively volatile class data so that we
   // are consistent across our output activities.
-  SampleSet snapshot;
-  SnapshotSample(&snapshot);
-  Count sample_count = snapshot.TotalCount();
+  scoped_ptr<SampleVector> snapshot = SnapshotSamples();
+  Count sample_count = snapshot->TotalCount();
 
-  WriteAsciiHeader(snapshot, sample_count, output);
+  WriteAsciiHeader(*snapshot, sample_count, output);
   output->append(newline);
 
   // Prepare to normalize graphical rendering of bucket contents.
   double max_size = 0;
   if (graph_it)
-    max_size = GetPeakBucketSize(snapshot);
+    max_size = GetPeakBucketSize(*snapshot);
 
   // Calculate space needed to print bucket range numbers.  Leave room to print
   // nearly the largest bucket range without sliding over the histogram.
   size_t largest_non_empty_bucket = bucket_count() - 1;
-  while (0 == snapshot.counts(largest_non_empty_bucket)) {
+  while (0 == snapshot->GetCountAtIndex(largest_non_empty_bucket)) {
     if (0 == largest_non_empty_bucket)
       break;  // All buckets are empty.
     --largest_non_empty_bucket;
@@ -605,7 +466,7 @@ void Histogram::WriteAsciiImpl(bool graph_it,
   // Calculate largest print width needed for any of our bucket range displays.
   size_t print_width = 1;
   for (size_t i = 0; i < bucket_count(); ++i) {
-    if (snapshot.counts(i)) {
+    if (snapshot->GetCountAtIndex(i)) {
       size_t width = GetAsciiBucketRange(i).size() + 1;
       if (width > print_width)
         print_width = width;
@@ -616,7 +477,7 @@ void Histogram::WriteAsciiImpl(bool graph_it,
   int64 past = 0;
   // Output the actual histogram graph.
   for (size_t i = 0; i < bucket_count(); ++i) {
-    Count current = snapshot.counts(i);
+    Count current = snapshot->GetCountAtIndex(i);
     if (!current && !PrintEmptyBucket(i))
       continue;
     remaining -= current;
@@ -624,9 +485,12 @@ void Histogram::WriteAsciiImpl(bool graph_it,
     output->append(range);
     for (size_t j = 0; range.size() + j < print_width + 1; ++j)
       output->push_back(' ');
-    if (0 == current && i < bucket_count() - 1 && 0 == snapshot.counts(i + 1)) {
-      while (i < bucket_count() - 1 && 0 == snapshot.counts(i + 1))
+    if (0 == current && i < bucket_count() - 1 &&
+        0 == snapshot->GetCountAtIndex(i + 1)) {
+      while (i < bucket_count() - 1 &&
+             0 == snapshot->GetCountAtIndex(i + 1)) {
         ++i;
+      }
       output->append("... ");
       output->append(newline);
       continue;  // No reason to plot emptiness.
@@ -641,17 +505,17 @@ void Histogram::WriteAsciiImpl(bool graph_it,
   DCHECK_EQ(sample_count, past);
 }
 
-double Histogram::GetPeakBucketSize(const SampleSet& snapshot) const {
+double Histogram::GetPeakBucketSize(const SampleVector& samples) const {
   double max = 0;
   for (size_t i = 0; i < bucket_count() ; ++i) {
-    double current_size = GetBucketSize(snapshot.counts(i), i);
+    double current_size = GetBucketSize(samples.GetCountAtIndex(i), i);
     if (current_size > max)
       max = current_size;
   }
   return max;
 }
 
-void Histogram::WriteAsciiHeader(const SampleSet& snapshot,
+void Histogram::WriteAsciiHeader(const SampleVector& samples,
                                  Count sample_count,
                                  string* output) const {
   StringAppendF(output,
@@ -659,9 +523,9 @@ void Histogram::WriteAsciiHeader(const SampleSet& snapshot,
                 histogram_name().c_str(),
                 sample_count);
   if (0 == sample_count) {
-    DCHECK_EQ(snapshot.sum(), 0);
+    DCHECK_EQ(samples.sum(), 0);
   } else {
-    double average = static_cast<float>(snapshot.sum()) / sample_count;
+    double average = static_cast<float>(samples.sum()) / sample_count;
 
     StringAppendF(output, ", average = %.1f", average);
   }
diff --git a/base/metrics/histogram.h b/base/metrics/histogram.h
index 6320db6..d86a397 100644
--- a/base/metrics/histogram.h
+++ b/base/metrics/histogram.h
@@ -68,8 +68,11 @@
 #include "base/compiler_specific.h"
 #include "base/gtest_prod_util.h"
 #include "base/logging.h"
+#include "base/memory/scoped_ptr.h"
 #include "base/metrics/bucket_ranges.h"
 #include "base/metrics/histogram_base.h"
+#include "base/metrics/histogram_samples.h"
+#include "base/metrics/sample_vector.h"
 #include "base/time.h"
 
 class Pickle;
@@ -338,8 +341,10 @@ class Lock;
 
 //------------------------------------------------------------------------------
 
-class BooleanHistogram;
 class BucketRanges;
+class SampleVector;
+
+class BooleanHistogram;
 class CustomHistogram;
 class Histogram;
 class LinearHistogram;
@@ -383,57 +388,6 @@ class BASE_EXPORT Histogram : public HistogramBase {
   };
 
   //----------------------------------------------------------------------------
-  // Statistic values, developed over the life of the histogram.
-
-  class BASE_EXPORT SampleSet {
-   public:
-    explicit SampleSet(size_t size);
-    SampleSet();
-    ~SampleSet();
-
-    void Resize(size_t size);
-
-    // Accessor for histogram to make routine additions.
-    void Accumulate(Sample value, Count count, size_t index);
-
-    // Accessor methods.
-    size_t size() const { return counts_.size(); }
-    Count counts(size_t i) const { return counts_[i]; }
-    Count TotalCount() const;
-    int64 sum() const { return sum_; }
-    int64 redundant_count() const { return redundant_count_; }
-
-    // Arithmetic manipulation of corresponding elements of the set.
-    void Add(const SampleSet& other);
-    void Subtract(const SampleSet& other);
-
-    bool Serialize(Pickle* pickle) const;
-    bool Deserialize(PickleIterator* iter);
-
-   protected:
-    // Actual histogram data is stored in buckets, showing the count of values
-    // that fit into each bucket.
-    Counts counts_;
-
-    // Save simple stats locally.  Note that this MIGHT get done in base class
-    // without shared memory at some point.
-    int64 sum_;         // sum of samples.
-
-   private:
-    // Allow tests to corrupt our innards for testing purposes.
-    FRIEND_TEST_ALL_PREFIXES(HistogramTest, CorruptSampleCounts);
-
-    // To help identify memory corruption, we reduntantly save the number of
-    // samples we've accumulated into all of our buckets.  We can compare this
-    // count to the sum of the counts in all buckets, and detect problems.  Note
-    // that due to races in histogram accumulation (if a histogram is indeed
-    // updated on several threads simultaneously), the tallies might mismatch,
-    // and also the snapshotting code may asynchronously get a mismatch (though
-    // generally either race based mismatch cause is VERY rare).
-    int64 redundant_count_;
-  };
-
-  //----------------------------------------------------------------------------
   // For a valid histogram, input should follow these restrictions:
   // minimum > 0 (if a minimum below 1 is specified, it will implicitly be
   //              normalized up to 1)
@@ -473,7 +427,8 @@ class BASE_EXPORT Histogram : public HistogramBase {
     Add(static_cast<int>(time.InMilliseconds()));
   }
 
-  void AddSampleSet(const SampleSet& sample);
+  void AddSamples(const HistogramSamples& samples);
+  bool AddSamplesFromPickle(PickleIterator* iter);
 
   // This method is an interface, used only by LinearHistogram.
   virtual void SetRangeDescriptions(const DescriptionPair descriptions[]);
@@ -491,19 +446,28 @@ class BASE_EXPORT Histogram : public HistogramBase {
   // Serialize the given snapshot of a Histogram into a String. Uses
   // Pickle class to flatten the object.
   static std::string SerializeHistogramInfo(const Histogram& histogram,
-                                            const SampleSet& snapshot);
+                                            const HistogramSamples& snapshot);
 
   // The following method accepts a list of pickled histograms and
   // builds a histogram and updates shadow copy of histogram data in the
   // browser process.
   static bool DeserializeHistogramInfo(const std::string& histogram_info);
 
+  // This constant if for FindCorruption. Since snapshots of histograms are
+  // taken asynchronously relative to sampling, and our counting code currently
+  // does not prevent race conditions, it is pretty likely that we'll catch a
+  // redundant count that doesn't match the sample count.  We allow for a
+  // certain amount of slop before flagging this as an inconsistency. Even with
+  // an inconsistency, we'll snapshot it again (for UMA in about a half hour),
+  // so we'll eventually get the data, if it was not the result of a corruption.
+  static const int kCommonRaceBasedCountMismatch;
+
   // Check to see if bucket ranges, counts and tallies in the snapshot are
   // consistent with the bucket ranges and checksums in our histogram.  This can
   // produce a false-alarm if a race occurred in the reading of the data during
   // a SnapShot process, but should otherwise be false at all times (unless we
   // have memory over-writes, or DRAM failures).
-  virtual Inconsistencies FindCorruption(const SampleSet& snapshot) const;
+  virtual Inconsistencies FindCorruption(const HistogramSamples& samples) const;
 
   //----------------------------------------------------------------------------
   // Accessors for factory constuction, serialization and testing.
@@ -517,7 +481,7 @@ class BASE_EXPORT Histogram : public HistogramBase {
 
   // Snapshot the current complete set of sample data.
   // Override with atomic/locked snapshot if needed.
-  virtual void SnapshotSample(SampleSet* sample) const;
+  virtual scoped_ptr<SampleVector> SnapshotSamples() const;
 
   virtual bool HasConstructionArguments(Sample minimum,
                                         Sample maximum,
@@ -553,11 +517,6 @@ class BASE_EXPORT Histogram : public HistogramBase {
   // Method to override to skip the display of the i'th bucket if it's empty.
   virtual bool PrintEmptyBucket(size_t index) const;
 
-  //----------------------------------------------------------------------------
-  // Methods to override to create histogram with different bucket widths.
-  //----------------------------------------------------------------------------
-  // Find bucket to increment for sample value.
-  virtual size_t BucketIndex(Sample value) const;
   // Get normalized size, relative to the ranges(i).
   virtual double GetBucketSize(Count current, size_t i) const;
 
@@ -566,12 +525,6 @@ class BASE_EXPORT Histogram : public HistogramBase {
   // be a name (or string description) given to the bucket.
   virtual const std::string GetAsciiBucketRange(size_t it) const;
 
-  //----------------------------------------------------------------------------
-  // Methods to override to create thread safe histogram.
-  //----------------------------------------------------------------------------
-  // Update all our internal data, including histogram
-  virtual void Accumulate(Sample value, Count count, size_t index);
-
  private:
   // Allow tests to corrupt our innards for testing purposes.
   FRIEND_TEST_ALL_PREFIXES(HistogramTest, CorruptBucketBounds);
@@ -589,12 +542,13 @@ class BASE_EXPORT Histogram : public HistogramBase {
                       const std::string& newline,
                       std::string* output) const;
 
-  // Find out how large the (graphically) the largest bucket will appear to be.
-  double GetPeakBucketSize(const SampleSet& snapshot) const;
+  // Find out how large (graphically) the largest bucket will appear to be.
+  double GetPeakBucketSize(const SampleVector& samples) const;
 
   // Write a common header message describing this histogram.
-  void WriteAsciiHeader(const SampleSet& snapshot,
-                        Count sample_count, std::string* output) const;
+  void WriteAsciiHeader(const SampleVector& samples,
+                        Count sample_count,
+                        std::string* output) const;
 
   // Write information about previous, current, and next buckets.
   // Information such as cumulative percentage, etc.
@@ -620,7 +574,7 @@ class BASE_EXPORT Histogram : public HistogramBase {
 
   // Finally, provide the state that changes with the addition of each new
   // sample.
-  SampleSet sample_;
+  scoped_ptr<SampleVector> samples_;
 
   DISALLOW_COPY_AND_ASSIGN(Histogram);
 };
diff --git a/base/metrics/histogram_flattener.h b/base/metrics/histogram_flattener.h
index 01a49fc..6dd169e 100644
--- a/base/metrics/histogram_flattener.h
+++ b/base/metrics/histogram_flattener.h
@@ -14,14 +14,16 @@
 
 namespace base {
 
+class HistogramSamples;
+
 // HistogramFlattener is an interface used by HistogramSnapshotManager, which
 // handles the logistics of gathering up available histograms for recording.
 // The implementors handle the exact lower level recording mechanism, or
 // error report mechanism.
 class BASE_EXPORT HistogramFlattener {
  public:
-  virtual void RecordDelta(const base::Histogram& histogram,
-                           const base::Histogram::SampleSet& snapshot) = 0;
+  virtual void RecordDelta(const Histogram& histogram,
+                           const HistogramSamples& snapshot) = 0;
 
   // Will be called each time a type of Inconsistenies is seen on a histogram,
   // during inspections done internally in HistogramSnapshotManager class.
diff --git a/base/metrics/histogram_samples.cc b/base/metrics/histogram_samples.cc
new file mode 100644
index 0000000..e375952
--- /dev/null
+++ b/base/metrics/histogram_samples.cc
@@ -0,0 +1,126 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/metrics/histogram_samples.h"
+
+#include "base/compiler_specific.h"
+#include "base/pickle.h"
+
+namespace base {
+
+namespace {
+
+class SampleCountPickleIterator : public SampleCountIterator {
+ public:
+  SampleCountPickleIterator(PickleIterator* iter);
+
+  virtual bool Done() const OVERRIDE;
+  virtual void Next() OVERRIDE;
+  virtual void Get(HistogramBase::Sample* min,
+                   HistogramBase::Sample* max,
+                   HistogramBase::Count* count) const OVERRIDE;
+ private:
+  PickleIterator* const iter_;
+
+  HistogramBase::Sample min_;
+  HistogramBase::Sample max_;
+  HistogramBase::Count count_;
+  bool is_done_;
+};
+
+SampleCountPickleIterator::SampleCountPickleIterator(PickleIterator* iter)
+    : iter_(iter),
+      is_done_(false) {
+  Next();
+}
+
+bool SampleCountPickleIterator::Done() const {
+  return is_done_;
+}
+
+void SampleCountPickleIterator::Next() {
+  DCHECK(!Done());
+  if (!iter_->ReadInt(&min_) ||
+      !iter_->ReadInt(&max_) ||
+      !iter_->ReadInt(&count_))
+    is_done_ = true;
+}
+
+void SampleCountPickleIterator::Get(HistogramBase::Sample* min,
+                                    HistogramBase::Sample* max,
+                                    HistogramBase::Count* count) const {
+  DCHECK(!Done());
+  *min = min_;
+  *max = max_;
+  *count = count_;
+}
+
+}  // namespace
+
+HistogramSamples::HistogramSamples() : sum_(0), redundant_count_(0) {}
+
+HistogramSamples::~HistogramSamples() {}
+
+void HistogramSamples::Add(const HistogramSamples& other) {
+  sum_ += other.sum();
+  redundant_count_ += other.redundant_count();
+  bool success = AddSubtractImpl(other.Iterator().get(), ADD);
+  DCHECK(success);
+}
+
+bool HistogramSamples::AddFromPickle(PickleIterator* iter) {
+  int64 sum;
+  HistogramBase::Count redundant_count;
+
+  if (!iter->ReadInt64(&sum) || !iter->ReadInt(&redundant_count))
+    return false;
+  sum_ += sum;
+  redundant_count_ += redundant_count;
+
+  SampleCountPickleIterator pickle_iter(iter);
+  return AddSubtractImpl(&pickle_iter, ADD);
+}
+
+void HistogramSamples::Subtract(const HistogramSamples& other) {
+  sum_ -= other.sum();
+  redundant_count_ -= other.redundant_count();
+  bool success = AddSubtractImpl(other.Iterator().get(), SUBTRACT);
+  DCHECK(success);
+}
+
+bool HistogramSamples::Serialize(Pickle* pickle) const {
+  if (!pickle->WriteInt64(sum_) || !pickle->WriteInt(redundant_count_))
+    return false;
+
+  HistogramBase::Sample min;
+  HistogramBase::Sample max;
+  HistogramBase::Count count;
+  for (scoped_ptr<SampleCountIterator> it = Iterator();
+       !it->Done();
+       it->Next()) {
+    it->Get(&min, &max, &count);
+    if (!pickle->WriteInt(min) ||
+        !pickle->WriteInt(max) ||
+        !pickle->WriteInt(count))
+      return false;
+  }
+  return true;
+}
+
+void HistogramSamples::IncreaseSum(int64 diff) {
+  sum_ += diff;
+}
+
+void HistogramSamples::IncreaseRedundantCount(HistogramBase::Count diff) {
+  redundant_count_ += diff;
+}
+
+SampleCountIterator::~SampleCountIterator() {}
+
+bool SampleCountIterator::GetBucketIndex(size_t* index) const {
+  DCHECK(!Done());
+  return false;
+}
+
+}  // namespace base
diff --git a/base/metrics/histogram_samples.h b/base/metrics/histogram_samples.h
new file mode 100644
index 0000000..778b11d
--- /dev/null
+++ b/base/metrics/histogram_samples.h
@@ -0,0 +1,89 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#ifndef BASE_METRICS_HISTOGRAM_SAMPLES_H_
+#define BASE_METRICS_HISTOGRAM_SAMPLES_H_
+
+#include "base/basictypes.h"
+#include "base/metrics/histogram_base.h"
+#include "base/memory/scoped_ptr.h"
+
+class Pickle;
+class PickleIterator;
+
+namespace base {
+
+class SampleCountIterator;
+
+// HistogramSamples is a container storing all samples of a histogram.
+class BASE_EXPORT HistogramSamples {
+ public:
+  HistogramSamples();
+  virtual ~HistogramSamples();
+
+  virtual void Accumulate(HistogramBase::Sample value,
+                          HistogramBase::Count count) = 0;
+  virtual HistogramBase::Count GetCount(HistogramBase::Sample value) const = 0;
+  virtual HistogramBase::Count TotalCount() const = 0;
+
+  virtual void Add(const HistogramSamples& other);
+
+  // Add from serialized samples.
+  virtual bool AddFromPickle(PickleIterator* iter);
+
+  virtual void Subtract(const HistogramSamples& other);
+
+  virtual scoped_ptr<SampleCountIterator> Iterator() const = 0;
+  virtual bool Serialize(Pickle* pickle) const;
+
+  // Accessor fuctions.
+  int64 sum() const { return sum_; }
+  HistogramBase::Count redundant_count() const { return redundant_count_; }
+
+ protected:
+  // Based on |instruction| type, add or subtract sample counts data from the
+  // iterator.
+  enum Instruction { ADD, SUBTRACT };
+  virtual bool AddSubtractImpl(SampleCountIterator* iter,
+                               Instruction instruction) = 0;
+
+  void IncreaseSum(int64 diff);
+  void IncreaseRedundantCount(HistogramBase::Count diff);
+
+ private:
+  int64 sum_;
+
+  // |redundant_count_| helps identify memory corruption. It redundantly stores
+  // the total number of samples accumulated in the histogram. We can compare
+  // this count to the sum of the counts (TotalCount() function), and detect
+  // problems. Note, depending on the implementation of different histogram
+  // types, there might be races during histogram accumulation and snapshotting
+  // that we choose to accept. In this case, the tallies might mismatch even
+  // when no memory corruption has happened.
+  HistogramBase::Count redundant_count_;
+};
+
+class BASE_EXPORT SampleCountIterator {
+ public:
+  virtual ~SampleCountIterator();
+
+  virtual bool Done() const = 0;
+  virtual void Next() = 0;
+
+  // Get the sample and count at current position.
+  // |min| |max| and |count| can be NULL if the value is not of interest.
+  // Requires: !Done();
+  virtual void Get(HistogramBase::Sample* min,
+                   HistogramBase::Sample* max,
+                   HistogramBase::Count* count) const = 0;
+
+  // Get the index of current histogram bucket.
+  // For histograms that don't use predefined buckets, it returns false.
+  // Requires: !Done();
+  virtual bool GetBucketIndex(size_t* index) const;
+};
+
+}  // namespace base
+
+#endif  // BASE_METRICS_HISTOGRAM_SAMPLES_H_
diff --git a/base/metrics/histogram_snapshot_manager.cc b/base/metrics/histogram_snapshot_manager.cc
index b75932c..ad054b1 100644
--- a/base/metrics/histogram_snapshot_manager.cc
+++ b/base/metrics/histogram_snapshot_manager.cc
@@ -4,10 +4,14 @@
 
 #include "base/metrics/histogram_snapshot_manager.h"
 
+#include "base/memory/scoped_ptr.h"
+#include "base/metrics/histogram_flattener.h"
+#include "base/metrics/histogram_samples.h"
 #include "base/metrics/statistics_recorder.h"
+#include "base/stl_util.h"
 
-using base::Histogram;
-using base::StatisticsRecorder;
+using std::map;
+using std::string;
 
 namespace base {
 
@@ -17,7 +21,9 @@ HistogramSnapshotManager::HistogramSnapshotManager(
   DCHECK(histogram_flattener_);
 }
 
-HistogramSnapshotManager::~HistogramSnapshotManager() {}
+HistogramSnapshotManager::~HistogramSnapshotManager() {
+  STLDeleteValues(&logged_samples_);
+}
 
 void HistogramSnapshotManager::PrepareDeltas(Histogram::Flags flag_to_set,
                                              bool record_only_uma) {
@@ -38,11 +44,10 @@ void HistogramSnapshotManager::PrepareDelta(const Histogram& histogram) {
   DCHECK(histogram_flattener_);
 
   // Get up-to-date snapshot of sample stats.
-  Histogram::SampleSet snapshot;
-  histogram.SnapshotSample(&snapshot);
+  scoped_ptr<HistogramSamples> snapshot(histogram.SnapshotSamples());
   const std::string& histogram_name = histogram.histogram_name();
 
-  int corruption = histogram.FindCorruption(snapshot);
+  int corruption = histogram.FindCorruption(*snapshot);
 
   // Crash if we detect that our histograms have been overwritten.  This may be
   // a fair distance from the memory smasher, but we hope to correlate these
@@ -59,54 +64,54 @@ void HistogramSnapshotManager::PrepareDelta(const Histogram& histogram) {
   // COUNT_LOW_ERROR and they never arise together, so we don't need to extract
   // bits from corruption.
   if (corruption) {
-    NOTREACHED();
+    DLOG(ERROR) << "Histogram: " << histogram_name
+                << " has data corruption: " << corruption;
     histogram_flattener_->InconsistencyDetected(
         static_cast<Histogram::Inconsistencies>(corruption));
-    // Don't record corrupt data to metrics survices.
-    if (NULL == inconsistencies_.get())
-      inconsistencies_.reset(new ProblemMap);
-    int old_corruption = (*inconsistencies_)[histogram_name];
+    // Don't record corrupt data to metrics services.
+    int old_corruption = inconsistencies_[histogram_name];
     if (old_corruption == (corruption | old_corruption))
       return;  // We've already seen this corruption for this histogram.
-    (*inconsistencies_)[histogram_name] |= corruption;
+    inconsistencies_[histogram_name] |= corruption;
     histogram_flattener_->UniqueInconsistencyDetected(
         static_cast<Histogram::Inconsistencies>(corruption));
     return;
   }
 
-  // Find the already recorded stats, or create an empty set. Remove from our
-  // snapshot anything that we've already recorded.
-  LoggedSampleMap::iterator it = logged_samples_.find(histogram_name);
-  Histogram::SampleSet* already_logged;
-  if (logged_samples_.end() == it) {
-    // Add new entry
-    already_logged = &logged_samples_[histogram.histogram_name()];
-    // Complete initialization.
-    already_logged->Resize(histogram.bucket_count());
+  HistogramSamples* to_log;
+  map<string, HistogramSamples*>::iterator it =
+      logged_samples_.find(histogram_name);
+  if (it == logged_samples_.end()) {
+    to_log = snapshot.release();
+
+    // This histogram has not been logged before, add a new entry.
+    logged_samples_[histogram_name] = to_log;
   } else {
-    already_logged = &(it->second);
-    int64 discrepancy(already_logged->TotalCount() -
-                      already_logged->redundant_count());
-    if (discrepancy) {
-      NOTREACHED();  // Already_logged has become corrupt.
-      int problem = static_cast<int>(discrepancy);
-      if (problem != discrepancy)
-        problem = INT_MAX;
-      histogram_flattener_->InconsistencyDetectedInLoggedCount(problem);
-      // With no valid baseline, we'll act like we've recorded everything in our
-      // snapshot.
-      already_logged->Subtract(*already_logged);
-      already_logged->Add(snapshot);
-    }
-    // Deduct any stats we've already logged from our snapshot.
-    snapshot.Subtract(*already_logged);
+    HistogramSamples* already_logged = it->second;
+    InspectLoggedSamplesInconsistency(*snapshot, already_logged);
+    snapshot->Subtract(*already_logged);
+    already_logged->Add(*snapshot);
+    to_log = snapshot.get();
   }
 
-  // Snapshot now contains only a delta to what we've already_logged.
-  if (snapshot.redundant_count() > 0) {
-    histogram_flattener_->RecordDelta(histogram, snapshot);
-    // Add new data into our running total.
-    already_logged->Add(snapshot);
+  if (to_log->redundant_count() > 0)
+    histogram_flattener_->RecordDelta(histogram, *to_log);
+}
+
+void HistogramSnapshotManager::InspectLoggedSamplesInconsistency(
+      const HistogramSamples& new_snapshot,
+      HistogramSamples* logged_samples) {
+  HistogramBase::Count discrepancy =
+      logged_samples->TotalCount() - logged_samples->redundant_count();
+  if (!discrepancy)
+    return;
+
+  histogram_flattener_->InconsistencyDetectedInLoggedCount(discrepancy);
+  if (discrepancy > Histogram::kCommonRaceBasedCountMismatch) {
+    // Fix logged_samples.
+    logged_samples->Subtract(*logged_samples);
+    logged_samples->Add(new_snapshot);
   }
 }
+
 }  // namespace base
diff --git a/base/metrics/histogram_snapshot_manager.h b/base/metrics/histogram_snapshot_manager.h
index a945ac7..0ec1ac3 100644
--- a/base/metrics/histogram_snapshot_manager.h
+++ b/base/metrics/histogram_snapshot_manager.h
@@ -9,12 +9,13 @@
 #include <string>
 
 #include "base/basictypes.h"
-#include "base/memory/scoped_ptr.h"
 #include "base/metrics/histogram.h"
-#include "base/metrics/histogram_flattener.h"
 
 namespace base {
 
+class HistogramSamples;
+class HistogramFlattener;
+
 // HistogramSnapshotManager handles the logistics of gathering up available
 // histograms for recording either to disk or for transmission (such as from
 // renderer to browser, or from browser to UMA upload). Since histograms can sit
@@ -30,23 +31,23 @@ class BASE_EXPORT HistogramSnapshotManager {
   // Snapshot all histograms, and ask |histogram_flattener_| to record the
   // delta. The arguments allow selecting only a subset of histograms for
   // recording, or to set a flag in each recorded histogram.
-  void PrepareDeltas(base::Histogram::Flags flags_to_set, bool record_only_uma);
+  void PrepareDeltas(Histogram::Flags flags_to_set, bool record_only_uma);
 
  private:
-  // Maintain a map of histogram names to the sample stats we've recorded.
-  typedef std::map<std::string, base::Histogram::SampleSet> LoggedSampleMap;
-  // List of histograms names, and their encontered corruptions.
-  typedef std::map<std::string, int> ProblemMap;
-
   // Snapshot this histogram, and record the delta.
-  void PrepareDelta(const base::Histogram& histogram);
+  void PrepareDelta(const Histogram& histogram);
+
+  // Try to detect and fix count inconsistency of logged samples.
+  void InspectLoggedSamplesInconsistency(
+      const HistogramSamples& new_snapshot,
+      HistogramSamples* logged_samples);
 
   // For histograms, track what we've already recorded (as a sample for
   // each histogram) so that we can record only the delta with the next log.
-  LoggedSampleMap logged_samples_;
+  std::map<std::string, HistogramSamples*> logged_samples_;
 
-  // List of histograms found corrupt to be corrupt, and their problems.
-  scoped_ptr<ProblemMap> inconsistencies_;
+  // List of histograms found to be corrupt, and their problems.
+  std::map<std::string, int> inconsistencies_;
 
   // |histogram_flattener_| handles the logistics of recording the histogram
   // deltas.
diff --git a/base/metrics/histogram_unittest.cc b/base/metrics/histogram_unittest.cc
index 7f23ede..f016487 100644
--- a/base/metrics/histogram_unittest.cc
+++ b/base/metrics/histogram_unittest.cc
@@ -4,6 +4,7 @@
 
 // Test of Histogram class
 
+#include <climits>
 #include <algorithm>
 #include <vector>
 
@@ -11,6 +12,7 @@
 #include "base/memory/scoped_ptr.h"
 #include "base/metrics/bucket_ranges.h"
 #include "base/metrics/histogram.h"
+#include "base/metrics/sample_vector.h"
 #include "base/metrics/statistics_recorder.h"
 #include "base/time.h"
 #include "testing/gtest/include/gtest/gtest.h"
@@ -206,14 +208,13 @@ TEST(HistogramTest, BoundsTest) {
   histogram->Add(10000);
 
   // Verify they landed in the underflow, and overflow buckets.
-  Histogram::SampleSet sample;
-  histogram->SnapshotSample(&sample);
-  EXPECT_EQ(2, sample.counts(0));
-  EXPECT_EQ(0, sample.counts(1));
+  scoped_ptr<SampleVector> samples = histogram->SnapshotSamples();
+  EXPECT_EQ(2, samples->GetCountAtIndex(0));
+  EXPECT_EQ(0, samples->GetCountAtIndex(1));
   size_t array_size = histogram->bucket_count();
   EXPECT_EQ(kBucketCount, array_size);
-  EXPECT_EQ(0, sample.counts(array_size - 2));
-  EXPECT_EQ(2, sample.counts(array_size - 1));
+  EXPECT_EQ(0, samples->GetCountAtIndex(array_size - 2));
+  EXPECT_EQ(2, samples->GetCountAtIndex(array_size - 1));
 
   vector<int> custom_ranges;
   custom_ranges.push_back(10);
@@ -227,15 +228,16 @@ TEST(HistogramTest, BoundsTest) {
   test_custom_histogram->Add(-50);
   test_custom_histogram->Add(100);
   test_custom_histogram->Add(1000);
+  test_custom_histogram->Add(INT_MAX);
 
   // Verify they landed in the underflow, and overflow buckets.
-  Histogram::SampleSet custom_sample;
-  test_custom_histogram->SnapshotSample(&custom_sample);
-  EXPECT_EQ(2, custom_sample.counts(0));
-  EXPECT_EQ(0, custom_sample.counts(1));
-  size_t custom_array_size = test_custom_histogram->bucket_count();
-  EXPECT_EQ(0, custom_sample.counts(custom_array_size - 2));
-  EXPECT_EQ(2, custom_sample.counts(custom_array_size - 1));
+  scoped_ptr<SampleVector> custom_samples =
+      test_custom_histogram->SnapshotSamples();
+  EXPECT_EQ(2, custom_samples->GetCountAtIndex(0));
+  EXPECT_EQ(0, custom_samples->GetCountAtIndex(1));
+  size_t bucket_count = test_custom_histogram->bucket_count();
+  EXPECT_EQ(0, custom_samples->GetCountAtIndex(bucket_count - 2));
+  EXPECT_EQ(3, custom_samples->GetCountAtIndex(bucket_count - 1));
 }
 
 // Check to be sure samples land as expected is "correct" buckets.
@@ -253,45 +255,45 @@ TEST(HistogramTest, BucketPlacementTest) {
   }
 
   // Check to see that the bucket counts reflect our additions.
-  Histogram::SampleSet sample;
-  histogram->SnapshotSample(&sample);
+  scoped_ptr<SampleVector> samples = histogram->SnapshotSamples();
   for (int i = 0; i < 8; i++)
-    EXPECT_EQ(i + 1, sample.counts(i));
+    EXPECT_EQ(i + 1, samples->GetCountAtIndex(i));
 }
 
 TEST(HistogramTest, CorruptSampleCounts) {
   Histogram* histogram(Histogram::FactoryGet(
       "Histogram", 1, 64, 8, Histogram::kNoFlags));  // As per header file.
 
-  EXPECT_EQ(0, histogram->sample_.redundant_count());
-  histogram->Add(20);  // Add some samples.
+  // Add some samples.
+  histogram->Add(20);
   histogram->Add(40);
-  EXPECT_EQ(2, histogram->sample_.redundant_count());
 
-  Histogram::SampleSet snapshot;
-  histogram->SnapshotSample(&snapshot);
-  EXPECT_EQ(Histogram::NO_INCONSISTENCIES, 0);
-  EXPECT_EQ(0, histogram->FindCorruption(snapshot));  // No default corruption.
-  EXPECT_EQ(2, snapshot.redundant_count());
+  scoped_ptr<SampleVector> snapshot = histogram->SnapshotSamples();
+  EXPECT_EQ(Histogram::NO_INCONSISTENCIES,
+            histogram->FindCorruption(*snapshot));
+  EXPECT_EQ(2, snapshot->redundant_count());
+  EXPECT_EQ(2, snapshot->TotalCount());
 
-  snapshot.counts_[3] += 100;  // Sample count won't match redundant count.
-  EXPECT_EQ(Histogram::COUNT_LOW_ERROR, histogram->FindCorruption(snapshot));
-  snapshot.counts_[2] -= 200;
-  EXPECT_EQ(Histogram::COUNT_HIGH_ERROR, histogram->FindCorruption(snapshot));
+  snapshot->counts_[3] += 100;  // Sample count won't match redundant count.
+  EXPECT_EQ(Histogram::COUNT_LOW_ERROR,
+            histogram->FindCorruption(*snapshot));
+  snapshot->counts_[2] -= 200;
+  EXPECT_EQ(Histogram::COUNT_HIGH_ERROR,
+            histogram->FindCorruption(*snapshot));
 
   // But we can't spot a corruption if it is compensated for.
-  snapshot.counts_[1] += 100;
-  EXPECT_EQ(0, histogram->FindCorruption(snapshot));
+  snapshot->counts_[1] += 100;
+  EXPECT_EQ(Histogram::NO_INCONSISTENCIES,
+            histogram->FindCorruption(*snapshot));
 }
 
 TEST(HistogramTest, CorruptBucketBounds) {
   Histogram* histogram(Histogram::FactoryGet(
       "Histogram", 1, 64, 8, Histogram::kNoFlags));  // As per header file.
 
-  Histogram::SampleSet snapshot;
-  histogram->SnapshotSample(&snapshot);
-  EXPECT_EQ(Histogram::NO_INCONSISTENCIES, 0);
-  EXPECT_EQ(0, histogram->FindCorruption(snapshot));  // No default corruption.
+  scoped_ptr<SampleVector> snapshot = histogram->SnapshotSamples();
+  EXPECT_EQ(Histogram::NO_INCONSISTENCIES,
+            histogram->FindCorruption(*snapshot));
 
   BucketRanges* bucket_ranges =
       const_cast<BucketRanges*>(histogram->bucket_ranges());
@@ -299,20 +301,20 @@ TEST(HistogramTest, CorruptBucketBounds) {
   bucket_ranges->set_range(1, bucket_ranges->range(2));
   bucket_ranges->set_range(2, tmp);
   EXPECT_EQ(Histogram::BUCKET_ORDER_ERROR | Histogram::RANGE_CHECKSUM_ERROR,
-            histogram->FindCorruption(snapshot));
+            histogram->FindCorruption(*snapshot));
 
   bucket_ranges->set_range(2, bucket_ranges->range(1));
   bucket_ranges->set_range(1, tmp);
-  EXPECT_EQ(0, histogram->FindCorruption(snapshot));
+  EXPECT_EQ(0, histogram->FindCorruption(*snapshot));
 
+  // Show that two simple changes don't offset each other
   bucket_ranges->set_range(3, bucket_ranges->range(3) + 1);
   EXPECT_EQ(Histogram::RANGE_CHECKSUM_ERROR,
-            histogram->FindCorruption(snapshot));
+            histogram->FindCorruption(*snapshot));
 
-  // Show that two simple changes don't offset each other
   bucket_ranges->set_range(4, bucket_ranges->range(4) - 1);
   EXPECT_EQ(Histogram::RANGE_CHECKSUM_ERROR,
-            histogram->FindCorruption(snapshot));
+            histogram->FindCorruption(*snapshot));
 
   // Repair histogram so that destructor won't DCHECK().
   bucket_ranges->set_range(3, bucket_ranges->range(3) - 1);
diff --git a/base/metrics/sample_vector.cc b/base/metrics/sample_vector.cc
new file mode 100644
index 0000000..327a1f2
--- /dev/null
+++ b/base/metrics/sample_vector.cc
@@ -0,0 +1,160 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include "base/metrics/sample_vector.h"
+
+#include "base/logging.h"
+#include "base/metrics/bucket_ranges.h"
+
+using std::vector;
+
+namespace base {
+
+typedef HistogramBase::Count Count;
+typedef HistogramBase::Sample Sample;
+
+SampleVector::SampleVector(const BucketRanges* bucket_ranges)
+    : counts_(bucket_ranges->size() - 1),
+      bucket_ranges_(bucket_ranges) {
+  CHECK_GE(bucket_ranges_->size(), 2u);
+}
+
+SampleVector::~SampleVector() {}
+
+void SampleVector::Accumulate(Sample value, Count count) {
+  size_t bucket_index = GetBucketIndex(value);
+  counts_[bucket_index] += count;
+  IncreaseSum(count * value);
+  IncreaseRedundantCount(count);
+}
+
+Count SampleVector::GetCount(Sample value) const {
+  size_t bucket_index = GetBucketIndex(value);
+  return counts_[bucket_index];
+}
+
+Count SampleVector::TotalCount() const {
+  Count count = 0;
+  for (size_t i = 0; i < counts_.size(); i++) {
+    count += counts_[i];
+  }
+  return count;
+}
+
+Count SampleVector::GetCountAtIndex(size_t bucket_index) const {
+  DCHECK(bucket_index >= 0 && bucket_index < counts_.size());
+  return counts_[bucket_index];
+}
+
+scoped_ptr<SampleCountIterator> SampleVector::Iterator() const {
+  return scoped_ptr<SampleCountIterator>(
+      new SampleVectorIterator(&counts_, bucket_ranges_));
+}
+
+bool SampleVector::AddSubtractImpl(SampleCountIterator* iter,
+                                   HistogramSamples::Instruction instruction) {
+  HistogramBase::Sample min;
+  HistogramBase::Sample max;
+  HistogramBase::Count count;
+
+  // Go through the iterator and add the counts into correct bucket.
+  size_t index = 0;
+  while (index < counts_.size() && !iter->Done()) {
+    iter->Get(&min, &max, &count);
+    if (min == bucket_ranges_->range(index) &&
+        max == bucket_ranges_->range(index + 1)) {
+      // Sample matches this bucket!
+      counts_[index] +=
+          (instruction ==  HistogramSamples::ADD) ? count : -count;
+      iter->Next();
+    } else if (min > bucket_ranges_->range(index)) {
+      // Sample is larger than current bucket range. Try next.
+      index++;
+    } else {
+      // Sample is smaller than current bucket range. We scan buckets from
+      // smallest to largest, so the sample value must be invalid.
+      return false;
+    }
+  }
+
+  return iter->Done();
+}
+
+// Use simple binary search.  This is very general, but there are better
+// approaches if we knew that the buckets were linearly distributed.
+size_t SampleVector::GetBucketIndex(Sample value) const {
+  size_t bucket_count = bucket_ranges_->size() - 1;
+  CHECK_GE(bucket_count, 1u);
+  CHECK_GE(value, bucket_ranges_->range(0));
+  CHECK_LT(value, bucket_ranges_->range(bucket_count));
+
+  size_t under = 0;
+  size_t over = bucket_count;
+  size_t mid;
+  do {
+    DCHECK_GE(over, under);
+    mid = under + (over - under)/2;
+    if (mid == under)
+      break;
+    if (bucket_ranges_->range(mid) <= value)
+      under = mid;
+    else
+      over = mid;
+  } while (true);
+
+  DCHECK_LE(bucket_ranges_->range(mid), value);
+  CHECK_GT(bucket_ranges_->range(mid + 1), value);
+  return mid;
+}
+
+SampleVectorIterator::SampleVectorIterator(const vector<Count>* counts,
+                                           const BucketRanges* bucket_ranges)
+    : counts_(counts),
+      bucket_ranges_(bucket_ranges),
+      index_(0) {
+  CHECK_GT(bucket_ranges_->size(), counts_->size());
+  SkipEmptyBuckets();
+}
+
+bool SampleVectorIterator::Done() const {
+  return index_ >= counts_->size();
+}
+
+void SampleVectorIterator::Next() {
+  DCHECK(!Done());
+  index_++;
+  SkipEmptyBuckets();
+}
+
+void SampleVectorIterator::Get(HistogramBase::Sample* min,
+                               HistogramBase::Sample* max,
+                               HistogramBase::Count* count) const {
+  DCHECK(!Done());
+  if (min != NULL)
+    *min = bucket_ranges_->range(index_);
+  if (max != NULL)
+    *max = bucket_ranges_->range(index_ + 1);
+  if (count != NULL)
+    *count = (*counts_)[index_];
+}
+
+bool SampleVectorIterator::GetBucketIndex(size_t* index) const {
+  DCHECK(!Done());
+  if (index != NULL)
+    *index = index_;
+  return true;
+}
+
+void SampleVectorIterator::SkipEmptyBuckets() {
+  if (Done())
+    return;
+
+  while (index_ < counts_->size()) {
+    if ((*counts_)[index_] != 0)
+      return;
+    index_++;
+  }
+}
+
+}  // namespace base
diff --git a/base/metrics/sample_vector.h b/base/metrics/sample_vector.h
new file mode 100644
index 0000000..8938d36
--- /dev/null
+++ b/base/metrics/sample_vector.h
@@ -0,0 +1,81 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// SampleVector implements HistogramSamples interface. It is used by all
+// Histogram based classes to store samples.
+
+#ifndef BASE_METRICS_SAMPLE_VECTOR_H_
+#define BASE_METRICS_SAMPLE_VECTOR_H_
+
+#include <vector>
+
+#include "base/compiler_specific.h"
+#include "base/gtest_prod_util.h"
+#include "base/memory/scoped_ptr.h"
+#include "base/metrics/histogram_base.h"
+#include "base/metrics/histogram_samples.h"
+
+namespace base {
+
+class BucketRanges;
+
+class BASE_EXPORT_PRIVATE SampleVector : public HistogramSamples {
+ public:
+  explicit SampleVector(const BucketRanges* bucket_ranges);
+  virtual ~SampleVector();
+
+  // HistogramSamples implementation:
+  virtual void Accumulate(HistogramBase::Sample value,
+                          HistogramBase::Count count) OVERRIDE;
+  virtual HistogramBase::Count GetCount(
+      HistogramBase::Sample value) const OVERRIDE;
+  virtual HistogramBase::Count TotalCount() const OVERRIDE;
+  virtual scoped_ptr<SampleCountIterator> Iterator() const OVERRIDE;
+
+  // Get count of a specific bucket.
+  HistogramBase::Count GetCountAtIndex(size_t bucket_index) const;
+
+ protected:
+  virtual bool AddSubtractImpl(
+      SampleCountIterator* iter,
+      HistogramSamples::Instruction instruction) OVERRIDE;
+
+  virtual size_t GetBucketIndex(HistogramBase::Sample value) const;
+
+ private:
+  FRIEND_TEST_ALL_PREFIXES(HistogramTest, CorruptSampleCounts);
+
+  std::vector<HistogramBase::Count> counts_;
+
+  // Shares the same BucketRanges with Histogram object.
+  const BucketRanges* const bucket_ranges_;
+
+  DISALLOW_COPY_AND_ASSIGN(SampleVector);
+};
+
+class BASE_EXPORT_PRIVATE SampleVectorIterator : public SampleCountIterator {
+ public:
+  SampleVectorIterator(const std::vector<HistogramBase::Count>* counts,
+                       const BucketRanges* bucket_ranges);
+
+  // SampleCountIterator implementation:
+  virtual bool Done() const OVERRIDE;
+  virtual void Next() OVERRIDE;
+  virtual void Get(HistogramBase::Sample* min,
+                   HistogramBase::Sample* max,
+                   HistogramBase::Count* count) const OVERRIDE;
+  virtual bool GetBucketIndex(size_t* index) const OVERRIDE;
+
+ private:
+  void SkipEmptyBuckets();
+
+  const std::vector<HistogramBase::Count>* counts_;
+  const BucketRanges* bucket_ranges_;
+
+  size_t index_;
+};
+
+}  // namespace base
+
+#endif  // BASE_METRICS_SAMPLE_VECTOR_H_
diff --git a/base/metrics/sample_vector_unittest.cc b/base/metrics/sample_vector_unittest.cc
new file mode 100644
index 0000000..13e5e8b
--- /dev/null
+++ b/base/metrics/sample_vector_unittest.cc
@@ -0,0 +1,265 @@
+// Copyright (c) 2012 The Chromium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+#include <vector>
+
+#include "base/memory/scoped_ptr.h"
+#include "base/metrics/bucket_ranges.h"
+#include "base/metrics/histogram.h"
+#include "base/metrics/sample_vector.h"
+#include "testing/gtest/include/gtest/gtest.h"
+
+using std::vector;
+
+namespace base {
+namespace {
+
+TEST(SampleVectorTest, AccumulateTest) {
+  // Custom buckets: [1, 5) [5, 10)
+  BucketRanges ranges(3);
+  ranges.set_range(0, 1);
+  ranges.set_range(1, 5);
+  ranges.set_range(2, 10);
+  SampleVector samples(&ranges);
+
+  samples.Accumulate(1, 200);
+  samples.Accumulate(2, -300);
+  EXPECT_EQ(-100, samples.GetCountAtIndex(0));
+
+  samples.Accumulate(5, 200);
+  EXPECT_EQ(200, samples.GetCountAtIndex(1));
+
+  EXPECT_EQ(600, samples.sum());
+  EXPECT_EQ(100, samples.redundant_count());
+  EXPECT_EQ(samples.TotalCount(), samples.redundant_count());
+
+  samples.Accumulate(5, -100);
+  EXPECT_EQ(100, samples.GetCountAtIndex(1));
+
+  EXPECT_EQ(100, samples.sum());
+  EXPECT_EQ(0, samples.redundant_count());
+  EXPECT_EQ(samples.TotalCount(), samples.redundant_count());
+}
+
+TEST(SampleVectorTest, AddSubtractTest) {
+  // Custom buckets: [0, 1) [1, 2) [2, 3) [3, INT_MAX)
+  BucketRanges ranges(5);
+  ranges.set_range(0, 0);
+  ranges.set_range(1, 1);
+  ranges.set_range(2, 2);
+  ranges.set_range(3, 3);
+  ranges.set_range(4, INT_MAX);
+
+  SampleVector samples1(&ranges);
+  samples1.Accumulate(0, 100);
+  samples1.Accumulate(2, 100);
+  samples1.Accumulate(4, 100);
+  EXPECT_EQ(600, samples1.sum());
+  EXPECT_EQ(300, samples1.TotalCount());
+  EXPECT_EQ(samples1.redundant_count(), samples1.TotalCount());
+
+  SampleVector samples2(&ranges);
+  samples2.Accumulate(1, 200);
+  samples2.Accumulate(2, 200);
+  samples2.Accumulate(4, 200);
+  EXPECT_EQ(1400, samples2.sum());
+  EXPECT_EQ(600, samples2.TotalCount());
+  EXPECT_EQ(samples2.redundant_count(), samples2.TotalCount());
+
+  samples1.Add(samples2);
+  EXPECT_EQ(100, samples1.GetCountAtIndex(0));
+  EXPECT_EQ(200, samples1.GetCountAtIndex(1));
+  EXPECT_EQ(300, samples1.GetCountAtIndex(2));
+  EXPECT_EQ(300, samples1.GetCountAtIndex(3));
+  EXPECT_EQ(2000, samples1.sum());
+  EXPECT_EQ(900, samples1.TotalCount());
+  EXPECT_EQ(samples1.redundant_count(), samples1.TotalCount());
+
+  samples1.Subtract(samples2);
+  EXPECT_EQ(100, samples1.GetCountAtIndex(0));
+  EXPECT_EQ(0, samples1.GetCountAtIndex(1));
+  EXPECT_EQ(100, samples1.GetCountAtIndex(2));
+  EXPECT_EQ(100, samples1.GetCountAtIndex(3));
+  EXPECT_EQ(600, samples1.sum());
+  EXPECT_EQ(300, samples1.TotalCount());
+  EXPECT_EQ(samples1.redundant_count(), samples1.TotalCount());
+}
+
+#if (!defined(NDEBUG) || defined(DCHECK_ALWAYS_ON)) && GTEST_HAS_DEATH_TEST
+TEST(SampleVectorDeathTest, BucketIndexTest) {
+  // 8 buckets with exponential layout:
+  // [0, 1) [1, 2) [2, 4) [4, 8) [8, 16) [16, 32) [32, 64) [64, INT_MAX)
+  BucketRanges ranges(9);
+  Histogram::InitializeBucketRanges(1, 64, 8, &ranges);
+  SampleVector samples(&ranges);
+
+  // Normal case
+  samples.Accumulate(0, 1);
+  samples.Accumulate(3, 2);
+  samples.Accumulate(64, 3);
+  EXPECT_EQ(1, samples.GetCount(0));
+  EXPECT_EQ(2, samples.GetCount(2));
+  EXPECT_EQ(3, samples.GetCount(65));
+
+  // Extreme case.
+  EXPECT_DEATH(samples.Accumulate(INT_MIN, 100), "");
+  EXPECT_DEATH(samples.Accumulate(-1, 100), "");
+  EXPECT_DEATH(samples.Accumulate(INT_MAX, 100), "");
+
+  // Custom buckets: [1, 5) [5, 10)
+  // Note, this is not a valid BucketRanges for Histogram because it does not
+  // have overflow buckets.
+  BucketRanges ranges2(3);
+  ranges2.set_range(0, 1);
+  ranges2.set_range(1, 5);
+  ranges2.set_range(2, 10);
+  SampleVector samples2(&ranges2);
+
+  // Normal case.
+  samples2.Accumulate(1, 1);
+  samples2.Accumulate(4, 1);
+  samples2.Accumulate(5, 2);
+  samples2.Accumulate(9, 2);
+  EXPECT_EQ(2, samples2.GetCount(1));
+  EXPECT_EQ(4, samples2.GetCount(5));
+
+  // Extreme case.
+  EXPECT_DEATH(samples2.Accumulate(0, 100), "");
+  EXPECT_DEATH(samples2.Accumulate(10, 100), "");
+}
+
+TEST(SampleVectorDeathTest, AddSubtractBucketNotMatchTest) {
+  // Custom buckets 1: [1, 3) [3, 5)
+  BucketRanges ranges1(3);
+  ranges1.set_range(0, 1);
+  ranges1.set_range(1, 3);
+  ranges1.set_range(2, 5);
+  SampleVector samples1(&ranges1);
+
+  // Custom buckets 2: [0, 1) [1, 3) [3, 6) [6, 7)
+  BucketRanges ranges2(5);
+  ranges2.set_range(0, 0);
+  ranges2.set_range(1, 1);
+  ranges2.set_range(2, 3);
+  ranges2.set_range(3, 6);
+  ranges2.set_range(4, 7);
+  SampleVector samples2(&ranges2);
+
+  samples2.Accumulate(1, 100);
+  samples1.Add(samples2);
+  EXPECT_EQ(100, samples1.GetCountAtIndex(0));
+
+  // Extra bucket in the beginning.
+  samples2.Accumulate(0, 100);
+  EXPECT_DEATH(samples1.Add(samples2), "");
+  EXPECT_DEATH(samples1.Subtract(samples2), "");
+
+  // Extra bucket in the end.
+  samples2.Accumulate(0, -100);
+  samples2.Accumulate(6, 100);
+  EXPECT_DEATH(samples1.Add(samples2), "");
+  EXPECT_DEATH(samples1.Subtract(samples2), "");
+
+  // Bucket not match: [3, 5) VS [3, 6)
+  samples2.Accumulate(6, -100);
+  samples2.Accumulate(3, 100);
+  EXPECT_DEATH(samples1.Add(samples2), "");
+  EXPECT_DEATH(samples1.Subtract(samples2), "");
+}
+
+// (!defined(NDEBUG) || defined(DCHECK_ALWAYS_ON)) && GTEST_HAS_DEATH_TEST
+#endif
+
+TEST(SampleVectorIteratorTest, IterateTest) {
+  BucketRanges ranges(5);
+  ranges.set_range(0, 0);
+  ranges.set_range(1, 1);
+  ranges.set_range(2, 2);
+  ranges.set_range(3, 3);
+  ranges.set_range(4, 4);
+
+  vector<HistogramBase::Count> counts(3);
+  counts[0] = 1;
+  counts[1] = 0;  // Iterator will bypass this empty bucket.
+  counts[2] = 2;
+
+  // BucketRanges can have larger size than counts.
+  SampleVectorIterator it(&counts, &ranges);
+  size_t index;
+
+  HistogramBase::Sample min;
+  HistogramBase::Sample max;
+  HistogramBase::Count count;
+  it.Get(&min, &max, &count);
+  EXPECT_EQ(0, min);
+  EXPECT_EQ(1, max);
+  EXPECT_EQ(1, count);
+  EXPECT_TRUE(it.GetBucketIndex(&index));
+  EXPECT_EQ(0u, index);
+
+  it.Next();
+  it.Get(&min, &max, &count);
+  EXPECT_EQ(2, min);
+  EXPECT_EQ(3, max);
+  EXPECT_EQ(2, count);
+  EXPECT_TRUE(it.GetBucketIndex(&index));
+  EXPECT_EQ(2u, index);
+
+  it.Next();
+  EXPECT_TRUE(it.Done());
+
+  // Create iterator from SampleVector.
+  SampleVector samples(&ranges);
+  samples.Accumulate(0, 0);
+  samples.Accumulate(1, 1);
+  samples.Accumulate(2, 2);
+  samples.Accumulate(3, 3);
+  scoped_ptr<SampleCountIterator> it2 = samples.Iterator();
+
+  int i;
+  for (i = 1; !it2->Done(); i++, it2->Next()) {
+    it2->Get(&min, &max, &count);
+    EXPECT_EQ(i, min);
+    EXPECT_EQ(i + 1, max);
+    EXPECT_EQ(i, count);
+
+    size_t index;
+    EXPECT_TRUE(it2->GetBucketIndex(&index));
+    EXPECT_EQ(static_cast<size_t>(i), index);
+  }
+  EXPECT_EQ(4, i);
+}
+
+#if (!defined(NDEBUG) || defined(DCHECK_ALWAYS_ON)) && GTEST_HAS_DEATH_TEST
+
+TEST(SampleVectorIteratorDeathTest, IterateDoneTest) {
+  BucketRanges ranges(5);
+  ranges.set_range(0, 0);
+  ranges.set_range(1, 1);
+  ranges.set_range(2, 2);
+  ranges.set_range(3, 3);
+  ranges.set_range(4, INT_MAX);
+  SampleVector samples(&ranges);
+
+  scoped_ptr<SampleCountIterator> it = samples.Iterator();
+
+  EXPECT_TRUE(it->Done());
+
+  HistogramBase::Sample min;
+  HistogramBase::Sample max;
+  HistogramBase::Count count;
+  EXPECT_DEATH(it->Get(&min, &max, &count), "");
+
+  EXPECT_DEATH(it->Next(), "");
+
+  samples.Accumulate(2, 100);
+  it = samples.Iterator();
+  EXPECT_FALSE(it->Done());
+}
+
+// (!defined(NDEBUG) || defined(DCHECK_ALWAYS_ON)) && GTEST_HAS_DEATH_TEST
+#endif
+
+}  // namespace
+}  // namespace base