Reduce the false positive rate for SafeBrowsing gethash requests

chrome/browser/safe_browsing/safe_browsing_database_bloom.cc

 // Copyright (c) 2006-2008 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 "chrome/browser/safe_browsing/safe_browsing_database_bloom.h"
 
 #include "base/compiler_specific.h"
 #include "base/file_util.h"
 #include "base/logging.h"
 #include "base/message_loop.h"
@@ skipping 12 matching lines @@
 using base::TimeDelta;
 
 // Database version.  If this is different than what's stored on disk, the
 // database is reset.
 static const int kDatabaseVersion = 6;
 
 // Don't want to create too small of a bloom filter initially while we're
 // downloading the data and then keep having to rebuild it.
 static const int kBloomFilterMinSize = 250000;
 
-// How many bits to use per item.  See the design doc for more information.
-static const int kBloomFilterSizeRatio = 13;
-
 // When we awake from a low power state, we try to avoid doing expensive disk
 // operations for a few minutes to let the system page itself in and settle
 // down.
 static const int kOnResumeHoldupMs = 5 * 60 * 1000;  // 5 minutes.
 
 // The maximum staleness for a cached entry.
 static const int kMaxStalenessMinutes = 45;
 
 // The bloom filter based file name suffix.
 static const FilePath::CharType kBloomFilterFileSuffix[] =
@@ skipping 1012 matching lines @@
   *filter = NULL;
   *new_add_count = 0;
 
   SQLITE_UNIQUE_STATEMENT(insert, *statement_cache_,
                           "INSERT INTO add_prefix VALUES (?,?)");
   if (!insert.is_valid()) {
     NOTREACHED();
     return false;
   }
 
+  // Determine the size of the new bloom filter. We will cap the maximum size at
+  // 2 MB to prevent an error from consuming large amounts of memory.
   int number_of_keys = std::max(add_count_, kBloomFilterMinSize);
-  int filter_size = number_of_keys * kBloomFilterSizeRatio;
+  int filter_size = std::min(number_of_keys * kBloomFilterSizeRatio,
+                             2 * 1024 * 1024 * 8);
   BloomFilter* new_filter = new BloomFilter(filter_size);
   SBPair* add = adds;
   int new_count = 0;
 
   while (add - adds < add_count_) {
     if (!adds_removed[add - adds]) {
       // Check to see if we have an AddDel for this chunk and skip writing it
       // if there is.
       if (add_del_cache_.find(add->chunk_id) != add_del_cache_.end()) {
         add++;
@@ skipping 270 matching lines @@
   UMA_HISTOGRAM_COUNTS("SB2.BuildWriteOperations",
                        static_cast<int>(io_after.WriteOperationCount -
                                         io_before.WriteOperationCount));
 #endif
   SB_DLOG(INFO) << "SafeBrowsingDatabaseImpl built bloom filter in "
                 << bloom_gen.InMilliseconds()
                 << " ms total.  prefix count: "<< add_count_;
   UMA_HISTOGRAM_LONG_TIMES("SB2.BuildFilter", bloom_gen);
   UMA_HISTOGRAM_COUNTS("SB2.AddPrefixes", add_count_);
   UMA_HISTOGRAM_COUNTS("SB2.SubPrefixes", subs);
+  UMA_HISTOGRAM_COUNTS("SB2.FilterSize", filter->size());
   int64 size_64;
   if (file_util::GetFileSize(filename_, &size_64))
     UMA_HISTOGRAM_COUNTS("SB2.DatabaseBytes", static_cast<int>(size_64));
 }
 
 void SafeBrowsingDatabaseBloom::GetCachedFullHashes(
     const std::vector<SBPrefix>* prefix_hits,
     std::vector<SBFullHashResult>* full_hits,
     Time last_update) {
   DCHECK(prefix_hits && full_hits);
@@ skipping 118 matching lines @@
   if (did_resume_) {
     PlatformThread::Sleep(kOnResumeHoldupMs);
     did_resume_ = false;
   }
 }
 
 // This database is always synchronous since we don't need to worry about
 // blocking any incoming reads.
 void SafeBrowsingDatabaseBloom::SetSynchronous() {
 }