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
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
|
// Copyright (c) 2010 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.h"
#include "base/command_line.h"
#include "base/file_util.h"
#include "base/histogram.h"
#include "base/time.h"
#include "base/message_loop.h"
#include "base/process_util.h"
#include "base/sha2.h"
#include "base/stats_counters.h"
#include "chrome/browser/safe_browsing/bloom_filter.h"
#include "chrome/browser/safe_browsing/safe_browsing_database_bloom.h"
#include "chrome/browser/safe_browsing/safe_browsing_store_file.h"
#include "chrome/browser/safe_browsing/safe_browsing_store_sqlite.h"
#include "chrome/common/chrome_switches.h"
#include "googleurl/src/gurl.h"
namespace {
// Filename suffix for the bloom filter.
const FilePath::CharType kBloomFilterFile[] = FILE_PATH_LITERAL(" Filter 2");
// The maximum staleness for a cached entry.
const int kMaxStalenessMinutes = 45;
// To save space, the incoming |chunk_id| and |list_id| are combined
// into an |encoded_chunk_id| for storage by shifting the |list_id|
// into the low-order bits. These functions decode that information.
int DecodeListId(const int encoded_chunk_id) {
return encoded_chunk_id & 1;
}
int DecodeChunkId(int encoded_chunk_id) {
return encoded_chunk_id >> 1;
}
int EncodeChunkId(int chunk, int list_id) {
DCHECK(list_id == 0 || list_id == 1);
return chunk << 1 | list_id;
}
// Generate the set of prefixes to check for |url|.
// TODO(shess): This function is almost the same as
// |CompareFullHashes()| in safe_browsing_util.cc, except that code
// does an early exit on match. Since match should be the infrequent
// case (phishing or malware found), consider combining this function
// with that one.
void PrefixesToCheck(const GURL& url, std::vector<SBPrefix>* prefixes) {
std::vector<std::string> hosts;
if (url.HostIsIPAddress()) {
hosts.push_back(url.host());
} else {
safe_browsing_util::GenerateHostsToCheck(url, &hosts);
}
std::vector<std::string> paths;
safe_browsing_util::GeneratePathsToCheck(url, &paths);
for (size_t i = 0; i < hosts.size(); ++i) {
for (size_t j = 0; j < paths.size(); ++j) {
SBFullHash full_hash;
base::SHA256HashString(hosts[i] + paths[j], &full_hash,
sizeof(SBFullHash));
prefixes->push_back(full_hash.prefix);
}
}
}
// Find the entries in |full_hashes| with prefix in |prefix_hits|, and
// add them to |full_hits| if not expired. "Not expired" is when
// either |last_update| was recent enough, or the item has been
// received recently enough. Expired items are not deleted because a
// future update may make them acceptable again.
//
// For efficiency reasons the code walks |prefix_hits| and
// |full_hashes| in parallel, so they must be sorted by prefix.
void GetCachedFullHashes(const std::vector<SBPrefix>& prefix_hits,
const std::vector<SBAddFullHash>& full_hashes,
std::vector<SBFullHashResult>* full_hits,
base::Time last_update) {
const base::Time expire_time =
base::Time::Now() - base::TimeDelta::FromMinutes(kMaxStalenessMinutes);
std::vector<SBPrefix>::const_iterator piter = prefix_hits.begin();
std::vector<SBAddFullHash>::const_iterator hiter = full_hashes.begin();
while (piter != prefix_hits.end() && hiter != full_hashes.end()) {
if (*piter < hiter->full_hash.prefix) {
++piter;
} else if (hiter->full_hash.prefix < *piter) {
++hiter;
} else {
if (expire_time < last_update ||
expire_time.ToTimeT() < hiter->received) {
SBFullHashResult result;
const int list_id = DecodeListId(hiter->chunk_id);
result.list_name = safe_browsing_util::GetListName(list_id);
result.add_chunk_id = DecodeChunkId(hiter->chunk_id);
result.hash = hiter->full_hash;
full_hits->push_back(result);
}
// Only increment |hiter|, |piter| might have multiple hits.
++hiter;
}
}
}
// Helper for |UpdateStarted()|. Separates |chunks| into malware and
// phishing vectors, and converts the results into range strings.
void GetChunkIds(const std::vector<int>& chunks,
std::string* malware_list, std::string* phishing_list) {
std::vector<int> malware_chunks;
std::vector<int> phishing_chunks;
for (std::vector<int>::const_iterator iter = chunks.begin();
iter != chunks.end(); ++iter) {
if (safe_browsing_util::MALWARE == DecodeListId(*iter)) {
malware_chunks.push_back(DecodeChunkId(*iter));
} else if (safe_browsing_util::PHISH == DecodeListId(*iter)) {
phishing_chunks.push_back(DecodeChunkId(*iter));
} else {
NOTREACHED();
}
}
std::sort(malware_chunks.begin(), malware_chunks.end());
std::vector<ChunkRange> malware_ranges;
ChunksToRanges(malware_chunks, &malware_ranges);
RangesToString(malware_ranges, malware_list);
std::sort(phishing_chunks.begin(), phishing_chunks.end());
std::vector<ChunkRange> phishing_ranges;
ChunksToRanges(phishing_chunks, &phishing_ranges);
RangesToString(phishing_ranges, phishing_list);
}
// Order |SBAddFullHash| on the prefix part. |SBAddPrefixLess()| from
// safe_browsing_store.h orders on both chunk-id and prefix.
bool SBAddFullHashPrefixLess(const SBAddFullHash& a, const SBAddFullHash& b) {
return a.full_hash.prefix < b.full_hash.prefix;
}
} // namespace
// Factory method.
// TODO(shess): Proposed staging of the rolling:
// - Ship "old" to dev channel to provide a safe fallback.
// - If that proves stable, change to "newsqlite". This changes the
// code which manipulates the data, without changing the data
// format. At this point all changes could be reverted without
// having to resync everyone's database from scratch.
// - If SafeBrowsingDatabaseNew proves stable, change the default to
// "newfile", which will change the file format. Changing back
// would require resync from scratch.
// - Once enough users are converted to "newfile", remove all of the
// redundent indirection classes and functions, perhaps leaving
// SafeBrowsingStoreSqlite for on-the-fly conversions.
// - Once there are few remaining SQLite-format users, remove
// SafeBrowsingStoreSqlite. Remaining users will resync their
// safe-browsing database from scratch. If users haven't sync'ed
// their database in months, this probably won't be more expensive
// than an incremental sync.
SafeBrowsingDatabase* SafeBrowsingDatabase::Create() {
const CommandLine& command_line = *CommandLine::ForCurrentProcess();
std::string value =
command_line.GetSwitchValueASCII(switches::kSafeBrowsingDatabaseStore);
if (!value.compare("newfile")) {
return new SafeBrowsingDatabaseNew(new SafeBrowsingStoreFile);
} else if (!value.compare("newsqlite")) {
return new SafeBrowsingDatabaseNew(new SafeBrowsingStoreSqlite);
} else {
DCHECK(value.empty() || !value.compare("old"));
// Default to the old implementation.
return new SafeBrowsingDatabaseBloom;
}
}
SafeBrowsingDatabase::~SafeBrowsingDatabase() {
}
// static
FilePath SafeBrowsingDatabase::BloomFilterForFilename(
const FilePath& db_filename) {
return FilePath(db_filename.value() + kBloomFilterFile);
}
SafeBrowsingDatabaseNew::SafeBrowsingDatabaseNew(SafeBrowsingStore* store)
: creation_loop_(MessageLoop::current()),
store_(store),
ALLOW_THIS_IN_INITIALIZER_LIST(reset_factory_(this)) {
DCHECK(store_.get());
}
SafeBrowsingDatabaseNew::SafeBrowsingDatabaseNew()
: creation_loop_(MessageLoop::current()),
store_(new SafeBrowsingStoreSqlite),
ALLOW_THIS_IN_INITIALIZER_LIST(reset_factory_(this)) {
DCHECK(store_.get());
}
SafeBrowsingDatabaseNew::~SafeBrowsingDatabaseNew() {
DCHECK_EQ(creation_loop_, MessageLoop::current());
}
void SafeBrowsingDatabaseNew::Init(const FilePath& filename) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
// NOTE: There is no need to grab the lock in this function, since
// until it returns, there are no pointers to this class on other
// threads. Then again, that means there is no possibility of
// contention on the lock...
AutoLock locked(lookup_lock_);
DCHECK(filename_.empty()); // Ensure we haven't been run before.
filename_ = filename;
store_->Init(
filename_,
NewCallback(this, &SafeBrowsingDatabaseNew::HandleCorruptDatabase));
full_hashes_.clear();
pending_hashes_.clear();
bloom_filter_filename_ = BloomFilterForFilename(filename_);
LoadBloomFilter();
}
bool SafeBrowsingDatabaseNew::ResetDatabase() {
DCHECK_EQ(creation_loop_, MessageLoop::current());
// Delete files on disk.
// TODO(shess): Hard to see where one might want to delete without a
// reset. Perhaps inline |Delete()|?
if (!Delete())
return false;
// Reset objects in memory.
{
AutoLock locked(lookup_lock_);
full_hashes_.clear();
pending_hashes_.clear();
prefix_miss_cache_.clear();
// TODO(shess): This could probably be |bloom_filter_.reset()|.
bloom_filter_ = new BloomFilter(BloomFilter::kBloomFilterMinSize *
BloomFilter::kBloomFilterSizeRatio);
}
return true;
}
bool SafeBrowsingDatabaseNew::ContainsUrl(
const GURL& url,
std::string* matching_list,
std::vector<SBPrefix>* prefix_hits,
std::vector<SBFullHashResult>* full_hits,
base::Time last_update) {
// Clear the results first.
matching_list->clear();
prefix_hits->clear();
full_hits->clear();
std::vector<SBPrefix> prefixes;
PrefixesToCheck(url, &prefixes);
if (prefixes.empty())
return false;
// Prevent changes to bloom filter and caches.
AutoLock locked(lookup_lock_);
if (!bloom_filter_.get())
return false;
// TODO(erikkay): Not filling in matching_list - is that OK?
size_t miss_count = 0;
for (size_t i = 0; i < prefixes.size(); ++i) {
if (bloom_filter_->Exists(prefixes[i])) {
prefix_hits->push_back(prefixes[i]);
if (prefix_miss_cache_.count(prefixes[i]) > 0)
++miss_count;
}
}
// If all the prefixes are cached as 'misses', don't issue a GetHash.
if (miss_count == prefix_hits->size())
return false;
// Find the matching full-hash results. |full_hashes_| are from the
// database, |pending_hashes_| are from GetHash requests between
// updates.
std::sort(prefix_hits->begin(), prefix_hits->end());
GetCachedFullHashes(*prefix_hits, full_hashes_, full_hits, last_update);
GetCachedFullHashes(*prefix_hits, pending_hashes_, full_hits, last_update);
return true;
}
// Helper to insert entries for all of the prefixes or full hashes in
// |entry| into the store.
void SafeBrowsingDatabaseNew::InsertAdd(int chunk_id, SBPrefix host,
const SBEntry* entry, int list_id) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
STATS_COUNTER("SB.HostInsert", 1);
const int encoded_chunk_id = EncodeChunkId(chunk_id, list_id);
const int count = entry->prefix_count();
DCHECK(!entry->IsSub());
if (!count) {
// No prefixes, use host instead.
STATS_COUNTER("SB.PrefixAdd", 1);
store_->WriteAddPrefix(encoded_chunk_id, host);
} else if (entry->IsPrefix()) {
// Prefixes only.
for (int i = 0; i < count; i++) {
const SBPrefix prefix = entry->PrefixAt(i);
STATS_COUNTER("SB.PrefixAdd", 1);
store_->WriteAddPrefix(encoded_chunk_id, prefix);
}
} else {
// Prefixes and hashes.
const base::Time receive_time = base::Time::Now();
for (int i = 0; i < count; ++i) {
const SBFullHash full_hash = entry->FullHashAt(i);
const SBPrefix prefix = full_hash.prefix;
STATS_COUNTER("SB.PrefixAdd", 1);
store_->WriteAddPrefix(encoded_chunk_id, prefix);
STATS_COUNTER("SB.PrefixAddFull", 1);
store_->WriteAddHash(encoded_chunk_id, receive_time, full_hash);
}
}
}
// Helper to iterate over all the entries in the hosts in |chunks| and
// add them to the store.
void SafeBrowsingDatabaseNew::InsertAddChunks(int list_id,
const SBChunkList& chunks) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
for (SBChunkList::const_iterator citer = chunks.begin();
citer != chunks.end(); ++citer) {
const int chunk_id = citer->chunk_number;
// The server can give us a chunk that we already have because
// it's part of a range. Don't add it again.
const int encoded_chunk_id = EncodeChunkId(chunk_id, list_id);
if (store_->CheckAddChunk(encoded_chunk_id))
continue;
store_->SetAddChunk(encoded_chunk_id);
for (std::deque<SBChunkHost>::const_iterator hiter = citer->hosts.begin();
hiter != citer->hosts.end(); ++hiter) {
// NOTE: Could pass |encoded_chunk_id|, but then inserting add
// chunks would look different from inserting sub chunks.
InsertAdd(chunk_id, hiter->host, hiter->entry, list_id);
}
}
}
// Helper to insert entries for all of the prefixes or full hashes in
// |entry| into the store.
void SafeBrowsingDatabaseNew::InsertSub(int chunk_id, SBPrefix host,
const SBEntry* entry, int list_id) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
STATS_COUNTER("SB.HostDelete", 1);
const int encoded_chunk_id = EncodeChunkId(chunk_id, list_id);
const int count = entry->prefix_count();
DCHECK(entry->IsSub());
if (!count) {
// No prefixes, use host instead.
STATS_COUNTER("SB.PrefixSub", 1);
const int add_chunk_id = EncodeChunkId(entry->chunk_id(), list_id);
store_->WriteSubPrefix(encoded_chunk_id, add_chunk_id, host);
} else if (entry->IsPrefix()) {
// Prefixes only.
for (int i = 0; i < count; i++) {
const SBPrefix prefix = entry->PrefixAt(i);
const int add_chunk_id =
EncodeChunkId(entry->ChunkIdAtPrefix(i), list_id);
STATS_COUNTER("SB.PrefixSub", 1);
store_->WriteSubPrefix(encoded_chunk_id, add_chunk_id, prefix);
}
} else {
// Prefixes and hashes.
for (int i = 0; i < count; ++i) {
const SBFullHash full_hash = entry->FullHashAt(i);
const int add_chunk_id =
EncodeChunkId(entry->ChunkIdAtPrefix(i), list_id);
STATS_COUNTER("SB.PrefixSub", 1);
store_->WriteSubPrefix(encoded_chunk_id, add_chunk_id, full_hash.prefix);
STATS_COUNTER("SB.PrefixSubFull", 1);
store_->WriteSubHash(encoded_chunk_id, add_chunk_id, full_hash);
}
}
}
// Helper to iterate over all the entries in the hosts in |chunks| and
// add them to the store.
void SafeBrowsingDatabaseNew::InsertSubChunks(int list_id,
const SBChunkList& chunks) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
for (SBChunkList::const_iterator citer = chunks.begin();
citer != chunks.end(); ++citer) {
const int chunk_id = citer->chunk_number;
// The server can give us a chunk that we already have because
// it's part of a range. Don't add it again.
const int encoded_chunk_id = EncodeChunkId(chunk_id, list_id);
if (store_->CheckSubChunk(encoded_chunk_id))
continue;
store_->SetSubChunk(encoded_chunk_id);
for (std::deque<SBChunkHost>::const_iterator hiter = citer->hosts.begin();
hiter != citer->hosts.end(); ++hiter) {
InsertSub(chunk_id, hiter->host, hiter->entry, list_id);
}
}
}
void SafeBrowsingDatabaseNew::InsertChunks(const std::string& list_name,
const SBChunkList& chunks) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
if (chunks.empty())
return;
const base::Time insert_start = base::Time::Now();
const int list_id = safe_browsing_util::GetListId(list_name);
store_->BeginChunk();
if (chunks.front().is_add) {
InsertAddChunks(list_id, chunks);
} else {
InsertSubChunks(list_id, chunks);
}
store_->FinishChunk();
UMA_HISTOGRAM_TIMES("SB2.ChunkInsert", base::Time::Now() - insert_start);
}
void SafeBrowsingDatabaseNew::DeleteChunks(
const std::vector<SBChunkDelete>& chunk_deletes) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
if (chunk_deletes.empty())
return;
const std::string& list_name = chunk_deletes.front().list_name;
const int list_id = safe_browsing_util::GetListId(list_name);
for (size_t i = 0; i < chunk_deletes.size(); ++i) {
std::vector<int> chunk_numbers;
RangesToChunks(chunk_deletes[i].chunk_del, &chunk_numbers);
for (size_t j = 0; j < chunk_numbers.size(); ++j) {
const int encoded_chunk_id = EncodeChunkId(chunk_numbers[j], list_id);
if (chunk_deletes[i].is_sub_del)
store_->DeleteSubChunk(encoded_chunk_id);
else
store_->DeleteAddChunk(encoded_chunk_id);
}
}
}
void SafeBrowsingDatabaseNew::CacheHashResults(
const std::vector<SBPrefix>& prefixes,
const std::vector<SBFullHashResult>& full_hits) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
// This is on the same thread as other updates, lock against
// |ContainsUrl()|.
AutoLock locked(lookup_lock_);
if (full_hits.empty()) {
prefix_miss_cache_.insert(prefixes.begin(), prefixes.end());
return;
}
// TODO(shess): SBFullHashResult and SBAddFullHash are very similar.
// Refactor to make them identical.
const base::Time now = base::Time::Now();
const size_t orig_size = pending_hashes_.size();
for (std::vector<SBFullHashResult>::const_iterator iter = full_hits.begin();
iter != full_hits.end(); ++iter) {
const int list_id = safe_browsing_util::GetListId(iter->list_name);
const int encoded_chunk_id = EncodeChunkId(iter->add_chunk_id, list_id);
pending_hashes_.push_back(SBAddFullHash(encoded_chunk_id, now, iter->hash));
}
// Sort new entries then merge with the previously-sorted entries.
std::vector<SBAddFullHash>::iterator
orig_end = pending_hashes_.begin() + orig_size;
std::sort(orig_end, pending_hashes_.end(), SBAddFullHashPrefixLess);
std::inplace_merge(pending_hashes_.begin(), orig_end, pending_hashes_.end(),
SBAddFullHashPrefixLess);
}
bool SafeBrowsingDatabaseNew::UpdateStarted(
std::vector<SBListChunkRanges>* lists) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
DCHECK(lists);
// If |BeginUpdate()| fails, reset the database.
if (!store_->BeginUpdate()) {
HandleCorruptDatabase();
return false;
}
SBListChunkRanges malware(safe_browsing_util::kMalwareList);
SBListChunkRanges phishing(safe_browsing_util::kPhishingList);
std::vector<int> add_chunks;
store_->GetAddChunks(&add_chunks);
GetChunkIds(add_chunks, &malware.adds, &phishing.adds);
std::vector<int> sub_chunks;
store_->GetSubChunks(&sub_chunks);
GetChunkIds(sub_chunks, &malware.subs, &phishing.subs);
lists->push_back(malware);
lists->push_back(phishing);
return true;
}
void SafeBrowsingDatabaseNew::UpdateFinished(bool update_succeeded) {
DCHECK_EQ(creation_loop_, MessageLoop::current());
// Unroll any partially-received transaction.
if (!update_succeeded) {
store_->CancelUpdate();
return;
}
// Copy out the pending add hashes. Copy rather than swapping in
// case |ContainsURL()| is called before the new filter is complete.
std::vector<SBAddFullHash> pending_add_hashes;
{
AutoLock locked(lookup_lock_);
pending_add_hashes.insert(pending_add_hashes.end(),
pending_hashes_.begin(), pending_hashes_.end());
}
std::vector<SBAddPrefix> add_prefixes;
std::vector<SBAddFullHash> add_full_hashes;
if (!store_->FinishUpdate(pending_add_hashes,
&add_prefixes, &add_full_hashes))
return;
// Measure the amount of IO during the bloom filter build.
base::IoCounters io_before, io_after;
base::ProcessHandle handle = base::Process::Current().handle();
scoped_ptr<base::ProcessMetrics> metric(
#if !defined(OS_MACOSX)
base::ProcessMetrics::CreateProcessMetrics(handle)
#else
// Getting stats only for the current process is enough, so NULL is fine.
base::ProcessMetrics::CreateProcessMetrics(handle, NULL)
#endif
);
// IoCounters are currently not supported on Mac, and may not be
// available for Linux, so we check the result and only show IO
// stats if they are available.
const bool got_counters = metric->GetIOCounters(&io_before);
const base::Time before = base::Time::Now();
// Create and populate |filter| from |add_prefixes|.
// TODO(shess): The bloom filter doesn't need to be a
// scoped_refptr<> for this code. Refactor that away.
const int filter_size =
BloomFilter::FilterSizeForKeyCount(add_prefixes.size());
scoped_refptr<BloomFilter> filter(new BloomFilter(filter_size));
for (size_t i = 0; i < add_prefixes.size(); ++i) {
filter->Insert(add_prefixes[i].prefix);
}
// This needs to be in sorted order by prefix for efficient access.
std::sort(add_full_hashes.begin(), add_full_hashes.end(),
SBAddFullHashPrefixLess);
// Swap in the newly built filter and cache.
{
AutoLock locked(lookup_lock_);
full_hashes_.swap(add_full_hashes);
pending_hashes_.clear();
prefix_miss_cache_.clear();
bloom_filter_.swap(filter);
}
const base::TimeDelta bloom_gen = base::Time::Now() - before;
// Persist the bloom filter to disk. Since only this thread changes
// |bloom_filter_|, there is no need to lock.
WriteBloomFilter();
// Gather statistics.
if (got_counters && metric->GetIOCounters(&io_after)) {
UMA_HISTOGRAM_COUNTS("SB2.BuildReadBytes",
static_cast<int>(io_after.ReadTransferCount -
io_before.ReadTransferCount));
UMA_HISTOGRAM_COUNTS("SB2.BuildWriteBytes",
static_cast<int>(io_after.WriteTransferCount -
io_before.WriteTransferCount));
UMA_HISTOGRAM_COUNTS("SB2.BuildReadOperations",
static_cast<int>(io_after.ReadOperationCount -
io_before.ReadOperationCount));
UMA_HISTOGRAM_COUNTS("SB2.BuildWriteOperations",
static_cast<int>(io_after.WriteOperationCount -
io_before.WriteOperationCount));
}
SB_DLOG(INFO) << "SafeBrowsingDatabaseImpl built bloom filter in "
<< bloom_gen.InMilliseconds()
<< " ms total. prefix count: "<< add_prefixes.size();
UMA_HISTOGRAM_LONG_TIMES("SB2.BuildFilter", bloom_gen);
UMA_HISTOGRAM_COUNTS("SB2.AddPrefixes", add_prefixes.size());
// TODO(shess): Push this line into |store_|? Or ignore?
// UMA_HISTOGRAM_COUNTS("SB2.SubPrefixes", subs);
UMA_HISTOGRAM_COUNTS("SB2.FilterSize", bloom_filter_->size());
int64 size_64;
if (file_util::GetFileSize(filename_, &size_64))
UMA_HISTOGRAM_COUNTS("SB2.DatabaseBytes", static_cast<int>(size_64));
}
void SafeBrowsingDatabaseNew::HandleCorruptDatabase() {
// Reset the database after the current task has unwound (but only
// reset once within the scope of a given task).
if (reset_factory_.empty())
MessageLoop::current()->PostTask(FROM_HERE,
reset_factory_.NewRunnableMethod(
&SafeBrowsingDatabaseNew::OnHandleCorruptDatabase));
}
void SafeBrowsingDatabaseNew::OnHandleCorruptDatabase() {
UMA_HISTOGRAM_COUNTS("SB2.HandleCorrupt", 1);
ResetDatabase();
DCHECK(false) << "SafeBrowsing database was corrupt and reset";
}
// TODO(shess): I'm not clear why this code doesn't have any
// real error-handling.
void SafeBrowsingDatabaseNew::LoadBloomFilter() {
DCHECK_EQ(creation_loop_, MessageLoop::current());
DCHECK(!bloom_filter_filename_.empty());
// If we're missing either of the database or filter files, we wait until the
// next update to generate a new filter.
// TODO(paulg): Investigate how often the filter file is missing and how
// expensive it would be to regenerate it.
int64 size_64;
if (!file_util::GetFileSize(filename_, &size_64) || size_64 == 0)
return;
if (!file_util::GetFileSize(bloom_filter_filename_, &size_64) ||
size_64 == 0) {
UMA_HISTOGRAM_COUNTS("SB2.FilterMissing", 1);
return;
}
const base::TimeTicks before = base::TimeTicks::Now();
bloom_filter_ = BloomFilter::LoadFile(bloom_filter_filename_);
SB_DLOG(INFO) << "SafeBrowsingDatabaseNew read bloom filter in "
<< (base::TimeTicks::Now() - before).InMilliseconds() << " ms";
if (!bloom_filter_.get())
UMA_HISTOGRAM_COUNTS("SB2.FilterReadFail", 1);
}
bool SafeBrowsingDatabaseNew::Delete() {
DCHECK_EQ(creation_loop_, MessageLoop::current());
const bool r1 = store_->Delete();
const bool r2 = file_util::Delete(bloom_filter_filename_, false);
return r1 && r2;
}
void SafeBrowsingDatabaseNew::WriteBloomFilter() {
DCHECK_EQ(creation_loop_, MessageLoop::current());
if (!bloom_filter_.get())
return;
const base::TimeTicks before = base::TimeTicks::Now();
const bool write_ok = bloom_filter_->WriteFile(bloom_filter_filename_);
SB_DLOG(INFO) << "SafeBrowsingDatabaseNew wrote bloom filter in " <<
(base::TimeTicks::Now() - before).InMilliseconds() << " ms";
if (!write_ok)
UMA_HISTOGRAM_COUNTS("SB2.FilterWriteFail", 1);
}
|