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
|
// 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 "chrome/browser/safe_browsing/client_side_model_loader.h"
#include "base/bind.h"
#include "base/command_line.h"
#include "base/memory/scoped_ptr.h"
#include "base/message_loop/message_loop.h"
#include "base/metrics/histogram.h"
#include "base/strings/string_number_conversions.h"
#include "base/strings/string_util.h"
#include "base/time/time.h"
#include "chrome/common/chrome_switches.h"
#include "chrome/common/safe_browsing/client_model.pb.h"
#include "chrome/common/safe_browsing/csd.pb.h"
#include "chrome/common/safe_browsing/safebrowsing_messages.h"
#include "components/variations/variations_associated_data.h"
#include "net/http/http_response_headers.h"
#include "net/http/http_status_code.h"
#include "net/url_request/url_fetcher.h"
#include "net/url_request/url_request_context_getter.h"
#include "net/url_request/url_request_status.h"
#include "url/gurl.h"
namespace safe_browsing {
// Model Loader strings
const size_t ModelLoader::kMaxModelSizeBytes = 150 * 1024;
const int ModelLoader::kClientModelFetchIntervalMs = 3600 * 1000;
const char ModelLoader::kClientModelUrlPrefix[] =
"https://ssl.gstatic.com/safebrowsing/csd/";
const char ModelLoader::kClientModelNamePattern[] =
"client_model_v5%s_variation_%d.pb";
const char ModelLoader::kClientModelFinchExperiment[] =
"ClientSideDetectionModel";
const char ModelLoader::kClientModelFinchParam[] =
"ModelNum";
// static
int ModelLoader::GetModelNumber() {
std::string num_str = variations::GetVariationParamValue(
kClientModelFinchExperiment, kClientModelFinchParam);
int model_number = 0;
if (!base::StringToInt(num_str, &model_number)) {
model_number = 0; // Default model
}
return model_number;
}
// static
std::string ModelLoader::FillInModelName(bool is_extended_reporting,
int model_number) {
return base::StringPrintf(kClientModelNamePattern,
is_extended_reporting ? "_ext" : "", model_number);
}
// static
bool ModelLoader::ModelHasValidHashIds(const ClientSideModel& model) {
const int max_index = model.hashes_size() - 1;
for (int i = 0; i < model.rule_size(); ++i) {
for (int j = 0; j < model.rule(i).feature_size(); ++j) {
if (model.rule(i).feature(j) < 0 ||
model.rule(i).feature(j) > max_index) {
return false;
}
}
}
for (int i = 0; i < model.page_term_size(); ++i) {
if (model.page_term(i) < 0 || model.page_term(i) > max_index) {
return false;
}
}
return true;
}
// Model name and URL are a function of is_extended_reporting and Finch.
ModelLoader::ModelLoader(base::Closure update_renderers_callback,
net::URLRequestContextGetter* request_context_getter,
bool is_extended_reporting)
: name_(FillInModelName(is_extended_reporting, GetModelNumber())),
url_(kClientModelUrlPrefix + name_),
update_renderers_callback_(update_renderers_callback),
request_context_getter_(request_context_getter),
weak_factory_(this) {
DCHECK(url_.is_valid());
}
// For testing only
ModelLoader::ModelLoader(base::Closure update_renderers_callback,
const std::string model_name)
: name_(model_name),
url_(kClientModelUrlPrefix + name_),
update_renderers_callback_(update_renderers_callback),
request_context_getter_(NULL),
weak_factory_(this) {
DCHECK(url_.is_valid());
}
ModelLoader::~ModelLoader() {
}
void ModelLoader::StartFetch() {
// Start fetching the model either from the cache or possibly from the
// network if the model isn't in the cache.
// TODO(nparker): If no profile needs this model, we shouldn't fetch it.
// Then only re-fetch when a profile setting changes to need it.
// This will save on the order of ~50KB/week/client of bandwidth.
fetcher_ = net::URLFetcher::Create(0 /* ID used for testing */, url_,
net::URLFetcher::GET, this);
fetcher_->SetRequestContext(request_context_getter_);
fetcher_->Start();
}
void ModelLoader::OnURLFetchComplete(const net::URLFetcher* source) {
DCHECK_EQ(fetcher_, source);
DCHECK_EQ(url_, source->GetURL());
std::string data;
source->GetResponseAsString(&data);
const bool is_success = source->GetStatus().is_success();
const int response_code = source->GetResponseCode();
// max_age is valid iff !0.
base::TimeDelta max_age;
if (is_success && net::HTTP_OK == response_code &&
source->GetResponseHeaders()) {
source->GetResponseHeaders()->GetMaxAgeValue(&max_age);
}
scoped_ptr<ClientSideModel> model(new ClientSideModel());
ClientModelStatus model_status;
if (!is_success || net::HTTP_OK != response_code) {
model_status = MODEL_FETCH_FAILED;
} else if (data.empty()) {
model_status = MODEL_EMPTY;
} else if (data.size() > kMaxModelSizeBytes) {
model_status = MODEL_TOO_LARGE;
} else if (!model->ParseFromString(data)) {
model_status = MODEL_PARSE_ERROR;
} else if (!model->IsInitialized() || !model->has_version()) {
model_status = MODEL_MISSING_FIELDS;
} else if (!ModelHasValidHashIds(*model)) {
model_status = MODEL_BAD_HASH_IDS;
} else if (model->version() < 0 ||
(model_.get() && model->version() < model_->version())) {
model_status = MODEL_INVALID_VERSION_NUMBER;
} else if (model_.get() && model->version() == model_->version()) {
model_status = MODEL_NOT_CHANGED;
} else {
// The model is valid => replace the existing model with the new one.
model_str_.assign(data);
model_.swap(model);
model_status = MODEL_SUCCESS;
}
EndFetch(model_status, max_age);
}
void ModelLoader::EndFetch(ClientModelStatus status, base::TimeDelta max_age) {
// We don't differentiate models in the UMA stats.
UMA_HISTOGRAM_ENUMERATION("SBClientPhishing.ClientModelStatus",
status,
MODEL_STATUS_MAX);
if (status == MODEL_SUCCESS) {
update_renderers_callback_.Run();
}
int delay_ms = kClientModelFetchIntervalMs;
// If the most recently fetched model had a valid max-age and the model was
// valid we're scheduling the next model update for after the max-age expired.
if (!max_age.is_zero() &&
(status == MODEL_SUCCESS || status == MODEL_NOT_CHANGED)) {
// We're adding 60s of additional delay to make sure we're past
// the model's age.
max_age += base::TimeDelta::FromMinutes(1);
delay_ms = max_age.InMilliseconds();
}
// Schedule the next model reload.
ScheduleFetch(delay_ms);
}
void ModelLoader::ScheduleFetch(int64 delay_ms) {
if (base::CommandLine::ForCurrentProcess()->HasSwitch(
switches::kSbDisableAutoUpdate))
return;
base::MessageLoop::current()->PostDelayedTask(
FROM_HERE,
base::Bind(&ModelLoader::StartFetch, weak_factory_.GetWeakPtr()),
base::TimeDelta::FromMilliseconds(delay_ms));
}
void ModelLoader::CancelFetcher() {
// Invalidate any scheduled request.
weak_factory_.InvalidateWeakPtrs();
// Cancel any request in progress.
fetcher_.reset();
}
} // namespace safe_browsing
|
