url_blacklist_manager.h revision 1320f92c476a1ad9d19dba2a48c72b75566198e9
1// Copyright 2014 The Chromium Authors. All rights reserved.
2// Use of this source code is governed by a BSD-style license that can be
3// found in the LICENSE file.
4
5#ifndef COMPONENTS_POLICY_CORE_BROWSER_URL_BLACKLIST_MANAGER_H_
6#define COMPONENTS_POLICY_CORE_BROWSER_URL_BLACKLIST_MANAGER_H_
7
8#include <map>
9#include <string>
10
11#include "base/basictypes.h"
12#include "base/callback_forward.h"
13#include "base/compiler_specific.h"
14#include "base/containers/hash_tables.h"
15#include "base/memory/ref_counted.h"
16#include "base/memory/scoped_ptr.h"
17#include "base/memory/weak_ptr.h"
18#include "base/prefs/pref_change_registrar.h"
19#include "components/policy/policy_export.h"
20#include "components/url_matcher/url_matcher.h"
21#include "url/gurl.h"
22
23class PrefService;
24
25namespace base {
26class ListValue;
27class SequencedTaskRunner;
28}
29
30namespace net {
31class URLRequest;
32}
33
34namespace user_prefs {
35class PrefRegistrySyncable;
36}
37
38namespace policy {
39
40// Contains a set of filters to block and allow certain URLs, and matches GURLs
41// against this set. The filters are currently kept in memory.
42class POLICY_EXPORT URLBlacklist {
43 public:
44  // This is meant to be bound to url_fixer::SegmentURL. See that function
45  // for documentation on the parameters and return value.
46  typedef std::string (*SegmentURLCallback)(const std::string&, url::Parsed*);
47
48  explicit URLBlacklist(SegmentURLCallback segment_url);
49  virtual ~URLBlacklist();
50
51  // Allows or blocks URLs matching one of the filters, depending on |allow|.
52  void AddFilters(bool allow, const base::ListValue* filters);
53
54  // URLs matching one of the |filters| will be blocked. The filter format is
55  // documented at
56  // http://www.chromium.org/administrators/url-blacklist-filter-format.
57  void Block(const base::ListValue* filters);
58
59  // URLs matching one of the |filters| will be allowed. If a URL is both
60  // Blocked and Allowed, Allow takes precedence.
61  void Allow(const base::ListValue* filters);
62
63  // Returns true if the URL is blocked.
64  bool IsURLBlocked(const GURL& url) const;
65
66  // Returns the number of items in the list.
67  size_t Size() const;
68
69  // Splits a URL filter into its components. A GURL isn't used because these
70  // can be invalid URLs e.g. "google.com".
71  // Returns false if the URL couldn't be parsed.
72  // The |host| is preprocessed so it can be passed to URLMatcher for the
73  // appropriate condition.
74  // The optional username and password are ignored.
75  // |match_subdomains| specifies whether the filter should include subdomains
76  // of the hostname (if it is one.)
77  // |port| is 0 if none is explicitly defined.
78  // |path| does not include query parameters.
79  // |query| contains the query parameters ('?' not included).
80  static bool FilterToComponents(SegmentURLCallback segment_url,
81                                 const std::string& filter,
82                                 std::string* scheme,
83                                 std::string* host,
84                                 bool* match_subdomains,
85                                 uint16* port,
86                                 std::string* path,
87                                 std::string* query);
88
89  // Creates a condition set that can be used with the |url_matcher|. |id| needs
90  // to be a unique number that will be returned by the |url_matcher| if the URL
91  // matches that condition set. |allow| indicates if it is a white-list (true)
92  // or black-list (false) filter.
93  static scoped_refptr<url_matcher::URLMatcherConditionSet> CreateConditionSet(
94      url_matcher::URLMatcher* url_matcher,
95      url_matcher::URLMatcherConditionSet::ID id,
96      const std::string& scheme,
97      const std::string& host,
98      bool match_subdomains,
99      uint16 port,
100      const std::string& path,
101      const std::string& query,
102      bool allow);
103
104 private:
105  struct FilterComponents;
106
107  // Returns true if |lhs| takes precedence over |rhs|.
108  static bool FilterTakesPrecedence(const FilterComponents& lhs,
109                                    const FilterComponents& rhs);
110
111  SegmentURLCallback segment_url_;
112  url_matcher::URLMatcherConditionSet::ID id_;
113  std::map<url_matcher::URLMatcherConditionSet::ID, FilterComponents> filters_;
114  scoped_ptr<url_matcher::URLMatcher> url_matcher_;
115
116  DISALLOW_COPY_AND_ASSIGN(URLBlacklist);
117};
118
119// Tracks the blacklist policies for a given profile, and updates it on changes.
120//
121// This class interacts with both the UI thread, where notifications of pref
122// changes are received from, and the IO thread, which owns it (in the
123// ProfileIOData) and checks for blacklisted URLs (from ChromeNetworkDelegate).
124//
125// It must be constructed on the UI thread, to set up |ui_weak_ptr_factory_| and
126// the prefs listeners.
127//
128// ShutdownOnUIThread must be called from UI before destruction, to release
129// the prefs listeners on the UI thread. This is done from ProfileIOData.
130//
131// Update tasks from the UI thread can post safely to the IO thread, since the
132// destruction order of Profile and ProfileIOData guarantees that if this
133// exists in UI, then a potential destruction on IO will come after any task
134// posted to IO from that method on UI. This is used to go through IO before
135// the actual update starts, and grab a WeakPtr.
136class POLICY_EXPORT URLBlacklistManager {
137 public:
138  // Returns true if the blacklist should be overridden for |url| and sets
139  // |block| to true if it should be blocked and false otherwise.
140  // |reason| is set to the exact reason for blocking |url| iff |block| is true.
141  typedef base::Callback<bool(const GURL& url, bool* block, int* reason)>
142      OverrideBlacklistCallback;
143
144  // Must be constructed on the UI thread.
145  // |background_task_runner| is used to build the blacklist in a background
146  // thread.
147  // |io_task_runner| must be backed by the IO thread.
148  // |segment_url| is used to break a URL spec into its components.
149  URLBlacklistManager(
150      PrefService* pref_service,
151      const scoped_refptr<base::SequencedTaskRunner>& background_task_runner,
152      const scoped_refptr<base::SequencedTaskRunner>& io_task_runner,
153      URLBlacklist::SegmentURLCallback segment_url,
154      OverrideBlacklistCallback override_blacklist);
155  virtual ~URLBlacklistManager();
156
157  // Must be called on the UI thread, before destruction.
158  void ShutdownOnUIThread();
159
160  // Returns true if |url| is blocked by the current blacklist. Must be called
161  // from the IO thread.
162  bool IsURLBlocked(const GURL& url) const;
163
164  // Returns true if |request| is blocked by the current blacklist.
165  // Only main frame and sub frame requests may be blocked; other sub resources
166  // or background downloads (e.g. extensions updates, sync, etc) are not
167  // filtered. The sync signin page is also not filtered.
168  // |reason| is populated with the exact reason for blocking the url if and
169  // only if the return value is true otherwise it is left untouched.
170  // Must be called from the IO thread.
171  bool IsRequestBlocked(const net::URLRequest& request, int* reason) const;
172
173  // Replaces the current blacklist. Must be called on the IO thread.
174  // Virtual for testing.
175  virtual void SetBlacklist(scoped_ptr<URLBlacklist> blacklist);
176
177  // Registers the preferences related to blacklisting in the given PrefService.
178  static void RegisterProfilePrefs(user_prefs::PrefRegistrySyncable* registry);
179
180 protected:
181  // Used to delay updating the blacklist while the preferences are
182  // changing, and execute only one update per simultaneous prefs changes.
183  void ScheduleUpdate();
184
185  // Updates the blacklist using the current preference values.
186  // Virtual for testing.
187  virtual void Update();
188
189  // Starts the blacklist update on the IO thread, using the filters in
190  // |block| and |allow|. Protected for testing.
191  void UpdateOnIO(scoped_ptr<base::ListValue> block,
192                  scoped_ptr<base::ListValue> allow);
193
194 private:
195  // ---------
196  // UI thread
197  // ---------
198
199  // Used to track the policies and update the blacklist on changes.
200  PrefChangeRegistrar pref_change_registrar_;
201  PrefService* pref_service_;  // Weak.
202
203  // Used to post tasks to a background thread.
204  scoped_refptr<base::SequencedTaskRunner> background_task_runner_;
205
206  // Used to post tasks to the IO thread.
207  scoped_refptr<base::SequencedTaskRunner> io_task_runner_;
208
209  // Used to break a URL into its components.
210  URLBlacklist::SegmentURLCallback segment_url_;
211
212  // Used to optionally skip blacklisting for some URLs.
213  OverrideBlacklistCallback override_blacklist_;
214
215  // ---------
216  // IO thread
217  // ---------
218
219  // Used to post tasks to the UI thread.
220  scoped_refptr<base::SequencedTaskRunner> ui_task_runner_;
221
222  // The current blacklist.
223  scoped_ptr<URLBlacklist> blacklist_;
224
225  // Used to post update tasks to the UI thread.
226  base::WeakPtrFactory<URLBlacklistManager> ui_weak_ptr_factory_;
227
228  // Used to get |weak_ptr_| to self on the IO thread.
229  base::WeakPtrFactory<URLBlacklistManager> io_weak_ptr_factory_;
230
231  DISALLOW_COPY_AND_ASSIGN(URLBlacklistManager);
232};
233
234}  // namespace policy
235
236#endif  // COMPONENTS_POLICY_CORE_BROWSER_URL_BLACKLIST_MANAGER_H_
237