summaryrefslogtreecommitdiffstats
path: root/extensions/spellcheck/hunspell/glue/RLBoxHunspell.cpp
blob: b3a1990ceb648a1b24b66db361d7ec463d5dda86 (plain)
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
/* -*- Mode: C++; tab-width: 20; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#include "mozilla/Assertions.h"
#ifdef MOZ_WASM_SANDBOXING_HUNSPELL
#  include "mozilla/ipc/LibrarySandboxPreload.h"
#endif
#include "RLBoxHunspell.h"
#include "mozHunspellRLBoxGlue.h"
#include "mozHunspellRLBoxHost.h"
#include "nsThread.h"

using namespace rlbox;
using namespace mozilla;

// Helper function for allocating and copying std::string into sandbox
static tainted_hunspell<char*> allocStrInSandbox(
    rlbox_sandbox_hunspell& aSandbox, const std::string& str) {
  size_t size = str.size() + 1;
  tainted_hunspell<char*> t_str = aSandbox.malloc_in_sandbox<char>(size);
  if (t_str) {
    rlbox::memcpy(aSandbox, t_str, str.c_str(), size);
  }
  return t_str;
}

/* static */
RLBoxHunspell* RLBoxHunspell::Create(const nsCString& affpath,
                                     const nsCString& dpath) {
  MOZ_DIAGNOSTIC_ASSERT(NS_IsMainThread());

  mozilla::UniquePtr<rlbox_sandbox_hunspell> sandbox(
      new rlbox_sandbox_hunspell());

#if defined(MOZ_WASM_SANDBOXING_HUNSPELL) && !defined(HAVE_64BIT_BUILD)
  // By default, the rlbox sandbox size is smaller on 32-bit builds than the max
  // 4GB. We may need to ask for a larger sandbox size for hunspell to
  // spellcheck in some locales See Bug 1739669 for more details

  // We first get the size of the dictionary. This is actually the first read we
  // try on dpath and it might fail for whatever filesystem reasons (invalid
  // path, unaccessible, ...).
  Result<int64_t, nsresult> dictSizeResult =
      mozHunspellFileMgrHost::GetSize(dpath);
  NS_ENSURE_TRUE(dictSizeResult.isOk(), nullptr);

  int64_t dictSize = dictSizeResult.unwrap();
  NS_ENSURE_TRUE(dictSize >= 0, nullptr);

  // Next, we compute the expected memory needed for hunspell spell checking.
  // This will vary based on the size of the dictionary file, which varies by
  // locale — so we size the sandbox by multiplying the file size by 4.8. This
  // allows the 1.5MB en_US dictionary to fit in an 8MB sandbox. See bug 1739669
  // and bug 1739761 for the analysis behind this.
  const uint64_t expectedMaxMemory = static_cast<uint64_t>(4.8 * dictSize);

  // Get a capacity of at least the expected size
  const w2c_mem_capacity capacity = get_valid_wasm2c_memory_capacity(
      expectedMaxMemory, true /* wasm's 32-bit memory */);

  bool success =
      sandbox->create_sandbox(/* shouldAbortOnFailure = */ false, &capacity);
#elif defined(MOZ_WASM_SANDBOXING_HUNSPELL)
  bool success = sandbox->create_sandbox(/* shouldAbortOnFailure = */ false);
#else
  sandbox->create_sandbox();
  const bool success = true;
#endif

  NS_ENSURE_TRUE(success, nullptr);

  mozilla::UniquePtr<rlbox_sandbox_hunspell, RLBoxDeleter> sandbox_initialized(
      sandbox.release());

  // Add the aff and dict files to allow list
  if (!affpath.IsEmpty()) {
    mozHunspellCallbacks::AllowFile(affpath);
  }
  if (!dpath.IsEmpty()) {
    mozHunspellCallbacks::AllowFile(dpath);
  }

  // TODO Bug 1788857: Verify error handling in case of inaccessible file
  return new RLBoxHunspell(std::move(sandbox_initialized), affpath, dpath);
}

RLBoxHunspell::RLBoxHunspell(
    mozilla::UniquePtr<rlbox_sandbox_hunspell, RLBoxDeleter> aSandbox,
    const nsCString& affpath, const nsCString& dpath)
    : mSandbox(std::move(aSandbox)), mHandle(nullptr) {
  // Register callbacks
  mCreateFilemgr =
      mSandbox->register_callback(mozHunspellCallbacks::CreateFilemgr);
  mGetLine = mSandbox->register_callback(mozHunspellCallbacks::GetLine);
  mGetLineNum = mSandbox->register_callback(mozHunspellCallbacks::GetLineNum);
  mDestructFilemgr =
      mSandbox->register_callback(mozHunspellCallbacks::DestructFilemgr);
  mHunspellToUpperCase =
      mSandbox->register_callback(mozHunspellCallbacks::ToUpperCase);
  mHunspellToLowerCase =
      mSandbox->register_callback(mozHunspellCallbacks::ToLowerCase);
  mHunspellGetCurrentCS =
      mSandbox->register_callback(mozHunspellCallbacks::GetCurrentCS);

  mSandbox->invoke_sandbox_function(RegisterHunspellCallbacks, mCreateFilemgr,
                                    mGetLine, mGetLineNum, mDestructFilemgr,
                                    mHunspellToUpperCase, mHunspellToLowerCase,
                                    mHunspellGetCurrentCS);

  // Copy the affpath and dpath into the sandbox
  // These allocations should definitely succeed as these are first allocations
  // inside the sandbox.
  tainted_hunspell<char*> t_affpath =
      allocStrInSandbox(*mSandbox, affpath.get());
  MOZ_RELEASE_ASSERT(t_affpath);

  tainted_hunspell<char*> t_dpath = allocStrInSandbox(*mSandbox, dpath.get());
  MOZ_RELEASE_ASSERT(t_dpath);

  // Create handle
  mHandle = mSandbox->invoke_sandbox_function(
      Hunspell_create, rlbox::sandbox_const_cast<const char*>(t_affpath),
      rlbox::sandbox_const_cast<const char*>(t_dpath));
  MOZ_RELEASE_ASSERT(mHandle);

  mSandbox->free_in_sandbox(t_dpath);
  mSandbox->free_in_sandbox(t_affpath);

  // Get dictionary encoding
  tainted_hunspell<char*> t_enc =
      mSandbox->invoke_sandbox_function(Hunspell_get_dic_encoding, mHandle);
  t_enc.copy_and_verify_string([&](std::unique_ptr<char[]> enc) {
    size_t len = std::strlen(enc.get());
    mDicEncoding = std::string(enc.get(), len);
  });
}

RLBoxHunspell::~RLBoxHunspell() {
  MOZ_DIAGNOSTIC_ASSERT(NS_IsMainThread());
  // Call hunspell's destroy which frees mHandle
  mSandbox->invoke_sandbox_function(Hunspell_destroy, mHandle);
  mHandle = nullptr;

  // Unregister callbacks
  mDestructFilemgr.unregister();
  mGetLineNum.unregister();
  mGetLine.unregister();
  mCreateFilemgr.unregister();
  mHunspellToUpperCase.unregister();
  mHunspellToLowerCase.unregister();
  mHunspellGetCurrentCS.unregister();

  // Clear any callback data and allow list
  mozHunspellCallbacks::Clear();
}

// Invoking hunspell with words larger than a certain size will cause the
// Hunspell sandbox to run out of memory. So we pick an arbitrary limit of
// 200000 here to ensure this doesn't happen.
static const size_t gWordSizeLimit = 200000;

int RLBoxHunspell::spell(const std::string& stdWord) {
  MOZ_DIAGNOSTIC_ASSERT(NS_IsMainThread());

  const int ok = 1;

  if (stdWord.length() >= gWordSizeLimit) {
    // Fail gracefully assuming the word is spelt correctly
    return ok;
  }

  // Copy word into the sandbox
  tainted_hunspell<char*> t_word = allocStrInSandbox(*mSandbox, stdWord);
  if (!t_word) {
    // Ran out of memory in the hunspell sandbox
    // Fail gracefully assuming the word is spelt correctly
    return ok;
  }

  // Check word
  int good = mSandbox
                 ->invoke_sandbox_function(
                     Hunspell_spell, mHandle,
                     rlbox::sandbox_const_cast<const char*>(t_word))
                 .copy_and_verify([](int good) { return good; });
  mSandbox->free_in_sandbox(t_word);
  return good;
}

const std::string& RLBoxHunspell::get_dict_encoding() const {
  return mDicEncoding;
}

// This function fails gracefully - if we run out of memory in the hunspell
// sandbox, we return empty suggestion list
std::vector<std::string> RLBoxHunspell::suggest(const std::string& stdWord) {
  MOZ_DIAGNOSTIC_ASSERT(NS_IsMainThread());

  if (stdWord.length() >= gWordSizeLimit) {
    return {};
  }

  // Copy word into the sandbox
  tainted_hunspell<char*> t_word = allocStrInSandbox(*mSandbox, stdWord);
  if (!t_word) {
    return {};
  }

  // Allocate suggestion list in the sandbox
  tainted_hunspell<char***> t_slst = mSandbox->malloc_in_sandbox<char**>();
  if (!t_slst) {
    // Free the earlier allocation
    mSandbox->free_in_sandbox(t_word);
    return {};
  }

  *t_slst = nullptr;

  // Get suggestions
  int nr = mSandbox
               ->invoke_sandbox_function(
                   Hunspell_suggest, mHandle, t_slst,
                   rlbox::sandbox_const_cast<const char*>(t_word))
               .copy_and_verify([](int nr) {
                 MOZ_RELEASE_ASSERT(nr >= 0);
                 return nr;
               });

  tainted_hunspell<char**> t_slst_ref = *t_slst;

  std::vector<std::string> suggestions;
  if (nr > 0 && t_slst_ref != nullptr) {
    // Copy suggestions from sandbox
    suggestions.reserve(nr);

    for (int i = 0; i < nr; i++) {
      tainted_hunspell<char*> t_sug = t_slst_ref[i];

      if (t_sug) {
        t_sug.copy_and_verify_string(
            [&](std::string sug) { suggestions.push_back(std::move(sug)); });
        // free the suggestion string allocated by the sandboxed hunspell
        mSandbox->free_in_sandbox(t_sug);
      }
    }

    // free the suggestion list allocated by the sandboxed hunspell
    mSandbox->free_in_sandbox(t_slst_ref);
  }

  mSandbox->free_in_sandbox(t_word);
  mSandbox->free_in_sandbox(t_slst);
  return suggestions;
}