net/http/http_chunked_decoder.cc

   1 // Copyright (c) 2011 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 // Derived from:
   6 //   mozilla/netwerk/protocol/http/src/nsHttpChunkedDecoder.cpp
   7 // The license block is:
   8 /* ***** BEGIN LICENSE BLOCK *****
   9  * Version: MPL 1.1/GPL 2.0/LGPL 2.1
  10  *
  11  * The contents of this file are subject to the Mozilla Public License Version
  12  * 1.1 (the "License"); you may not use this file except in compliance with
  13  * the License. You may obtain a copy of the License at
  14  * http://www.mozilla.org/MPL/
  15  *
  16  * Software distributed under the License is distributed on an "AS IS" basis,
  17  * WITHOUT WARRANTY OF ANY KIND, either express or implied. See the License
  18  * for the specific language governing rights and limitations under the
  19  * License.
  20  *
  21  * The Original Code is Mozilla.
  22  *
  23  * The Initial Developer of the Original Code is
  24  * Netscape Communications.
  25  * Portions created by the Initial Developer are Copyright (C) 2001
  26  * the Initial Developer. All Rights Reserved.
  27  *
  28  * Contributor(s):
  29  *   Darin Fisher <darin@netscape.com> (original author)
  30  *
  31  * Alternatively, the contents of this file may be used under the terms of
  32  * either the GNU General Public License Version 2 or later (the "GPL"), or
  33  * the GNU Lesser General Public License Version 2.1 or later (the "LGPL"),
  34  * in which case the provisions of the GPL or the LGPL are applicable instead
  35  * of those above. If you wish to allow use of your version of this file only
  36  * under the terms of either the GPL or the LGPL, and not to allow others to
  37  * use your version of this file under the terms of the MPL, indicate your
  38  * decision by deleting the provisions above and replace them with the notice
  39  * and other provisions required by the GPL or the LGPL. If you do not delete
  40  * the provisions above, a recipient may use your version of this file under
  41  * the terms of any one of the MPL, the GPL or the LGPL.
  42  *
  43  * ***** END LICENSE BLOCK ***** */
  44
  45 #include "net/http/http_chunked_decoder.h"
  46
  47 #include "base/logging.h"
  48 #include "base/strings/string_number_conversions.h"
  49 #include "base/strings/string_piece.h"
  50 #include "base/strings/string_util.h"
  51 #include "net/base/net_errors.h"
  52
  53 namespace net {
  54
  55 // Absurdly long size to avoid imposing a constraint on chunked encoding
  56 // extensions.
  57 const size_t HttpChunkedDecoder::kMaxLineBufLen = 16384;
  58
  59 HttpChunkedDecoder::HttpChunkedDecoder()
  60     : chunk_remaining_(0),
  61       chunk_terminator_remaining_(false),
  62       reached_last_chunk_(false),
  63       reached_eof_(false),
  64       bytes_after_eof_(0) {
  65 }
  66
  67 int HttpChunkedDecoder::FilterBuf(char* buf, int buf_len) {
  68   int result = 0;
  69
  70   while (buf_len) {
  71     if (chunk_remaining_) {
  72       int num = std::min(chunk_remaining_, buf_len);
  73
  74       buf_len -= num;
  75       chunk_remaining_ -= num;
  76
  77       result += num;
  78       buf += num;
  79
  80       // After each chunk's data there should be a CRLF
  81       if (!chunk_remaining_)
  82         chunk_terminator_remaining_ = true;
  83       continue;
  84     } else if (reached_eof_) {
  85       bytes_after_eof_ += buf_len;
  86       break;  // Done!
  87     }
  88
  89     int bytes_consumed = ScanForChunkRemaining(buf, buf_len);
  90     if (bytes_consumed < 0)
  91       return bytes_consumed; // Error
  92
  93     buf_len -= bytes_consumed;
  94     if (buf_len)
  95       memmove(buf, buf + bytes_consumed, buf_len);
  96   }
  97
  98   return result;
  99 }
 100
 101 int HttpChunkedDecoder::ScanForChunkRemaining(const char* buf, int buf_len) {
 102   DCHECK_EQ(0, chunk_remaining_);
 103   DCHECK_GT(buf_len, 0);
 104
 105   int bytes_consumed = 0;
 106
 107   size_t index_of_lf = base::StringPiece(buf, buf_len).find('\n');
 108   if (index_of_lf != base::StringPiece::npos) {
 109     buf_len = static_cast<int>(index_of_lf);
 110     if (buf_len && buf[buf_len - 1] == '\r')  // Eliminate a preceding CR.
 111       buf_len--;
 112     bytes_consumed = static_cast<int>(index_of_lf) + 1;
 113
 114     // Make buf point to the full line buffer to parse.
 115     if (!line_buf_.empty()) {
 116       line_buf_.append(buf, buf_len);
 117       buf = line_buf_.data();
 118       buf_len = static_cast<int>(line_buf_.size());
 119     }
 120
 121     if (reached_last_chunk_) {
 122       if (buf_len)
 123         DVLOG(1) << "ignoring http trailer";
 124       else
 125         reached_eof_ = true;
 126     } else if (chunk_terminator_remaining_) {
 127       if (buf_len) {
 128         DLOG(ERROR) << "chunk data not terminated properly";
 129         return ERR_INVALID_CHUNKED_ENCODING;
 130       }
 131       chunk_terminator_remaining_ = false;
 132     } else if (buf_len) {
 133       // Ignore any chunk-extensions.
 134       size_t index_of_semicolon = base::StringPiece(buf, buf_len).find(';');
 135       if (index_of_semicolon != base::StringPiece::npos)
 136         buf_len = static_cast<int>(index_of_semicolon);
 137
 138       if (!ParseChunkSize(buf, buf_len, &chunk_remaining_)) {
 139         DLOG(ERROR) << "Failed parsing HEX from: " <<
 140             std::string(buf, buf_len);
 141         return ERR_INVALID_CHUNKED_ENCODING;
 142       }
 143
 144       if (chunk_remaining_ == 0)
 145         reached_last_chunk_ = true;
 146     } else {
 147       DLOG(ERROR) << "missing chunk-size";
 148       return ERR_INVALID_CHUNKED_ENCODING;
 149     }
 150     line_buf_.clear();
 151   } else {
 152     // Save the partial line; wait for more data.
 153     bytes_consumed = buf_len;
 154
 155     // Ignore a trailing CR
 156     if (buf[buf_len - 1] == '\r')
 157       buf_len--;
 158
 159     if (line_buf_.length() + buf_len > kMaxLineBufLen) {
 160       DLOG(ERROR) << "Chunked line length too long";
 161       return ERR_INVALID_CHUNKED_ENCODING;
 162     }
 163
 164     line_buf_.append(buf, buf_len);
 165   }
 166   return bytes_consumed;
 167 }
 168
 169
 170 // While the HTTP 1.1 specification defines chunk-size as 1*HEX
 171 // some sites rely on more lenient parsing.
 172 // http://www.yahoo.com/, for example, pads chunk-size with trailing spaces
 173 // (0x20) to be 7 characters long, such as "819b   ".
 174 //
 175 // A comparison of browsers running on WindowsXP shows that
 176 // they will parse the following inputs (egrep syntax):
 177 //
 178 // Let \X be the character class for a hex digit: [0-9a-fA-F]
 179 //
 180 //   RFC 2616: ^\X+$
 181 //        IE7: ^\X+[^\X]*$
 182 // Safari 3.1: ^[\t\r ]*\X+[\t ]*$
 183 //  Firefox 3: ^[\t\f\v\r ]*[+]?(0x)?\X+[^\X]*$
 184 // Opera 9.51: ^[\t\f\v ]*[+]?(0x)?\X+[^\X]*$
 185 //
 186 // Our strategy is to be as strict as possible, while not breaking
 187 // known sites.
 188 //
 189 //         Us: ^\X+[ ]*$
 190 bool HttpChunkedDecoder::ParseChunkSize(const char* start, int len, int* out) {
 191   DCHECK_GE(len, 0);
 192
 193   // Strip trailing spaces
 194   while (len && start[len - 1] == ' ')
 195     len--;
 196
 197   // Be more restrictive than HexStringToInt;
 198   // don't allow inputs with leading "-", "+", "0x", "0X"
 199   base::StringPiece chunk_size(start, len);
 200   if (chunk_size.find_first_not_of("0123456789abcdefABCDEF")
 201       != base::StringPiece::npos) {
 202     return false;
 203   }
 204
 205   int parsed_number;
 206   bool ok = base::HexStringToInt(chunk_size, &parsed_number);
 207   if (ok && parsed_number >= 0) {
 208     *out = parsed_number;
 209     return true;
 210   }
 211   return false;
 212 }
 213
 214 }  // namespace net