1*b1cdbd2cSJim Jagielski /**************************************************************
2*b1cdbd2cSJim Jagielski *
3*b1cdbd2cSJim Jagielski * Licensed to the Apache Software Foundation (ASF) under one
4*b1cdbd2cSJim Jagielski * or more contributor license agreements. See the NOTICE file
5*b1cdbd2cSJim Jagielski * distributed with this work for additional information
6*b1cdbd2cSJim Jagielski * regarding copyright ownership. The ASF licenses this file
7*b1cdbd2cSJim Jagielski * to you under the Apache License, Version 2.0 (the
8*b1cdbd2cSJim Jagielski * "License"); you may not use this file except in compliance
9*b1cdbd2cSJim Jagielski * with the License. You may obtain a copy of the License at
10*b1cdbd2cSJim Jagielski *
11*b1cdbd2cSJim Jagielski * http://www.apache.org/licenses/LICENSE-2.0
12*b1cdbd2cSJim Jagielski *
13*b1cdbd2cSJim Jagielski * Unless required by applicable law or agreed to in writing,
14*b1cdbd2cSJim Jagielski * software distributed under the License is distributed on an
15*b1cdbd2cSJim Jagielski * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16*b1cdbd2cSJim Jagielski * KIND, either express or implied. See the License for the
17*b1cdbd2cSJim Jagielski * specific language governing permissions and limitations
18*b1cdbd2cSJim Jagielski * under the License.
19*b1cdbd2cSJim Jagielski *
20*b1cdbd2cSJim Jagielski *************************************************************/
21*b1cdbd2cSJim Jagielski
22*b1cdbd2cSJim Jagielski
23*b1cdbd2cSJim Jagielski
24*b1cdbd2cSJim Jagielski // MARKER(update_precomp.py): autogen include statement, do not remove
25*b1cdbd2cSJim Jagielski #include "precompiled_tools.hxx"
26*b1cdbd2cSJim Jagielski
27*b1cdbd2cSJim Jagielski #include "md5.hxx"
28*b1cdbd2cSJim Jagielski
29*b1cdbd2cSJim Jagielski #include <cstddef>
30*b1cdbd2cSJim Jagielski #include <stdio.h>
31*b1cdbd2cSJim Jagielski
32*b1cdbd2cSJim Jagielski #include <tools/string.hxx>
33*b1cdbd2cSJim Jagielski
34*b1cdbd2cSJim Jagielski #ifdef WNT
35*b1cdbd2cSJim Jagielski #define FILE_OPEN_READ "rb"
36*b1cdbd2cSJim Jagielski #else
37*b1cdbd2cSJim Jagielski #define FILE_OPEN_READ "r"
38*b1cdbd2cSJim Jagielski #endif
39*b1cdbd2cSJim Jagielski
40*b1cdbd2cSJim Jagielski // Extended calc_md5_checksum to recognize Windows executables and libraries. To
41*b1cdbd2cSJim Jagielski // create the same md5 checksum for a (code/data) identical file it ignores a different
42*b1cdbd2cSJim Jagielski // date and header checksum. Please see crashrep/source/win32/soreport.cpp
43*b1cdbd2cSJim Jagielski // where the same method is also used. The crash reporter uses the MD5
44*b1cdbd2cSJim Jagielski // checksums to transfer them to the crash database. You have to make sure that both
45*b1cdbd2cSJim Jagielski // methods use the same algorithm otherwise there could be problems with stack reports.
46*b1cdbd2cSJim Jagielski
normalize_pe_image(sal_uInt8 * buffer,size_t nBufferSize)47*b1cdbd2cSJim Jagielski void normalize_pe_image(sal_uInt8* buffer, size_t nBufferSize)
48*b1cdbd2cSJim Jagielski {
49*b1cdbd2cSJim Jagielski const int OFFSET_PE_OFFSET = 0x3c;
50*b1cdbd2cSJim Jagielski const int OFFSET_COFF_TIMEDATESTAMP = 4;
51*b1cdbd2cSJim Jagielski const int PE_SIGNATURE_SIZE = 4;
52*b1cdbd2cSJim Jagielski const int COFFHEADER_SIZE = 20;
53*b1cdbd2cSJim Jagielski const int OFFSET_PE_OPTIONALHEADER_CHECKSUM = 64;
54*b1cdbd2cSJim Jagielski
55*b1cdbd2cSJim Jagielski // Check the header part of the file buffer
56*b1cdbd2cSJim Jagielski if (buffer[0] == sal_uInt8('M') && buffer[1] == sal_uInt8('Z'))
57*b1cdbd2cSJim Jagielski {
58*b1cdbd2cSJim Jagielski unsigned long PEHeaderOffset = (long)buffer[OFFSET_PE_OFFSET];
59*b1cdbd2cSJim Jagielski if (PEHeaderOffset < nBufferSize-4)
60*b1cdbd2cSJim Jagielski {
61*b1cdbd2cSJim Jagielski if ( buffer[PEHeaderOffset+0] == sal_uInt8('P') &&
62*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+1] == sal_uInt8('E') &&
63*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+2] == 0 &&
64*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+3] == 0 )
65*b1cdbd2cSJim Jagielski {
66*b1cdbd2cSJim Jagielski PEHeaderOffset += PE_SIGNATURE_SIZE;
67*b1cdbd2cSJim Jagielski if (PEHeaderOffset+OFFSET_COFF_TIMEDATESTAMP < nBufferSize-4)
68*b1cdbd2cSJim Jagielski {
69*b1cdbd2cSJim Jagielski // Set timedatestamp and checksum fields to a normalized
70*b1cdbd2cSJim Jagielski // value to enforce the same MD5 checksum for identical
71*b1cdbd2cSJim Jagielski // Windows executables/libraries.
72*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+OFFSET_COFF_TIMEDATESTAMP+0] = 0;
73*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+OFFSET_COFF_TIMEDATESTAMP+1] = 0;
74*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+OFFSET_COFF_TIMEDATESTAMP+2] = 0;
75*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+OFFSET_COFF_TIMEDATESTAMP+3] = 0;
76*b1cdbd2cSJim Jagielski }
77*b1cdbd2cSJim Jagielski
78*b1cdbd2cSJim Jagielski if (PEHeaderOffset+COFFHEADER_SIZE+OFFSET_PE_OPTIONALHEADER_CHECKSUM < nBufferSize-4)
79*b1cdbd2cSJim Jagielski {
80*b1cdbd2cSJim Jagielski // Set checksum to a normalized value
81*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+COFFHEADER_SIZE+OFFSET_PE_OPTIONALHEADER_CHECKSUM] = 0;
82*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+COFFHEADER_SIZE+OFFSET_PE_OPTIONALHEADER_CHECKSUM+1] = 0;
83*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+COFFHEADER_SIZE+OFFSET_PE_OPTIONALHEADER_CHECKSUM+2] = 0;
84*b1cdbd2cSJim Jagielski buffer[PEHeaderOffset+COFFHEADER_SIZE+OFFSET_PE_OPTIONALHEADER_CHECKSUM+3] = 0;
85*b1cdbd2cSJim Jagielski }
86*b1cdbd2cSJim Jagielski }
87*b1cdbd2cSJim Jagielski }
88*b1cdbd2cSJim Jagielski }
89*b1cdbd2cSJim Jagielski }
90*b1cdbd2cSJim Jagielski
calc_md5_checksum(const char * filename,ByteString & aChecksum)91*b1cdbd2cSJim Jagielski rtlDigestError calc_md5_checksum( const char *filename, ByteString &aChecksum )
92*b1cdbd2cSJim Jagielski {
93*b1cdbd2cSJim Jagielski const size_t BUFFER_SIZE = 0x1000;
94*b1cdbd2cSJim Jagielski const size_t MINIMAL_SIZE = 512;
95*b1cdbd2cSJim Jagielski
96*b1cdbd2cSJim Jagielski sal_uInt8 checksum[RTL_DIGEST_LENGTH_MD5];
97*b1cdbd2cSJim Jagielski rtlDigestError error = rtl_Digest_E_None;
98*b1cdbd2cSJim Jagielski
99*b1cdbd2cSJim Jagielski FILE *fp = fopen( filename, FILE_OPEN_READ );
100*b1cdbd2cSJim Jagielski
101*b1cdbd2cSJim Jagielski if ( fp )
102*b1cdbd2cSJim Jagielski {
103*b1cdbd2cSJim Jagielski rtlDigest digest = rtl_digest_createMD5();
104*b1cdbd2cSJim Jagielski
105*b1cdbd2cSJim Jagielski if ( digest )
106*b1cdbd2cSJim Jagielski {
107*b1cdbd2cSJim Jagielski size_t nBytesRead;
108*b1cdbd2cSJim Jagielski sal_uInt8 buffer[BUFFER_SIZE];
109*b1cdbd2cSJim Jagielski bool bHeader(true);
110*b1cdbd2cSJim Jagielski
111*b1cdbd2cSJim Jagielski while ( rtl_Digest_E_None == error &&
112*b1cdbd2cSJim Jagielski 0 != (nBytesRead = fread( buffer, 1, sizeof(buffer), fp )) )
113*b1cdbd2cSJim Jagielski {
114*b1cdbd2cSJim Jagielski if (bHeader)
115*b1cdbd2cSJim Jagielski {
116*b1cdbd2cSJim Jagielski bHeader = false;
117*b1cdbd2cSJim Jagielski if (nBytesRead >= MINIMAL_SIZE && buffer[0] == sal_uInt8('M') && buffer[1] == sal_uInt8('Z') )
118*b1cdbd2cSJim Jagielski normalize_pe_image(buffer, nBytesRead);
119*b1cdbd2cSJim Jagielski }
120*b1cdbd2cSJim Jagielski
121*b1cdbd2cSJim Jagielski error = rtl_digest_updateMD5( digest, buffer, nBytesRead );
122*b1cdbd2cSJim Jagielski }
123*b1cdbd2cSJim Jagielski
124*b1cdbd2cSJim Jagielski if ( rtl_Digest_E_None == error )
125*b1cdbd2cSJim Jagielski {
126*b1cdbd2cSJim Jagielski error = rtl_digest_getMD5( digest, checksum, sizeof(checksum) );
127*b1cdbd2cSJim Jagielski }
128*b1cdbd2cSJim Jagielski
129*b1cdbd2cSJim Jagielski rtl_digest_destroyMD5( digest );
130*b1cdbd2cSJim Jagielski
131*b1cdbd2cSJim Jagielski for ( std::size_t i = 0; i < sizeof(checksum); i++ )
132*b1cdbd2cSJim Jagielski {
133*b1cdbd2cSJim Jagielski if ( checksum[i] < 16 )
134*b1cdbd2cSJim Jagielski aChecksum.Append( "0" );
135*b1cdbd2cSJim Jagielski aChecksum += ByteString::CreateFromInt32( checksum[i], 16 );
136*b1cdbd2cSJim Jagielski }
137*b1cdbd2cSJim Jagielski }
138*b1cdbd2cSJim Jagielski
139*b1cdbd2cSJim Jagielski fclose( fp );
140*b1cdbd2cSJim Jagielski }
141*b1cdbd2cSJim Jagielski else
142*b1cdbd2cSJim Jagielski error = rtl_Digest_E_Unknown;
143*b1cdbd2cSJim Jagielski
144*b1cdbd2cSJim Jagielski return error;
145*b1cdbd2cSJim Jagielski }
146