PE文件格式分析心得
PE文件格式分析心得
PE文件格式最近好像炒得沸沸扬扬，由于我正在做一个这样的程序，索性将自己的一点心得写出来与大家同享。
PE文件头分两大部分：
1：DOS ‘MZ’ HEADER
2：IMAGE_NT_HEADERS
其中IMAGE_NT_HEADERS中包含
PE signature
IMAGE_FILE_HEADER
IMAGE_OPTIONAL_HEADER（其中包含Data Direcotry）
文件头后紧跟着为
Section Table (array of IMAGE_SECTION_HEADERs)
在DELPHI的windows.pad中已经有定义的有：
TImageDosHeader;
TImageNtHeaders;
TImageSectionHeader; { size of TIm..der is $28 }
定义变量后按住Ctrl可以察看具体的项目，这里我就不多说了，这方面的东西也很多。
而其他的如TImageResourceDirectory等，在DELPHI中却没有定义，察看其他资料，我在这里给出他们的结构和简单说明：
以下是我写的PEDump.exe的类型说明：
type
PIMAGE_RESOURCE_DIRECTORY = ^TImageResourceDirectory;
_IMAGE_RESOURCE_DIRECTORY = packed record
Characteristics:DWORD;
TimeDateStamp:DWORD;
MajorVersion:WORD;
MinorVersion:WORD;
NumberOfNamedEntries:WORD;
NumberOfIdEntries:WORD;
end;
TImageResourceDirectory = _IMAGE_RESOURCE_DIRECTORY;
{ 资源目录的格式说明 }
PIMAGE_RESOURCE_DIRECTORY_ENTRY = ^TImageResourceDirectoryEntry;
_IMAGE_RESOURCE_DIRECTORY_ENTRY = packed record
Name:DWORD; { NameOffset:31,NameIsString:1 }
// Id:WORD;
OffsetToData:DWORD; { OffsetToDirectory:31,DataIsDirectory:1 }
end;
TImageResourceDirectoryEntry = _IMAGE_RESOURCE_DIRECTORY_ENTRY;
{ 资源目录进入点的格式说明 }
PIMAGE_RESOURCE_DIRECTORY_STRING = ^TImageResourceDirectoryString;
_IMAGE_RESOURCE_DIRECTORY_STRING = packed record
Length:WORD;
NameString:CHAR;
end;
TImageResourceDirectoryString = _IMAGE_RESOURCE_DIRECTORY_STRING;
{ 资源目录名的格式说明 }
PIMAGE_RESOURCE_DIR_STRING_U = ^TImageResourceDirStringU;
_IMAGE_RESOURCE_DIR_STRING_U = packed record
Length:WORD;
NameString:WCHAR;
end;
TImageResourceDirStringU = _IMAGE_RESOURCE_DIR_STRING_U;
{ unicode形式的资源目录名的格式说明 }
PIMAGE_RESOURCE_DATA_ENTRY = ^TImageResourceDataEntry;
_IMAGE_RESOURCE_DATA_ENTRY = packed record
OffsetToData:DWORD;
Size:DWORD;
CodePage:DWORD;
Reserved:DWORD;
end;
TImageResourceDataEntry = _IMAGE_RESOURCE_DATA_ENTRY;
{ 资源目录数据进入点的格式说明 }
const
IMAGE_RESOURCE_NAME_IS_STRING = $80000000;
{ 检测TImageResourceDirectoryEntry.Name的最高为是否设立，
是则说明剩下的31位指向IMAGE_RESOURCE_DIR_STRING_U的偏移，
否则说明剩下的31位为一个整数ID。 }
IMAGE_RESOURCE_DATA_IS_DIRECTORY = $80000000;
{ 检测TImageResourceDirectoryEntry.OffsetToData的最高为是否设立，
是则说明剩下的31位指向另一个IMAGE_RESOURCE_DIRECTORY的偏移，
否则说明剩下的31位指向IMAGE_RESOURCE_DATA_ENTRY的偏移。 }
{ 以下是文件属性具体值常量说明 }
{ File Characteristics }
IMAGE_FILE_RELOCS_STRIPPED = $0001; // Relocation info stripped from file.
IMAGE_FILE_EXECUTABLE_IMAGE = $0002; // File is executable (i.e. no unresolved externel references).
IMAGE_FILE_LINE_NUMS_STRIPPED = $0004; // Line nunbers stripped from file.
IMAGE_FILE_LOCAL_SYMS_STRIPPED = $0008; // Local symbols stripped from file.
IMAGE_FILE_AGGRESIVE_WS_TRIM = $0010; // Agressively trim working set
IMAGE_FILE_LARGE_ADDRESS_AWARE = $0020; // App can handle >2gb addresses
IMAGE_FILE_BYTES_REVERSED_LO = $0080; // Bytes of machine word are reversed.
IMAGE_FILE_32BIT_MACHINE = $0100; // 32 bit word machine.
IMAGE_FILE_DEBUG_STRIPPED = $0200; // Debugging info stripped from file in .DBG file
IMAGE_FILE_REMOVABLE_RUN_FROM_SWAP = $0400; // If Image is on removable media, copy and run from the swap file.
IMAGE_FILE_NET_RUN_FROM_SWAP = $0800; // If Image is on Net, copy and run from the swap file.
IMAGE_FILE_SYSTEM = $1000; // System File.
IMAGE_FILE_DLL = $2000; // File is a DLL.
IMAGE_FILE_UP_SYSTEM_ONLY = $4000; // File should only be run on a UP machine
IMAGE_FILE_BYTES_REVERSED_HI = $8000; // Bytes of machine word are reversed.
{ 以下是文件头机器属性值的具体说明 }
{ Machine }
IMAGE_FILE_MACHINE_UNKNOWN = $0;
IMAGE_FILE_MACHINE_I386 = $014c; // Intel 386.
IMAGE_FILE_MACHINE_R3000 = $0162; // MIPS little-endian, $160 big-endian
IMAGE_FILE_MACHINE_R4000 = $0166; // MIPS little-endian
IMAGE_FILE_MACHINE_R10000 = $0168; // MIPS little-endian
IMAGE_FILE_MACHINE_WCEMIPSV2 = $0169; // MIPS little-endian WCE v2
IMAGE_FILE_MACHINE_ALPHA = $0184; // Alpha_AXP
IMAGE_FILE_MACHINE_SH3 = $01a2; // SH3 little-endian
IMAGE_FILE_MACHINE_SH3E = $01a4; // SH3E little-endian
IMAGE_FILE_MACHINE_SH4 = $01a6; // SH4 little-endian
IMAGE_FILE_MACHINE_SH5 = $01a8; // SH5
IMAGE_FILE_MACHINE_ARM = $01c0; // ARM Little-Endian
IMAGE_FILE_MACHINE_THUMB = $01c2;
IMAGE_FILE_MACHINE_ARM33 = $01d3;
IMAGE_FILE_MACHINE_POWERPC = $01F0; // IBM PowerPC Little-Endian
IMAGE_FILE_MACHINE_IA64 = $0200; // Intel 64
IMAGE_FILE_MACHINE_MIPS16 = $0266; // MIPS
IMAGE_FILE_MACHINE_ALPHA64 = $0284; // ALPHA64
IMAGE_FILE_MACHINE_MIPSFPU = $0366; // MIPS
IMAGE_FILE_MACHINE_MIPSFPU16 = $0466; // MIPS
// IMAGE_FILE_MACHINE_AXP64 IMAGE_FILE_MACHINE_ALPHA64
IMAGE_FILE_MACHINE_AMD64 = $0500; // AMD K8
IMAGE_FILE_MACHINE_TRICORE = $0520; // Infineon
IMAGE_FILE_MACHINE_CEF = $0CEF;
{ 以下是SECTION的属性值具体说明 }
{ Section characteristics }
// IMAGE_SCN_TYPE_REG = $00000000; // Reserved.
// IMAGE_SCN_TYPE_DSECT = $00000001; // Reserved.
// IMAGE_SCN_TYPE_NOLOAD = $00000002; // Reserved.
// IMAGE_SCN_TYPE_GROUP = $00000004; // Reserved.
IMAGE_SCN_TYPE_NO_PAD = $00000008; // Reserved.
// IMAGE_SCN_TYPE_COPY = $00000010; // Reserved.
IMAGE_SCN_CNT_CODE = $00000020; // Section contains code.
IMAGE_SCN_CNT_INITIALIZED_DATA = $00000040; // Section contains initialized data.
IMAGE_SCN_CNT_UNINITIALIZED_DATA = $00000080; // Section contains uninitialized data.
IMAGE_SCN_LNK_OTHER = $00000100; // Reserved.
IMAGE_SCN_LNK_INFO = $00000200; // Section contains comments or some other type of information.
// IMAGE_SCN_TYPE_OVER = $00000400; // Reserved.
IMAGE_SCN_LNK_REMOVE = $00000800; // Section contents will not become part of image.
IMAGE_SCN_LNK_COMDAT = $00001000; // Section contents comdat.
// = $00002000; // Reserved.
// IMAGE_SCN_MEM_PROTECTED - Obsolete = $00004000;
IMAGE_SCN_NO_DEFER_SPEC_EXC = $00004000; // Reset speculative exceptions handling bits in the TLB entries for this section.
IMAGE_SCN_GPREL = $00008000; // Section content can be accessed relative to GP
IMAGE_SCN_MEM_FARDATA = $00008000;
// IMAGE_SCN_MEM_SYSHEAP - Obsolete = $00010000;
IMAGE_SCN_MEM_PURGEABLE = $00020000;
IMAGE_SCN_MEM_16BIT = $00020000;
IMAGE_SCN_MEM_LOCKED = $00040000;
IMAGE_SCN_MEM_PRELOAD = $00080000;
IMAGE_SCN_ALIGN_1BYTES = $00100000; //
IMAGE_SCN_ALIGN_2BYTES = $00200000; //
IMAGE_SCN_ALIGN_4BYTES = $00300000; //
IMAGE_SCN_ALIGN_8BYTES = $00400000; //
IMAGE_SCN_ALIGN_16BYTES = $00500000; // Default alignment if no others are specified.
IMAGE_SCN_ALIGN_32BYTES = $00600000; //
IMAGE_SCN_ALIGN_64BYTES = $00700000; //
IMAGE_SCN_ALIGN_128BYTES = $00800000; //
IMAGE_SCN_ALIGN_256BYTES = $00900000; //
IMAGE_SCN_ALIGN_512BYTES = $00A00000; //
IMAGE_SCN_ALIGN_1024BYTES = $00B00000; //
IMAGE_SCN_ALIGN_2048BYTES = $00C00000; //
IMAGE_SCN_ALIGN_4096BYTES = $00D00000; //
IMAGE_SCN_ALIGN_8192BYTES = $00E00000; //
// Unused = $00F00000;
IMAGE_SCN_ALIGN_MASK = $00F00000;
IMAGE_SCN_LNK_NRELOC_OVFL = $01000000; // Section contains extended relocations.
IMAGE_SCN_MEM_DISCARDABLE = $02000000; // Section can be discarded.
IMAGE_SCN_MEM_NOT_CACHED = $04000000; // Section is not cachable.
IMAGE_SCN_MEM_NOT_PAGED = $08000000; // Section is not pageable.
IMAGE_SCN_MEM_SHARED = $10000000; // Section is shareable.
IMAGE_SCN_MEM_EXECUTE = $20000000; // Section is executable.
IMAGE_SCN_MEM_READ = $40000000; // Section is readable.
IMAGE_SCN_MEM_WRITE = $80000000; // Section is writeable.
我写了检测是否包含此属性的函数
function BeTrue(fg:Cardinal,Value):Boolean;
begin
Result:=fg and not Value=0;
end;
如果fg的属性值在Value中，则为True,否则为False;
例如 BeTrue(IMAGE_FILE_RELOCS_STRIPPED,PENTHead.FileHeader.Characteristics);
至于资源目录的读取，至少需要两重循环来定位，具体实现就要靠你的算法功力了：）
好了，差不多了，再具体一点我也说不上了，我做了一个PEDump的程序，
在我的主页http://mantousoft.51.net可以下载，程序中有更详细的说明，相信能帮助想了解这方面的东西的朋友。
以上是我的个人小心得，欢迎转载，转载请保持完整，谢谢！
