Hello again, been a few days since I have posted anything at Evilzone now. I have been reading a lot of papers on various things related to assembly, disassembly, CPU architecture, PE/COFF structure documentation and so on. But now am back for further questions!
Also, for those of you who have been reading my other topics. I will update them with solutions once they are solved, I am still working on translating bytes into assembly instructions and so on. Should be doing some progress tonight/tomorrow.
Now, to my question. Or rather, verification.
I just want someone with experience with this to verify that I got things right. When or if this topic/my question gets verified to be correct it can probably be used as guidance for others to
So. I will just write up what I have read/interpriated about the PE structure below this text block. Then someone who knows his shit read thought it and see if I got anything wrong. There might be a question or two in the text to. Thanks in advance
-------------------------------
The PE(Portable Executable) StructureAny PE file will consist of the following structures(roughly, we will go in more details later). List will be in the order as they will be in the file structure.
- DOS MZ Header. This is the very first bytes in any PE, it states various things about the PE. (In-depth details will be put along side the structure(s) below)
- DOS Stub. This is 100 bytes++ with machine code just to output "This program cannot run in DOS mode". This is here to avoid programs from running in DOS Mode. This part of the PE is not required tho, you can tell your compiler to not include this.
- PE Signature, part of the PE Header. Its the first 4 bytes in the IMAGE_NT_HEADERS. Its there to verify that this is in fact a PE. The signature for a valid PE file is the ASCII letters 'PE' followed by 2x 0-bytes
- PE Header(s). This is the vital part of a PE, it contains information about the entire PE structure. It also contains two inner headers and one inner inner header. They are listed below.
- PE Header - File Header. This header contains further vital information about the file. Details will be listed in the header structures below.
- PE Header - Optional Header. Despite its name, this header is NOT optional at all. This is the largest header in the PE Header. Once again, this header contains tons of information regarding the PE layout/structure. Among other things, the PE entry point. (The first 'line' of code that will be executed once the program runs)
- PE Header - Optional Header - DataDirectory. The DataDirectory header is an array of the IMAGE_DATA_DIRECTORY structure that contains information about the various directories in the PE. Import Directory, Export Directory, Resource Directory and so on. These directories contains information about where to locate other information, in short.
- Section Header(s). Section header structures. One structure for each of the sections in the file. Normally like 4-6 sections. Structure contains section name, size, address and so on. More info on sections 4 points down. The number of sections are set in PE Headers -> File Header -> NumberOfSections
- Data Directories. These are the directories mentioned above, containing various information. Resources, imported DLL's and so on. (I haven't read so much about all of these yet)
- Data Directories ...
- [...]
- Sections. Sections are large chunks of bytes that can be anything. Code, resources, icon(s), information and so on. Each section have its own "rules", some are read-only, some are executable and some are data - read and writable. The sections are typically called ".text", ".data", ".rsrc", ".reloc". There can be up to 16 sections. The number of sections are set by a value in the File Header. Each section have a structure of IMAGE_SECTION_HEADER about itself containing its name, size, address, offset and "rules" or Characteristics.
- Sections ...
- [...]
- EOF. End Of File
The structures (C/++ syntax)
DOS MZ Headertypedef struct _IMAGE_DOS_HEADER { // DOS .EXE header
USHORT e_magic; // Magic number // 2 bytes
USHORT e_cblp; // Bytes on last page of file // 2 bytes
USHORT e_cp; // Pages in file // 2 bytes
USHORT e_crlc; // Relocations // 2 bytes
USHORT e_cparhdr; // Size of header in paragraphs // 2 bytes
USHORT e_minalloc; // Minimum extra paragraphs needed // 2 bytes
USHORT e_maxalloc; // Maximum extra paragraphs needed // 2 bytes
USHORT e_ss; // Initial (relative) SS value // 2 bytes
USHORT e_sp; // Initial SP value // 2 bytes
USHORT e_csum; // Checksum // 2 bytes
USHORT e_ip; // Initial IP value // 2 bytes
USHORT e_cs; // Initial (relative) CS value // 2 bytes
USHORT e_lfarlc; // File address of relocation table // 2 bytes
USHORT e_ovno; // Overlay number // 2 bytes
USHORT e_res[4]; // Reserved words // 2*4 bytes
USHORT e_oemid; // OEM identifier (for e_oeminfo) // 2 bytes
USHORT e_oeminfo; // OEM information; e_oemid specific // 2 bytes
USHORT e_res2[10]; // Reserved words // 2 * 10 bytes
LONG e_lfanew; // File address of new exe header // 4 bytes
} IMAGE_DOS_HEADER, *PIMAGE_DOS_HEADER;
Total bytes: 64
e_lfanew points to the offset of the PE Header, this is because of the DOS header. Because its optional and it can be as small or large as you want/your compiler wants. The offset of the PE Header can vary.
Couldn't find any MS.com link :S
PE Headertypedef struct _IMAGE_NT_HEADERS { // Includes the PE signature, File Header and Optional Header
DWORD Signature; // The PE signature // 4 bytes
IMAGE_FILE_HEADER FileHeader; // The File Header structure // 20 bytes
IMAGE_OPTIONAL_HEADER OptionalHeader; // The Optional Header structure // 224 bytes
} IMAGE_NT_HEADERS, *PIMAGE_NT_HEADERS;
Total bytes: 248 (including the other structures)(Maximum, can be less depending on the amount of DataDirectories in the Optional Header
http://msdn.microsoft.com/en-us/library/ms680336%28v=vs.85%29.aspxFile Header
typedef struct _IMAGE_FILE_HEADER {
WORD Machine; // Machine type (alpha/motorola/...)(001, 076 == I386) // 2 bytes
WORD NumberOfSections; // Number of sections in the file // 2 bytes
DWORD TimeDateStamp; // The time that the file was created // 4 bytes
DWORD PointerToSymbolTable; // The file offset of the COFF symbol table // 4 bytes
// This field is only used in OBJ files and PE files
// with COFF debug info
DWORD NumberOfSymbols; // The number of symbols in the COFF symbol table // 4 bytes
WORD SizeOfOptionalHeader; // Size of the OptionalHeader structure // 2 bytes
WORD Characteristics; // Flags of the file (EXE/DLL/System file/...) // 2 bytes
} IMAGE_FILE_HEADER, *PIMAGE_FILE_HEADER;
Total bytes: 20
http://msdn.microsoft.com/en-us/library/ms680313%28v=vs.85%29.aspxOptional Header
typedef struct _IMAGE_OPTIONAL_HEADER {
WORD Magic; // 2 bytes identifying the state of the file. // 2 bytes
BYTE MajorLinkerVersion; // Linker major version number. // 1 bytes
BYTE MinorLinkerVersion; // Linker minor version number. // 1 bytes
DWORD SizeOfCode; // Size of the code section OR Sum of all code sections (multiple sections). // 4 bytes
DWORD SizeOfInitializedData; // Size of the initialized data OR .... multiple data sections. // 4 bytes
DWORD SizeOfUninitializedData; // Size of the uninitialized data section (BSS) OR ... multiple BBS sections. // 4 bytes
DWORD AddressOfEntryPoint; // Address of entry point (RVA of the 1st instruction to be executed) // 4 bytes
DWORD BaseOfCode; // Address (RVA) of beginning of code section. // 4 bytes
DWORD BaseOfData; // Address (RVA) of beginning of data section. // 4 bytes
DWORD ImageBase; // The *preferred* load address of the file (default is 0x00400000). // 4 bytes
DWORD SectionAlignment; // Alignment (in as strings) of sections when loaded into memory. // 4 bytes
DWORD FileAlignment; // Alignment (in as strings) of sections in the file (multiplies of 512 as strings). // 4 bytes
WORD MajorOperatingSystemVersion; // Major version number of required OS. // 2 bytes
WORD MinorOperatingSystemVersion; // Minor version number of required OS. // 2 bytes
WORD MajorImageVersion; // Major version number of image. // 2 bytes
WORD MinorImageVersion; // Minor version number of image. // 2 bytes
WORD MajorSubsystemVersion; // Major version number of subsystem. // 2 bytes
WORD MinorSubsystemVersion; // Minor version number of subsystem. // 2 bytes
DWORD Win32VersionValue; // Dunno! But I guess for future use. // 4 bytes
DWORD SizeOfImage; // Total size of the PE image in memory (All Headers & Sections aligned to SectionAlignment). // 4 bytes
DWORD SizeOfHeaders; // Size of all headers & section table. (The file offset of the first section in the PE file) // 4 bytes
DWORD CheckSum; // Image file checksum. (computing algorithm is in IMAGHELP.DLL) // 4 bytes
WORD Subsystem; // Target subsystem of the PE file. (Mostly GUI & CUI) // 2 bytes
WORD DllCharacteristics; // Flags used to indicate if a DLL image includes EPs. // 2 bytes
DWORD SizeOfStackReserve; // Size of stack to reserve. // 4 bytes
DWORD SizeOfStackCommit; // Size of stack to commit. // 4 bytes
DWORD SizeOfHeapReserve; // Size of local heap space to reserve. // 4 bytes
DWORD SizeOfHeapCommit; // Size of local heap space to commit. // 4 bytes
DWORD LoaderFlags; // Choose to (break/debug/run normally (default)) on load. // 4 bytes
DWORD NumberOfRvaAndSizes; // The length of the DataDirectory array that follows. // 4 bytes
IMAGE_DATA_DIRECTORY DataDirectory[IMAGE_NUMBEROF_DIRECTORY_ENTRIES]; // An array of IMAGE_DATA_DIRECTORY, lenght = 16 // 8 bytes * 16
} IMAGE_OPTIONAL_HEADER, *PIMAGE_OPTIONAL_HEADER;
Total max bytes: 224
http://msdn.microsoft.com/en-us/library/ms680339%28v=vs.85%29.aspxDataDirectorytypedef struct _IMAGE_DATA_DIRECTORY {
DWORD VirtualAddress; // RVA of the location of the directory // 4 bytes
DWORD Size; // Size of the directory // 4 bytes
} IMAGE_DATA_DIRECTORY, *PIMAGE_DATA_DIRECTORY;
* IMAGE_NUMBEROF_DIRECTORY_ENTRIES
Additionally I am guessing there is structures for each of the directories, export, import, resource, exception, security, basereloc, debug, copyright. etc. etc
http://msdn.microsoft.com/en-us/library/ms680305%28v=vs.85%29.aspxSection Header(s)
typedef struct _IMAGE_SECTION_HEADER {
BYTE Name[IMAGE_SIZEOF_SHORT_NAME]; // Simply the name of the section (ANSI name) // 8 bytes
union {
DWORD PhysicalAddress; //? //
DWORD VirtualSize; // Size of the file when its mapped to memory. Must be multiple of 4096 // 4 bytes
} Misc;
DWORD VirtualAddress; // The RVAof the file where it should be mapped to memory // 4 bytes
DWORD SizeOfRawData; // The size of the section, rounded to the next multiply of the FileAlighment // 4 bytes
DWORD PointerToRawData; // The offset in the file of the beginning og the section // 4 bytes
DWORD PointerToRelocations; // Not used // 4 bytes
DWORD PointerToLinenumbers; // Not used // 4 bytes
WORD NumberOfRelocations; // Not used // 2 bytes
WORD NumberOfLinenumbers; // File-based offset of the line number table. (Not really used) // 2 bytes
DWORD Characteristics; // Flags for section charactaristics as follow below // 4 bytes
} IMAGE_SECTION_HEADER, *PIMAGE_SECTION_HEADER;
Total bytes: 40
There is one of this structure for each section in the PE
http://msdn.microsoft.com/en-us/library/ms680341%28v=vs.85%29.aspxNow, I believe this is all the headers and whatnot. Except the directory structures, import, export and so on. Am I rit?
