I
import
Unregistered / Unconfirmed
GUEST, unregistred user!
PORTABLE EXECUTABLE FORMAT
Author: Micheal J. O'Leary
Preface
This document was edited and released by Microsoft Developer
Support. It describes the binary portable executable format for NT.
The information is provided at this point because we feel it will
make the work of application development easier. Unfortunately, the
information in this document may change before the final release of
Windows NT. Microsoft is NOT committing to stay with these formats
by releasing this document. Questions or follow-ups for any of the
information presented here should be posted to CompuServe MSWIN32
forum, section 6.
--Steve Firebaugh
Microsoft Developer Support
Contents
1. Overview
2. PE Header
3. Object Table
4. Image Pages
5. Exports
5.1 Export Directory Table
5.2 Export Address Table
5.3 Export Name Table Pointers
5.4 Export Ordinal Table
5.5 Export Name Table
6. Imports
6.1 Import Directory Table
6.2 Import Lookup Table
6.3 Hint-Name Table
6.4 Import Address Table
7. Thread Local Storage
7.1 Thread Local Storage Directory Table
7.2 Thread Local Storage CallBack Table
8. Resources
8.1 Resource Directory Table
8.2 Resource Example
9. Fixup Table
9.1 Fixup Block
10. Debug Information
10.1 Debug Directory
1. Overview
< < Base of Image Header
DOS 2 Compatible
EXE Header
unused
OEM Identifier
OEM Info
DOS 2.0 Section
Offset to (for DOS compatibility only)
PE Header
DOS 2.0 Stub
Program &
Reloc. Table
unused
Aligned on 8 byte boundary
PE Header
Object Table
Image Pages
import info
export info
fixup info
resource info
debug info
Figure
1. A typical 32-bit Portable EXE File Layout
2. PE Header
SIGNATURE BYTES
CPU TYPE
# OBJECTS
TIME/DATE STAMP
RESERVED
RESERVED
NT HDR SIZE
FLAGS
RESERVED
LMAJORLMINOR
RESERVED
RESERVED
RESERVED
ENTRYPOINT RVA
RESERVED
RESERVED
IMAGE BASE
OBJECT ALIGN
FILE ALIGN
OS MAJOR
OS MINOR
USER MAJOR
USER MINOR
SUBSYS MAJOR
SUBSYS MINOR
RESERVED
IMAGE SIZE
HEADER SIZE
FILE CHECKSUM
SUBSYSTEM
DLL FLAGS
STACK RESERVE SIZE
STACK COMMIT SIZE
HEAP RESERVE SIZE
HEAP COMMIT SIZE
RESERVED
# INTERESTING RVA/SIZES
EXPORT TABLE RVA
TOTAL EXPORT DATA SIZE
IMPORT TABLE RVA
TOTAL IMPORT DATA SIZE
RESOURCE TABLE RVA
TOTAL RESOURCE DATA SIZE
EXCEPTION TABLE RVA
TOTAL EXCEPTION DATA SIZE
SECURITY TABLE RVA
TOTAL SECURITY DATA SIZE
FIXUP TABLE RVA
TOTAL FIXUP DATA SIZE
DEBUG TABLE RVA
TOTAL DEBUG DIRECTORIES
IMAGE DESCRIPTION RVA
TOTAL DESCRIPTION SIZE
MACHINE SPECIFIC RVA
MACHINE SPECIFIC SIZE
THREAD LOCAL STORAGE RVA
TOTAL TLS SIZE
Figure 2. PE Header
Notes:
o A VA is a virtual address that is already biased by the Image
Base found in the PE Header. A RVA is a virtual address that is
relative to the Image Base.
o An RVA in the PE Header which has a value of zero indicates the
field isn't used.
o Image pages are aligned and zero padded to a File Align
boundary. The bases of all other tables and structures must be
aligned on DWORD (4 byte) boundary. Thus, all VA's and RVA's
must be on a 32 bit boundary. All table and structure fields
must be aligned on their "natural" boundaries, with the possible
exception of the Debug Info.
SIGNATURE BYTES = DB * 4.
Current value is "PE/0/0". Thats PE followed by two zeros (nulls).
CPU TYPE = DW CPU Type.
This field specifies the type of CPU compatibility required by this
image to run. The values are:
o 0000h __unknown
o 014Ch __80386
o 014Dh __80486
o 014Eh __80586
o 0162h __MIPS Mark I (R2000, R3000)
o 0163h __MIPS Mark II (R6000)
o 0166h __MIPS Mark III (R4000)
# OBJECTS = DW Number of object entries.
This field specifies the number of entries in the Object Table.
TIME/DATE STAMP = DD Used to store the time and date the file was
created or modified by the linker.
NT HDR SIZE = DW This is the number of remaining bytes in the NT
header that follow the FLAGS field.
FLAGS = DW Flag bits for the image.
The flag bits have the following definitons:
o 0000h __Program image.
o 0002h __Image is executable.
If this bit isn't set, then it indicates that either errors
where detected at link time or that the image is being
incrementally linked and therefore can't be loaded.
o 0200h __Fixed.
Indicates that if the image can't be loaded at the Image Base,
then don't load it.
o 2000h __Library image.
LMAJOR/LMINOR = DB Linker major/minor version number.
ENTRYPOINT RVA = DD Entrypoint relative virtual address.
The address is relative to the Image Base. The address is the
starting address for program images and the library initialization
and library termination address for library images.
IMAGE BASE = DD The virtual base of the image.
This will be the virtual address of the first byte of the file (Dos
Header). This must be a multiple of 64K.
OBJECT ALIGN = DD The alignment of the objects. This must be a power
of 2 between 512 and 256M inclusive. The default is 64K.
FILE ALIGN = DD Alignment factor used to align image pages. The
alignment factor (in bytes) used to align the base of the image pages
and to determine the granularity of per-object trailing zero pad.
Larger alignment factors will cost more file space; smaller alignment
factors will impact demand load performance, perhaps significantly.
Of the two, wasting file space is preferable. This value should be a
power of 2 between 512 and 64K inclusive.
OS MAJOR/MINOR = DW OS version number required to run this image.
USER MAJOR/MINOR # = DW User major/minor version number.
This is useful for differentiating between revisions of
images/dynamic linked libraries. The values are specified at link
time by the user.
SUBSYS MAJOR/MINOR # = DW Subsystem major/minor version number.
IMAGE SIZE = DD The virtual size (in bytes) of the image.
This includes all headers. The total image size must be a multiple
of Object Align.
HEADER SIZE = DD Total header size.
The combined size of the Dos Header, PE Header and Object Table.
FILE CHECKSUM = DD Checksum for entire file. Set to 0 by the linker.
SUBSYSTEM = DW NT Subsystem required to run this image.
The values are:
o 0000h __Unknown
o 0001h __Native
o 0002h __Windows GUI
o 0003h __Windows Character
o 0005h __OS/2 Character
o 0007h __Posix Character
DLL FLAGS = DW Indicates special loader requirements.
This flag has the following bit values:
o 0001h __Per-Process Library Initialization.
o 0002h __Per-Process Library Termination.
o 0004h __Per-Thread Library Initialization.
o 0008h __Per-Thread Library Termination.
All other bits are reserved for future use and should be set to zero.
STACK RESERVE SIZE = DD Stack size needed for image.
The memory is reserved, but only the STACK COMMIT SIZE is committed.
The next page of the stack is a 'guarded page'. When the application
hits the guarded page, the guarded page becomes valid, and the next
page becomes the guarded page. This continues until the RESERVE SIZE
is reached.
STACK COMMIT SIZE = DD Stack commit size.
HEAP RESERVE SIZE = DD Size of local heap to reserve.
HEAP COMMIT SIZE = DD Amount to commit in local heap.
# INTERESTING VA/SIZES = DD Indicates the size of the VA/SIZE array
that follows.
EXPORT TABLE RVA = DD Relative Virtual Address of the Export Table.
This address is relative to the Image Base.
IMPORT TABLE RVA = DD Relative Virtual Address of the Import Table.
This address is relative to the Image Base.
RESOURCE TABLE RVA = DD Relative Virtual Address of the Resource
Table. This address is relative to the Image Base.
EXCEPTION TABLE RVA = DD Relative Virtual Address of the Exception
Table. This address is relative to the Image Base.
SECURITY TABLE RVA = DD Relative Virtual Address of the Security
Table. This address is relative to the Image Base.
FIXUP TABLE RVA = DD Relative Virtual Address of the Fixup Table.
This address is relative to the Image Base.
DEBUG TABLE RVA = DD Relative Virtual Address of the Debug Table.
This address is relative to the Image Base.
IMAGE DESCRIPTION RVA = DD Relative Virtual Address of the
description string specified in the module definiton file.
MACHINE SPECIFIC RVA = DD Relative Virtual Address of a machine
specific value. This address is relative to the Image Base.
TOTAL EXPORT DATA SIZE = DD Total size of the export data.
TOTAL IMPORT DATA SIZE = DD Total size of the import data.
TOTAL RESOURCE DATA SIZE = DD Total size of the resource data.
TOTAL EXCEPTION DATA SIZE = DD Total size of the exception data.
TOTAL SECURITY DATA SIZE = DD Total size of the security data.
TOTAL FIXUP DATA SIZE = DD Total size of the fixup data.
TOTAL DEBUG DIRECTORIES = DD Total number of debug directories.
TOTAL DESCRIPTION SIZE = DD Total size of the description data.
MACHINE SPECIFIC SIZE = DD A machine specific value.
3. Object Table
The number of entries in the Object Table is given by the # Objects
field in the PE Header. Entries in the Object Table are numbered
starting from one. The object table immediately follows the PE
Header. The code and data memory object entries are in the order
chosen by the linker. The virtual addresses for objects must be
assigned by the linker such that they are in ascending order and
adjacent, and must be a multiple of Object Align in the PE header.
Each Object Table entry has the following format:
OBJECT NAME
VIRTUAL SIZE
RVA
PHYSICAL SIZE
PHYSICAL OFFSET
RESERVED
RESERVED
RESERVED
OBJECT FLAGS
Figure 3. Object Table
OBJECT NAME = DB * 8 Object name. This is an eight-byte null-padded
ASCII string representing the object name.
VIRTUAL SIZE = DD Virtual memory size. The size of the object that
will be allocated when the object is loaded. Any difference between
PHYSICAL SIZE and VIRTUAL SIZE is zero filled.
RVA = DD Relative Virtual Address. The virtual address the object is
currently relocated to, relative to the Image Base. Each Object's
virtual address space consumes a multiple of Object Align (power of 2
between 512 and 256M inclusive. Default is 64K), and immediately
follows the previous Object in the virtual address space (the virtual
address space for a image must be dense).
PHYSICAL SIZE = DD Physical file size of initialized data. The size
of the initialized data in the file for the object. The physical
size must be a multiple of the File Align field in the PE Header, and
must be less than or equal to the Virtual Size.
PHYSICAL OFFSET = DD Physical offset for object's first page. This
offset is relative to beginning of the EXE file, and is aligned on a
multiple of the File Align field in the PE Header. The offset is
used as a seek value.
OBJECT FLAGS = DD Flag bits for the object. The object flag bits
have the following definitions:
o 000000020h __Code object.
o 000000040h __Initialized data object.
o 000000080h __Uninitialized data object.
o 040000000h __Object must not be cached.
o 080000000h __Object is not pageable.
o 100000000h __Object is shared.
o 200000000h __Executable object.
o 400000000h __Readable object.
o 800000000h __Writeable object.
All other bits are reserved for future use and should be set to zero.
4. Image Pages
The Image Pages section contains all initialized data for all
objects. The seek offset for the first page in each object is
specified in the object table and is aligned on a File Align
boundary. The objects are ordered by the RVA. Every object begins
on a multiple of Object Align.
5. Exports
A typical file layout for the export information follows:
DIRECTORY TABLE
ADDRESS TABLE
NAME PTR TABLE
ORDINAL TABLE
NAME STRINGS
Figure 4. Export File Layout
5.1 Export Directory Table
The export information begins with the Export Directory Table which
describes the remainder of the export information. The Export
Directory Table contains address information that is used to resolve
fixup references to the entry points within this image.
EXPORT FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
NAME RVA
ORDINAL BASE
# EAT ENTRIES
# NAME PTRS
ADDRESS TABLE RVA
NAME PTR TABLE RVA
ORDINAL TABLE RVA
Figure 5. Export Directory Table Entry
EXPORT FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the export data was created.
MAJOR/MINOR VERSION = DW A user settable major/minor version number.
NAME RVA = DD Relative Virtual Address of the Dll asciiz Name.
This is the address relative to the Image Base.
ORDINAL BASE = DD First valid exported ordinal.
This field specifies the starting ordinal number for the export
address table for this image. Normally set to 1.
# EAT ENTRIES = DD Indicates number of entries in the Export Address
Table.
# NAME PTRS = DD This indicates the number of entries in the Name Ptr
Table (and parallel Ordinal Table).
ADDRESS TABLE RVA = DD Relative Virtual Address of the Export Address
Table.
This address is relative to the Image Base.
NAME TABLE RVA = DD Relative Virtual Address of the Export Name Table
Pointers.
This address is relative to the beginning of the Image Base. This
table is an array of RVA's with # NAMES entries.
ORDINAL TABLE RVA = DD Relative Virtual Address of Export Ordinals
Table Entry.
This address is relative to the beginning of the Image Base.
5.2 Export Address Table
The Export Address Table contains the address of exported entrypoints
and exported data and absolutes. An ordinal number is used to index
the Export Address Table. The ORDINAL BASE must be subracted from the
ordinal number before indexing into this table.
Export Address Table entry formats are described below:
EXPORTED RVA
Figure 6. Export Address Table Entry
EXPORTED RVA = DD Export address.
This field contains the relative virtual address of the exported
entry (relative to the Image Base).
5.3 Export Name Table Pointers
The export name table pointers array contains address into the Export
Name Table. The pointers are 32-bits each, and are relative to the
Image Base. The pointers are ordered lexically to allow binary
searches.
5.4 Export Ordinal Table
The Export Name Table Pointers and the Export Ordinal Table form two
parallel arrays, separated to allow natural field alignment. The
export ordinal table array contains the Export Address Table ordinal
numbers associated with the named export referenced by corresponding
Export Name Table Pointers.
The ordinals are 16-bits each, and already include the Ordinal Base
stored in the Export Directory Table.
5.5 Export Name Table
The export name table contains optional ASCII names for exported
entries in the image. These tables are used with the array of Export
Name Table Pointers and the array of Export Ordinals to translate a
procedure name string into an ordinal number by searching for a
matching name string. The ordinal number is used to locate the entry
point information in the export address table.
Import references by name require the Export Name Table Pointers
table to be binary searched to find the matching name, then the
corresponding Export Ordinal Table is known to contain the entry
point ordinal number. Import references by ordinal number provide
the fastest lookup since searching the name table is not required.
Each name table entry has the following format:
ASCII STRING ::: :::::::: '/0'
Figure 7. Export Name Table Entry
ASCII STRING = DB ASCII String.
The string is case sensitive and is terminated by a null byte.
6. Imports
A typical file layout for the import information follows:
DIRECTORY TABLE
NULL DIR ENTRY
DLL1 LOOKUP TABLE
NULL
DLL2 LOOKUP TABLE
NULL
DLL3 LOOKUP TABLE
NULL
HINT-NAME TABLE
DLL1 ADDRESS TABLE
NULL
DLL2 ADDRESS TABLE
NULL
DLL3 ADDRESS TABLE
NULL
Figure 8. Import File Layout
6.1 Import Directory Table
The import information begins with the Import Directory Table which
describes the remainder of the import information. The Import
Directory Table contains address information that is used to resolve
fixup references to the entry points within a DLL image. The import
directory table consists of an array of Import Directory Entries, one
entry for each DLL this image references. The last directory entry is
empty (NULL) which indicates the end of the directory table.
An Import Directory Entry has the following format:
IMPORT FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
NAME RVA
IMPORT LOOKUP TABLE RVA
IMPORT ADDRESS TABLE RVA
Figure 9. Import Directory Entry
IMPORT FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the import data was pre-snapped or
zero if not pre-snapped.
MAJOR/MINOR VERSION = DW The major/minor version number of the dll
being referenced.
NAME RVA = DD Relative Virtual Address of the Dll asciiz Name.
This is the address relative to the Image Base.
IMPORT LOOKUP TABLE RVA = DD This field contains the address of the
start of the import lookup table for this image. The address is
relative to the beginning of the Image Base.
IMPORT ADDRESS TABLE RVA = DD This field contains the address of the
start of the import addresses for this image. The address is
relative to the beginning of the Image Base.
6.2 Import Lookup Table
The Import Lookup Table is an array of ordinal or hint/name RVA's for
each DLL. The last entry is empty (NULL) which indicates the end of
the table.
The last element is empty.
3
0
1
0
ORDINAL
#/HINT-NAME
TABLE RVA
Figure 10. Import Address Table Format
ORDINAL/HINT-NAME TABLE RVA = 31-bits (mask = 7fffffffh) Ordinal
Number or Name Table RVA.
If the import is by ordinal, this field contains a 31 bit ordinal
number. If the import is by name, this field contains a 31 bit
address relative to the Image Base to the Hint-Name Table.
O = 1-bit (mask = 80000000h) Import by ordinal flag.
o 00000000h __Import by name.
o 80000000h __Import by ordinal.
6.3 Hint-Name Table
The Hint-Name Table format follows:
HINT
ASCII STRING |||||||||||||||||||| '/0'
PAD
The PAD field is optional.
Figure 11. Import Hint-Name Table
HINT = DW Hint into Export Name Table Pointers.
The hint value is used to index the Export Name Table Pointers array,
allowing faster by-name imports. If the hint is incorrect, then a
binary search is performed on the Export Name Ptr Table.
ASCII STRING = DB ASCII String.
The string is case sensitive and is terminated by a null byte.
PAD = DB Zero pad byte.
A trailing zero pad byte appears after the trailing null byte if
necessary to align the next entry on an even boundary.
The loader overwrites the import address table when loading the image
with the 32-bit address of the import.
6.4 Import Address Table
The Import Address Table is an array of addresses of the imported
routines for each DLL. The last entry is empty (NULL) which indicates
the end of the table.
7. Thread Local Storage
Thread local storage is a special contiguous block of data. Each
thread will gets its own block upon creation of the thread.
The file layout for thread local storage follows:
DIRECTORY TABLE
TLS DATA
INDEX VARIABLE
CALLBACK ADDRESSES
Figure 12. Thread Local Storage Layout
7.1 Thread Local Storage Directory Table
The Thread Local Storage Directory Table contains address information
that is used to describe the rest of TLS.
The Thread Local Storage Directory Table has the following format:
START DATA BLOCK VA
END DATA BLOCK VA
INDEX VA
CALLBACK TABLE VA
Figure 13. Thread Local Storage Directory Table
START DATA BLOCK VA = DD Virtual Address of the start of the thread
local storage data block.
END DATA BLOCK VA = DD Virtual Address of the end of the thread local
storage data block.
INDEX VA = DD Virtual Address of the index variable used to access
the thread local storage data block.
CALLBACK TABLE VA = DD Virtual Address of the callback table.
7.2 Thread Local Storage CallBack Table
The Thread Local Storage Callbacks is an array of Virtual Address of
functions to be called by the loader after thread creation and thread
termination. The last entry is empty (NULL) which indicates the end
of the table.
The Thread Local Storage CallBack Table has the following format:
FUNCTION1 VA
FUNCTION2 VA
NULL
Figure 14. Thread Local Storage CallBack Table
8. Resources
Resources are indexed by a multiple level binary-sorted tree
structure. The overall design can incorporate 2**31 levels, however,
NT uses only three: the highest is TYPE, then NAME, then LANGUAGE.
A typical file layout for the resource information follows:
RESOURCE DIRECTORY
RESOURCE DATA
Figure 15. Resource File Layout
The Resource directory is made up of the following tables:
8.1 Resource Directory Table
RESOURCE FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
# NAME ENTRY
# ID ENTRY
RESOURCE DIR ENTRIES
Figure 16. Resource Table Entry
RESOURCE FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the resource data was created by the
resource compiler.
MAJOR/MINOR VERSION = DW A user settable major/minor version number.
# NAME ENTRY = DW The number of name entries.
This field contains the number of entries at the beginning of the
array of directory entries which have actual string names associated
with them.
# ID ENTRY = DW The number of ID integer entries.
This field contains the number of 32-bit integer IDs as their names
in the array of directory entries.
The resource directory is followed by a variable length array of
directory entries. # NAME ENTRY is the number of entries at the
beginning of the array that have actual names associated with each
entry. The entires are in ascending order, case insensitive strings.
# ID ENTRY identifies the number of entries that have 32-bit integer
IDs as their name. These entries are also sorted in ascending order.
This structure allows fast lookup by either name or number, but for
any given resource entry only one form of lookup is supported, not
both. This is consistent with the syntax of the .RC file and the .RES
file.
The array of directory entries have the following format:
3
0
1
NAME RVA/INTEGER ID
E
DATA ENTRY RVA/SUBDIR RVA
Figure 17. Resource Directory Entry
INTERGER ID = DD ID.
This field contains a integer ID field to identify a resource.
NAME RVA = DD Name RVA address.
This field contains a 31-bit address relative to the beginning of the
Image Base to a Resource Directory String Entry.
E = 1-bit (mask 80000000h) Unescape bit.
This bit is zero for unescaped Resource Data Entries.
DATA RVA = 31-bits (mask 7fffffffh) Data entry address.
This field contains a 31-bit address relative to the beginning of the
Image Base to a Resource Data Entry.
E = 1-bit (mask 80000000h) Escape bit.
This bit is 1 for escaped Subdirectory Entry.
DATA RVA = 31-bits (mask 7fffffffh) Directory entries.
This field contains a 31-bit address relative to the beginning of the
Image Base to Subdirectory Entry.
Each resource directory string entry has the following format:
LENGTH UNICODE STRING
Figure 18. Resource Directory String Entry
LENGTH = DW Length of string.
UNICODE STRING = DW UNICODE String.
All of these string objects are stored together after the last
resource directory entry and before the first resource data object.
This minimizes the impact of these variable length objects on the
alignment of the fixed size directory entry objects. The length needs
to be word aligned.
Each Resource Data Entry has the following format:
DATA RVA
SIZE
CODEPAGE
RESERVED
Figure 19. Resource Data Entry
DATA RVA = DD Address of Resource Data.
This field contains 32-bit virtaul address of the resource data
(relative to the Image Base).
SIZE = DD Size of Resource Data.
This field contains the size of the resource data for this resource.
CODEPAGE = DD Codepage.
RESERVED = DD Reserved - must be zero.
Each resource data entry describes a leaf node in the resource
directory tree. It contains an address which is relative to the
beginning of Image Base, a size field that gives the number of bytes
of data at that address, a CodePage that should be used when decoding
code point values within the resource data. Typically for new
applications the code page would be the unicode code page.
8.2 Resource Example
The following is an example for an app. which wants to use the following
data
as resources:
TypeId# NameId# Language ID Resource Data
00000001 00000001 0 00010001
00000001 00000001 1 10010001
00000001 00000002 0 00010002
00000001 00000003 0 00010003
00000002 00000001 0 00020001
00000002 00000002 0 00020002
00000002 00000003 0 00020003
00000002 00000004 0 00020004
00000009 00000001 0 00090001
00000009 00000009 0 00090009
00000009 00000009 1 10090009
00000009 00000009 2 20090009
Then the Resource Directory in the Portable format looks like:
Offset Data
0000: 00000000 00000000 00000000 00030000 (3 entries in this directory)
0010: 00000001 80000028 (TypeId #1, Subdirectory at offset 0x28)
0018: 00000002 80000050 (TypeId #2, Subdirectory at offset 0x50)
0020: 00000009 80000080 (TypeId #9, Subdirectory at offset 0x80)
0028: 00000000 00000000 00000000 00030000 (3 entries in this directory)
0038: 00000001 800000A0 (NameId #1, Subdirectory at offset 0xA0)
0040: 00000002 00000108 (NameId #2, data desc at offset 0x108)
0048: 00000003 00000118 (NameId #3, data desc at offset 0x118)
0050: 00000000 00000000 00000000 00040000 (4 entries in this directory)
0060: 00000001 00000128 (NameId #1, data desc at offset 0x128)
0068: 00000002 00000138 (NameId #2, data desc at offset 0x138)
0070: 00000003 00000148 (NameId #3, data desc at offset 0x148)
0078: 00000004 00000158 (NameId #4, data desc at offset 0x158)
0080: 00000000 00000000 00000000 00020000 (2 entries in this directory)
0090: 00000001 00000168 (NameId #1, data desc at offset 0x168)
0098: 00000009 800000C0 (NameId #9, Subdirectory at offset 0xC0)
00A0: 00000000 00000000 00000000 00020000 (2 entries in this directory)
00B0: 00000000 000000E8 (Language ID 0, data desc at offset 0xE8
00B8: 00000001 000000F8 (Language ID 1, data desc at offset 0xF8
00C0: 00000000 00000000 00000000 00030000 (3 entries in this directory)
00D0: 00000001 00000178 (Language ID 0, data desc at offset 0x178
00D8: 00000001 00000188 (Language ID 1, data desc at offset 0x188
00E0: 00000001 00000198 (Language ID 2, data desc at offset 0x198
00E8: 000001A8 (At offset 0x1A8, for TypeId #1, NameId #1, Language id #0
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
00F8: 000001AC (At offset 0x1AC, for TypeId #1, NameId #1, Language id #1
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0108: 000001B0 (At offset 0x1B0, for TypeId #1, NameId #2,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0118: 000001B4 (At offset 0x1B4, for TypeId #1, NameId #3,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0128: 000001B8 (At offset 0x1B8, for TypeId #2, NameId #1,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0138: 000001BC (At offset 0x1BC, for TypeId #2, NameId #2,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0148: 000001C0 (At offset 0x1C0, for TypeId #2, NameId #3,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0158: 000001C4 (At offset 0x1C4, for TypeId #2, NameId #4,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0168: 000001C8 (At offset 0x1C8, for TypeId #9, NameId #1,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0178: 000001CC (At offset 0x1CC, for TypeId #9, NameId #9, Language id #0
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0188: 000001D0 (At offset 0x1D0, for TypeId #9, NameId #9, Language id #1
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0198: 000001D4 (At offset 0x1D4, for TypeId #9, NameId #9, Language id #2
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
And the data for the resources will look like:
01A8: 00010001
01AC: 10010001
01B0: 00010002
01B4: 00010003
01B8: 00020001
01BC: 00020002
01C0: 00020003
01C4: 00020004
01C8: 00090001
01CC: 00090009
01D0: 10090009
01D4: 20090009
9. Fixup Table
The Fixup Table contains entries for all fixups in the image. The
Total Fixup Data Size in the PE Header is the number of bytes in the
fixup table. The fixup table is broken into blocks of fixups. Each
block represents the fixups for a 4K page.
Fixups that are resolved by the linker do not need to be processed by
the loader, unless the load image can't be loaded at the Image Base
specified in the PE Header.
9.1 Fixup Block
Fixup blocks have the following format:
PAGE RVA
BLOCK SIZE
TYPE/OFFSET
TYPE/OFFSET
TYPE/OFFSET
Figure 20. Fixup Block Format
To apply a fixup, a delta needs to be calculated. The 32-bit delta
is the difference between the preferred base, and the base where the
image is actually loaded. If the image is loaded at its preferred
base, the delta would be zero, and thus the fixups would not have to
be applied. Each block must start on a DWORD boundary. The ABSOLUTE
fixup type can be used to pad a block.
PAGE RVA = DD Page RVA. The image base plus the page rva is added to
each offset to create the virtual address of where the fixup needs to
be applied.
BLOCK SIZE = DD Number of bytes in the fixup block. This includes the
PAGE RVA and SIZE fields.
TYPE/OFFSET is defined as:
1
1
0
5
1
TYPE
OFFSET
Figure 21. Fixup Record Format
TYPE = 4-bit fixup type. This value has the following definitions:
o 0h __ABSOLUTE. This is a NOP. The fixup is skipped.
o 1h __HIGH. Add the high 16-bits of the delta to the 16-bit field
at Offset. The 16-bit field represents the high value of a 32-
bit word.
o 2h __LOW. Add the low 16-bits of the delta to the 16-bit field
at Offset. The 16-bit field represents the low half value of a
32-bit word. This fixup will only be emitted for a RISC machine
when the image Object Align isn't the default of 64K.
o 3h __HIGHLOW. Apply the 32-bit delta to the 32-bit field at
Offset.
o 4h __HIGHADJUST. This fixup requires a full 32-bit value. The
high 16-bits is located at Offset, and the low 16-bits is
located in the next Offset array element (this array element is
included in the SIZE field). The two need to be combined into a
signed variable. Add the 32-bit delta. Then add 0x8000 and
store the high 16-bits of the signed variable to the 16-bit
field at Offset.
o 5h __MIPSJMPADDR.
All other values are reserved.
10. Debug Information
The debug information is defined by the debugger and is not
controlled by the portable EXE format or linker. The only data
defined by the portable EXE format is the Debug Directory Table.
10.1 Debug Directory
The debug directory table consists of one or more entries that have
the following format:
DEBUG FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
DEBUG TYPE
DATA SIZE
DATA RVA
DATA SEEK
Figure 22. Debug Directory Entry
DEBUG FLAGS = DD Set to zero for now.
TIME/DATE STAMP = DD Time/Date the debug data was created.
MAJOR/MINOR VERSION = DW Version stamp.
This stamp can be used to determine the version of the debug data.
DEBUG TYPE = DD Format type.
To support multiple debuggers, this field determines the format of
the debug information. This value has the following definitions:
o 0001h __Image contains COFF symbolics.
o 0001h __Image contains CodeView symbolics.
o 0001h __Image contains FPO symbolics.
DATA SIZE = DD The number of bytes in the debug data. This is the
size of the actual debug data and does not include the debug
directory.
DATA RVA = DD The relative virtual address of the debug data. This
address is relative to the beginning of the Image Base.
DATA SEEK = DD The seek value from the beginning of the file to the
debug data.
If the image contains more than one type of debug information, then
the next debug directory will immediately follow the first debug
directory.
Author: Micheal J. O'Leary
Preface
This document was edited and released by Microsoft Developer
Support. It describes the binary portable executable format for NT.
The information is provided at this point because we feel it will
make the work of application development easier. Unfortunately, the
information in this document may change before the final release of
Windows NT. Microsoft is NOT committing to stay with these formats
by releasing this document. Questions or follow-ups for any of the
information presented here should be posted to CompuServe MSWIN32
forum, section 6.
--Steve Firebaugh
Microsoft Developer Support
Contents
1. Overview
2. PE Header
3. Object Table
4. Image Pages
5. Exports
5.1 Export Directory Table
5.2 Export Address Table
5.3 Export Name Table Pointers
5.4 Export Ordinal Table
5.5 Export Name Table
6. Imports
6.1 Import Directory Table
6.2 Import Lookup Table
6.3 Hint-Name Table
6.4 Import Address Table
7. Thread Local Storage
7.1 Thread Local Storage Directory Table
7.2 Thread Local Storage CallBack Table
8. Resources
8.1 Resource Directory Table
8.2 Resource Example
9. Fixup Table
9.1 Fixup Block
10. Debug Information
10.1 Debug Directory
1. Overview
< < Base of Image Header
DOS 2 Compatible
EXE Header
unused
OEM Identifier
OEM Info
DOS 2.0 Section
Offset to (for DOS compatibility only)
PE Header
DOS 2.0 Stub
Program &
Reloc. Table
unused
Aligned on 8 byte boundary
PE Header
Object Table
Image Pages
import info
export info
fixup info
resource info
debug info
Figure
1. A typical 32-bit Portable EXE File Layout
2. PE Header
SIGNATURE BYTES
CPU TYPE
# OBJECTS
TIME/DATE STAMP
RESERVED
RESERVED
NT HDR SIZE
FLAGS
RESERVED
LMAJORLMINOR
RESERVED
RESERVED
RESERVED
ENTRYPOINT RVA
RESERVED
RESERVED
IMAGE BASE
OBJECT ALIGN
FILE ALIGN
OS MAJOR
OS MINOR
USER MAJOR
USER MINOR
SUBSYS MAJOR
SUBSYS MINOR
RESERVED
IMAGE SIZE
HEADER SIZE
FILE CHECKSUM
SUBSYSTEM
DLL FLAGS
STACK RESERVE SIZE
STACK COMMIT SIZE
HEAP RESERVE SIZE
HEAP COMMIT SIZE
RESERVED
# INTERESTING RVA/SIZES
EXPORT TABLE RVA
TOTAL EXPORT DATA SIZE
IMPORT TABLE RVA
TOTAL IMPORT DATA SIZE
RESOURCE TABLE RVA
TOTAL RESOURCE DATA SIZE
EXCEPTION TABLE RVA
TOTAL EXCEPTION DATA SIZE
SECURITY TABLE RVA
TOTAL SECURITY DATA SIZE
FIXUP TABLE RVA
TOTAL FIXUP DATA SIZE
DEBUG TABLE RVA
TOTAL DEBUG DIRECTORIES
IMAGE DESCRIPTION RVA
TOTAL DESCRIPTION SIZE
MACHINE SPECIFIC RVA
MACHINE SPECIFIC SIZE
THREAD LOCAL STORAGE RVA
TOTAL TLS SIZE
Figure 2. PE Header
Notes:
o A VA is a virtual address that is already biased by the Image
Base found in the PE Header. A RVA is a virtual address that is
relative to the Image Base.
o An RVA in the PE Header which has a value of zero indicates the
field isn't used.
o Image pages are aligned and zero padded to a File Align
boundary. The bases of all other tables and structures must be
aligned on DWORD (4 byte) boundary. Thus, all VA's and RVA's
must be on a 32 bit boundary. All table and structure fields
must be aligned on their "natural" boundaries, with the possible
exception of the Debug Info.
SIGNATURE BYTES = DB * 4.
Current value is "PE/0/0". Thats PE followed by two zeros (nulls).
CPU TYPE = DW CPU Type.
This field specifies the type of CPU compatibility required by this
image to run. The values are:
o 0000h __unknown
o 014Ch __80386
o 014Dh __80486
o 014Eh __80586
o 0162h __MIPS Mark I (R2000, R3000)
o 0163h __MIPS Mark II (R6000)
o 0166h __MIPS Mark III (R4000)
# OBJECTS = DW Number of object entries.
This field specifies the number of entries in the Object Table.
TIME/DATE STAMP = DD Used to store the time and date the file was
created or modified by the linker.
NT HDR SIZE = DW This is the number of remaining bytes in the NT
header that follow the FLAGS field.
FLAGS = DW Flag bits for the image.
The flag bits have the following definitons:
o 0000h __Program image.
o 0002h __Image is executable.
If this bit isn't set, then it indicates that either errors
where detected at link time or that the image is being
incrementally linked and therefore can't be loaded.
o 0200h __Fixed.
Indicates that if the image can't be loaded at the Image Base,
then don't load it.
o 2000h __Library image.
LMAJOR/LMINOR = DB Linker major/minor version number.
ENTRYPOINT RVA = DD Entrypoint relative virtual address.
The address is relative to the Image Base. The address is the
starting address for program images and the library initialization
and library termination address for library images.
IMAGE BASE = DD The virtual base of the image.
This will be the virtual address of the first byte of the file (Dos
Header). This must be a multiple of 64K.
OBJECT ALIGN = DD The alignment of the objects. This must be a power
of 2 between 512 and 256M inclusive. The default is 64K.
FILE ALIGN = DD Alignment factor used to align image pages. The
alignment factor (in bytes) used to align the base of the image pages
and to determine the granularity of per-object trailing zero pad.
Larger alignment factors will cost more file space; smaller alignment
factors will impact demand load performance, perhaps significantly.
Of the two, wasting file space is preferable. This value should be a
power of 2 between 512 and 64K inclusive.
OS MAJOR/MINOR = DW OS version number required to run this image.
USER MAJOR/MINOR # = DW User major/minor version number.
This is useful for differentiating between revisions of
images/dynamic linked libraries. The values are specified at link
time by the user.
SUBSYS MAJOR/MINOR # = DW Subsystem major/minor version number.
IMAGE SIZE = DD The virtual size (in bytes) of the image.
This includes all headers. The total image size must be a multiple
of Object Align.
HEADER SIZE = DD Total header size.
The combined size of the Dos Header, PE Header and Object Table.
FILE CHECKSUM = DD Checksum for entire file. Set to 0 by the linker.
SUBSYSTEM = DW NT Subsystem required to run this image.
The values are:
o 0000h __Unknown
o 0001h __Native
o 0002h __Windows GUI
o 0003h __Windows Character
o 0005h __OS/2 Character
o 0007h __Posix Character
DLL FLAGS = DW Indicates special loader requirements.
This flag has the following bit values:
o 0001h __Per-Process Library Initialization.
o 0002h __Per-Process Library Termination.
o 0004h __Per-Thread Library Initialization.
o 0008h __Per-Thread Library Termination.
All other bits are reserved for future use and should be set to zero.
STACK RESERVE SIZE = DD Stack size needed for image.
The memory is reserved, but only the STACK COMMIT SIZE is committed.
The next page of the stack is a 'guarded page'. When the application
hits the guarded page, the guarded page becomes valid, and the next
page becomes the guarded page. This continues until the RESERVE SIZE
is reached.
STACK COMMIT SIZE = DD Stack commit size.
HEAP RESERVE SIZE = DD Size of local heap to reserve.
HEAP COMMIT SIZE = DD Amount to commit in local heap.
# INTERESTING VA/SIZES = DD Indicates the size of the VA/SIZE array
that follows.
EXPORT TABLE RVA = DD Relative Virtual Address of the Export Table.
This address is relative to the Image Base.
IMPORT TABLE RVA = DD Relative Virtual Address of the Import Table.
This address is relative to the Image Base.
RESOURCE TABLE RVA = DD Relative Virtual Address of the Resource
Table. This address is relative to the Image Base.
EXCEPTION TABLE RVA = DD Relative Virtual Address of the Exception
Table. This address is relative to the Image Base.
SECURITY TABLE RVA = DD Relative Virtual Address of the Security
Table. This address is relative to the Image Base.
FIXUP TABLE RVA = DD Relative Virtual Address of the Fixup Table.
This address is relative to the Image Base.
DEBUG TABLE RVA = DD Relative Virtual Address of the Debug Table.
This address is relative to the Image Base.
IMAGE DESCRIPTION RVA = DD Relative Virtual Address of the
description string specified in the module definiton file.
MACHINE SPECIFIC RVA = DD Relative Virtual Address of a machine
specific value. This address is relative to the Image Base.
TOTAL EXPORT DATA SIZE = DD Total size of the export data.
TOTAL IMPORT DATA SIZE = DD Total size of the import data.
TOTAL RESOURCE DATA SIZE = DD Total size of the resource data.
TOTAL EXCEPTION DATA SIZE = DD Total size of the exception data.
TOTAL SECURITY DATA SIZE = DD Total size of the security data.
TOTAL FIXUP DATA SIZE = DD Total size of the fixup data.
TOTAL DEBUG DIRECTORIES = DD Total number of debug directories.
TOTAL DESCRIPTION SIZE = DD Total size of the description data.
MACHINE SPECIFIC SIZE = DD A machine specific value.
3. Object Table
The number of entries in the Object Table is given by the # Objects
field in the PE Header. Entries in the Object Table are numbered
starting from one. The object table immediately follows the PE
Header. The code and data memory object entries are in the order
chosen by the linker. The virtual addresses for objects must be
assigned by the linker such that they are in ascending order and
adjacent, and must be a multiple of Object Align in the PE header.
Each Object Table entry has the following format:
OBJECT NAME
VIRTUAL SIZE
RVA
PHYSICAL SIZE
PHYSICAL OFFSET
RESERVED
RESERVED
RESERVED
OBJECT FLAGS
Figure 3. Object Table
OBJECT NAME = DB * 8 Object name. This is an eight-byte null-padded
ASCII string representing the object name.
VIRTUAL SIZE = DD Virtual memory size. The size of the object that
will be allocated when the object is loaded. Any difference between
PHYSICAL SIZE and VIRTUAL SIZE is zero filled.
RVA = DD Relative Virtual Address. The virtual address the object is
currently relocated to, relative to the Image Base. Each Object's
virtual address space consumes a multiple of Object Align (power of 2
between 512 and 256M inclusive. Default is 64K), and immediately
follows the previous Object in the virtual address space (the virtual
address space for a image must be dense).
PHYSICAL SIZE = DD Physical file size of initialized data. The size
of the initialized data in the file for the object. The physical
size must be a multiple of the File Align field in the PE Header, and
must be less than or equal to the Virtual Size.
PHYSICAL OFFSET = DD Physical offset for object's first page. This
offset is relative to beginning of the EXE file, and is aligned on a
multiple of the File Align field in the PE Header. The offset is
used as a seek value.
OBJECT FLAGS = DD Flag bits for the object. The object flag bits
have the following definitions:
o 000000020h __Code object.
o 000000040h __Initialized data object.
o 000000080h __Uninitialized data object.
o 040000000h __Object must not be cached.
o 080000000h __Object is not pageable.
o 100000000h __Object is shared.
o 200000000h __Executable object.
o 400000000h __Readable object.
o 800000000h __Writeable object.
All other bits are reserved for future use and should be set to zero.
4. Image Pages
The Image Pages section contains all initialized data for all
objects. The seek offset for the first page in each object is
specified in the object table and is aligned on a File Align
boundary. The objects are ordered by the RVA. Every object begins
on a multiple of Object Align.
5. Exports
A typical file layout for the export information follows:
DIRECTORY TABLE
ADDRESS TABLE
NAME PTR TABLE
ORDINAL TABLE
NAME STRINGS
Figure 4. Export File Layout
5.1 Export Directory Table
The export information begins with the Export Directory Table which
describes the remainder of the export information. The Export
Directory Table contains address information that is used to resolve
fixup references to the entry points within this image.
EXPORT FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
NAME RVA
ORDINAL BASE
# EAT ENTRIES
# NAME PTRS
ADDRESS TABLE RVA
NAME PTR TABLE RVA
ORDINAL TABLE RVA
Figure 5. Export Directory Table Entry
EXPORT FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the export data was created.
MAJOR/MINOR VERSION = DW A user settable major/minor version number.
NAME RVA = DD Relative Virtual Address of the Dll asciiz Name.
This is the address relative to the Image Base.
ORDINAL BASE = DD First valid exported ordinal.
This field specifies the starting ordinal number for the export
address table for this image. Normally set to 1.
# EAT ENTRIES = DD Indicates number of entries in the Export Address
Table.
# NAME PTRS = DD This indicates the number of entries in the Name Ptr
Table (and parallel Ordinal Table).
ADDRESS TABLE RVA = DD Relative Virtual Address of the Export Address
Table.
This address is relative to the Image Base.
NAME TABLE RVA = DD Relative Virtual Address of the Export Name Table
Pointers.
This address is relative to the beginning of the Image Base. This
table is an array of RVA's with # NAMES entries.
ORDINAL TABLE RVA = DD Relative Virtual Address of Export Ordinals
Table Entry.
This address is relative to the beginning of the Image Base.
5.2 Export Address Table
The Export Address Table contains the address of exported entrypoints
and exported data and absolutes. An ordinal number is used to index
the Export Address Table. The ORDINAL BASE must be subracted from the
ordinal number before indexing into this table.
Export Address Table entry formats are described below:
EXPORTED RVA
Figure 6. Export Address Table Entry
EXPORTED RVA = DD Export address.
This field contains the relative virtual address of the exported
entry (relative to the Image Base).
5.3 Export Name Table Pointers
The export name table pointers array contains address into the Export
Name Table. The pointers are 32-bits each, and are relative to the
Image Base. The pointers are ordered lexically to allow binary
searches.
5.4 Export Ordinal Table
The Export Name Table Pointers and the Export Ordinal Table form two
parallel arrays, separated to allow natural field alignment. The
export ordinal table array contains the Export Address Table ordinal
numbers associated with the named export referenced by corresponding
Export Name Table Pointers.
The ordinals are 16-bits each, and already include the Ordinal Base
stored in the Export Directory Table.
5.5 Export Name Table
The export name table contains optional ASCII names for exported
entries in the image. These tables are used with the array of Export
Name Table Pointers and the array of Export Ordinals to translate a
procedure name string into an ordinal number by searching for a
matching name string. The ordinal number is used to locate the entry
point information in the export address table.
Import references by name require the Export Name Table Pointers
table to be binary searched to find the matching name, then the
corresponding Export Ordinal Table is known to contain the entry
point ordinal number. Import references by ordinal number provide
the fastest lookup since searching the name table is not required.
Each name table entry has the following format:
ASCII STRING ::: :::::::: '/0'
Figure 7. Export Name Table Entry
ASCII STRING = DB ASCII String.
The string is case sensitive and is terminated by a null byte.
6. Imports
A typical file layout for the import information follows:
DIRECTORY TABLE
NULL DIR ENTRY
DLL1 LOOKUP TABLE
NULL
DLL2 LOOKUP TABLE
NULL
DLL3 LOOKUP TABLE
NULL
HINT-NAME TABLE
DLL1 ADDRESS TABLE
NULL
DLL2 ADDRESS TABLE
NULL
DLL3 ADDRESS TABLE
NULL
Figure 8. Import File Layout
6.1 Import Directory Table
The import information begins with the Import Directory Table which
describes the remainder of the import information. The Import
Directory Table contains address information that is used to resolve
fixup references to the entry points within a DLL image. The import
directory table consists of an array of Import Directory Entries, one
entry for each DLL this image references. The last directory entry is
empty (NULL) which indicates the end of the directory table.
An Import Directory Entry has the following format:
IMPORT FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
NAME RVA
IMPORT LOOKUP TABLE RVA
IMPORT ADDRESS TABLE RVA
Figure 9. Import Directory Entry
IMPORT FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the import data was pre-snapped or
zero if not pre-snapped.
MAJOR/MINOR VERSION = DW The major/minor version number of the dll
being referenced.
NAME RVA = DD Relative Virtual Address of the Dll asciiz Name.
This is the address relative to the Image Base.
IMPORT LOOKUP TABLE RVA = DD This field contains the address of the
start of the import lookup table for this image. The address is
relative to the beginning of the Image Base.
IMPORT ADDRESS TABLE RVA = DD This field contains the address of the
start of the import addresses for this image. The address is
relative to the beginning of the Image Base.
6.2 Import Lookup Table
The Import Lookup Table is an array of ordinal or hint/name RVA's for
each DLL. The last entry is empty (NULL) which indicates the end of
the table.
The last element is empty.
3
0
1
0
ORDINAL
#/HINT-NAME
TABLE RVA
Figure 10. Import Address Table Format
ORDINAL/HINT-NAME TABLE RVA = 31-bits (mask = 7fffffffh) Ordinal
Number or Name Table RVA.
If the import is by ordinal, this field contains a 31 bit ordinal
number. If the import is by name, this field contains a 31 bit
address relative to the Image Base to the Hint-Name Table.
O = 1-bit (mask = 80000000h) Import by ordinal flag.
o 00000000h __Import by name.
o 80000000h __Import by ordinal.
6.3 Hint-Name Table
The Hint-Name Table format follows:
HINT
ASCII STRING |||||||||||||||||||| '/0'
PAD
The PAD field is optional.
Figure 11. Import Hint-Name Table
HINT = DW Hint into Export Name Table Pointers.
The hint value is used to index the Export Name Table Pointers array,
allowing faster by-name imports. If the hint is incorrect, then a
binary search is performed on the Export Name Ptr Table.
ASCII STRING = DB ASCII String.
The string is case sensitive and is terminated by a null byte.
PAD = DB Zero pad byte.
A trailing zero pad byte appears after the trailing null byte if
necessary to align the next entry on an even boundary.
The loader overwrites the import address table when loading the image
with the 32-bit address of the import.
6.4 Import Address Table
The Import Address Table is an array of addresses of the imported
routines for each DLL. The last entry is empty (NULL) which indicates
the end of the table.
7. Thread Local Storage
Thread local storage is a special contiguous block of data. Each
thread will gets its own block upon creation of the thread.
The file layout for thread local storage follows:
DIRECTORY TABLE
TLS DATA
INDEX VARIABLE
CALLBACK ADDRESSES
Figure 12. Thread Local Storage Layout
7.1 Thread Local Storage Directory Table
The Thread Local Storage Directory Table contains address information
that is used to describe the rest of TLS.
The Thread Local Storage Directory Table has the following format:
START DATA BLOCK VA
END DATA BLOCK VA
INDEX VA
CALLBACK TABLE VA
Figure 13. Thread Local Storage Directory Table
START DATA BLOCK VA = DD Virtual Address of the start of the thread
local storage data block.
END DATA BLOCK VA = DD Virtual Address of the end of the thread local
storage data block.
INDEX VA = DD Virtual Address of the index variable used to access
the thread local storage data block.
CALLBACK TABLE VA = DD Virtual Address of the callback table.
7.2 Thread Local Storage CallBack Table
The Thread Local Storage Callbacks is an array of Virtual Address of
functions to be called by the loader after thread creation and thread
termination. The last entry is empty (NULL) which indicates the end
of the table.
The Thread Local Storage CallBack Table has the following format:
FUNCTION1 VA
FUNCTION2 VA
NULL
Figure 14. Thread Local Storage CallBack Table
8. Resources
Resources are indexed by a multiple level binary-sorted tree
structure. The overall design can incorporate 2**31 levels, however,
NT uses only three: the highest is TYPE, then NAME, then LANGUAGE.
A typical file layout for the resource information follows:
RESOURCE DIRECTORY
RESOURCE DATA
Figure 15. Resource File Layout
The Resource directory is made up of the following tables:
8.1 Resource Directory Table
RESOURCE FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
# NAME ENTRY
# ID ENTRY
RESOURCE DIR ENTRIES
Figure 16. Resource Table Entry
RESOURCE FLAGS = DD Currently set to zero.
TIME/DATE STAMP = DD Time/Date the resource data was created by the
resource compiler.
MAJOR/MINOR VERSION = DW A user settable major/minor version number.
# NAME ENTRY = DW The number of name entries.
This field contains the number of entries at the beginning of the
array of directory entries which have actual string names associated
with them.
# ID ENTRY = DW The number of ID integer entries.
This field contains the number of 32-bit integer IDs as their names
in the array of directory entries.
The resource directory is followed by a variable length array of
directory entries. # NAME ENTRY is the number of entries at the
beginning of the array that have actual names associated with each
entry. The entires are in ascending order, case insensitive strings.
# ID ENTRY identifies the number of entries that have 32-bit integer
IDs as their name. These entries are also sorted in ascending order.
This structure allows fast lookup by either name or number, but for
any given resource entry only one form of lookup is supported, not
both. This is consistent with the syntax of the .RC file and the .RES
file.
The array of directory entries have the following format:
3
0
1
NAME RVA/INTEGER ID
E
DATA ENTRY RVA/SUBDIR RVA
Figure 17. Resource Directory Entry
INTERGER ID = DD ID.
This field contains a integer ID field to identify a resource.
NAME RVA = DD Name RVA address.
This field contains a 31-bit address relative to the beginning of the
Image Base to a Resource Directory String Entry.
E = 1-bit (mask 80000000h) Unescape bit.
This bit is zero for unescaped Resource Data Entries.
DATA RVA = 31-bits (mask 7fffffffh) Data entry address.
This field contains a 31-bit address relative to the beginning of the
Image Base to a Resource Data Entry.
E = 1-bit (mask 80000000h) Escape bit.
This bit is 1 for escaped Subdirectory Entry.
DATA RVA = 31-bits (mask 7fffffffh) Directory entries.
This field contains a 31-bit address relative to the beginning of the
Image Base to Subdirectory Entry.
Each resource directory string entry has the following format:
LENGTH UNICODE STRING
Figure 18. Resource Directory String Entry
LENGTH = DW Length of string.
UNICODE STRING = DW UNICODE String.
All of these string objects are stored together after the last
resource directory entry and before the first resource data object.
This minimizes the impact of these variable length objects on the
alignment of the fixed size directory entry objects. The length needs
to be word aligned.
Each Resource Data Entry has the following format:
DATA RVA
SIZE
CODEPAGE
RESERVED
Figure 19. Resource Data Entry
DATA RVA = DD Address of Resource Data.
This field contains 32-bit virtaul address of the resource data
(relative to the Image Base).
SIZE = DD Size of Resource Data.
This field contains the size of the resource data for this resource.
CODEPAGE = DD Codepage.
RESERVED = DD Reserved - must be zero.
Each resource data entry describes a leaf node in the resource
directory tree. It contains an address which is relative to the
beginning of Image Base, a size field that gives the number of bytes
of data at that address, a CodePage that should be used when decoding
code point values within the resource data. Typically for new
applications the code page would be the unicode code page.
8.2 Resource Example
The following is an example for an app. which wants to use the following
data
as resources:
TypeId# NameId# Language ID Resource Data
00000001 00000001 0 00010001
00000001 00000001 1 10010001
00000001 00000002 0 00010002
00000001 00000003 0 00010003
00000002 00000001 0 00020001
00000002 00000002 0 00020002
00000002 00000003 0 00020003
00000002 00000004 0 00020004
00000009 00000001 0 00090001
00000009 00000009 0 00090009
00000009 00000009 1 10090009
00000009 00000009 2 20090009
Then the Resource Directory in the Portable format looks like:
Offset Data
0000: 00000000 00000000 00000000 00030000 (3 entries in this directory)
0010: 00000001 80000028 (TypeId #1, Subdirectory at offset 0x28)
0018: 00000002 80000050 (TypeId #2, Subdirectory at offset 0x50)
0020: 00000009 80000080 (TypeId #9, Subdirectory at offset 0x80)
0028: 00000000 00000000 00000000 00030000 (3 entries in this directory)
0038: 00000001 800000A0 (NameId #1, Subdirectory at offset 0xA0)
0040: 00000002 00000108 (NameId #2, data desc at offset 0x108)
0048: 00000003 00000118 (NameId #3, data desc at offset 0x118)
0050: 00000000 00000000 00000000 00040000 (4 entries in this directory)
0060: 00000001 00000128 (NameId #1, data desc at offset 0x128)
0068: 00000002 00000138 (NameId #2, data desc at offset 0x138)
0070: 00000003 00000148 (NameId #3, data desc at offset 0x148)
0078: 00000004 00000158 (NameId #4, data desc at offset 0x158)
0080: 00000000 00000000 00000000 00020000 (2 entries in this directory)
0090: 00000001 00000168 (NameId #1, data desc at offset 0x168)
0098: 00000009 800000C0 (NameId #9, Subdirectory at offset 0xC0)
00A0: 00000000 00000000 00000000 00020000 (2 entries in this directory)
00B0: 00000000 000000E8 (Language ID 0, data desc at offset 0xE8
00B8: 00000001 000000F8 (Language ID 1, data desc at offset 0xF8
00C0: 00000000 00000000 00000000 00030000 (3 entries in this directory)
00D0: 00000001 00000178 (Language ID 0, data desc at offset 0x178
00D8: 00000001 00000188 (Language ID 1, data desc at offset 0x188
00E0: 00000001 00000198 (Language ID 2, data desc at offset 0x198
00E8: 000001A8 (At offset 0x1A8, for TypeId #1, NameId #1, Language id #0
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
00F8: 000001AC (At offset 0x1AC, for TypeId #1, NameId #1, Language id #1
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0108: 000001B0 (At offset 0x1B0, for TypeId #1, NameId #2,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0118: 000001B4 (At offset 0x1B4, for TypeId #1, NameId #3,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0128: 000001B8 (At offset 0x1B8, for TypeId #2, NameId #1,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0138: 000001BC (At offset 0x1BC, for TypeId #2, NameId #2,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0148: 000001C0 (At offset 0x1C0, for TypeId #2, NameId #3,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0158: 000001C4 (At offset 0x1C4, for TypeId #2, NameId #4,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0168: 000001C8 (At offset 0x1C8, for TypeId #9, NameId #1,
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0178: 000001CC (At offset 0x1CC, for TypeId #9, NameId #9, Language id #0
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0188: 000001D0 (At offset 0x1D0, for TypeId #9, NameId #9, Language id #1
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
0198: 000001D4 (At offset 0x1D4, for TypeId #9, NameId #9, Language id #2
00000004 (4 bytes of data)
00000000 (codepage)
00000000 (reserved)
And the data for the resources will look like:
01A8: 00010001
01AC: 10010001
01B0: 00010002
01B4: 00010003
01B8: 00020001
01BC: 00020002
01C0: 00020003
01C4: 00020004
01C8: 00090001
01CC: 00090009
01D0: 10090009
01D4: 20090009
9. Fixup Table
The Fixup Table contains entries for all fixups in the image. The
Total Fixup Data Size in the PE Header is the number of bytes in the
fixup table. The fixup table is broken into blocks of fixups. Each
block represents the fixups for a 4K page.
Fixups that are resolved by the linker do not need to be processed by
the loader, unless the load image can't be loaded at the Image Base
specified in the PE Header.
9.1 Fixup Block
Fixup blocks have the following format:
PAGE RVA
BLOCK SIZE
TYPE/OFFSET
TYPE/OFFSET
TYPE/OFFSET
Figure 20. Fixup Block Format
To apply a fixup, a delta needs to be calculated. The 32-bit delta
is the difference between the preferred base, and the base where the
image is actually loaded. If the image is loaded at its preferred
base, the delta would be zero, and thus the fixups would not have to
be applied. Each block must start on a DWORD boundary. The ABSOLUTE
fixup type can be used to pad a block.
PAGE RVA = DD Page RVA. The image base plus the page rva is added to
each offset to create the virtual address of where the fixup needs to
be applied.
BLOCK SIZE = DD Number of bytes in the fixup block. This includes the
PAGE RVA and SIZE fields.
TYPE/OFFSET is defined as:
1
1
0
5
1
TYPE
OFFSET
Figure 21. Fixup Record Format
TYPE = 4-bit fixup type. This value has the following definitions:
o 0h __ABSOLUTE. This is a NOP. The fixup is skipped.
o 1h __HIGH. Add the high 16-bits of the delta to the 16-bit field
at Offset. The 16-bit field represents the high value of a 32-
bit word.
o 2h __LOW. Add the low 16-bits of the delta to the 16-bit field
at Offset. The 16-bit field represents the low half value of a
32-bit word. This fixup will only be emitted for a RISC machine
when the image Object Align isn't the default of 64K.
o 3h __HIGHLOW. Apply the 32-bit delta to the 32-bit field at
Offset.
o 4h __HIGHADJUST. This fixup requires a full 32-bit value. The
high 16-bits is located at Offset, and the low 16-bits is
located in the next Offset array element (this array element is
included in the SIZE field). The two need to be combined into a
signed variable. Add the 32-bit delta. Then add 0x8000 and
store the high 16-bits of the signed variable to the 16-bit
field at Offset.
o 5h __MIPSJMPADDR.
All other values are reserved.
10. Debug Information
The debug information is defined by the debugger and is not
controlled by the portable EXE format or linker. The only data
defined by the portable EXE format is the Debug Directory Table.
10.1 Debug Directory
The debug directory table consists of one or more entries that have
the following format:
DEBUG FLAGS
TIME/DATE STAMP
MAJOR VERSION
MINOR VERSION
DEBUG TYPE
DATA SIZE
DATA RVA
DATA SEEK
Figure 22. Debug Directory Entry
DEBUG FLAGS = DD Set to zero for now.
TIME/DATE STAMP = DD Time/Date the debug data was created.
MAJOR/MINOR VERSION = DW Version stamp.
This stamp can be used to determine the version of the debug data.
DEBUG TYPE = DD Format type.
To support multiple debuggers, this field determines the format of
the debug information. This value has the following definitions:
o 0001h __Image contains COFF symbolics.
o 0001h __Image contains CodeView symbolics.
o 0001h __Image contains FPO symbolics.
DATA SIZE = DD The number of bytes in the debug data. This is the
size of the actual debug data and does not include the debug
directory.
DATA RVA = DD The relative virtual address of the debug data. This
address is relative to the beginning of the Image Base.
DATA SEEK = DD The seek value from the beginning of the file to the
debug data.
If the image contains more than one type of debug information, then
the next debug directory will immediately follow the first debug
directory.