Wine Developer's Guide

Table of Contents
I. Developing Wine
1. Debugging Wine
1.1. Introduction
1.2. WineDbg modes of invocation
1.3. Using the Wine Debugger
1.4. Useful memory addresses
1.5. Configuration
1.6. WineDbg Expressions and Variables
1.7. WineDbg Command Reference
1.8. Other debuggers
2. Debug Logging
2.1. Debugging classes
2.2. Debugging channels
2.3. Are we debugging?
2.4. Helper functions
2.5. Controlling the debugging output
2.6. A Few Notes on Style
3. Other debugging techniques
3.1. Understanding undocumented APIs
3.2. How to do regression testing using Git
3.3. Which code has been tested?
4. Coding Practice
4.1. Patch Format
4.2. Some notes about style
4.3. Quality Assurance
4.4. Porting Wine to new Platforms
4.5. Adding New Languages
5. Writing Conformance tests
5.1. Introduction
5.2. What to test for?
5.3. Running the tests in Wine
5.4. Cross-compiling the tests with MinGW-w64
5.5. Building and running the tests on Windows
5.6. Inside a test
5.7. Writing good error messages
5.8. Handling platform issues
6. Documenting Wine
6.1. An Overview Of Wine Documentation
6.2. Writing Wine API Documentation
6.3. The Wine DocBook System
II. Wine Architecture
7. Overview
7.1. Wine Overview
7.2. Standard Windows Architectures
7.3. Wine architecture
8. Kernel modules
8.1. The Wine initialization process
8.2. Detailed memory management
8.3. Multi-processing in Wine
8.4. Multi-threading in Wine
8.5. Structured Exception Handling
8.6. File management
8.7. NTDLL module
8.8. KERNEL32 Module
9. Windowing system
9.1. USER32 Module
9.2. X Windows System interface
10. COM in Wine
10.1. Writing COM Components for Wine
10.2. A brief introduction to DCOM in Wine
11. Wine and OpenGL
11.1. What is needed to have OpenGL support in Wine
11.2. How it all works
11.3. Known problems
12. Outline of DirectDraw Architecture
12.1. DirectDraw inheritance tree
12.2. DirectDrawSurface inheritance tree
12.3. Interface Thunks
12.4. Logical Object Layout
12.5. Creating Objects
13. Wine and Multimedia
13.1. Overview
13.2. Multimedia architecture
13.3. Low level layers
13.4. Mid level drivers (MCI)
13.5. High level layers
13.6. MS ACM DLLs
13.7. MS Video DLLs
13.8. Multimedia configuration
List of Tables
1-1. WineDbg misc. commands
1-2. WineDbg flow control commands
1-3. WineDbg break & watch points
1-4. WineDbg stack manipulation
1-5. WineDbg directory & source file manipulation
1-6. WineDbg list command examples
1-7. WineDbg displays
1-8. WineDbg dissassembly
1-9. WineDbg memory management
1-10. WineDbg Win32 objects management
1-11. WineDbg debug channels management
1-12. WineDbg debug channels management
1-13. Debuggers comparison
7-1. Wine executables
7-2. Memory layout (Windows and Wine)
8-1. DOS, Win32 and NT paths equivalences
8-2. File systems' properties
8-3. Mapping of Windows device names into Unix device names
8-4. Function consoles implementation comparison
8-5. Console registry settings
13-1. Wine MCI drivers
13-2. Wine ACM drivers
13-3. Wine VIDC drivers
13-4. Wine multimedia configuration scheme

Chapter 1. Debugging Wine

1.1. Introduction

1.1.1. Processes and threads: in underlying OS and in Windows

Before going into the depths of debugging in Wine, here's a small overview of process and thread handling in Wine. It has to be clear that there are two different beasts: processes/threads from the Unix point of view and processes/threads from a Windows point of view.

Each Windows thread is implemented as a Unix thread, meaning that all threads of a same Windows process share the same (unix) address space.

In the following:

  • W-process means a process in Windows terminology

  • U-process means a process in Unix terminology

  • W-thread means a thread in Windows terminology

A W-process is made of one or several W-threads. Each W-thread is mapped to one and only one U-process. All U-processes of a same W-process share the same address space.

Each Unix process can be identified by two values:

  • the Unix process id (upid in the following)

  • the Windows thread id (tid)

Each Windows process has also a Windows process id (wpid in the following). It must be clear that upid and wpid are different and shall not be used instead of the other.

Wpid and tid are defined (Windows) system wide. They must not be confused with process or thread handles which, as any handle, are indirections to system objects (in this case processes or threads). A same process can have several different handles on the same kernel object. The handles can be defined as local (the values is only valid in a process), or system wide (the same handle can be used by any W-process).

1.1.2. Wine, debugging and WineDbg

When talking of debugging in Wine, there are at least two levels to think of:

  • the Windows debugging API.

  • the Wine integrated debugger, dubbed winedbg.

Wine implements most of the Windows debugging API. The first part of the debugging APIs (in KERNEL32.DLL) allows a W-process, called the debugger, to control the execution of another W-process, the debuggee. To control means stopping/resuming execution, enabling/disabling single stepping, setting breakpoints, reading/writing debuggee memory... Another part of the debugging APIs resides in DBGHELP.DLL (and its ancestor IMAGEHLP.DLL) and lets a debugger look into symbols and types from any module (if the module has been compiled with the proper options).

winedbg is a Winelib application making use of these APIs (KERNEL32.DLL debugging API and class="libraryfile"DBGHELP.DLL) to allow debugging both any Wine or Winelib application as well as Wine itself (kernel and all DLLs).

1.2. WineDbg modes of invocation

1.2.1. Starting a process

Any application (either a Windows native executable, or a Winelib application) can be run through winedbg. Command line options and tricks are the same as for wine:

winedbg telnet.exe
winedbg hl.exe -windowed

1.2.2. Attaching

winedbg can also be launched without any command line argument: winedbg is started without any attached process. You can get a list of running W-processes (and their wpid) using the info process command, and then, with the attach command, pick up the wpid of the W-process you want to debug. This is a neat feature as it allows you to debug an already started application.

1.2.3. On exceptions

When something goes wrong, Windows tracks this as an exception. Exceptions exist for segmentation violation, stack overflow, division by zero, etc.

When an exception occurs, Wine checks if the W-process is debugged. If so, the exception event is sent to the debugger, which takes care of it: end of the story. This mechanism is part of the standard Windows debugging API.

If the W-process is not debugged, Wine tries to launch a debugger. This debugger (normally winedbg, see III Configuration for more details), at startup, attaches to the W-process which generated the exception event. In this case, you are able to look at the causes of the exception, and either fix the causes (and continue further the execution) or dig deeper to understand what went wrong.

If winedbg is the standard debugger, the pass and cont commands are the two ways to let the process go further for the handling of the exception event.

To be more precise on the way Wine (and Windows) generates exception events, when a fault occurs (segmentation violation, stack overflow...), the event is first sent to the debugger (this is known as a first chance exception). The debugger can give two answers:


the debugger had the ability to correct what's generated the exception, and is now able to continue process execution.


the debugger couldn't correct the cause of the first chance exception. Wine will now try to walk the list of exception handlers to see if one of them can handle the exception. If no exception handler is found, the exception is sent once again to the debugger to indicate the failure of the exception handling.

Note: since some of Wine code uses exceptions and try/catch blocks to provide some functionality, winedbg can be entered in such cases with segv exceptions. This happens, for example, with IsBadReadPtr function. In that case, the pass command shall be used, to let the handling of the exception to be done by the catch block in IsBadReadPtr.

1.2.4. Interrupting

You can stop the debugger while it's running by hitting Ctrl-C in its window. This will stop the debugged process, and let you manipulate the current context.

1.2.5. Quitting

Wine supports the new XP APIs, allowing for a debugger to detach from a program being debugged (see detach command).

1.3. Using the Wine Debugger

This section describes where to start debugging Wine. If at any point you get stuck and want to ask for help, please read the How to Report A Bug section of the Wine Users Guide for information on how to write useful bug reports.

1.3.1. Crashes

These usually show up like this:

Unhandled exception: page fault on write access to 0x00000000 in 32-bit code (0x0043369e).
Register dump:
 CS:0023 SS:002b DS:002b ES:002b FS:0063 GS:006b
 EIP:0043369e ESP:0b3ee90c EBP:0b3ee938 EFLAGS:00010246(  R- --  I  Z- -P- )
 EAX:00000072 EBX:7b8acff4 ECX:00000000 EDX:6f727265
 ESI:7ba3b37c EDI:7ffa0000
Stack dump:
0x0b3ee90c:  7b82ced8 00000000 7ba3b348 7b884401
0x0b3ee91c:  7b883cdc 00000008 00000000 7bc36e7b
0x0b3ee92c:  7b8acff4 7b82ceb9 7b8acff4 0b3eea18
0x0b3ee93c:  7b82ce82 00000000 00000000 00000000
0x0b3ee94c:  00000000 0b3ee968 70d7ed7b 70c50000
0x0b3ee95c:  00000000 0b3eea40 7b87fd40 7b82d0d0
=>0 0x0043369e in elementclient (+0x3369e) (0x0b3ee938)
  1 0x7b82ce82 CONSOLE_SendEventThread+0xe1(pmt=0x0(nil)) [/usr/src/debug/wine-1.5.14/dlls/kernel32/console.c:1989] in kernel32 (0x0b3eea18)
  2 0x7bc76320 call_thread_func_wrapper+0xb() in ntdll (0x0b3eea28)
  3 0x7bc7916e call_thread_func+0x7d(entry=0x7b82cda0, arg=0x0(nil), frame=0xb3eeb18) [/usr/src/debug/wine-1.5.14/dlls/ntdll/signal_i386.c:2522] in ntdll (0x0b3eeaf8)
  4 0x7bc762fe RtlRaiseException+0x21() in ntdll (0x0b3eeb18)
  5 0x7bc7f3da start_thread+0xe9(info=0x7ffa0fb8) [/usr/src/debug/wine-1.5.14/dlls/ntdll/thread.c:408] in ntdll (0x0b3ef368)
  6 0xf7597adf start_thread+0xce() in (0x0b3ef468)
0x0043369e: movl	%edx,0x0(%ecx)
Module	Address			Debug info	Name (143 modules)
PE	  340000-  3af000	Deferred        speedtreert
PE	  3b0000-  3d6000	Deferred        ftdriver
PE	  3e0000-  3e6000	Deferred        immwrapper
PE	  400000-  b87000	Export          elementclient
PE	  b90000-  e04000	Deferred        elementskill
PE	  e10000-  e42000	Deferred        ifc22
PE	10000000-10016000	Deferred        zlibwapi
ELF	41f75000-41f7e000	Deferred
ELF	41ff9000-42012000	Deferred
PE	48080000-480a8000	Deferred        msls31
PE	65340000-653d2000	Deferred        oleaut32
PE	70200000-70294000	Deferred        wininet
PE	702b0000-70328000	Deferred        urlmon
PE	70440000-704cf000	Deferred        mlang
PE	70bd0000-70c34000	Deferred        shlwapi
PE	70c50000-70ef3000	Deferred        mshtml
PE	71930000-719b8000	Deferred        shdoclc
PE	78130000-781cb000	Deferred        msvcr80
ELF	79afb000-7b800000	Deferred
ELF	7b800000-7ba3d000	Dwarf           kernel32<elf>
  \-PE	7b810000-7ba3d000	\               kernel32
ELF	7bc00000-7bcd5000	Dwarf           ntdll<elf>
  \-PE	7bc10000-7bcd5000	\               ntdll
ELF	7bf00000-7bf04000	Deferred        <wine-loader>
ELF	7c288000-7c400000	Deferred
PE	7c420000-7c4a7000	Deferred        msvcp80
ELF	7c56d000-7c5b6000	Deferred        dinput<elf>
  \-PE	7c570000-7c5b6000	\               dinput
ELF	7c5b6000-7c600000	Deferred
ELF	7c70e000-7c715000	Deferred
ELF	7c715000-7c77e000	Deferred
ELF	7c77e000-7c7e5000	Deferred
ELF	7c7e5000-7c890000	Deferred
PE	7c7f0000-7c890000	Deferred        krnl386.exe16
ELF	7c890000-7c900000	Deferred        ieframe<elf>
  \-PE	7c8a0000-7c900000	\               ieframe
ELF	7ca00000-7ca1a000	Deferred        rasapi32<elf>
  \-PE	7ca10000-7ca1a000	\               rasapi32
ELF	7ca1a000-7ca21000	Deferred
ELF	7ca21000-7ca25000	Deferred
ELF	7ca25000-7ca2d000	Deferred
ELF	7ca2d000-7ca5a000	Deferred
ELF	7cd5d000-7cd9c000	Deferred
ELF	7cd9c000-7cdea000	Deferred
ELF	7cdfe000-7ce23000	Deferred        iphlpapi<elf>
  \-PE	7ce00000-7ce23000	\               iphlpapi
ELF	7cff1000-7cffd000	Deferred
ELF	7d60d000-7d629000	Deferred        wsock32<elf>
  \-PE	7d610000-7d629000	\               wsock32
ELF	7d80d000-7d828000	Deferred
ELF	7d8cf000-7d8db000	Deferred
ELF	7d8db000-7d903000	Deferred        winepulse<elf>
  \-PE	7d8e0000-7d903000	\               winepulse
ELF	7d95c000-7d966000	Deferred
ELF	7d966000-7d96d000	Deferred
ELF	7d96d000-7d992000	Deferred        mmdevapi<elf>
  \-PE	7d970000-7d992000	\               mmdevapi
ELF	7d9b3000-7d9d0000	Deferred        msimtf<elf>
  \-PE	7d9c0000-7d9d0000	\               msimtf
ELF	7d9d0000-7d9e5000	Deferred
PE	7d9e0000-7d9e5000	Deferred        comm.drv16
ELF	7da83000-7db5f000	Deferred
ELF	7db60000-7db63000	Deferred
ELF	7db63000-7db78000	Deferred
PE	7db70000-7db78000	Deferred        system.drv16
ELF	7db98000-7dca1000	Deferred        opengl32<elf>
  \-PE	7dbb0000-7dca1000	\               opengl32
ELF	7dca1000-7dcb6000	Deferred        vdmdbg<elf>
  \-PE	7dcb0000-7dcb6000	\               vdmdbg
ELF	7dcce000-7dd04000	Deferred        uxtheme<elf>
  \-PE	7dcd0000-7dd04000	\               uxtheme
ELF	7dd04000-7dd0a000	Deferred
ELF	7dd0a000-7dd15000	Deferred
ELF	7dd16000-7dd1f000	Deferred
ELF	7dd24000-7dd38000	Deferred        psapi<elf>
  \-PE	7dd30000-7dd38000	\               psapi
ELF	7dd78000-7dda1000	Deferred
ELF	7dda1000-7ddd6000	Deferred
ELF	7ddd6000-7dde6000	Deferred
ELF	7dde6000-7ddef000	Deferred
ELF	7ddef000-7de11000	Deferred
ELF	7de11000-7df49000	Deferred
ELF	7df49000-7df5b000	Deferred
ELF	7df5b000-7df75000	Deferred
ELF	7df75000-7e005000	Deferred        winex11<elf>
  \-PE	7df80000-7e005000	\               winex11
ELF	7e005000-7e0a5000	Deferred
ELF	7e0a5000-7e0c5000	Deferred
ELF	7e0c5000-7e0ea000	Deferred
ELF	7e123000-7e1eb000	Deferred        crypt32<elf>
  \-PE	7e130000-7e1eb000	\               crypt32
ELF	7e1eb000-7e235000	Deferred        dsound<elf>
  \-PE	7e1f0000-7e235000	\               dsound
ELF	7e235000-7e2a7000	Deferred        ddraw<elf>
  \-PE	7e240000-7e2a7000	\               ddraw
ELF	7e2a7000-7e3e3000	Deferred        wined3d<elf>
  \-PE	7e2b0000-7e3e3000	\               wined3d
ELF	7e3e3000-7e417000	Deferred        d3d8<elf>
  \-PE	7e3f0000-7e417000	\               d3d8
ELF	7e417000-7e43b000	Deferred        imm32<elf>
  \-PE	7e420000-7e43b000	\               imm32
ELF	7e43b000-7e46f000	Deferred        ws2_32<elf>
  \-PE	7e440000-7e46f000	\               ws2_32
ELF	7e46f000-7e49a000	Deferred        msacm32<elf>
  \-PE	7e470000-7e49a000	\               msacm32
ELF	7e49a000-7e519000	Deferred        rpcrt4<elf>
  \-PE	7e4b0000-7e519000	\               rpcrt4
ELF	7e519000-7e644000	Deferred        ole32<elf>
  \-PE	7e530000-7e644000	\               ole32
ELF	7e644000-7e6f7000	Deferred        winmm<elf>
  \-PE	7e650000-7e6f7000	\               winmm
ELF	7e6f7000-7e7fa000	Deferred        comctl32<elf>
  \-PE	7e700000-7e7fa000	\               comctl32
ELF	7e7fa000-7ea23000	Deferred        shell32<elf>
  \-PE	7e810000-7ea23000	\               shell32
ELF	7ea23000-7eaf9000	Deferred        gdi32<elf>
  \-PE	7ea30000-7eaf9000	\               gdi32
ELF	7eafb000-7eaff000	Deferred
ELF	7eaff000-7eb09000	Deferred
ELF	7eb09000-7eb0f000	Deferred
ELF	7eb0f000-7eb18000	Deferred
ELF	7eb18000-7eb32000	Deferred        version<elf>
  \-PE	7eb20000-7eb32000	\               version
ELF	7eb32000-7ec87000	Deferred        user32<elf>
  \-PE	7eb40000-7ec87000	\               user32
ELF	7ec87000-7ecf1000	Deferred        advapi32<elf>
  \-PE	7ec90000-7ecf1000	\               advapi32
ELF	7ecf1000-7ed8f000	Deferred        msvcrt<elf>
  \-PE	7ed00000-7ed8f000	\               msvcrt
ELF	7ef8f000-7ef9c000	Deferred
ELF	7ef9c000-7efc7000	Deferred
ELF	7efc8000-7efe5000	Deferred
ELF	7efe5000-7f000000	Deferred        crtdll<elf>
  \-PE	7eff0000-7f000000	\               crtdll
ELF	f73d0000-f73d4000	Deferred
ELF	f73d4000-f73d8000	Deferred
ELF	f73da000-f73df000	Deferred
ELF	f73df000-f7591000	Dwarf 
ELF	f7591000-f75ab000	Dwarf 
ELF	f75ab000-f76ef000	Dwarf 
ELF	f7722000-f7728000	Deferred
ELF	f7729000-f774a000	Deferred
ELF	f774a000-f774b000	Deferred        [vdso].so
process  tid      prio (all id:s are in hex)
00000008 (D) C:\Perfect World Entertainment\Perfect World International\element\elementclient.exe
	00000031    0 <==
	00000035   15
	00000012    0
	00000021    0
	00000045    0
	00000044    0
	00000043    0
	00000038   15
	00000037    0
	00000036   15
	00000034    0
	00000033    0
	00000032    0
	00000027    0
	00000009    0
0000000e services.exe
	0000000b    0
	00000020    0
	00000017    0
	00000010    0
	0000000f    0
00000014 winedevice.exe
	0000001e    0
	0000001b    0
	00000016    0
	00000015    0
0000001c plugplay.exe
	00000022    0
	0000001f    0
	0000001d    0
00000023 explorer.exe
	00000024    0

Steps to debug a crash. You may stop at any step, but please report the bug and provide as much of the information gathered to the bug report as feasible.

  1. Get the reason for the crash. This is usually a page fault, an unimplemented function in Wine, or the like. When reporting a crash, report this whole crashdump even if it doesn't make sense to you.

    (In this case it is page fault on write access to 0x00000000. Most likely Wine passed NULL to the application or the like.)

  2. Determine the cause of the crash. Since this is usually a primary/secondary reaction to a failed or misbehaving Wine function, rerun Wine with the WINEDEBUG=+relay environment variable set. This will generate quite a lot of output, but usually the reason is located in the last calls. Those lines usually look like this:

    000d:Call advapi32.RegOpenKeyExW(00000090,7eb94da0 L"Patterns",00000000,00020019,0033f968) ret=7eb39af8
    ^^^^      ^^^^^^^^ ^^^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^ ^^^^^^^^^^^ ^^^^^^^^^^^^^^^^^^^^^^^^^^  ^^^^^^^^^^^^
    |         |        |             |                 |           |                           |Return address.
    |         |        |             |                 |           |More arguments.
    |         |        |             |                 |Textual parameter.
    |         |        |             |Arguments.
    |         |        |Function called.
    |         |The module of the called function.
    |The thread in which the call was made.
    000d:Ret  advapi32.RegOpenKeyExW() retval=00000000 ret=7eb39af8
                                       |Return value is 32-bit and has the value 0.
  3. If you have found a misbehaving Wine function, try to find out why it misbehaves. Find the function in the source code. Try to make sense of the arguments passed. Usually there is a WINE_DEFAULT_DEBUG_CHANNEL(channel); at the beginning of the source file. Rerun wine with the WINEDEBUG=+xyz,+relay environment variable set.

    Occasionally there are additional debug channels defined at the beginning of the source file in the form WINE_DECLARE_DEBUG_CHANNEL(channel); if so the offending function may also use one of these alternate channels. Look through the the function for TRACE_(channel)(" ... /n"); and add any additional channels to the command line.

  4. Additional information on how to debug using the internal debugger can be found in programs/winedbg/README.

  5. If this information isn't clear enough or if you want to know more about what's happening in the function itself, try running wine with WINEDEBUG=+all, which dumps ALL included debug information in wine. It is often necessary to limit the debug output produced. That can be done by piping the output through grep, or alternatively with registry keys. See Section 1.5.3 for more information.

  6. If even that isn't enough, add more debug output for yourself into the functions you find relevant. See The section on Debug Logging in this guide for more information. You might also try to run the program in gdb instead of using the Wine debugger. If you do that, use handle SIGSEGV nostop noprint to disable the handling of seg faults inside gdb (needed for Win16).

  7. You can also set a breakpoint for that function. Start wine using winedbg instead of wine. Once the debugger is running enter break RegOpenKeyExW (replace by function you want to debug, case is relevant) to set a breakpoint. Then use continue to start normal program-execution. Wine will stop if it reaches the breakpoint. If the program isn't yet at the crashing call of that function, use continue again until you are about to enter that function. You may now proceed with single-stepping the function until you reach the point of crash. Use the other debugger commands to print registers and the like.

1.3.2. Program hangs, nothing happens

Start the program with winedbg instead of wine. When the program locks up switch to the winedbg terminal and press Ctrl-C. This will stop the program and let you debug the program as you would for a crash.

1.3.3. Program reports an error with a message box

Sometimes programs are reporting failure using more or less nondescript message boxes. We can debug this using the same method as Crashes, but there is one problem... For setting up a message box the program also calls Wine producing huge chunks of debug code.

Since the failure happens usually directly before setting up the message box you can start winedbg and set a breakpoint at MessageBoxA (called by win16 and win32 programs) and proceed with continue. With WINEDEBUG=+all Wine will now stop directly before setting up the message box. Proceed as explained above.

You can also run wine using WINEDEBUG=+relay wine program.exe 2>&1 | less -i and in less search for "MessageBox".

1.3.4. Disassembling programs

You may also try to disassemble the offending program to check for undocumented features and/or use of them.

The best, freely available, disassembler for Win16 programs is Windows Codeback, archive name (e.g.

Disassembling win32 programs is possible using e.g. GoVest by Ansgar Trimborn. It can be found here.

You can also use the newer and better Interactive Disassembler (IDA) from DataRescue. Take a look in the AppDB for links to various versions of IDA.

Another popular disassembler is Windows Disassembler 32 from URSoft. Look for a file called (or similar) on or It seems that Windows Disassembler 32 currently has problems working correctly under Wine, so use IDA or GoVest.

Also of considerable fame amongst disassemblers is SoftIce from NuMega. That software has since been acquired by CompuWare and made part of their Windows driver development suite. Newer versions of SoftIce needs to run as a Windows Service and therefore won't currently work under Wine.

If nothing works for you, you might try one of the disassemblers found in Google directory.

Understanding disassembled code is mostly a question of exercise. Most code out there uses standard C function entries (for it is usually written in C). Win16 function entries usually look like that:

push bp
mov bp, sp
... function code ..
retf XXXX 	<--------- XXXX is number of bytes of arguments

This is a FAR function with no local storage. The arguments usually start at [bp+6] with increasing offsets. Note, that [bp+6] belongs to the rightmost argument, for exported win16 functions use the PASCAL calling convention. So, if we use strcmp(a,b) with a and b both 32-bit variables b would be at [bp+6] and a at [bp+10].

Most functions make also use of local storage in the stackframe:

enter 0086, 00
... function code ...
retf XXXX

This does mostly the same as above, but also adds 0x86 bytes of stackstorage, which is accessed using [bp-xx]. Before calling a function, arguments are pushed on the stack using something like this:

push word ptr [bp-02]   <- will be at [bp+8]
push di                 <- will be at [bp+6]

Here first the selector and then the offset to the passed string are pushed.

1.3.5. Sample debugging session

Let's debug the infamous Word SHARE.EXE message box:

|marcus@jet $ wine winword.exe
|            +---------------------------------------------+
|            | !  You must leave Windows and load SHARE.EXE|
|            |    before starting Word.                    |
|            +---------------------------------------------+
|marcus@jet $ WINEDEBUG=+relay,-debug wine winword.exe
|Win16 task 'winword': Breakpoint 1 at 0x01d7:0x001a
|Call WPROCS.24: TASK_RESCHEDULE() ret=00b7:1456 ds=0927
|Ret  WPROCS.24: TASK_RESCHEDULE() retval=0x8672 ret=00b7:1456 ds=0927
|     AX=0000 BX=3cb4 CX=1f40 DX=0000 SI=0000 DI=0927 BP=0000 ES=11f7
|Loading symbols: /home/marcus/wine/wine...
|Stopped on breakpoint 1 at 0x01d7:0x001a
|In 16 bit mode.
|Wine-dbg>break MessageBoxA                          <---- Set Breakpoint
|Breakpoint 2 at 0x40189100 (MessageBoxA [msgbox.c:190])
|Wine-dbg>c                                            <---- Continue
|Call KERNEL.91: INITTASK() ret=0157:0022 ds=08a7
|     AX=0000 BX=3cb4 CX=1f40 DX=0000 SI=0000 DI=08a7 ES=11d7 EFL=00000286
|...                                                   <----- Much debug output
|Call KERNEL.136: GETDRIVETYPE(0x0000) ret=060f:097b ds=0927
                               ^^^^^^ Drive 0 (A:)
|Ret  KERNEL.136: GETDRIVETYPE() retval=0x0002 ret=060f:097b ds=0927
                                        ^^^^^^  DRIVE_REMOVEABLE
						(It is a floppy diskdrive.)

|Call KERNEL.136: GETDRIVETYPE(0x0001) ret=060f:097b ds=0927
                               ^^^^^^ Drive 1 (B:)
|Ret  KERNEL.136: GETDRIVETYPE() retval=0x0000 ret=060f:097b ds=0927
                                        ^^^^^^  DRIVE_CANNOTDETERMINE
						(I don't have drive B: assigned)

|Call KERNEL.136: GETDRIVETYPE(0x0002) ret=060f:097b ds=0927
                               ^^^^^^^ Drive 2 (C:)
|Ret  KERNEL.136: GETDRIVETYPE() retval=0x0003 ret=060f:097b ds=0927
                                        ^^^^^^ DRIVE_FIXED
                                               (specified as a hard disk)

|Call KERNEL.97: GETTEMPFILENAME(0x00c3,0x09278364"doc",0x0000,0927:8248) ret=060f:09b1 ds=0927
                                 ^^^^^^           ^^^^^        ^^^^^^^^^
                                 |                |            |buffer for fname
                                 |                |temporary name ~docXXXX.tmp
                                 |Force use of Drive C:.

|Warning: GetTempFileName returns 'C:~doc9281.tmp', which doesn't seem to be writable.
|Please check your configuration file if this generates a failure.

Whoops, it even detects that something is wrong!

|Ret  KERNEL.97: GETTEMPFILENAME() retval=0x9281 ret=060f:09b1 ds=0927
                                          ^^^^^^ Temporary storage ID

|Call KERNEL.74: OPENFILE(0x09278248"C:~doc9281.tmp",0927:82da,0x1012) ret=060f:09d8 ds=0927
                                    ^^^^^^^^^^^^^^^^ ^^^^^^^^^ ^^^^^^^
                                    |filename        |OFSTRUCT |open mode:


This fails, since my C: drive is in this case mounted readonly.

|Ret  KERNEL.74: OPENFILE() retval=0xffff ret=060f:09d8 ds=0927
                                   ^^^^^^ HFILE_ERROR16, yes, it failed.

|Call USER.1: MESSAGEBOX(0x0000,0x09278376"You must close Windows and load SHARE.EXE before you start Word.",0x00000000,0x1030) ret=060f:084f ds=0927

And MessageBox'ed.

|Stopped on breakpoint 2 at 0x40189100 (MessageBoxA [msgbox.c:190])
|190     {		<- the sourceline
In 32 bit mode.

The code seems to find a writable harddisk and tries to create a file there. To work around this bug, you can define C: as a network drive, which is ignored by the code above.

1.3.6. Debugging Tips

Here are some additional debugging tips:

  • If you have a program crashing at such an early loader phase that you can't use the Wine debugger normally, but Wine already executes the program's start code, then you may use a special trick. You should do a

    WINEDEBUG=+relay wine program
    to get a listing of the functions the program calls in its start function. Now you do a
    winedbg winfile.exe

    This way, you get into winedbg. Now you can set a breakpoint on any function the program calls in the start function and just type c to bypass the eventual calls of Winfile to this function until you are finally at the place where this function gets called by the crashing start function. Now you can proceed with your debugging as usual.

  • If you try to run a program and it quits after showing an error message box, the problem can usually be identified in the return value of one of the functions executed before MessageBox(). That's why you should re-run the program with e.g.

    WINEDEBUG=+relay wine program_name &>relmsg
    Then do a more relmsg and search for the last occurrence of a call to the string "MESSAGEBOX". This is a line like
    Call USER.1: MESSAGEBOX(0x0000,0x01ff1246 "Runtime error 219 at 0004:1056.",0x00000000,0x1010) ret=01f7:2160 ds=01ff
    In my example the lines before the call to MessageBox() look like that:
    Call KERNEL.96: FREELIBRARY(0x0347) ret=01cf:1033 ds=01ff
    Ret  KERNEL.96: FREELIBRARY() retval=0x0001 ret=01cf:1033 ds=01ff
    Call KERNEL.96: FREELIBRARY(0x036f) ret=01cf:1043 ds=01ff
    Ret  KERNEL.96: FREELIBRARY() retval=0x0001 ret=01cf:1043 ds=01ff
    Call KERNEL.96: FREELIBRARY(0x031f) ret=01cf:105c ds=01ff
    Ret  KERNEL.96: FREELIBRARY() retval=0x0001 ret=01cf:105c ds=01ff
    Call USER.171: WINHELP(0x02ac,0x01ff05b4 "COMET.HLP",0x0002,0x00000000) ret=01cf:1070 ds=01ff
    Call WPROCS.24: TASK_RESCHEDULE() ret=00a7:0a2d ds=002b
    Ret  WPROCS.24: TASK_RESCHEDULE() retval=0x0000 ret=00a7:0a2d ds=002b
    Ret  USER.171: WINHELP() retval=0x0001 ret=01cf:1070 ds=01ff
    Call KERNEL.96: FREELIBRARY(0x01be) ret=01df:3e29 ds=01ff
    Ret  KERNEL.96: FREELIBRARY() retval=0x0000 ret=01df:3e29 ds=01ff
    Call KERNEL.52: FREEPROCINSTANCE(0x02cf00ba) ret=01f7:1460 ds=01ff
    Ret  KERNEL.52: FREEPROCINSTANCE() retval=0x0001 ret=01f7:1460 ds=01ff
    Call USER.1: MESSAGEBOX(0x0000,0x01ff1246 "Runtime error 219 at 0004:1056.",0x00000000,0x1010) ret=01f7:2160 ds=01ff

    I think that the call to MessageBox() in this example is not caused by a wrong result value of some previously executed function (it's happening quite often like that), but instead the message box complains about a runtime error at 0x0004:0x1056.

    As the segment value of the address is only 4, I think that that is only an internal program value. But the offset address reveals something quite interesting: offset 1056 is very close to the return address of FREELIBRARY():

    Call KERNEL.96: FREELIBRARY(0x031f) ret=01cf:105c ds=01ff
    Provided that segment 0x0004 is indeed segment 0x1cf, we now we can use IDA to disassemble the part that caused the error. We just have to find the address of the call to FreeLibrary(). Some lines before that the runtime error occurred. But be careful! In some cases you don't have to disassemble the main program, but instead some DLL called by it in order to find the correct place where the runtime error occurred. That can be determined by finding the origin of the segment value (in this case 0x1cf).

  • If you have created a relay file of some crashing program and want to set a breakpoint at a certain location which is not yet available as the program loads the breakpoint segment during execution, you may set a breakpoint to GetVersion16/32 as those functions are called very often.

    Then do a c until you are able to set this breakpoint without error message.

1.3.7. Some basic debugger usages

After starting your program with

winedbg myprog.exe

the program loads and you get a prompt at the program starting point. Then you can set breakpoints:

b RoutineName      (by routine name) OR
b *0x812575        (by address)

Then you hit c (continue) to run the program. It stops at the breakpoint. You can type

step               (to step one line) OR
stepi              (to step one machine instruction at a time;
                    here, it helps to know the basic 386
                    instruction set)
info reg           (to see registers)
info stack         (to see hex values in the stack)
info local         (to see local variables)
list line number   (to list source code)
x variable name    (to examine a variable; only works if code
                    is not compiled with optimization)
x 0x4269978        (to examine a memory location)
?                  (help)
q                  (quit)

By hitting Enter, you repeat the last command.

1.3.8. Useful programs

Some useful programs:

GoVest: is available from

Simple win32 disassembler that works well with Wine.


IDA Pro is highly recommended, but is not free. DataRescue does however make trial versions available.

Take a look in the AppDB for links to various versions of IDA.


Traces DOS calls (Int 21h, DPMI, ...). Use it with Windows to correct file management problems etc.


Dumps the imports and exports of a PE (Portable Executable) DLL.

winedump: (included in wine tree)

Dumps the imports and exports of a PE (Portable Executable) DLL.

1.4. Useful memory addresses

Wine uses several different kinds of memory addresses.

Win32/"normal" Wine addresses/Linux: linear addresses.

Linear addresses can be everything from 0x0 up to 0xffffffff. In Wine on Linux they are often around e.g. 0x08000000, 0x00400000 (std. Win32 program load address), 0x40000000. Every Win32 process has its own private 4GB address space (that is, from 0x0 up to 0xffffffff).

Win16 "enhanced mode": segmented addresses.

These are the "normal" Win16 addresses, called SEGPTR. They have a segment:offset notation, e.g. 0x01d7:0x0012. The segment part usually is a "selector", which always has the lowest 3 bits set. Some sample selectors are 0x1f7, 0x16f, 0x8f. If these bits are set except for the lowest bit, as e.g. with 0x1f6,xi then it might be a handle to global memory. Just set the lowest bit to get the selector in these cases. A selector kind of "points" to a certain linear (see above) base address. It has more or less three important attributes: segment base address, segment limit, segment access rights.


Selector 0x1f7 (0x40320000, 0x0000ffff, r-x) So 0x1f7 has a base address of 0x40320000, the segment's last address is 0x4032ffff (limit 0xffff), and it's readable and executable. So an address of 0x1f7:0x2300 would be the linear address of 0x40322300.

DOS/Win16 "standard mode"

They, too, have a segment:offset notation. But they are completely different from "normal" Win16 addresses, as they just represent at most 1MB of memory: the segment part can be anything from 0 to 0xffff, and it's the same with the offset part.

Now the strange thing is the calculation that's behind these addresses: just calculate segment*16 + offset in order to get a "linear DOS" address. So e.g. 0x0f04:0x3628 results in 0xf040 + 0x3628 = 0x12668. And the highest address you can get is 0xfffff (1MB), of course.

1.5. Configuration

1.5.1. Windows Debugging configuration

The Windows debugging API uses a registry entry to know which debugger to invoke when an unhandled exception occurs (see On exceptions for some details). Two values in key

[MACHINE\\Software\\Microsoft\\Windows NT\\CurrentVersion\\AeDebug]

determine the behavior:


This is the command line used to launch the debugger (it uses two printf formats (%ld) to pass context dependent information to the debugger). You should put here a complete path to your debugger (winedbg can of course be used, but any other Windows debugging API aware debugger will do). The path to the debugger you choose to use must be reachable via one of the DOS drives configured under /dosdevices in your WINEPREFIX or ~/.wine folder.


If this value is zero, a message box will ask the user if he/she wishes to launch the debugger when an unhandled exception occurs. Otherwise, the debugger is automatically started.

A regular Wine registry looks like:

[MACHINE\\Software\\Microsoft\\Windows NT\\CurrentVersion\\AeDebug] 957636538
"Debugger"="winedbg %ld %ld"

Note 1: Creating this key is mandatory. Not doing so will not fire the debugger when an exception occurs.

Note 2: wineinstall (available in Wine source) sets up this correctly. However, due to some limitation of the registry installed, if a previous Wine installation exists, it's safer to remove the whole

[MACHINE\\Software\\Microsoft\\Windows NT\\CurrentVersion\\AeDebug]

key before running again wineinstall to regenerate this key.

1.5.2. WineDbg configuration

winedbg can be configured through a number of options. Those options are stored in the registry, on a per user basis. The key is (in my registry)


Those options can be read/written while inside winedbg, as part of the debugger expressions. To refer to one of these options, its name must be prefixed by a $ sign. For example,

set $BreakAllThreadsStartup = 1

sets the option BreakAllThreadsStartup to TRUE.

All the options are read from the registry when winedbg starts (if no corresponding value is found, a default value is used), and are written back to the registry when winedbg exits (hence, all modifications to those options are automatically saved when winedbg terminates).

Here's the list of all options:


Set to TRUE if at all threads start-up the debugger stops set to FALSE if only at the first thread startup of a given process the debugger stops. FALSE by default.


Set to TRUE if the debugger stops when a critical section times out (5 minutes); TRUE by default.


Set to TRUE if when winedbg attaches to an existing process after an unhandled exception, winedbg shall be entered on the first attach event. Since the attach event is meaningless in the context of an exception event (the next event which is the exception event is of course relevant), that option is likely to be FALSE.


An exception can generate two debug events. The first one is passed to the debugger (known as a first chance) just after the exception. The debugger can then decide either to resume execution (see winedbg's cont command) or pass the exception up to the exception handler chain in the program (if it exists) (winedbg implements this through the pass command). If none of the exception handlers takes care of the exception, the exception event is sent again to the debugger (known as last chance exception). You cannot pass on a last exception. When the BreakOnFirstChance exception is TRUE, then winedbg is entered for both first and last chance exceptions (to FALSE, it's only entered for last chance exceptions).


Set to TRUE if the debugger, when looking up for a symbol from its name, displays all the thunks with that name. The default value (FALSE) allows not to have to choose between a symbol and all the import thunks from all the DLLs using that symbols.

1.5.3. Configuring +relay behaviour

When setting WINEDEBUG to +relay and debugging, you might get a lot of output. You can limit the output by configuring the value RelayExclude in the registry, located under the key [HKCU\\Software\\Wine\\Debug]

Set the value of RelayExclude to a semicolon-separated list of calls to exclude, e.g. "RtlEnterCriticalSection;RtlLeaveCriticalSection;kernel32.97;kernel32.98".

RelayInclude is an option similar to RelayExclude, except that functions listed here will be the only ones included in the output.

If your application runs too slow with +relay to get meaningful output and you're stuck with multi-GB relay log files, but you're not sure what to exclude, here's a trick to get you started. First, run your application for a minute or so, piping its output to a file on disk:

WINEDEBUG=+relay wine appname.exe &>relay.log
Then run this command to see which calls are performed the most:
awk -F'(' '{print $1}' < relay.log | awk '{print $2}' | sort | uniq -c | sort
Exclude the bottom-most calls with RelayExclude after making sure that they are irrelevant, then run your application again.

1.6. WineDbg Expressions and Variables

1.6.1. Expressions

Expressions in Wine Debugger are mostly written in a C form. However, there are a few discrepancies:

  • Identifiers can take a '!' in their names. This allow mainly to access symbols from different DLLs like USER32!CreateWindowExA.

  • In cast operation, when specifying a structure or an union, you must use the struct or union keyword (even if your program uses a typedef).

When specifying an identifier by its name, if several symbols with the same name exist, the debugger will prompt for the symbol you want to use. Pick up the one you want from its number.

In lots of cases, you can also use regular expressions for looking for a symbol.

winedbg defines its own set of variables. The configuration variables from above are part of them. Some others include:


ID of the W-thread currently examined by the debugger


ID of the W-thread currently examined by the debugger


All CPU registers are also available, using '$' as a prefix. You can use info regs to get a list of available CPU registers.

The $ThreadId and $ProcessId variables can be handy to set conditional breakpoints on a given thread or process.

1.7. WineDbg Command Reference

1.7.1. Misc

Table 1-1. WineDbg misc. commands

abortaborts the debugger
quitexits the debugger
attach N attach to a W-process (N is its ID, numeric or hexadecimal (0xN)). IDs can be obtained using the info process command. Note the info process command returns hexadecimal values.
detachdetach from a W-process.
helpprints some help on the commands
help infoprints some help on info commands

1.7.2. Flow control

Table 1-2. WineDbg flow control commands

cont, c continue execution until next breakpoint or exception.
passpass the exception event up to the filter chain.
step, s continue execution until next "C" line of code (enters function call)
next, n continue execution until next "C" line of code (doesn't enter function call)
stepi, si execute next assembly instruction (enters function call)
nexti, ni execute next assembly instruction (doesn't enter function call)
finish, f execute until current function is exited

cont, step, next, stepi, nexti can be postfixed by a number (N), meaning that the command must be executed N times.

1.7.3. Breakpoints, watch points

Table 1-3. WineDbg break & watch points

enable Nenables (break|watch)point N
disable Ndisables (break|watch)point N
delete Ndeletes (break|watch)point N
cond N removes any existing condition to (break|watch)point N
cond N expr adds condition expr to (break|watch)point N. expr will be evaluated each time the breakpoint is hit. If the result is a zero value, the breakpoint isn't triggered.
break * N adds a breakpoint at address N
break id adds a breakpoint at the address of symbol id
break id N adds a breakpoint at line N inside symbol id
break N adds a breakpoint at line N of current source file
break adds a breakpoint at current $PC address
watch * N adds a watch command (on write) at address N (on 4 bytes)
watch id adds a watch command (on write) at the address of symbol id
info break lists all (break|watch)points (with state)

You can use the symbol EntryPoint to stand for the entry point of the DLL.

When setting a break/watch-point by id, if the symbol cannot be found (for example, the symbol is contained in a not yet loaded module), winedbg will recall the name of the symbol and will try to set the breakpoint each time a new module is loaded (until it succeeds).

1.7.4. Stack manipulation

Table 1-4. WineDbg stack manipulation

btprint calling stack of current thread
bt N print calling stack of thread of ID N (note: this doesn't change the position of the current frame as manipulated by the up and dn commands)
up goes up one frame in current thread's stack
up N goes up N frames in current thread's stack
dn goes down one frame in current thread's stack
dn N goes down N frames in current thread's stack
frame N set N as the current frame for current thread's stack
info local prints information on local variables for current function frame

1.7.5. Directory & source file manipulation

Table 1-5. WineDbg directory & source file manipulation

show dir prints the list of dirs where source files are looked for
dir pathname adds pathname to the list of dirs where to look for source files
dir deletes the list of dirs where to look for source files
symbolfile pathnameloads external symbol definition
symbolfile pathname N loads external symbol definition (applying an offset of N to addresses)
list lists 10 source lines forwards from current position
list - lists 10 source lines backwards from current position
list N lists 10 source lines from line N in current file
list path:N lists 10 source lines from line N in file path
list idlists 10 source lines of function id
list * Nlists 10 source lines from address N

You can specify the end target (to change the 10 lines value) using the ','. For example:

Table 1-6. WineDbg list command examples

list 123, 234 lists source lines from line 123 up to line 234 in current file
list foo.c:1, 56 lists source lines from line 1 up to 56 in file foo.c

1.7.6. Displaying

A display is an expression that's evaluated and printed after the execution of any winedbg command.

winedbg will automatically detect if the expression you entered contains a local variable. If so, display will only be shown if the context is still in the same function as the one the debugger was in when the display expression was entered.

Table 1-7. WineDbg displays

info display lists the active displays
display print the active displays' values (as done each time the debugger stops)
display expradds a display for expression expr
display /fmt expr adds a display for expression expr. Printing evaluated expr is done using the given format (see print command for more on formats)
del display N, undisplay N deletes display N

1.7.7. Disassembly

Table 1-8. WineDbg dissassembly

disasdisassemble from current position
disas exprdisassemble from address expr
disas expr, expr disassembles code between addresses specified by the two exprs

1.7.8. Memory (reading, writing, typing)

Table 1-9. WineDbg memory management

x exprexamines memory at expr address
x /fmt expr examines memory at expr address using format fmt
print expr prints the value of expr (possibly using its type)
print /fmt expr prints the value of expr using format fmt
set lval=expr writes the value of expr in lval
whatis exprprints the C type of expression expr
set ! symbol_picker interactive when printing a value, if several symbols are found, ask the user which one to pick (default)
set ! symbol_picker scoped when printing a value, give precedence to local symbols over global symbols

fmt is either letter or count letter (without a space between count and letter), where letter can be

san ASCII string
ua Unicode UTF16 string
iinstructions (disassemble)
x32-bit unsigned hexadecimal integer
d32-bit signed decimal integer
w16-bit unsigned hexadecimal integer
ccharacter (only printable 0x20-0x7f are actually printed)
b8-bit unsigned hexadecimal integer

1.7.9. Information on Wine internals

Table 1-10. WineDbg Win32 objects management

info class lists all Windows classes registered in Wine
info class idprints information on Windows class id
info share lists all the dynamic libraries loaded in the debugged program (including .so files, NE and PE DLLs)
info share N prints information on module at address N
info regs prints the value of the CPU registers
info all-regs prints the value of the CPU and Floating Point registers
info segment N prints information on segment N (i386 only)
info segment lists all allocated segments (i386 only)
info stack prints the values on top of the stack
info map lists all virtual mappings used by the debugged program
info map N lists all virtual mappings used by the program of wpid N
info wnd Nprints information of Window of handle N
info wnd lists all the window hierarchy starting from the desktop window
info process lists all w-processes in Wine session
info threadlists all w-threads in Wine session
info exception lists the exception frames (starting from current stack frame)

1.7.10. Debug channels

It is possible to turn on and off debug messages as you are debugging using the set command (only for debug channels specified in WINEDEBUG environment variable). See Chapter 2 for more details on debug channels.

Table 1-11. WineDbg debug channels management

set + warn channel turn on warn on channel
set + channelturn on warn/fixme/err/trace on channel
set - channel turn off warn/fixme/err/trace on channel
set - fixme turn off the "fixme" class

1.8. Other debuggers

1.8.1. GDB mode

WineDbg can act as a remote monitor for GDB. This allows to use all the power of GDB, but while debugging wine and/or any Win32 application. To enable this mode, just add --gdb to winedbg command line. You'll end up on a GDB prompt. You'll have to use the GDB commands (not WineDbg ones).

However, some limitation in GDB while debugging wine (see below) don't appear in this mode:

  • GDB will correctly present Win32 thread information and breakpoint behavior

  • Moreover, it also provides support for the Dwarf II debug format (which became the default format (instead of stabs) in gcc 3.1).

A few Wine extensions available through the monitor command.

Table 1-12. WineDbg debug channels management

monitor wndlists all window in the Wine session
monitor proc lists all processes in the Wine session
monitor mem displays memory mapping of debugged process

1.8.2. Graphical frontends to gdb

This section will describe how you can debug Wine using the GDB mode of winedbg and some graphical front ends to GDB for those of you who really like graphical debuggers. DDD

Use the following steps, in this order:

  1. Start the Wine debugger with a command line like:

    winedbg --gdb --no-start name_of_exe_to_debug.exe [optional parameters]

  2. Start ddd

  3. In ddd, use Open File or Open Program to point to the Wine executable.

  4. In the output of above command, there's a line like

    target remote localhost:12345
    Copy that line and paste into ddd command pane (the one with the (gdb) prompt)

The program should now be loaded and up and running. kdbg

Use the following steps, in this order:

  1. Start the Wine debugger with a command line like:

    winedbg --gdb --no-start name_of_exe_to_debug.exe [optional parameters]

  2. In the output of above command, there's a line like

    target remote localhost:12345
    Start kdbg with
    kdbg -r localhost:12345 wine
    localhost:12345 is not a fixed value, but has been printed in first step. "wine" should also be the full path to the Wine executable.

The program should now be loaded and up and running.

1.8.3. Using other Unix debuggers

You can also use other debuggers (like gdb), but you must be aware of a few items:

You need to attach the unix debugger to the correct unix process (representing the correct windows thread) (you can "guess" it from a ps fax command for example. When running the emulator, usually the first two upids are for the Windows application running the desktop, the first thread of the application is generally the third upid; when running a Winelib program, the first thread of the application is generally the first upid)

Note: If you plan to used gdb for a multi-threaded Wine application (native or Winelib), then gdb will be able to handle the multiple threads directly only if:

  • Wine is running on the pthread model (it won't work in the kthread one). See the Wine architecture documentation for further details.

  • gdb supports the multi-threading (you need at least version 5.0 for that).

In the unfortunate case (no direct thread support in gdb because one of the above conditions is false), you'll have to spawn a different gdb session for each Windows thread you wish to debug (which means no synchronization for debugging purposes between the various threads).

Here's how to get info about the current execution status of a certain Wine process:

Change into your Wine source dir and enter:

$ gdb wine
Switch to another console and enter ps ax | grep wine to find all wine processes. Inside gdb, repeat for all Wine processes:
(gdb) attach wpid
with wpid being the process ID of one of the Wine processes. Use
(gdb) bt
to get the backtrace of the current Wine process, i.e. the function call history. That way you can find out what the current process is doing right now. And then you can use several times:
(gdb) n
or maybe even
(gdb) b SomeFunction
(gdb) c

to set a breakpoint at a certain function and continue up to that function. Finally you can enter

(gdb) detach
to detach from the Wine process.

1.8.4. Using other Windows debuggers

You can use any Windows debugging API compliant debugger with Wine. Some reports have been made of success with VisualStudio debugger (in remote mode, only the hub runs in Wine). GoVest fully runs in Wine.

1.8.5. Main differences between winedbg and regular Unix debuggers

Table 1-13. Debuggers comparison

WineDbg debugs a Windows process: the various threads will be handled by the same WineDbg session, and a breakpoint will be triggered for any thread of the W-process gdb debugs a Windows thread: a separate gdb session is needed for each thread of a Windows process and a breakpoint will be triggered only for the w-thread debugged
WineDbg supports debug information from stabs (standard Unix format) and C, CodeView, .DBG (Microsoft) GDB supports debug information from stabs (standard Unix format) and Dwarf II.

Chapter 2. Debug Logging

To better manage the large volume of debugging messages that Wine can generate, we divide the messages on a component basis, and classify them based on the severity of the reported problem. Therefore a message belongs to a channel and a class respectively.

This section will describe the debugging classes, how you can create a new debugging channel, what the debugging API is, and how you can control the debugging output. A picture is worth a thousand words, so here are a few examples of the debugging API in action:

ERR("lock_count == 0 ... please report\n");
FIXME("Unsupported RTL style!\n");
WARN(": file seems to be truncated!\n");
TRACE("[%p]: new horz extent = %d\n", hwnd, extent );
MESSAGE( "Could not create graphics driver '%s'\n", buffer );

2.1. Debugging classes

A debugging class categorizes a message based on the severity of the reported problem. There is a fixed set of classes, and you must carefully choose the appropriate one for your messages. There are five classes of messages:


Messages in this class are meant to signal unimplemented features, known bugs, etc. They serve as a constant and active reminder of what needs to be done.


Messages in this class indicate serious errors in Wine, such as as conditions that should never happen by design.


These are warning messages. You should report a warning when something unwanted happens, and the function cannot deal with the condition. This is seldom used since proper functions can usually report failures back to the caller. Think twice before making the message a warning.


These are detailed debugging messages that are mainly useful to debug a component. These are turned off unless explicitly enabled.


There messages are intended for the end user. They do not belong to any channel. As with warnings, you will seldom need to output such messages.

2.2. Debugging channels

Each component is assigned a debugging channel. The identifier of the channel must be a valid C identifier (reserved words like int or static are permitted). To use a new channel, simply use it in your code. It will be picked up automatically by the build process.

Typically, a file contains code pertaining to only one component, and as such, there is only one channel to output to. You can declare a default channel for the file using the WINE_DEFAULT_DEBUG_CHANNEL() macro:

#include "wine/debug.h"


    FIXME("some unimplemented feature", ...);
    if (zero != 0)
        ERR("This should never be non-null: %d", zero);

In rare situations there is a need to output to more than one debug channel per file. In such cases, you need to declare all the additional channels at the top of the file, and use the _-version of the debugging macros:

#include "wine/debug.h"


    FIXME("this one goes to xxx channel");
    FIXME_(yyy)("Some other msg for the yyy channel");
    WARN_(zzz)("And yet another msg on another channel!");

2.3. Are we debugging?

To test whether the debugging channel xxx is enabled, use the TRACE_ON, WARN_ON, FIXME_ON, or ERR_ON macros. For example:

You should normally need to test only if TRACE_ON, all the others are very seldom used. With careful coding, you can avoid the use of these macros, which is generally desired.

2.4. Helper functions

Resource identifiers can be either strings or numbers. To make life a bit easier for outputting these beasts (and to help you avoid the need to build the message in memory), I introduced a new function called debugres.

The function is defined in wine/debug.h and has the following prototype:

LPSTR debugres(const void *id);

It takes a pointer to the resource id and returns a nicely formatted string of the identifier (which can be a string or a number, depending on the value of the high word). Numbers are formatted as such:


while strings as:


Simply use it in your code like this:

#include "wine/debug.h"


TRACE("resource is %s", debugres(myresource));

Many times strings need to be massaged before output: they may be NULL, contain control characters, or they may be too long. Similarly, Unicode strings need to be converted to ASCII for usage with the debugging API. For all this, you can use the debugstr_[aw]n? family of functions:

    FIXME("(%s): stub\n", debugstr_a(s));

2.5. Controlling the debugging output

It is possible to turn on and off debugging output from within the debugger using the set command. Please see the WineDbg Command Reference section (Section 1.7.10) for how to do this.

You can do the same using the task manager (taskmgr) and selecting your application in the applications list. Right clicking on the application, and selecting the debug option in the popup menu, will let you select the modifications you want on the debug channels.

Another way to conditionally log debug output (e.g. in case of very large installers which may create gigabytes of log output) is to create a pipe:

$ mknod /tmp/debug_pipe p

and then to run wine like that:

$ WINEDEBUG=+relay,+snoop wine setup.exe &>/tmp/debug_pipe

Since the pipe is initially blocking (and thus wine as a whole), you have to activate it by doing:

$ cat /tmp/debug_pipe
(press Ctrl-C to stop pasting the pipe content)

Once you are about to approach the problematic part of the program, you just do:

$ cat /tmp/debug_pipe >/tmp/wine.log

to capture specifically the part that interests you from the pipe without wasting excessive amounts of HDD space and slowing down installation considerably.

The WINEDEBUG environment variable controls the output of the debug messages. It has the following syntax: WINEDEBUG= [yyy]#xxx[,[yyy1]#xxx1]*

  • where # is either + or -

  • when the optional class argument (yyy) is not present, then the statement will enable(+)/disable(-) all messages for the given channel (xxx) on all classes. For example:


    enables all messages on the reg channel and disables all messages on the file channel.

  • when the optional class argument (yyy) is present, then the statement will enable (+)/disable(-) messages for the given channel (xxx) only on the given class. For example:


    enables trace messages on the reg channel and disables warning messages on the file channel.

  • also, the pseudo-channel all is also supported and it has the intuitive semantics:

    WINEDEBUG=+all      -- enables all debug messages
    WINEDEBUG=-all      -- disables all debug messages
    WINEDEBUG=yyy+all   -- enables debug messages for class yyy on all channels
    WINEDEBUG=yyy-all   -- disables debug messages for class yyy on all channels

    So, for example:

    WINEDEBUG=warn-all  -- disables all warning messages

Also, note that at the moment:

  • the FIXME and ERR classes are enabled by default

  • the TRACE and WARN classes are disabled by default

2.6. A Few Notes on Style

This new scheme makes certain things more consistent but there is still room for improvement by using a common style of debug messages. Before I continue, let me note that the output format is the following:

yyy:xxx:fff message
yyy = the class (fixme, err, warn, trace)
xxx = the channel (atom, win, font, etc)
fff = the function name

These fields are output automatically. All you have to provide is the message part.

So here are some ideas:

  • do not include the name of the function: it is included automatically

  • if you want to output the parameters of the function, do it as the first thing and include them in parentheses, like this:

    TRACE("(%d, %p, ...)\n", par1, par2, ...);

  • if you want to name a parameter, use = :

    TRACE("(fd=%d, file=%s): stub\n", fd, name);

  • for stubs, you should output a FIXME message. I suggest this style:

    FIXME("(%x, %d, ...): stub\n", par1, par2, ...);

  • try to output one line per message. That is, the format string should contain only one '\n' and it should always appear at the end of the string.

  • if the output string needs to be dynamically constructed, render it in memory before outputting it:

    char buffer[128] = "";
    if (flags & FLAG_A) strcat(buffer, "FLAG_A ");
    if (flags & FLAG_B) strcat(buffer, "FLAG_B ");
    if (flags & FLAG_C) strcat(buffer, "FLAG_C ");
    TRACE("flags = %s\n", buffer);
    Most of the time however, it is better to create a helper function that renders to a temporary buffer:
    static const char *dbgstr_flags(int flags)
        char buffer[128] = "";
        if (flags & FLAG_A) strcat(buffer, "FLAG_A ");
        if (flags & FLAG_B) strcat(buffer, "FLAG_B ");
        if (flags & FLAG_C) strcat(buffer, "FLAG_C ");
        return wine_dbg_sprintf("flags = %s\n\n", buffer);
    TRACE("flags = %s\n", dbgstr_flags(flags));

Chapter 3. Other debugging techniques

3.1. Understanding undocumented APIs

Some background: On the i386 class of machines, stack entries are usually dword (4 bytes) in size, little-endian. The stack grows downward in memory. The stack pointer, maintained in the esp register, points to the last valid entry; thus, the operation of pushing a value onto the stack involves decrementing esp and then moving the value into the memory pointed to by esp (i.e., push p in assembly resembles *(--esp) = p; in C). Removing (popping) values off the stack is the reverse (i.e., pop p corresponds to p = *(esp++); in C).

In the stdcall calling convention, arguments are pushed onto the stack right-to-left. For example, the C call myfunction(40, 20, 70, 30); is expressed in Intel assembly as:

push 30
push 70
push 20
push 40
call myfunction
The called function is responsible for removing the arguments off the stack. Thus, before the call to myfunction, the stack would look like:
         [local variable or temporary]
         [local variable or temporary]
esp ->    40
After the call returns, it should look like:
         [local variable or temporary]
esp ->   [local variable or temporary]

To restore the stack to this state, the called function must know how many arguments to remove (which is the number of arguments it takes). This is a problem if the function is undocumented.

One way to attempt to document the number of arguments each function takes is to create a wrapper around that function that detects the stack offset. Essentially, each wrapper assumes that the function will take a large number of arguments. The wrapper copies each of these arguments into its stack, calls the actual function, and then calculates the number of arguments by checking esp before and after the call.

The main problem with this scheme is that the function must actually be called from another program. Many of these functions are seldom used. An attempt was made to aggressively query each function in a given library (ntdll.dll) by passing 64 arguments, all 0, to each function. Unfortunately, Windows NT quickly goes to a blue screen of death, even if the program is run from a non-administrator account.

Another method that has been much more successful is to attempt to figure out how many arguments each function is removing from the stack. This instruction, ret hhll (where hhll is the number of bytes to remove, i.e. the number of arguments times 4), contains the bytes 0xc2 ll hh in memory. It is a reasonable assumption that few, if any, functions take more than 16 arguments; therefore, simply searching for hh == 0 and ll < 0x40 starting from the address of a function yields the correct number of arguments most of the time.

Of course, this is not without errors. ret 00ll is not the only instruction that can can have the byte sequence 0xc2 ll 0x0; for example, push 0x000040c2 has the byte sequence 0x68 0xc2 0x40 0x0 0x0, which matches the above. Properly, the utility should look for this sequence only on an instruction boundary; unfortunately, finding instruction boundaries on an i386 requires implementing a full disassembler -- quite a daunting task. Besides, the probability of having such a byte sequence that is not the actual return instruction is fairly low.

Much more troublesome is the non-linear flow of a function. For example, consider the following two functions:

    jmp  somefunction1_impl

    ret  0004

    ret  0008
In this case, we would incorrectly detect both somefunction1 and somefunction2 as taking only a single argument, whereas somefunction1 really takes two arguments.

With these limitations in mind, it is possible to implement more stubs in Wine and, eventually, the functions themselves.

3.2. How to do regression testing using Git

A problem that can happen sometimes is "it used to work before, now it doesn't anymore...". Here is a step by step procedure to try to pinpoint when the problem occurred. This is NOT for casual users.

  1. Clone the "Git" repository from winehq. It's more than 100Mb, so it may take some time with a slow Internet connection.

  2. If you found that something broke between wine-1.1.42 and wine-1.1.44 (these are [WWW] release tags). To start regression testing we run:

    git bisect start
    git bisect good wine-1.1.42
    git bisect bad wine-1.1.44

    If you have exact date/time instead of a release you will need to use sha1 IDs from git log.

  3. Having told Git when things were working and when they broke, it will automatically "position" your source tree to the middle. So all you need to do is build the source:

    ./configure && make clean && make depend && make
    ./wine 'c:\test.exe'

    If the version of Wine that Git picked still has the bug, run:

    git bisect bad
    and if it does not, run:
    git bisect good

    When you run this command, Git will checkout a new version of Wine for you to rebuild, so repeat this step again. When the regression has been isolated, git will inform you.

    To find out what's left to test, try:

    git bisect visualize.

  4. When you have found the bad patch and want to go back to the current HEAD run:

    git bisect reset

  5. If you find the patch that is the cause of the problem, you have almost won; report about it to Wine Bugzilla or subscribe to wine-devel and post it there. There is a chance that the author will jump in to suggest a fix; or there is always the possibility to look hard at the patch until it is coerced to reveal where is the bug :-)

3.3. Which code has been tested?

Deciding what code should be tested next can be a difficult decision. And in any given project, there is always code that isn't tested where bugs could be lurking. This section goes over how to identify these sections using a tool called gcov.

To use gcov on wine, do the following:

  1. In order to activate code coverage in the wine source code, when running make set EXTRACFLAGS and LDFLAGS like so: make EXTRACFLAGS=--coverage LDFLAGS=--coverage. Note that this can be done at any directory level. Since compile and run time are significantly increased by these flags, you may want to only use these flags inside a given dll directory.

  2. Run any application or test suite.

  3. Run gcov on the file which you would like to know more about code coverage.

The following is an example situation when using gcov to determine the coverage of a file could be helpful. We'll use the dlls/advapi32/registry.c file. At one time the code in this file was not fully tested (as it may still be). For example at the time of this writing, the function RegSetValueW had the following lines in it:

if (name && name[0])  /* need to create the subkey */
    if ((ret = RegCreateKeyW( hkey, name, &subkey )) != ERROR_SUCCESS) return ret;

Currently there are a few tests written to test this function. However, these tests don't check that everything is correct. Using gcov and directed tests, we can validate the correctness of this line of code. First, we see what has been tested already by running gcov on the file. To do this, do the following:

cd dlls/advapi32
make clean && make EXTRACFLAGS=--coverage LDFLAGS=--coverage
cd tests
make test
cd ..
gcov registry.c
less registry.c.gcov

The interesting code part looks like this in registry.c.gcov:

    4: 1273:    if (name && name[0])  /* need to create the subkey */
    -: 1274:    {
#####: 1275:        if ((ret = RegCreateKeyW( hkey, name, &subkey )) != ERROR_SUCCESS) return ret;
    -: 1276:    }
gcov output consists of three components: the number of times a line was run, the line number, and the actual text of the line. Note: If a line is optimized out by the compiler, it will appear as if it was never run. Line 1275 is never executed, most likely because name is never passed to the function. In order to validate this line, we need to do two things. First, we must write the test:
ret = RegSetValueW(hkey_main, name1W, REG_SZ, string1W, sizeof(string1W));
ok(ret == ERROR_SUCCESS, "RegSetValueW failed: %d, GLE=%d\n", ret, GetLastError());
test_hkey_main_Value_A(name1A, string1A, sizeof(string1A));
test_hkey_main_Value_W(name1W, string1W, sizeof(string1W));
Once we add in this test case, we now want to know if the line in question is run by this test and works as expected. You should be in the same directory as in the previous command example. The only difference is that we have to remove the *.gcda files in order to start the count over. (If we leave the files then the number of times the line is run is just added, e.g. line 545 below would be run 19 times.) We remove the *.gcov files because they are out of date and need to be recreated:
rm *.gcda *.gcov
cd tests
make test
cd ..
cd ..
gcov registry.c
less registry.c.gcov

The interesting code part looks like this in registry.c.gcov:

 5: 1273:    if (name && name[0])  /* need to create the subkey */
 -: 1274:    {
 1: 1275:        if ((ret = RegCreateKeyW( hkey, name, &subkey )) != ERROR_SUCCESS) return ret;
 -: 1276:    }

Based on gcov, we now know that line 1275 is executed once. And since all of our other tests have remain unchanged, we can assume that the one time it is executed is to satisfy the testcase we added where we check for it. Thus we have validated a line of code. While this is a cursory example, it demonstrates the potential usefulness of this tool.

For a further in depth description of gcov, the official gcc compiler suite page for gcov is

Chapter 4. Coding Practice

This chapter describes the relevant coding practices in Wine, that you should be aware of before doing any serious development in Wine.

4.1. Patch Format

Patches are submitted via email to the Wine patches mailing list, . Your patch should include:

  • A meaningful subject (very short description of patch)

  • A long (paragraph) description of what was wrong and what is now better (recommended).

  • A change log entry (short description of what was changed).

  • The patch in "Git" format

To generate a patch using Git, first commit it to your local tree.

If you are adding a file, use git add. Commit the change using git commit. The commit message will be sent with your patch, and recorded in the commit log.

After committing the patch, you can extract it using git format-patch and send it to wine-patches using git imap-send or simply attaching it to your mail manually.

4.2. Some notes about style

There are a few conventions about coding style that have been adopted over the years of development. The rational for these "rules" is explained for each one.

  • No HTML mail, since patches should be in-lined and HTML turns the patch into garbage. Also it is considered bad etiquette as it uglifies the message, and is not viewable by many of the subscribers.

  • Only one change set per patch. Patches should address only one bug/problem at a time. If a lot of changes need to be made then it is preferred to break it into a series of patches. This makes it easier to find regressions.

  • Tabs are not forbidden but discouraged. A tab is defined as 8 characters and the usual amount of indentation is 4 characters.

  • C++ style comments are forbidden since some compilers choke on them.

  • Commenting out a block of code is usually done by using if (0). For example:

    /* note about reason for commenting block */
    if (0) {
    code /* comments */

    The reason for using this method is that it does not require that you edit comments that may be inside the block of code and it makes sure the code gets maintained.

  • Patches should be in-lined (if you can configure your email client to not wrap lines), or attached as plain text attachments so they can be read inline. This may mean some more work for you. However it allows others to review your patch easily and decreases the chances of it being overlooked or forgotten.

  • Code is usually limited to 100 columns. It generally makes code easier to read.

  • If the patch fixes a bug in Bugzilla please provide a link to the bug in the comments of the patch. Also, mark the bug as "resolved fixed" after your patch gets committed.

4.2.1. Inline attachments with Outlook Express

Outlook Express is notorious for mangling attachments. Giving the patch a .txt extension and attaching will solve the problem for most mailers including Outlook. Also, there is a way to enable Outlook Express to send .diff attachments.

You need the following two things to make it work.

  1. Make sure that .diff files have "\r\n" line endings, because if OE detects that there are none it switches to quoted-printable format attachments.

  2. Using regedit add key Content Type with value text/plain to the .diff extension under HKEY_CLASSES_ROOT (same as for .txt extension). This tells OE to use Content-Type: text/plain instead of application/octet-stream.

Item #1 is important. After you hit the Send button, go to Outbox and using Properties verify the message source to make sure that the mail has the correct format. You might want to send several test emails to yourself too.

4.2.2. Alexandre's Bottom Line

"The basic rules are: no attachments, no MIME crap, no line wrapping, a single patch per mail. Basically if I can't do cat raw_mail | patch -p0 it's in the wrong format."

4.3. Quality Assurance

(Or, "How do I get Alexandre to apply my patch quickly so I can build on it and it will not go stale?")

Make sure your patch applies to the current Git HEAD revisions. If a bunch of patches are committed that may affect whether your patch will apply cleanly then verify that your patch does apply! git fetch; git rebase origin is your friend!

Patches must not break building Wine or cause test failures. Adding dead code should also be avoided.

Save yourself some embarrassment and run your patched code against more than just your current test example. Experience will tell you how much effort to apply here. If there are any conformance tests for the code you're working on, run them and make sure they still pass after your patch is applied. Running tests can be done by running make test. You may need to run make testclean to undo the results of a previous test run. See the "testing" guide for more details on Wine conformance tests.

4.4. Porting Wine to new Platforms

This document provides a few tips on porting Wine to your favorite (UNIX-based) operating system.

4.4.1. Why ifdef MyOS is probably a mistake.

Operating systems change. Maybe yours doesn't have the foo.h header, but maybe a future version will have it. If you want to #include <foo.h>, it doesn't matter what operating system you are using; it only matters whether foo.h is there.

Furthermore, operating systems change names or "fork" into several ones. An #ifdef MyOS will break over time.

If you use the feature of autoconf -- the Gnu auto-configuration utility -- wisely, you will help future porters automatically because your changes will test for features, not names of operating systems. A feature can be many things:

  • existence of a header file

  • existence of a library function

  • existence of libraries

  • bugs in header files, library functions, the compiler, ...

You will need GNU autoconf, which you can get from your friendly GNU mirror. This program takes Wine file and produces a configure shell script that users use to configure Wine to their system.

There are exceptions to the "avoid #ifdef MyOS" rule. Wine, for example, needs the internals of the signal stack -- that cannot easily be described in terms of features. Moreover, you cannot use autoconf HAVE_* symbols in Wine headers, as these may be used by Winelib users who may not be using a configure script.

Let's now turn to specific porting problems and how to solve them.

4.4.2. MyOS doesn't have the foo.h header!

This first step is to make autoconf check for this header. In you add a segment like this in the section that checks for header files (search for "header files"):


If your operating system supports a header file with the same contents but a different name, say bar.h, add a check for that also.

Now you can change

#include <foo.h>


#ifdef HAVE_FOO_H
#include <foo.h>
#elif defined (HAVE_BAR_H)
#include <bar.h>

If your system doesn't have a corresponding header file even though it has the library functions being used, you might have to add an #else section to the conditional. Avoid this if you can.

You will also need to add #undef HAVE_FOO_H (etc.) to include/

Finish up with autoconf and ./configure.

4.4.3. MyOS doesn't have the bar function!

A typical example of this is the memmove function. To solve this problem you would add memmove to the list of functions that autoconf checks for. In you search for AC_CHECK_FUNCS and add memmove (you will notice that someone already did this for this particular function).

Secondly, you will also need to add #undef HAVE_BAR to include/

The next step depends on the nature of the missing function.

Case 1:

It's easy to write a complete implementation of the function (memmove belongs to this case).

You add your implementation in misc/port.c surrounded by #ifndef HAVE_MEMMOVE and #endif.

You might have to add a prototype for your function. If so, include/miscemu.h might be the place. Don't forget to protect that definition by #ifndef HAVE_MEMMOVE and #endif also!

Case 2:

A general implementation is hard, but Wine is only using a special case.

An example is the various wait calls used in SIGNAL_child from loader/signal.c. Here we have a multi-branch case on features:

#ifdef HAVE_THIS
#elif defined (HAVE_THAT)
#elif defined (HAVE_SOMETHING_ELSE)

Note that this is very different from testing on operating systems. If a new version of your operating systems comes out and adds a new function, this code will magically start using it.

Finish up with autoconf and ./configure.

4.5. Adding New Languages

This section documents the procedure for adding a new language to the list of languages that Wine can display system menus and forms in. Adding new translations is not hard as it requires no programming knowledge or special skills.

First add the language ID (most often like xx, e.g. fr for French, or xx_YY for country-specific variations, e.g. pt_BR for Brazilian Portuguese ; look in tools/wmc/po.c for a list) in the LINGUAS section of, and in po/LINGUAS.

Language dependent resources reside in .po files in the po/ directory. Simply create the language-specific .po file using the wine.pot template:

cp po/wine.pot po/xx.po
and start translating (note the file must use UTF-8 encoding).

When adding a new language, also make sure the parameters defined in dlls/kernel32/nls/*.nls fit your local habits and language.

Chapter 5. Writing Conformance tests

5.1. Introduction

The Windows API follows no standard, it is itself a de facto standard, and deviations from that standard, even small ones, often cause applications to crash or misbehave in some way.

The question becomes, "How do we ensure compliance with that standard?". The answer is, "By using the API documentation available to us and backing that up with conformance tests." Furthermore, a conformance test suite is the most accurate (if not necessarily the most complete) form of API documentation and can be used to supplement the Windows API documentation.

Writing a conformance test suite for more than 10000 APIs is no small undertaking. Fortunately it can prove very useful to the development of Wine way before it is complete.

  • The conformance test suite must run on Windows. This is necessary to provide a reasonable way to verify its accuracy. Furthermore the tests must pass successfully on all Windows platforms (tests not relevant to a given platform should be skipped).

    A consequence of this is that the test suite will provide a great way to detect variations in the API between different Windows versions. For instance, this can provide insights into the differences between the, often undocumented, Win9x and NT Windows families.

    However, one must remember that the goal of Wine is to run Windows applications on Linux, not to be a clone of any specific Windows version. So such variations must only be tested for when relevant to that goal.

  • Writing conformance tests is also an easy way to discover bugs in Wine. Of course, before fixing the bugs discovered in this way, one must first make sure that the new tests do pass successfully on at least the Wine Test Bot base VMs.

    Bugs discovered this way should also be easier to fix. Unlike some mysterious application crashes, when a conformance test fails, the expected behavior and APIs tested for are known thus greatly simplifying the diagnosis.

  • To detect regressions. Simply running the test suite regularly in Wine turns it into a great tool to detect regressions. When a test fails, one immediately knows what was the expected behavior and which APIs are involved. Thus regressions caught this way should be detected earlier, because it is easy to run all tests on a regular basis, and be easier to fix because of the reduced diagnosis work.

  • Tests written in advance of the Wine development (possibly even by non Wine developers) can also simplify the work of the future implementer by making it easier for him to check the correctness of his code.

  • Conformance tests will also come in handy when testing Wine on new (or not as widely used) architectures such as FreeBSD, Solaris x86 or even non-x86 systems. Even when the port does not involve any significant change in the thread management, exception handling or other low-level aspects of Wine, new architectures can expose subtle bugs that can be hard to diagnose when debugging regular (complex) applications.

5.2. What to test for?

The first thing to test for is the documented behavior of APIs and such as CreateFile. For instance one can create a file using a long pathname, check that the behavior is correct when the file already exists, try to open the file using the corresponding short pathname, convert the filename to Unicode and try to open it using CreateFileW, and all other things which are documented and that applications rely on.

While the testing framework is not specifically geared towards this type of tests, it is also possible to test the behavior of Windows messages. To do so, create a window, preferably a hidden one so that it does not steal the focus when running the tests, and send messages to that window or to controls in that window. Then, in the message procedure, check that you receive the expected messages and with the correct parameters.

For instance you could create an edit control and use WM_SETTEXT to set its contents, possibly check length restrictions, and verify the results using WM_GETTEXT. Similarly one could create a listbox and check the effect of LB_DELETESTRING on the number of items of the list, selected items list, highlighted item, etc. For concrete examples, see dlls/user32/tests/win.c and the related tests.

However, undocumented behavior should not be tested for unless there is an application that relies on this behavior, and in that case the test should mention that application, or unless one can strongly expect applications to rely on this behavior, typically APIs that return the required buffer size when the buffer pointer is NULL.

5.3. Running the tests in Wine

The simplest way to run the tests in Wine is to type 'make test' in the Wine sources top level directory. This will run all the Wine conformance tests.

The tests for a specific Wine library are located in a 'tests' subdirectory of the library directory. Each test is contained in a file (e.g. dlls/kernel32/tests/thread.c). Each file itself contains many checks concerning one or more related APIs.

So to run all the tests related to a given Wine library, go to the corresponding tests directory and type make test. This will compile the tests, run them, and create a xxx.ok file for each test that passes successfully. And if you only want to run the tests contained in the thread.c file of the kernel library, you would do:

$ cd dlls/kernel32/tests
$ make thread.ok

Note that if the test has already been run and is up to date (i.e. if neither the kernel library nor the thread.c file has changed since the thread.ok file was created), then make will say so. To force the test to be re-run, delete the thread.ok file, and run the make command again.

You can also run tests manually using a command similar to the following:

$ ../../../tools/runtest -q -M kernel32.dll -p thread.c
$ ../../../tools/runtest -P wine -p thread.c
thread.c: 86 tests executed, 5 marked as todo, 0 failures.
The -P option defines the platform that is currently being tested and is used in conjunction with the todo statements (see below). Remove the -q option if you want the testing framework to report statistics about the number of successful and failed tests. Run runtest -h for more details.

5.4. Cross-compiling the tests with MinGW-w64

5.4.1. Setup of the MinGW-w64 cross-compiling environment

Here are some instructions to setup MinGW-w64 on different Linux distributions and *BSD. Debian GNU/Linux like deb systems

This includes Debian GNU/Linux, Ubuntu, etc. The following step should probably work on any deb based system.

Run apt-get install mingw-w64. Red Hat Linux like rpm systems

This includes Fedora, Red Hat Enterprise Linux, Mandrake, most probably SuSE Linux too, etc. But this list isn't exhaustive; The following step should probably work on any rpm based system.

Run yum install mingw64-gcc. *BSD

The *BSD systems have in their ports collection only a port for the MinGW cross-compiling environment. Please see the documentation of your system about how to build and install a port.

5.4.2. Compiling the tests

Having the cross-compiling environment set up the generation of the Windows executables is easy by using the Wine build system.

If you had already run configure, then delete config.cache and re-run configure. You can then run make crosstest. To sum up:

$ rm config.cache
$ ./configure
$ make crosstest

5.5. Building and running the tests on Windows

5.5.1. Using pre-compiled binaries

The simplest solution is to visit and download the latest version of winetest at the bottom of the page. This executable contains all the Wine conformance tests, runs them and reports the results.

You can also get the older versions from Paul Millar's website.

5.5.2. With Visual C++

  • If you are using Visual Studio 6, make sure you have the processor pack. The processor pack fixes error C2520: conversion from unsigned __int64 to double not implemented, use signed __int64. However note that the processor pack is incompatible with Visual Studio 6.0 Standard Edition, and with the Visual Studio 6 Service Pack 6. If you are using Visual Studio 7 or greater you don't need it. In either case it is recommended to use the most recent compatible Visual Studio service pack. If are using Visual Studio Express and you need specific libraries like ntdll.lib that don't ship with Visual Studio Express, you will need the Windows Driver Development Kit (DDK) which is part of the Windows Developer Kit.

  • get the Wine sources

  • Run msvcmaker to generate Visual C++ project files for the tests. msvcmaker is a perl script so you may be able to run it on Windows.

    $ ./tools/winapi/msvcmaker --no-wine

  • If the previous steps were done on your Linux development machine, make the Wine sources accessible to the Windows machine on which you are going to compile them. Typically you would do this using Samba but copying them altogether would work too.

  • On the Windows machine, open the winetest.dsw workspace. This will load each test's project. For each test there are two configurations: one compiles the test with the Wine headers, and the other uses the Microsoft headers.

  • If you choose the "Win32 MSVC Headers" configuration, most of the tests will not compile with the regular Visual Studio headers. So to use this configuration, download and install a recent Platform SDK as well as the latest DirectX SDK. Then, configure Visual Studio to use these SDK headers and libraries. Alternately you could go to the Project->Settings... menu and modify the settings appropriately, but you would then have to redo this whenever you rerun msvcmaker.

  • Open the Build->Batch build... menu and select the tests and build configurations you want to build. Then click on Build.

  • To run a specific test from Visual C++, go to Project->Settings.... There select that test's project and build configuration and go to the Debug tab. There type the name of the specific test to run (e.g. thread) in the Program arguments field. Validate your change by clicking on Ok and start the test by clicking the red exclamation mark (or hitting F5 or any other usual method).

  • You can also run the tests from the command line. You will find them in either Output\Win32_Wine_Headers or Output\Win32_MSVC_Headers depending on the build method. So to run the kernel path tests you would do:

    C:\> cd dlls\kernel\tests\Output\Win32_MSVC_Headers
    C:\wine\dlls\kernel\tests\Output\Win32_MSVC_Headers> kernel32_test path

5.5.3. With MinGW-w64

Wine build system already has support for building tests with a MinGW-w64 cross-compiler. See the section above called "Setup of the MinGW-w64 cross-compiling environment" for instructions on how to set things up. When you have a MinGW-w64 environment installed all you need to do is rerun configure and it should detect the MinGW-w64 compiler and tools. Then run make crosstest to start building the tests.

5.5.4. Standalone, using the Microsoft C++ Toolkit

Sometimes it's nice to be able to build a new unit test on Windows without Wine, and without buying Microsoft Visual C++. Here's the simplest way to do that on a Windows system:

  • Download and install the free-as-in-beer Microsoft C++ Toolkit and the Microsoft Platform SDK.

  • Make a wine directory underneath your work directory, and copy the file wine/test.h from the Wine source tree there (you can download this file from the latest revision at;a=blob_plain;f=include/wine/test.h).

  • Copy some existing test from the Wine source tree, or create your test program (say, mytest.c) using Notepad, being sure to begin it with #include <wine/test.h> following the usual Wine test style.

  • Finally, in a command prompt window, compile the test with the command

    C:\your\work\dir> cl -I. -DSTANDALONE -D_X86_ mytest.c

  • Once that's working, try running the program under Wine without recompiling it. See? No Wine source required at all, save for that one header, wine/test.h.

  • If you want to use the Microsoft C++ Toolkit under Wine, see CL Howto for some tips on making it easy to use it from the Linux command line.

5.6. Inside a test

When writing new checks you can either modify an existing test file or add a new one. If your tests are related to the tests performed by an existing file, then add them to that file. Otherwise create a new .c file in the tests directory and add that file to the CTESTS variable in

A new test file will look something like the following:

#include <wine/test.h>
#include <winbase.h>

/* Maybe auxiliary functions and definitions here */

   /* Write your checks there or put them in functions you will call from
    * there

The test entry point is the START_TEST section. This is where execution will start. You can put all your tests in that section but it may be better to split related checks in functions you will call from the START_TEST section. The parameter to START_TEST must match the name of the C file. So in the above example the C file would be called paths.c.

Tests should start by including the wine/test.h header. This header will provide you access to all the testing framework functions. You can then include the windows header you need, but make sure to not include any Unix or Wine specific header: tests must compile on Windows.

You can use trace to print informational messages. Note that these messages will only be printed if runtest -v is being used.

trace("testing GlobalAddAtomA\n");

Then just call functions and use ok to make sure that they behaved as expected:

ATOM atom = GlobalAddAtomA( "foobar" );
ok( GlobalFindAtomA( "foobar" ) == atom, "could not find atom foobar\n" );
ok( GlobalFindAtomA( "FOOBAR" ) == atom, "could not find atom FOOBAR\n" );
The first parameter of ok is an expression which must evaluate to true if the test was successful. The next parameter is a printf-compatible format string which is displayed in case the test failed, and the following optional parameters depend on the format string.

5.7. Writing good error messages

The message that is printed when a test fails is extremely important.

Someone will take your test, run it on a Windows platform that you don't have access to, and discover that it fails. They will then post an email with the output of the test, and in particular your error message. Someone, maybe you, will then have to figure out from this error message why the test failed.

If the error message contains all the relevant information that will be easy. If not, then it will require modifying the test, finding someone to compile it on Windows, sending the modified version to the original tester and waiting for his reply. In other words, it will be long and painful.

So how do you write a good error message? Let's start with an example of a bad error message:

   "GetThreadPriorityBoost Failed\n");

This will yield:

thread.c:123: Test failed: GetThreadPriorityBoost Failed

Did you notice how the error message provides no information about why the test failed? We already know from the line number exactly which test failed. In fact the error message gives strictly no information that cannot already be obtained by reading the code. In other words it provides no more information than an empty string!

Let's look at how to rewrite it:

BOOL rc;
ok(rc!=0 && disabled==0,"rc=%d error=%ld disabled=%d\n",
This will yield:
thread.c:123: Test failed: rc=0 error=120 disabled=0

When receiving such a message, one would check the source, see that it's a call to GetThreadPriorityBoost, that the test failed not because the API returned the wrong value, but because it returned an error code. Furthermore we see that GetLastError() returned 120 which winerror.h defines as ERROR_CALL_NOT_IMPLEMENTED. So the source of the problem is obvious: this Windows platform (here Windows 98) does not support this API and thus the test must be modified to detect such a condition and skip the test.

So a good error message should provide all the information which cannot be obtained by reading the source, typically the function return value, error codes, and any function output parameter. Even if more information is needed to fully understand a problem, systematically providing the above is easy and will help cut down the number of iterations required to get to a resolution.

It may also be a good idea to dump items that may be hard to retrieve from the source, like the expected value in a test if it is the result of an earlier computation, or comes from a large array of test values (e.g. index 112 of _pTestStrA in vartest.c). In that respect, for some tests you may want to define a macro such as the following:

#define eq(received, expected, label, type) \
        ok((received) == (expected), "%s: got " type " instead of " type "\n", (label),(received),(expected))


eq( b, curr_val, "SPI_{GET,SET}BEEP", "%d" );

5.8. Handling platform issues

Some checks may be written before they pass successfully in Wine. Without some mechanism, such checks would potentially generate hundred of known failures for months each time the tests are being run. This would make it hard to detect new failures caused by a regression. or to detect that a patch fixed a long standing issue.

Thus the Wine testing framework has the concept of platforms and groups of checks can be declared as expected to fail on some of them. In the most common case, one would declare a group of tests as expected to fail in Wine. To do so, use the following construct:

todo_wine {
    SetLastError( 0xdeadbeef );
    ok( GlobalAddAtomA(0) == 0 && GetLastError() == 0xdeadbeef, "failed to add atom 0\n" );
On Windows the above check would be performed normally, but on Wine it would be expected to fail, and not cause the failure of the whole test. However. If that check were to succeed in Wine, it would cause the test to fail, thus making it easy to detect when something has changed that fixes a bug. Also note that todo checks are accounted separately from regular checks so that the testing statistics remain meaningful.

When writing tests you will also encounter differences between Windows 9x and Windows NT platforms. Such differences should be treated differently from the platform issues mentioned above. In particular you should remember that the goal of Wine is not to be a clone of any specific Windows version but to run Windows applications on Unix.

So, if an API returns a different error code on Windows 9x and Windows NT, your check should just verify that Wine returns one or the other:

ok ( GetLastError() == WIN9X_ERROR || GetLastError() == NT_ERROR, ...);

If an API is only present on some Windows platforms, then use LoadLibrary and GetProcAddress to check if it is implemented and invoke it. Remember, tests must run on all Windows platforms. Similarly, conformance tests should not try to correlate the Windows version returned by GetVersion with whether given APIs are implemented or not. Again, the goal of Wine is to run Windows applications (which do not do such checks), and not be a clone of a specific Windows version.

Chapter 6. Documenting Wine

This chapter describes how you can help improve Wine documentation.

Like most large scale volunteer projects, Wine is strongest in areas that are rewarding for its volunteers to work in. The majority of contributors send code patches either fixing bugs, adding new functionality or otherwise improving the software components of the distribution. A lesser number contribute in other ways, such as reporting bugs and regressions, creating tests, providing organizational assistance, or helping to document Wine.

Documentation is important for many reasons, and is often the key to the end user having a successful experience in installing, setting up and using software. Because Wine is a complicated, evolving entity, providing quality up to date documentation is vital to encourage more people to persevere with using and contributing to the project. The following sections describe in detail how to go about adding to or updating Wine existing documentation.

6.1. An Overview Of Wine Documentation

The Wine source code tree comes with a large amount of documentation in the documentation/ subdirectory. This used to be a collection of text files culled from various places such as the Wine Weekly News and the wine-devel mailing list, but was reorganized some time ago into a number of books, each of which is marked up using SGML. You are reading one of these books (the Wine Developer's Guide) right now.

Since being reorganized, the books have been updated and extended regularly. In their current state they provide a good framework which over time can be expanded and kept up to date. This means that most of the time when further documentation is added, it is a simple matter of updating the content of an already existing file. The books available at the time of writing are:

  • The Wine User Guide. This book contains information for end users on installing, configuring and running Wine.

  • The Wine Developer's Guide. This book contains information and guidelines for developers and contributors to the Wine project.

  • The Winelib User's Guide. This book contains information for developers using Winelib to port Win32 applications to Unix.

To obtain a copy of the Wine documentation refer to the Documentation Git tree.

Another source of documentation is the Wine API Guide. This is generated information taken from special comments placed in the Wine source code. When you update or add new API calls to Wine you should consider documenting them so that developers can determine what the API does and how it should be used.

The next sections describe how to create Wine API documentation and how to work with SGML so you can add to the existing books.

6.2. Writing Wine API Documentation

6.2.1. Introduction to API Documentation

Wine includes a large amount of documentation on the API functions it implements. There are several reasons to want to document the Win32 API:

  • To allow Wine developers to know what each function should do, should they need to update or fix it.

  • To allow Winelib users to understand the functions that are available to their applications.

  • To provide an alternative source of free documentation on the Win32 API.

  • To provide more accurate documentation where the existing documentation is accidentally or deliberately vague or misleading.

To this end, a semi formalized way of producing documentation from the Wine source code has evolved. Since the primary users of API documentation are Wine developers themselves, documentation is usually inserted into the source code in the form of comments and notes. Good things to include in the documentation of a function include:

  • The purpose of the function.

  • The parameters of the function and their purpose.

  • The return value of the function, in success as well as failure cases.

  • Additional notes such as interaction with other parts of the system, differences between Wine and Win32 implementations, errors in MSDN documentation, undocumented cases and bugs that Wine corrects or is compatible with.

Good documentation helps developers be aware of the effects of making changes. It also allows good tests to be written which cover all of the documented cases.

Note that you do not need to be a programmer to update the documentation in Wine. If you would like to contribute to the project, patches that improve the API documentation are welcome. The following describes how to format any documentation that you write so that the Wine documentation generator can extract it and make it available to other developers and users.

In general, if you did not write the function in question, you should be wary of adding comments to other peoples code. It is quite possible you may misunderstand or misrepresent what the original author intended! Adding API documentation on the other hand can be done by anybody, since in most cases there is plenty of information about what a function is supposed to do (if it isn't obvious) available in books and articles on the internet.

A final warning concerns copyright and must be noted. If you read MSDN or any publication in order to find out what an API call does, you must be aware that the text you are reading is copyrighted and in most cases cannot legally be reproduced without the authors permission. If you copy verbatim any information from such sources and submit it for inclusion into Wine, you open yourself up to potential legal liability. You must ensure that anything you submit is your own work, although it can be based on your understanding gleaned from reading other peoples work.

6.2.2. Basic API Documentation

The general form of an API comment in Wine is a block comment immediately before a function is implemented in the source code. General comments within a function body or at the top of an implementation file are ignored by the API documentation generator. Such comments are for the benefit of developers only, for example to explain what the source code is doing or to describe something that may not be obvious to the person reading the source code.

The following text uses the function PathRelativePathToA() from SHLWAPI.DLL as an example. You can find this function in the Wine source code tree in the file dlls/shlwapi/path.c.

The first line of the comment gives the name of the function, the DLL that the function is exported from, and its export ordinal number. This is the simplest (and most common type of) comment:

 * PathRelativePathToW   [SHLWAPI.@]

The functions name and the DLL name are obvious. The ordinal number takes one of two forms: Either @ as in the above, or a number if the export is exported by ordinal. You can see which to use by looking at the DLL .spec file. If the line on which the function is listed begins with a number, use it, otherwise use the @ symbol, which indicates that this function is imported only by name.

Note also that round or square brackets can be used, and whitespace between the name and the DLL/ordinal is free form. Thus the following is equally valid:

 *		PathRelativePathToW	(SHLWAPI.@)

This basic comment will not get processed into documentation, since it contains no information. In order to produce documentation for the function, We must add some of the information listed above.

First we add a description of the function. This can be as long as you like, but typically contains only a brief description of what the function is meant to do in general terms. It is free form text:

 * PathRelativePathToW   [SHLWAPI.@]
 * Create a relative path from one path to another.

To be truly useful however we must document the parameters to the function. There are two methods for doing this: In the comment, or in the function prototype.

Parameters documented in the comment should be formatted as follows:

 * PathRelativePathToW   [SHLWAPI.@]
 * Create a relative path from one path to another.
 *  lpszPath   [O] Destination for relative path
 *  lpszFrom   [I] Source path
 *  dwAttrFrom [I] File attribute of source path
 *  lpszTo     [I] Destination path
 *  dwAttrTo   [I] File attributes of destination path

The parameters section starts with PARAMS on its own line. Each parameter is listed in the order they appear in the functions prototype, first with the parameters name, followed by its input/output status, followed by a free form text description of the comment.

The input/output status tells the programmer whether the value will be modified by the function (an output parameter), or only read (an input parameter). The status must be enclosed in square brackets to be recognized, otherwise, or if it is absent, anything following the parameter name is treated as the parameter description. This field is case insensitive and can be any of the following: [I], [In], [O], [Out], [I/O], [In/Out].

Following the description and parameters come a number of optional sections, all in the same format. A section is defined as the section name, which is an all upper case section name on its own line, followed by free form text. You can create any sections you like, however for consistency it is recommended you use the following section names:

  1. NOTES. Anything that needs to be noted about the function such as special cases and the effects of input arguments.

  2. BUGS. Any bugs in the function that exist "by design", i.e. those that will not be fixed or exist for compatibility with Windows.

  3. TODO. Any unhandled cases or missing functionality in the Wine implementation of the function.

  4. FIXME. Things that should be updated or addressed in the implementation of the function at some future date (perhaps dependent on other parts of Wine). Note that if this information is only relevant to Wine developers then it should probably be placed in the relevant code section instead.

Following or before the optional sections comes the RETURNS section which describes the return value of the function. This is free form text but should include what is returned on success as well as possible error return codes. Note that this section must be present for documentation to be generated for your comment.

Our final documentation looks like the following:

 * PathRelativePathToW   [SHLWAPI.@]
 * Create a relative path from one path to another.
 *  lpszPath   [O] Destination for relative path
 *  lpszFrom   [I] Source path
 *  dwAttrFrom [I] File attribute of source path
 *  lpszTo     [I] Destination path
 *  dwAttrTo   [I] File attributes of destination path
 *  TRUE  If a relative path can be formed. lpszPath contains the new path
 *  FALSE If the paths are not relative or any parameters are invalid
 *  lpszTo should be at least MAX_PATH in length.
 *  Calling this function with relative paths for lpszFrom or lpszTo may
 *  give erroneous results.
 *  The Win32 version of this function contains a bug where the lpszTo string
 *  may be referenced 1 byte beyond the end of the string. As a result random
 *  garbage may be written to the output path, depending on what lies beyond
 *  the last byte of the string. This bug occurs because of the behaviour of
 *  PathCommonPrefix() (see notes for that function), and no workaround seems
 *  possible with Win32.
 *  This bug has been fixed here, so for example the relative path from "\\"
 *  to "\\" is correctly determined as "." in this implementation.

6.2.3. Advanced API Documentation

There is no markup language for formatting API comments, since they should be easily readable by any developer working on the source file. A number of constructs are treated specially however, and are noted here. You can use these constructs to enhance the usefulness of the generated documentation by making it easier to read and referencing related documents.

Any valid C identifier that ends with () is taken to be an API function and is formatted accordingly. When generating documentation, this text will become a link to that API call, if the output type supports hyperlinks or their equivalent.

Similarly, any interface name starting with a 'I' and followed by the words reference or object becomes a link to the documentation of that object.

Where an Ascii and Unicode version of a function are available, it is recommended that you document only the Unicode version and have the Ascii version refer to the Unicode one, as follows:

 * PathRelativePathToA   [SHLWAPI.@]
 * See PathRelativePathToW.

Alternately you may use the following form:

 * PathRelativePathToA   [SHLWAPI.@]
 * Unicode version of PathRelativePathToW.

You may also use this construct in any other section, such as NOTES.

Any numbers and text in quotes ("") are highlighted.

Words in all uppercase are assumed to be API constants and are highlighted. If you want to emphasize something in the documentation, put it in a section by itself rather than making it upper case.

Blank lines in a section cause a new paragraph to be started. Blank lines at the start and end of sections are ignored.

Any comment line starting with '*' or '|' is treated as raw text and is not pre-processed before being output. This should be used for code listings, tables and any text that should remain unformatted.

Any line starting with a single word followed by a colon (':') is assumed to be case listing and is emphasized and put in its own paragraph. This is most often used for return values, as in the example section below.

 *  Success: TRUE. Something happens that is documented here.
 *  Failure: FALSE. The reasons why this call can fail are listed here.

Any line starting with a '-' is put into a paragraph by itself. This allows lists to avoid being run together.

If you are in doubt as to how your comment will look, try generating the API documentation and checking the output.

6.2.4. Extra API Documentation

Simply documenting the API calls available provides a great deal of information to developers working with the Win32 API. However additional documentation is needed before the API Guide can be considered truly useful or comprehensive. For example, COM objects that are available for developers use should be documented, along with the interfaces that those objects export. Also, it would be helpful to document each DLL, to provide some structure to the documentation.

To facilitate providing extra documentation, you can create comments that provide extra documentation on functions, or on keywords such as the name of a COM interface or a type definition.

These items are generated using the same formatting rules as described earlier. The only difference is the first line of the comment, which indicates to the generator that the documentation is supplemental and does not describe an export from the DLL being processed.

Let's assume you have implemented a COM interface that you want to document; we'll use the name IExample as an example here. Your comment would look like the following (assuming you are exporting this object from EXAMPLE.DLL):

 * IExample   {EXAMPLE}
 * The IExample object provides lots of interesting functionality.
 * ...

Format this documentation exactly as you would a standard export. The only difference is the use of curly brackets to mark this documentation as supplemental. The generator will output this documentation using the name given before the DLL name, and will link to it from the main DLL page. In addition, if you have referred to the comment name in other documentation using "IExample interface", "IExample object", or "IExample()", those references will point to this documentation.

If you document you COM interfaces this way then all following extra comments that follow in the same source file that begin with the same document title will be added as references to this comment before it is output. For an example of this see dlls/oleaut32/safearray.c. This uses an extra comment to document The SafeArray functions and link them together under one heading.

As a special case, if you use the DLL name as the comment name, the comment will be treated as documentation on the DLL itself. When the documentation for the DLL is processed, the contents of the comment will be placed before the generated statistics, exports and other information that makes up a DLL documentation page.

6.2.5. Generating API Documentation

Having edited or added new API documentation to a source code file, you should generate the documentation to ensure that the result is what you expected. Wine includes a tool (slightly misleadingly) called in the tools/ directory which is used to generate the documentation from the source code.

You can run manually for testing purposes; it is a fairly simple perl script which parses .c files to create output in several formats. If you wish to try this you may want to run it with no arguments, which will cause it to print usage information.

An easier way is to use Wine build system. To create man pages for a given DLL, just type make man from within the dlls directory or type make manpages in the root directory of the Wine source tree. You can then check that a man page was generated for your function, it should be present in the documentation/man3w directory with the same name as the function.

Once you have generated the man pages from the source code, running make install will install them for you. By default they are installed in section 3w of the manual, so they don't conflict with any existing man page names. So, to read the man page you should use man -S 3w name. Alternately you can edit /etc/man.config and add 3w to the list of search paths given in the MANSECT variable.

You can also generate HTML output for the API documentation, in this case the make command is make htmlpages in the dll directory, or from the root. The output will be placed by default under documentation/html. Similarly you can create SGML/XML Docbook source code to produce the Wine API Guide with the command make sgmlpages/make xmlpages respectively.

6.3. The Wine DocBook System

6.3.1. Writing Documentation with DocBook

DocBook is a flavour of SGML (Standard Generalized Markup Language), a syntax for marking up the contents of documents. HTML is another very common flavour of SGML; DocBook markup looks very similar to HTML markup, although the names of the markup tags differ. Getting Started

Why SGML?: The simple answer to that is that SGML allows you to create multiple formats of a given document from a single source. Currently it is used to create HTML, PDF, PS (PostScript) and text versions of the Wine books.

What do I need?: You need the SGML tools. There are various places where you can get them. The most generic way of getting them is from their source as discussed below.

Quick instructions: These are the basic steps to create the Wine books from the SGML source.

  1. Go to

  2. Download all of the sgmltools packages

  3. Install them all and build them (./configure; make; make install)

  4. Switch to your toplevel wine-docs directory

  5. Run ./configure

  6. run make html

  7. View en/wineusr-guide.html, en/winedev-guide.html, etc. in your favorite browser

For more information on building the documentation, please see the README file at the toplevel wine-docs directory. Getting SGML for various distributions

Most Linux distributions have everything you need already bundled up in package form. Unfortunately, each distribution seems to handle its SGML environment differently, installing it into different paths, and naming its packages according to its own whims. SGML on Red Hat

The following packages seem to be sufficient for Red Hat 7.1. You will want to be careful about the order in which you install the RPMs.

  • sgml-common-*.rpm

  • openjade-*.rpm

  • perl-SGMLSpm-*.rpm

  • docbook-dtd*.rpm

  • docbook-style-dsssl-*.rpm

  • tetex-*.rpm

  • jadetex-*.rpm

  • docbook-utils-*.rpm

You can also install ghostscript to view the PostScript format output. SGML on Debian

This is not a definitive list yet, but it seems you might need the following packages:

  • docbook

  • docbook-dsssl

  • docbook-utils

  • docbook-xml

  • docbook-xsl

  • sgml-base

  • sgml-data

  • tetex-base

  • tetex-bin

  • jade

  • jadetex Terminology

SGML markup contains a number of syntactical elements that serve different purposes in the markup. We'll run through the basics here to make sure we're on the same page when we refer to SGML semantics.

The basic currency of SGML is the tag. A simple tag consists of a pair of angle brackets and the name of the tag. For example, the para tag would appear in an SGML document as <para>. This start tag indicates that the immediately following text should be classified according to the tag. In regular SGML, each opening tag must have a matching end tag to show where the tag contents ends. End tags begin with "</" markup, e.g. </para>.

The combination of a start tag, contents, and an end tag is called an element. SGML elements can be nested inside of each other, or contain only text, or may be a combination of both text and other elements, although in most cases it is better to limit your elements to one or the other.

The XML (eXtensible Markup Language) specification, a modern subset of the SGML specification, adds a so-called empty tag, for elements that contain no text content. The entire element is a single tag, ending with "/>", e.g. <xref/>. However, use of this tag style restricts you to XML DocBook processing, and your document may no longer compile with SGML-only processing systems.

Often a processing system will need more information about an element than you can provide with just tags. SGML allows you to add extra "hints" in the form of SGML attributes to pass along this information. The most common use of attributes in DocBook is giving specific elements a name, or an ID, so you can refer to it from elsewhere. This ID can be used for many things, including file-naming for HTML output, hyper-linking to specific parts of the document, and even pulling text from that element (see the <xref> tag).

An SGML attribute appears inside the start tag, between the '<' and '>' brackets. For example, if you wanted to set the id attribute of the <book> element to "mybook", you would create a start tag like this:

<book id="mybook">

Notice that the contents of the attribute are enclosed in quote marks. These quotes are optional in SGML, but mandatory in XML. It's a good habit to use quotes, as it will make it much easier to migrate your documents to an XML processing system later on.

You can also specify more than one attribute in a single tag:

<book id="mybook" status="draft">

Another commonly used type of SGML markup is the entity. An entity lets you associate a block of text with a name. You declare the entity once, at the beginning of your document, and can invoke it as many times as you like throughout the document. You can use entities as shorthand, or to make it easier to maintain certain phrases in a central location, or even to insert the contents of an entire file into your document.

An entity in your document is always surrounded by the '&' and ';' characters. One entity you'll need sooner or later is the one for the '<' character. Since SGML expects all tags to begin with a '<', the '<' is a reserved character. To use it in your document (as I am doing here), you must insert it with the &lt; entity. Each time the SGML processor encounters &lt;, it will place a literal '<' in the output document. Similarly you must use the &gt; and &amp; entities for the '>' and '&' characters.

The final term you'll need to know when writing simple DocBook documents is the DTD (Document Type Declaration). The DTD defines the flavour of SGML a given document is written in. It lists all the legal tag names, like <book>, <para>, and so on, and declares how those tags are allowed to be used together. For example, it doesn't make sense to put a <book> element inside a <para> paragraph element -- only the reverse makes sense.

The DTD thus defines the legal structure of the document. It also declares which attributes can be used with which tags. The SGML processing system can use the DTD to make sure the document is laid out properly before attempting to process it. SGML-aware text editors like Emacs can also use the DTD to guide you while you write, offering you choices about which tags you can add in different places in the document, and beeping at you when you try to add a tag where it doesn't belong.

Generally, you will declare which DTD you want to use as the first line of your SGML document. In the case of DocBook, you will use something like this:

<!doctype book PUBLIC "-//OASIS//DTD DocBook V4.5//EN" []>

Note that you must specify your toplevel element inside the doctype declaration. If you were writing an article rather than a book, you might use this declaration instead:

<!doctype article PUBLIC "-//OASIS//DTD DocBook V4.5//EN" []>
</article> The Document

Once you're comfortable with SGML, creating a DocBook document is quite simple and straightforward. Even though DocBook contains over 300 different tags, you can usually get by with only a small subset of those tags. Most of them are for inline formatting, rather than for document structuring. Furthermore, the common tags have short, intuitive names.

Below is a (completely nonsensical) example to illustrate how a simple document might be laid out. Notice that all <chapter> and <sect1> elements have id attributes. This is not mandatory, but is a good habit to get into, as DocBook is commonly converted into HTML, with a separate generated file for each <book>, <chapter>, and/or <sect1> element. If the given element has an id attribute, the processor will typically name the file accordingly. Thus, the below document might result in index.html, chapter-one.html, blobs.html, and so on.

Also notice the text marked off with "<!-- " and " -->" strings. These denote SGML comments. SGML processors will completely ignore anything between these markers, similar to "/*" and "*/" comments in C source code.

<!doctype book PUBLIC "-//OASIS//DTD DocBook V4.5//EN" []>
<book id="index">
    <title>A Poet's Guide to Nonsense</title>

  <chapter id="chapter-one">
    <title>Blobs and Gribbles</title>

    <!-- This section contains only one major topic -->
    <sect1 id="blobs">
      <title>The Story Behind Blobs</title>
        Blobs are often mistaken for ice cubes and rain

    <!-- This section contains embedded sub-sections -->
    <sect1 id="gribbles">
      <title>Your Friend the Gribble</title>
        A Gribble is a cute, unassuming little fellow...

      <sect2 id="gribble-temperament">
        <title>Gribble Temperament</title>
          When left without food for several days...

      <sect2 id="gribble-appearance">
        <title>Gribble Appearance</title>
          Most Gribbles have a shock of white fur running from...

  <chapter id="chapter-two">

    <sect1 id="dretch-pools">
      <title>Dretch Pools</title>

        When most poets think of Dretch Pools, they tend to...
</book> Common Elements

Once you get used to the syntax of SGML, the next hurdle in writing DocBook documentation is to learn the many DocBook-specific tag names, and when to use them. DocBook was created for technical documentation, and as such, the tag names and document structure are slanted towards the needs of such documentation.

To cover its target audience, DocBook declares a wide variety of specialized tags, including tags for formatting source code (with somewhat of a C/C++ bias), computer prompts, GUI application features, keystrokes, and so on. DocBook also includes tags for universal formatting needs, like headers, footnotes, tables, and graphics.

We won't cover all of these elements here (over 300 DocBook tags exist!), but we will cover the basics. To learn more about the other tags, check out the official DocBook guide at To see how they are used in practice, download the SGML source for this manual (the Wine Developer's Guide) and browse through it, comparing it to the generated HTML (or PostScript or PDF).

There are often many correct ways to mark up a given piece of text, and you may have to make guesses about which tag to use. Sometimes you'll have to make compromises. However, remember that it is possible to further customize the output of the SGML processors. If you don't like the way a certain tag looks in HTML, that doesn't mean you should choose a different tag based on its output formatting. The processing stylesheets can be altered to fix the formatting of that same tag everywhere in the document (not just in the place you're working on). For example, if you're frustrated that the <systemitem> tag doesn't produce any formatting by default, you should fix the stylesheets, not change the valid <systemitem> tag to, for example, an <emphasis> tag.

Here are the common SGML elements:

Structural Elements


The book is the most common toplevel element, and is probably the one you should use for your document.


If you want to group more than one book into a single unit, you can place them all inside a set. This is useful when you want to bundle up documentation in alternate ways. We do this with the Wine documentation, using <book> to put each Wine guide into a separate directory (see documentation/wine-devel.sgml, etc.).


A <chapter> element includes a single entire chapter of the book.


If the chapters in your book fall into major categories or groupings (as in the Wine Developer's Guide), you can place each collection of chapters into a <part> element.


DocBook has many section elements to divide the contents of a chapter into smaller chunks. The encouraged approach is to use the numbered section tags, <sect1>, <sect2>, <sect3>, <sect4>, and <sect5> (if necessary). These tags must be nested in order: you can't place a <sect3> directly inside a <sect1>. You have to nest the <sect3> inside a <sect2>, and so forth. Documents with these explicit section groupings are easier for SGML processors to deal with, and lead to better organized documents. DocBook also supplies a <section> element which you can nest inside itself, but its use is discouraged in favor of the numbered section tags.


The title of a book, chapter, part, section, etc. In most of the major structural elements, like <chapter>, <part>, and the various section tags, <title> is mandatory. In other elements like <book> and <note>, it's optional.


The basic unit of text is the paragraph, represented by the <para> tag. This is probably the tag you'll use most often. In fact, in a simple document, you can probably get away with using only <book>, <chapter>, <title>, and <para>.


For shorter, more targeted documents, like topic pieces and whitepapers, you can use <article> as your toplevel element.

Inline Formatting Elements


The name of a file. You can optionally set the class attribute to Directory, HeaderFile, and SymLink to further classify the filename.


Literal text entered by the user.


Literal text output by the computer.


A catch-all element for literal computer data. Its use is somewhat vague; try to use a more specific tag if possible, like <userinput> or <computeroutput>.


An inline quotation. This tag typically inserts quotation marks for you, so you would write <quote>This is a quote</quote> rather than "This is a quote". This usage may be a little bulkier, but it does allow for automated formatting of all quoted material in the document. Thus, if you wanted all quotations to appear in italic, you could make the change once in your stylesheet, rather than doing a search and replace throughout the document. For larger chunks of quoted text, you can use <blockquote>.


Insert a side note for the reader. By default, the SGML processor usually prefixes the content with "Note:". You can change this text by adding a <title> element. Thus, to add a visible FIXME comment to the documentation, you might write:

  <para>This is an example note...</para>

The results will look something like this:

EXAMPLE: This is an example note...


Used for inserting SGML tags, etc., into a SGML document without resorting to a lot of entity quoting, e.g. &lt;. You can change the appearance of the text with the class attribute. Some common values of this are starttag, endtag, attribute, attvalue, and even sgmlcomment. See this SGML file, documentation/documentation.sgml, for examples.


The text used for a computer prompt, for example a shell prompt, or command-line application prompt.


Meta-text that should be replaced by the user, not typed in literally, e.g. in command descriptions and --help outputs.


A programming constant, e.g. MAX_PATH.


A symbolic value replaced, for example, by a pre-processor. This applies primarily to C macros, but may have other uses. Use the <constant> tag instead of <symbol> where appropriate.


A programming function name.


Programming language parameters you pass with a function.


Parameters you pass to a command-line executable.


Variable name, typically in a programming language.


Programming language types, e.g. from a typedef definition. May have other uses, too.


The name of a C-language struct declaration, e.g. sockaddr.


A field inside a C struct.


An executable binary, e.g. wine or ls.


An environment variable, e.g. $PATH.


A generic catch-all for system-related things, like OS names, computer names, system resources, etc.


An email address. The SGML processor will typically add extra formatting characters, and even a mailto: link for HTML pages. Usage: <email></email>


Special emphasis for introducing a new term. Can also be linked to a <glossary> entry, if desired.

Item Listing Elements


For bulleted lists, no numbering. You can tweak the layout with SGML attributes.


A numbered list; the SGML processor will insert the numbers for you. You can suggest numbering styles with the numeration attribute.


A very simple list of items, often inlined. Control the layout with the type attribute.


A list of terms with definitions or descriptions, like this very list!

Block Text Quoting Elements


Quote a block of source code. Typically highlighted in the output and set off from normal text.


Quote a block of visible computer output, like the output of a command or chunks of debug logs.

Hyperlink Elements


Generic hypertext link, used for pointing to other sections within the current document. You supply the visible text for the link, plus the name of the id attribute of the element that you want to link to. For example:

<link linkend="configuring-wine">the section on configuring wine</link>
<sect2 id="configuring-wine">


In-document hyperlink that can generate its own text. Similar to the <link> tag, you use the linkend attribute to specify which target element you want to jump to:

<xref linkend="configuring-wine">
<sect2 id="configuring-wine">

By default, most SGML processors will auto generate some generic text for the <xref> link, like "Section 2.3.1". You can use the endterm attribute to grab the visible text content of the hyperlink from another element:

<xref linkend="configuring-wine" endterm="config-title">
<sect2 id="configuring-wine">
  <title id="config-title">Configuring Wine</title>

This would create a link to the configuring-wine element, displaying the text of the config-title element for the hyperlink. Most often, you'll add an id attribute to the <title> of the section you're linking to, as above, in which case the SGML processor will use the target's title text for the link text.

Alternatively, you can use an xreflabel attribute in the target element tag to specify the link text:

<sect1 id="configuring-wine" xreflabel="Configuring Wine">

Note: <xref> is an empty element. You don't need a closing tag for it (this is defined in the DTD). In SGML documents, you should use the form <xref>, while in XML documents you should use <xref/>.


An invisible tag, used for inserting id attributes into a document to link to arbitrary places (i.e., when it's not close enough to link to the top of an element).


Hyperlink in URL form, e.g.


Indirect hyperlink; can be used for linking to external documents. Not often used in practice.

6.3.2. Editing SGML Documents

You can write SGML/DocBook documents in any text editor you might find although some editors are more friendly for this task than others.

The most commonly used open source SGML editor is Emacs, with the PSGML mode, or extension. Emacs does not supply a GUI or WYSIWYG (What You See Is What You Get) interface, but it does provide many helpful shortcuts for creating SGML, as well as automatic formatting, validity checking, and the ability to create your own macros to simplify complex, repetitive actions.

Chapter 7. Overview

Brief overview of Wine architecture...

7.1. Wine Overview

7.1.1. Foreword

Wine is often used as a recursive acronym, standing for "Wine Is Not an Emulator". Sometimes it is also known to be used for "Windows Emulator". In a way, both meanings are correct, only seen from different perspectives. The first meaning says that Wine is not a virtual machine, it does not emulate a CPU, and you are not supposed to install Windows nor any Windows device drivers on top of it; rather, Wine is an implementation of the Windows API, and can be used as a library to port Windows applications to Unix. The second meaning, obviously, is that to Windows binaries (.exe files), Wine does look like Windows, and emulates its behaviour and quirks rather closely.

"Emulator": The "Emulator" perspective should not be thought of as if Wine is a typical inefficient emulation layer that means Wine can't be anything but slow - the faithfulness to the badly designed Windows API may of course impose a minor overhead in some cases, but this is both balanced out by the higher efficiency of the Unix platforms Wine runs on, and that other possible abstraction libraries (like Motif, GTK+, CORBA, etc) has a runtime overhead typically comparable to Wine's.

7.1.2. Executables

Wine main task is to run Windows executables under non Windows operating systems. It supports different types of executables:

  • DOS executable. Those are even older programs, using the DOS format (either .com or .exe (the later being also called MZ)).

  • Windows NE executable, also called 16-bit. They were the native processes run by Windows 2.x and 3.x. NE stands for New Executable <g>.

  • Windows PE executable. These are programs were introduced in Windows 95 (and became the native formats for all later Windows versions), even if 16-bit applications were still supported. PE stands for Portable Executable, in a sense where the format of the executable (as a file) is independent of the CPU (even if the content of the file - the code - is CPU dependent).

  • Winelib executable. These are applications, written using the Windows API, but compiled as a Unix executable. Wine provides the tools to create such executables.

Let's quickly review the main differences for the supported executables:

Table 7-1. Wine executables

  DOS (.COM or .EXE)Win16 (NE)Win32 (PE)Win64 (PE)Winelib
MultitaskingOnly one application at a time (except for TSR)CooperativePreemptivePreemptivePreemptive
Address space One MB of memory, where each application is loaded and unloaded. All 16-bit applications share a single address space, protected mode. Each application has it's own address space. Requires MMU support from CPU. Each application has it's own address space. Requires MMU support from CPU. Each application has it's own address space. Requires MMU support from CPU.
Windows API No Windows API but the DOS API (like Int 21h traps). Will call the 16-bit Windows API. Will call the 32-bit Windows API. Will call the 64-bit Windows API. Will call the 32/64-bit Windows API, and possibly also the Unix APIs.
Code (CPU level) Only available on x86 in real mode. Code and data are in segmented forms, with 16-bit offsets. Processor is in real mode. Only available on IA-32 architectures, code and data are in segmented forms, with 16-bit offsets (hence the 16-bit name). Processor is in protected mode. Available (with NT) on several CPUs, including IA-32. On this CPU, uses a flat memory model with 32-bit offsets (hence the 32-bit name). Only available on AMD64 architectures. Flat model, with 32/64-bit addresses.
Multi-threadingNot available.Not available. Available. Available. Available, but must use the Win32/64 APIs for threading and synchronization, not the Unix ones.

Wine deals with this issue by launching a separate Wine process (which is in fact a Unix process) for each Win32 process, but not for Win16 tasks. Win16 tasks are run as different intersynchronized Unix-threads in the same dedicated Wine process; this Wine process is commonly known as a WOW process (Windows on Windows), referring to a similar mechanism used by Windows NT.

Synchronization between the Win16 tasks running in the WOW process is normally done through the Win16 mutex - whenever one of them is running, it holds the Win16 mutex, keeping the others from running. When the task wishes to let the other tasks run, the thread releases the Win16 mutex, and one of the waiting threads will then acquire it and let its task run.

winevdm is the Wine process dedicated to running the Win16 processes. Note that several instances of this process could exist, as Windows has support for different VDM (Virtual Dos Machines) in order to have Win16 processes running in different address spaces. Wine also uses the same architecture to run DOS programs (in this case, the DOS emulation is provided by KRNL386.EXE).

7.2. Standard Windows Architectures

7.2.1. Windows 9x architecture

The windows architecture (Win 9x way) looks like this:

+---------------------+        \
|     Windows EXE     |         } application
+---------------------+        /

+---------+ +---------+        \
| Windows | | Windows |         \ application & system DLLs
|   DLL   | |   DLL   |         /
+---------+ +---------+        /

+---------+ +---------+      \
|  GDI32  | | USER32  |       \
|   DLL   | |   DLL   |        \
+---------+ +---------+         } core system DLLs
+---------------------+        /
|     Kernel32 DLL    |       /
+---------------------+      / 

+---------------------+        \
|     Win9x kernel    |         } kernel space
+---------------------+        /

+---------------------+       \
|  Windows low-level  |        \ drivers (kernel space)
|       drivers       |        /
+---------------------+       /

7.2.2. Windows NT architecture

The windows architecture (Windows NT way) looks like the following drawing. Note the new DLL (NTDLL) which allows implementing different subsystems (as win32); kernel32 in NT architecture implements the Win32 subsystem on top of NTDLL.

+---------------------+                      \
|     Windows EXE     |                       } application
+---------------------+                      /

+---------+ +---------+                      \
| Windows | | Windows |                       \ application & system DLLs
|   DLL   | |   DLL   |                       /
+---------+ +---------+                      /

+---------+ +---------+     +-----------+   \
|  GDI32  | |  USER32 |     |           |    \
|   DLL   | |   DLL   |     |           |     \
+---------+ +---------+     |           |      \ core system DLLs
+---------------------+     |           |      / (on the left side)
|    Kernel32 DLL     |     | Subsystem |     /
|  (Win32 subsystem)  |     |Posix, OS/2|    /
+---------------------+     +-----------+   / 

|               NTDLL.DLL               |

+---------------------------------------+     \
|               NT kernel               |      } NT kernel (kernel space)
+---------------------------------------+     /
+---------------------------------------+     \
|       Windows low-level drivers       |      } drivers (kernel space)
+---------------------------------------+     /

Note also (not depicted in schema above) that the 16-bit applications are supported in a specific subsystem. Some basic differences between the Win9x and the NT architectures include:

  • Several subsystems (Win32, Posix...) can be run on NT, while not on Win 9x.

  • Win 9x roots its architecture in 16-bit systems, while NT is truly a 32-bit system.

  • The drivers model and interfaces in Win 9x and NT are different (even if Microsoft tried to bridge the gap with some support of WDM drivers in Win 98 and above).

7.3. Wine architecture

7.3.1. Global picture

Wine implementation is closer to the Windows NT architecture, even if several subsystems are not implemented yet (remind also that 16-bit support is implemented in a 32-bit Windows EXE, not as a subsystem). Here's the overall picture:

+---------------------+                                  \
|     Windows EXE     |                                   } application
+---------------------+                                  /

+---------+ +---------+                                  \
| Windows | | Windows |                                   \ application & system DLLs
|   DLL   | |   DLL   |                                   /
+---------+ +---------+                                  /

+---------+ +---------+     +-----------+  +--------+  \
|  GDI32  | |  USER32 |     |           |  |        |   \
|   DLL   | |   DLL   |     |           |  |  Wine  |    \
+---------+ +---------+     |           |  | Server |     \ core system DLLs
+---------------------+     |           |  |        |     / (on the left side)
|    Kernel32 DLL     |     | Subsystem |  | NT-like|    /
|  (Win32 subsystem)  |     |Posix, OS/2|  | Kernel |   /
+---------------------+     +-----------+  |        |  / 
                                           |        |
+---------------------------------------+  |        |
|                 NTDLL                 |  |        |
+---------------------------------------+  +--------+

+---------------------------------------+               \
|            Wine executable            |                } unix executable
+---------------------------------------+               /
+---------------------------------------------------+   \
|                   Wine drivers                    |    } Wine specific DLLs
+---------------------------------------------------+   /

+------------+    +------------+     +--------------+   \
|    libc    |    |   libX11   |     |  other libs  |    } unix shared libraries
+------------+    +------------+     +--------------+   /  (user space)

+---------------------------------------------------+   \
|         Unix kernel (Linux,*BSD,Solaris,OS/X)     |    } (Unix) kernel space
+---------------------------------------------------+   /
+---------------------------------------------------+   \
|                 Unix device drivers               |    } Unix drivers (kernel space)
+---------------------------------------------------+   /

Wine must at least completely replace the "Big Three" DLLs (KERNEL/KERNEL32, GDI/GDI32, and USER/USER32), which all other DLLs are layered on top of. But since Wine is (for various reasons) leaning towards the NT way of implementing things, the NTDLL is another core DLL to be implemented in Wine, and many KERNEL32 and ADVAPI32 features will be implemented through the NTDLL.

As of today, no real subsystem (apart the Win32 one) has been implemented in Wine.

The Wine server provides the backbone for the implementation of the core DLLs. It mainly implements inter-process synchronization and object sharing. It can be seen, from a functional point of view, as a NT kernel (even if the APIs and protocols used between Wine DLLs and the Wine server are Wine-specific).

Wine uses the Unix drivers to access the various hardware pieces on the box. However, in some cases, Wine will provide a driver (in Windows sense) to a physical hardware device. This driver will be a proxy to the Unix driver (this is the case, for example, for the graphical part with X11 or Mac drivers, audio with OSS or ALSA drivers...).

All DLLs provided by Wine try to stick as much as possible to the exported APIs from the Windows platforms. There are rare cases where this is not the case, and have been properly documented (Wine DLLs export some Wine specific APIs). Usually, those are prefixed with __wine.

Let's now review in greater details all of those components.

7.3.2. The Wine server

The Wine server is among the most confusing concepts in Wine. What is its function in Wine? Well, to be brief, it provides Inter-Process Communication (IPC), synchronization, and process/thread management. When the Wine server launches, it creates a Unix socket for the current host based on (see below) your home directory's .wine subdirectory (or wherever the WINEPREFIX environment variable points to) - all Wine processes launched later connects to the Wine server using this socket. If a Wine server was not already running, the first Wine process will start up the Wine server in auto-terminate mode (i.e. the Wine server will then terminate itself once the last Wine process has terminated).

The master socket mentioned above is created within the /tmp directory with a name that reflects the configuration directory. This means that there can actually be several separate copies of the Wine server running; one per combination of user and configuration directory. Note that you should not have several users using the same configuration directory at the same time; they will have different copies of the Wine server running and this could well lead to problems with the registry information that they are sharing.

Every thread in each Wine process has its own request buffer, which is shared with the Wine server. When a thread needs to synchronize or communicate with any other thread or process, it fills out its request buffer, then writes a command code through the socket. The Wine server handles the command as appropriate, while the client thread waits for a reply. In some cases, like with the various WaitFor??? synchronization primitives, the server handles it by marking the client thread as waiting and does not send it a reply before the wait condition has been satisfied.

The Wine server itself is a single and separate Unix process and does not have its own threading - instead, it is built on top of a large poll() loop that alerts the Wine server whenever anything happens, such as a client having sent a command, or a wait condition having been satisfied. There is thus no danger of race conditions inside the Wine server itself - it is often called upon to do operations that look completely atomic to its clients.

Because the Wine server needs to manage processes, threads, shared handles, synchronization, and any related issues, all the clients' Win32 objects are also managed by the Wine server, and the clients must send requests to the Wine server whenever they need to know any Win32 object handle's associated Unix file descriptor (in which case the Wine server duplicates the file descriptor, transmits it back to the client, and leaves it to the client to close the duplicate when the client has finished with it).

7.3.3. Wine builtin DLLs: about Relays, Thunks, and DLL descriptors

This section mainly applies to builtin DLLs (DLLs provided by Wine). See section Section 7.3.4 for the details on native vs. builtin DLL handling.

Loading a Windows binary into memory isn't that hard by itself, the hard part is all those various DLLs and entry points it imports and expects to be there and function as expected; this is, obviously, what the entire Wine implementation is all about. Wine contains a range of DLL implementations. You can find the DLLs implementation in the dlls/ directory.

Each DLL (at least, the 32-bit version, see below) is implemented in a Unix shared library. The file name of this shared library is the module name of the DLL with a suffix (or or any other relevant extension depending on the DLL type). This shared library contains the code itself for the DLL, as well as some more information, as the DLL resources and a Wine specific DLL descriptor.

The DLL descriptor, when the DLL is instantiated, is used to create an in-memory PE header, which will provide access to various information about the DLL, including but not limited to its entry point, its resources, its sections, its debug information...

The DLL descriptor and entry point table is generated by the winebuild tool, taking DLL specification files with the extension .spec as input. Resources (after compilation by wrc) or message tables (after compilation by wmc) are also added to the descriptor by winebuild.

When an application module wants to import a DLL, Wine will look:

  • through its list of registered DLLs (in fact, both the already loaded DLLs, and the already loaded shared libraries which have registered a DLL descriptor). Since, the DLL descriptor is automatically registered when the shared library is loaded - remember, registration call is put inside a shared library constructor.

  • If it's not registered, Wine will look for it on disk, building the shared library name from the DLL module name. Directory searched for are specified by the WINEDLLPATH environment variable.

  • Failing that, it will look for a real Windows .DLL file to use, and look through its imports, etc) and use the loading of native DLLs.

After the DLL has been identified (assuming it's still a native one), it's mapped into memory using a dlopen() call. Note that Wine doesn't use the shared library mechanisms for resolving and/or importing functions between two shared libraries (for two DLLs). The shared library is only used for providing a way to load a piece of code on demand. This piece of code, thanks to the DLL descriptor, will provide the same type of information a native DLL would. Wine can then use the same code for native and builtin DLL to handle imports/exports.

Wine also relies on the dynamic loading features of the Unix shared libraries to relocate the DLLs if needed (the same DLL can be loaded at different addresses in two different processes, and even in two consecutive runs of the same executable if the order in which the DLLs are loaded differs).

The DLL descriptor is registered in the Wine realm using some tricks. The winebuild tool, while creating the code for DLL descriptor, also creates a constructor, that will be called when the shared library is loaded into memory. This constructor will actually register the descriptor to the Wine DLL loader. Hence, before the dlopen call returns, the DLL descriptor will be known and registered. This also helps to deal with the cases where there are still dependencies (at the ELF shared lib level, not at the embedded DLL level) between different shared libraries: the embedded DLLs will be properly registered, and even loaded (from a Windows point of view).

Since Wine is 32-bit code itself, and if the compiler supports Windows calling convention, stdcall (gcc does), Wine can resolve imports into Win32 code by substituting the addresses of the Wine handlers directly without any thunking layer in between. This eliminates the overhead most people associate with "emulation", and is what the applications expect anyway.

However, if the user specified WINEDEBUG=+relay, a thunk layer is inserted between the application imports and the Wine handlers (actually the export table of the DLL is modified, and a thunk is inserted in the table); this layer is known as "relay" because all it does is print out the arguments/return values (by using the argument lists in the DLL descriptor's entry point table), then pass the call on, but it's invaluable for debugging misbehaving calls into Wine code. A similar mechanism also exists between Windows DLLs - Wine can optionally insert thunk layers between them, by using WINEDEBUG=+snoop, but since no DLL descriptor information exists for non-Wine DLLs, this is less reliable and may lead to crashes.

For Win16 code, there is no way around thunking - Wine needs to relay between 16-bit and 32-bit code. These thunks switch between the app's 16-bit stack and Wine's 32-bit stack, copies and converts arguments as appropriate (integer sizes differ, pointers are segmented in 16-bit but are 32-bit linear values in 32-bit), and handles the Win16 mutex. Some finer control can be obtained on the conversion, see winebuild reference manual for the details. Suffices to say that the kind of intricate stack content juggling this results in, is not exactly suitable study material for beginners.

7.3.4. Wine/Windows DLLs

This document mainly deals with the status of current DLL support by Wine. winecfg currently supports settings to change the load order of DLLs. The load order depends on several issues, which results in different settings for various DLLs. Pros of Native DLLs

Native DLLs of course guarantee 100% compatibility for routines they implement. For example, using the native USER DLL would maintain a virtually perfect and Windows 95-like look for window borders, dialog controls, and so on. Using the built-in Wine version of this library, on the other hand, would produce a display that does not precisely mimic that of Windows 95. Such subtle differences can be engendered in other important DLLs, such as the common controls library COMMCTRL or the common dialogs library COMMDLG, when built-in Wine DLLs outrank other types in load order.

More significant, less aesthetically-oriented problems can result if the built-in Wine version of the SHELL DLL is loaded before the native version of this library. SHELL contains routines such as those used by installer utilities to create desktop shortcuts. Some installers might fail when using Wine's built-in SHELL. Cons of Native DLLs

Not every application performs better under native DLLs. If a library tries to access features of the rest of the system that are not fully implemented in Wine, the native DLL might work much worse than the corresponding built-in one, if at all. For example, the native Windows GDI library must be paired with a Windows display driver, which of course is not present under Unix and Wine.

Finally, occasionally built-in Wine DLLs implement more features than the corresponding native Windows DLLs. Probably the most important example of such behavior is the integration of Wine with X provided by Wine's built-in USER DLL. Should the native Windows USER library take load-order precedence, such features as the ability to use the clipboard or drag-and-drop between Wine windows and X windows will be lost. Deciding Between Native and Built-In DLLs

Clearly, there is no one rule-of-thumb regarding which load-order to use. So, you must become familiar with what specific DLLs do and which other DLLs or features a given library interacts with, and use this information to make a case-by-case decision. Load Order for DLLs

The default load order follows this algorithm: for all DLLs which have a fully-functional Wine implementation, or where the native DLL is known not to work, the built-in library will be loaded first. In all other cases, the native DLL takes load-order precedence.

See The Wine User Guide for information on how to change the settings.

7.3.5. Memory management

Every Win32 process in Wine has its own dedicated native process on the host system, and therefore its own address space. This section explores the layout of the Windows address space and how it is emulated.

Firstly, a quick recap of how virtual memory works. Physical memory in RAM chips is split into frames, and the memory that each process sees is split into pages. Each process has its own 4 gigabytes of address space (4 GB being the maximum space addressable with a 32-bit pointer). Pages can be mapped or unmapped: attempts to access an unmapped page cause an EXCEPTION_ACCESS_VIOLATION which has the easily recognizable code of 0xC0000005. Any page can be mapped to any frame, therefore you can have multiple addresses which actually "contain" the same memory. Pages can also be mapped to things like files or swap space, in which case accessing that page will cause a disk access to read the contents into a free frame. Initial layout (in Windows)

When a Win32 process starts, it does not have a clear address space to use as it pleases. Many pages are already mapped by the operating system. In particular, the EXE file itself and any DLLs it needs are mapped into memory, and space has been reserved for the stack and a couple of heaps (zones used to allocate memory to the app from). Some of these things need to be at a fixed address, and others can be placed anywhere.

The EXE file itself is usually mapped at address 0x400000 and up: indeed, most EXEs have their relocation records stripped which means they must be loaded at their base address and cannot be loaded at any other address.

DLLs are internally much the same as EXE files but they have relocation records, which means that they can be mapped at any address in the address space. Remember we are not dealing with physical memory here, but rather virtual memory which is different for each process. Therefore OLEAUT32.DLL may be loaded at one address in one process, and a totally different one in another. Ensuring all the functions loaded into memory can find each other is the job of the Windows dynamic linker, which is a part of NTDLL.

So, we have the EXE and its DLLs mapped into memory. Two other very important regions also exist: the stack and the process heap. The process heap is simply the equivalent of the libc malloc arena on UNIX: it's a region of memory managed by the OS which malloc/HeapAlloc partitions and hands out to the application. Windows applications can create several heaps but the process heap always exists.

Windows 9x also implements another kind of heap: the shared heap. The shared heap is unusual in that anything allocated from it will be visible in every other process. Comparison

So far we've assumed the entire 4 gigs of address space is available for the application. In fact that's not so: only the lower 2 gigs are available, the upper 2 gigs are on Windows NT used by the operating system and hold the kernel (from 0x80000000). Why is the kernel mapped into every address space? Mostly for performance: while it's possible to give the kernel its own address space too - this is what Ingo Molnar's 4G/4G VM split patch does for Linux - it requires that every system call into the kernel switches address space. As that is a fairly expensive operation (requires flushing the translation lookaside buffers etc) and syscalls are made frequently it's best avoided by keeping the kernel mapped at a constant position in every processes address space.

Basically, the comparison of memory mappings looks as follows:

Table 7-2. Memory layout (Windows and Wine)

AddressWindows 9xWindows NTLinux

On Windows 9x, in fact only the upper gigabyte (0xC0000000 and up) is used by the kernel, the region from 2 to 3 gigs is a shared area used for loading system DLLs and for file mappings. The bottom 2 gigs on both NT and 9x are available for the programs memory allocation and stack.

7.3.6. Wine drivers

Wine will not allow running native Windows drivers under Unix. This comes mainly because (look at the generic architecture schemas) Wine doesn't implement the kernel features of Windows (kernel here really means the kernel, not the KERNEL32 DLL), but rather sets up a proxy layer on top of the Unix kernel to provide the NTDLL and KERNEL32 features. This means that Wine doesn't provide the inner infrastructure to run native drivers, either from the Win9x family or from the NT family.

In other words, Wine will only be able to provide access to a specific device, if and only if, 1/ this device is supported in Unix (there is Unix-driver to talk to it), 2/ Wine has implemented the proxy code to make the glue between the API of a Windows driver, and the Unix interface of the Unix driver.

Wine, however, tries to implement in the various DLLs needing to access devices to do it through the standard Windows APIs for device drivers in user space. This is for example the case for the multimedia drivers, where Wine loads Wine builtin DLLs to talk to the OSS interface, or the ALSA interface. Those DLLs implement the same interface as any user space audio driver in Windows.

Chapter 8. Kernel modules

This section covers the kernel modules. As already stated, Wine implements the NT architecture, hence provides NTDLL for the core kernel functions, and KERNEL32, which is the implementation of the basis of the Win32 subsystem, on top of NTDLL.

This chapter is made of two types of material (depending of their point of view). Some items will be tackled from a global point of view and then, when needed, explaining the split of work between NTDLL and KERNEL32; some others will be tackled from a DLL point of view (NTDLL or KERNEL32). The choice is made so that the output is more readable and understandable. At least, that's the intent (sigh).

8.1. The Wine initialization process

Wine has a rather complex startup procedure, so unlike many programs the best place to begin exploring the code-base is not in fact at the main() function but instead at some of the more straightforward DLLs that exist on the periphery such as MSI, the widget library (in USER and COMCTL32) etc. The purpose of this section is to document and explain how Wine starts up from the moment the user runs wine myprogram.exe to the point at which myprogram gets control.

8.1.1. First Steps

The actual wine binary that the user runs does not do very much, in fact it is only responsible for checking the threading model in use (NPTL vs LinuxThreads) and then invoking a new binary which performs the next stage in the startup sequence. See the beginning of this chapter for more information on this check and why it's necessary. You can find this code in loader/glibc.c. The result of this check is an exec of wine, potentially (on Linux) via the preloader.

The Wine preloader is found in loader/preloader.c, and is required in order to impose a Win32 style address space layout upon the newly created Win32 process. The details of what this does is covered in the address space layout chapter. The preloader is a statically linked ELF binary which is passed the name of the actual Wine binary to run along with the arguments the user passed in from the command line. The preloader is an unusual program: it does not have a main() function. In standard ELF applications, the entry point is actually at a symbol named _start: this is provided by the standard gcc infrastructure and normally jumps to __libc_start_main() which initializes glibc before passing control to the main function as defined by the programmer.

The preloader takes control direct from the entry point for a few reasons. Firstly, it is required that glibc is not initialized twice: the result of such behaviour is undefined and subject to change without notice. Secondly, it's possible that as part of initializing glibc, the address space layout could be changed - for instance, any call to malloc() will initialize a heap arena which modifies the VM mappings. Finally, glibc does not return to _start() at any point, so by reusing it we avoid the need to recreate the ELF bootstrap stack (env, argv, auxiliary array etc).

The preloader is responsible for two things: protecting important regions of the address space so the dynamic linker does not map shared libraries into them, and once that is done loading the real Wine binary off disk, linking it and starting it up. Normally all this is done automatically by glibc and the kernel but as we intercepted this process by using a static binary it's up to us to restart the process. The bulk of the code in the preloader is about loading wine and off disk, linking them together, then starting the dynamic linking process.

One of the last things the preloader does before jumping into the dynamic linker is scan the symbol table of the loaded Wine binary and set the value of a global variable directly: this is a more efficient way of passing information to the main Wine program than flattening the data structures into an environment variable or command line parameter then unpacking it on the other side, but it achieves pretty much the same thing. The global variable set points to the preload descriptor table, which contains the VMA regions protected by the preloader. This allows Wine to unmap them once the dynamic linker has been run, so leaving gaps we can initialize properly later on.

8.1.2. Starting the emulator

The process of starting up the emulator itself is mostly one of chaining through various initializer functions defined in the core libraries and DLLs: libwine, then NTDLL, then KERNEL32.

The wine binary has a main() function, defined in loader/main.c, so after the preloader has run we start here. This passes the information provided by the preloader into libwine and then calls wine_init(), defined in libs/wine/loader.c. This is where the emulation really starts: wine_init() can, with the correct preparation, be called from programs other than the wine loader itself.

wine_init() does some very basic setup tasks such as initializing the debugging infrastructure, yet more address space manipulation (see the information on the 4G/4G VM split in the address space chapter), before loading NTDLL - the core of both Wine and the Windows NT series - and jumping to the __wine_process_init() function defined in dlls/ntdll/loader.c

This function is responsible for initializing the primary Win32 environment. In thread_init(), it sets up the TEB, the wineserver connection for the main thread and the process heap. See the beginning of this chapter for more information on this.

Finally, it loads and jumps to __wine_kernel_init() in KERNEL32.DLL: this is defined in dlls/kernel32/process.c. This is where the bulk of the work is done. The KERNEL32 initialization code retrieves the startup info for the process from the server, initializes the registry, sets up the drive mapping system and locale data, then begins loading the requested application itself. Each process has a STARTUPINFO block that can be passed into CreateProcess specifying various things like how the first window should be displayed: this is sent to the new process via the wineserver.

After determining the type of file given to Wine by the user (a Win32 EXE file, a Win16 EXE, a Winelib app etc), the program is loaded into memory (which may involve loading and initializing other DLLs, the bulk of Wines startup code), before control reaches the end of __wine_kernel_init(). This function ends with the new process stack being initialized, and start_process being called on the new stack. Nearly there!

The final element of initializing Wine is starting the newly loaded program itself. start_process() sets up the SEH backstop handler, calls LdrInitializeThunk() which performs the last part of the process initialization (such as performing relocations and calling the DllMain() with PROCESS_ATTACH), grabs the entry point of the executable and then on this line:

ExitProcess( entry( peb ) );

... jumps to the entry point of the program. At this point the users program is running and the API provided by Wine is ready to be used. When entry returns, the ExitProcess API will be used to initialize a graceful shutdown.

8.2. Detailed memory management

As already explained in previous chapter (see Section 7.3.5 for the details), Wine creates every 32-bit Windows process in its own 32-bit address space. Wine also tries to map at the relevant addresses what Windows would do. There are however a few nasty bits to look at.

8.2.1. Implementation

Wine (with a bit of black magic) is able to map the main module at it's desired address (likely 0x400000), to create the process heap, its stack (as a Windows executable can ask for a specific stack size), Wine simply use the initial stack of the ELF executable for its initialisation, but creates a new stack (as a Win32 one) for the main thread of the executable. Wine also tries to map all native DLLs at their desired address, so that no relocation has to be performed.

Wine also implements the shared heap so native win9x DLLs can be used. This heap is always created at the SYSTEM_HEAP_BASE address or 0x80000000 and defaults to 16 megabytes in size.

There are a few other magic locations. The bottom 64k of memory is deliberately left unmapped to catch null pointer dereferences. The region from 64k to 1mb+64k are reserved for DOS compatibility and contain various DOS data structures. Finally, the address space also contains mappings for the Wine binary itself, any native libraries Wine is using, the glibc malloc arena and so on.

8.2.2. Laying out the address space

Up until about the start of 2004, the Linux address space very much resembled the Windows 9x layout: the kernel sat in the top gigabyte, the bottom pages were unmapped to catch null pointer dereferences, and the rest was free. The kernels mmap algorithm was predictable: it would start by mapping files at low addresses and work up from there.

The development of a series of new low level patches violated many of these assumptions, and resulted in Wine needing to force the Win32 address space layout upon the system. This section looks at why and how this is done.

The exec-shield patch increases security by randomizing the kernels mmap algorithms. Rather than consistently choosing the same addresses given the same sequence of requests, the kernel will now choose randomized addresses. Because the Linux dynamic linker ( loads DSOs into memory by using mmap, this means that DSOs are no longer loaded at predictable addresses, so making it harder to attack software by using buffer overflows. It also attempts to relocate certain binaries into a special low area of memory known as the ASCII armor so making it harder to jump into them when using string based attacks.

Prelink is a technology that enhances startup times by precalculating ELF global offset tables then saving the results inside the native binaries themselves. By grid fitting each DSO into the address space, the dynamic linker does not have to perform as many relocations so allowing applications that heavily rely on dynamic linkage to be loaded into memory much quicker. Complex C++ applications such as Mozilla, OpenOffice and KDE can especially benefit from this technique.

The 4G VM split patch was developed by Ingo Molnar. It gives the Linux kernel its own address space, thereby allowing processes to access the maximum addressable amount of memory on a 32-bit machine: 4 gigabytes. It allows people with lots of RAM to fully utilise that in any given process at the cost of performance: the reason behind giving the kernel a part of each processes address space was to avoid the overhead of switching on each syscall.

Each of these changes alter the address space in a way incompatible with Windows. Prelink and exec-shield mean that the libraries Wine uses can be placed at any point in the address space: typically this meant that a library was sitting in the region that the EXE you wanted to run had to be loaded (remember that unlike DLLs, EXE files cannot be moved around in memory). The 4G VM split means that programs could receive pointers to the top gigabyte of address space which some are not prepared for (they may store extra information in the high bits of a pointer, for instance). In particular, in combination with exec-shield this one is especially deadly as it's possible the process heap could be allocated beyond ADDRESS_SPACE_LIMIT which causes Wine initialization to fail.

The solution to these problems is for Wine to reserve particular parts of the address space so that areas that we don't want the system to use will be avoided. We later on (re/de)allocate those areas as needed. One problem is that some of these mappings are put in place automatically by the dynamic linker: for instance any libraries that Wine is linked to (like libc, libwine, libpthread etc) will be mapped into memory before Wine even gets control. In order to solve that, Wine overrides the default ELF initialization sequence at a low level and reserves the needed areas by using direct syscalls into the kernel (i.e. without linking against any other code to do it) before restarting the standard initialization and letting the dynamic linker continue. This is referred to as the preloader and is found in loader/preloader.c.

Once the usual ELF boot sequence has been completed, some native libraries may well have been mapped above the 3gig limit: however, this doesn't matter as 3G is a Windows limit, not a Linux limit. We still have to prevent the system from allocating anything else above there (like the heap or other DLLs) though so Wine performs a binary search over the upper gig of address space in order to iteratively fill in the holes with MAP_NORESERVE mappings so the address space is allocated but the memory to actually back it is not. This code can be found in libs/wine/mmap.c:reserve_area.

8.3. Multi-processing in Wine

Let's take a closer look at the way Wine loads and run processes in memory.

8.3.1. Starting a process from command line

When starting a Wine process from command line (we'll get later on to the differences between NE, PE and Winelib executables), there are a couple of things Wine need to do first. A first executable is run to check the threading model of the underlying OS (see Section 8.4 for the details) and will start the real Wine loader corresponding to the choosen threading model.

Then Wine graps a few elements from the Unix world: the environment, the program arguments. Then the is loaded into memory using the standard shared library dynamic loader. When loaded, NTDLL will mainly first create a decent Windows environment:

  • create a PEB (Process Environment Block) and a TEB (Thread Environment Block).

  • set up the connection to the Wine server - and eventually launching the Wine server if none runs

  • create the process heap

Then Kernel32 is loaded (but now using the Windows dynamic loading capabilities) and a Wine specific entry point is called __wine_kernel_init. This function will actually handle all the logic of the process loading and execution, and will never return from it's call.

__wine_kernel_init will undergo the following tasks:

  • initialization of program arguments from Unix program arguments

  • lookup of executable in the file system

  • If the file is not found, then an error is printed and the Wine loader stops.

  • We'll cover the non-PE file type later on, so assume for now it's a PE file. The PE module is loaded in memory using the same mechanisms as for a native DLLs (mainly mapping the file data and code sections into memory, and handling relocation if needed). Note that the dependencies on the module are not resolved at this point.

  • A new stack is created, which size is given in the PE header, and this stack is made the one of the running thread (which is still the only one in the process). The stack used at startup will no longer be used.

  • Which this new stack, ntdll.LdrInitializeThunk is called which performs the remaining initialization parts, including resolving all the DLL imports on the PE module, and doing the init of the TLS slots.

  • Control can now be passed to the EntryPoint of the PE module, which will let the executable run.

8.3.2. Creating a child process from a running process

The steps used are closely link to what is done in the previous case.

There are however a few points to look at a bit more closely. The inner implementation creates the child process using the fork() and exec() calls. This means that we don't need to check again for the threading model, we can use what the parent (or the grand-parent process...) started from command line has found.

The Win32 process creation allows to pass a lot of information between the parent and the child. This includes object handles, windows title, console parameters, environment strings... Wine makes use of both the standard Unix inheritance mechanisms (for environment for example) and the Wine server (to pass from parent to child a chunk of data containing the relevant information).

The previously described loading mechanism will check in the Wine server if such a chunk exists, and, if so, will perform the relevant initialization.

Some further synchronization is also put in place: a parent will wait until the child has started, or has failed. The Wine server is also used to perform those tasks.

8.3.3. Starting a Winelib process

Before going into the gory details, let's first go back to what a Winelib application is. It can be either a regular Unix executable, or a more specific Wine beast. This later form in fact creates two files for a given executable (say foo.exe). The first one, named foo will be a symbolic link to the Wine loader (wine). The second one, named, is the equivalent of the files we've already described for DLLs. As in Windows, an executable is, among other things, a module with its import and export information, as any DLL, it makes sense Wine uses the same mechanisms for loading native executables and DLLs.

When starting a Winelib application from the command line (say with foo arg1 arg2), the Unix shell will execute foo as a Unix executable. Since this is in fact the Wine loader, Wine will fire up. However, it will notice that it hasn't been started as wine but as foo, and hence, will try to load (using Unix shared library mechanism) the second file Wine will recognize a 32-bit module (with its descriptor) embedded in the shared library, and once the shared library loaded, it will proceed the same path as when loading a standard native PE executable.

Wine needs to implement this second form of executable in order to maintain the order of initialization of some elements in the executable. One particular issue is when dealing with global C++ objects. In standard Unix executable, the call of the constructor to such objects is stored in the specific section of the executable (.init not to name it). All constructors in this section are called before the main() or WinMain function is called. Creating a Wine executable using the first form mentionned above will let those constructors being called before Wine gets a chance to initialize itself. So, any constructor using a Windows API will fail, because Wine infrastructure isn't in place. The use of the second form for Winelib executables ensures that we do the initialization using the following steps:

  • initialize the Wine infrastructure

  • load the executable into memory

  • handle the import sections for the executable

  • call the global object constructors (if any). They now can properly call the Windows APIs

  • call the executable entry point

The attentive reader would have noted that the resolution of imports for the executable is done, as for a DLL, when the executable/DLL descriptor is registered. However, this is done also by adding a specific constructor in the .init section. For the above describe scheme to function properly, this constructor must be the first constructor to be called, before all the other constructors, generated by the executable itself. The Wine build chain takes care of that, and also generating the executable/DLL descriptor for the Winelib executable.

8.4. Multi-threading in Wine

This section will assume you understand the basics of multithreading. If not there are plenty of good tutorials available on the net to get you started.

Threading in Wine is somewhat complex due to several factors. The first is the advanced level of multithreading support provided by Windows - there are far more threading related constructs available in Win32 than the Linux equivalent (pthreads). The second is the need to be able to map Win32 threads to native Linux threads which provides us with benefits like having the kernel schedule them without our intervention. While it's possible to implement threading entirely without kernel support, doing so is not desirable on most platforms that Wine runs on.

8.4.1. Threading support in Win32

Win32 is an unusually thread friendly API. Not only is it entirely thread safe, but it provides many different facilities for working with threads. These range from the basics such as starting and stopping threads, to the extremely complex such as injecting threads into other processes and COM inter-thread marshalling.

One of the primary challenges of writing Wine code therefore is ensuring that all our DLLs are thread safe, free of race conditions and so on. This isn't simple - don't be afraid to ask if you aren't sure whether a piece of code is thread safe or not!

Win32 provides many different ways you can make your code thread safe however the most common are critical section and the interlocked functions. Critical sections are a type of mutex designed to protect a geographic area of code. If you don't want multiple threads running in a piece of code at once, you can protect them with calls to EnterCriticalSection() and LeaveCriticalSection(). The first call to EnterCriticalSection() by a thread will lock the section and continue without stopping. If another thread calls it then it will block until the original thread calls LeaveCriticalSection() again.

It is therefore vitally important that if you use critical sections to make some code thread-safe, that you check every possible codepath out of the code to ensure that any held sections are left. Code like this:

if (res != ERROR_SUCCESS) return res;

is extremely suspect in a function that also contains a call to EnterCriticalSection(). Be careful.

If a thread blocks while waiting for another thread to leave a critical section, you will see an error from the RtlpWaitForCriticalSection() function, along with a note of which thread is holding the lock. This only appears after a certain timeout, normally a few seconds. It's possible the thread holding the lock is just being really slow which is why Wine won't terminate the app like a non-checked build of Windows would, but the most common cause is that for some reason a thread forgot to call LeaveCriticalSection(), or died while holding the lock (perhaps because it was in turn waiting for another lock). This doesn't just happen in Wine code: a deadlock while waiting for a critical section could be due to a bug in the app triggered by a slight difference in the emulation.

Another popular mechanism available is the use of functions like InterlockedIncrement() and InterlockedExchange(). These make use of native CPU abilities to execute a single instruction while ensuring any other processors on the system cannot access memory, and allow you to do common operations like add/remove/check a variable in thread-safe code without holding a mutex. These are useful for reference counting especially in free-threaded (thread safe) COM objects.

Finally, the usage of TLS slots are also popular. TLS stands for thread-local storage, and is a set of slots scoped local to a thread which you can store pointers in. Look on MSDN for the TlsAlloc() function to learn more about the Win32 implementation of this. Essentially, the contents of a given slot will be different in each thread, so you can use this to store data that is only meaningful in the context of a single thread. On recent versions of Linux the __thread keyword provides a convenient interface to this functionality - a more portable API is exposed in the pthread library. However, these facilities are not used by Wine, rather, we implement Win32 TLS entirely ourselves.

8.4.2. The Win32 thread environment

All Win32 code, whether from a native EXE/DLL or in Wine itself, expects certain constructs to be present in its environment. This section explores what those constructs are and how Wine sets them up. The lack of this environment is one thing that makes it hard to use Wine code directly from standard Linux applications - in order to interact with Win32 code a thread must first be "adopted" by Wine.

The first thing Win32 code requires is the TEB or "Thread Environment Block". This is an internal (undocumented) Windows structure associated with every thread which stores a variety of things such as TLS slots, a pointer to the threads message queue, the last error code and so on. You can see the definition of the TEB in include/thread.h, or at least what we know of it so far. Being internal and subject to change, the layout of the TEB has had to be reverse engineered from scratch.

A pointer to the TEB is stored in the %fs register and can be accessed using NtCurrentTeb() from within Wine code. %fs actually stores a selector, and setting it therefore requires modifying the processes local descriptor table (LDT) - the code to do this is in lib/wine/ldt.c.

The TEB is required by nearly all Win32 code run in the Wine environment, as any wineserver RPC will use it, which in turn implies that any code which could possibly block for instance by using a critical section) needs it. The TEB also holds the SEH exception handler chain as the first element, so if disassembling you see code like this:

movl %esp, %fs:0

... then you are seeing the program set up an SEH handler frame. All threads must have at least one SEH entry, which normally points to the backstop handler which is ultimately responsible for popping up the all-too-familiar "This program has performed an illegal operation and will be terminated" message. On Wine we just drop straight into the debugger. A full description of SEH is out of the scope of this section, however there are some good articles in MSJ if you are interested.

All Win32-aware threads must have a wineserver connection. Many different APIs require the ability to communicate with the wineserver. In turn, the wineserver must be aware of Win32 threads in order to be able to accurately report information to other parts of the program and do things like route inter-thread messages, dispatch APCs (asynchronous procedure calls) and so on. Therefore a part of thread initialization is initializing the thread server-side. The result is not only correct information in the server, but a set of file descriptors the thread can use to communicate with the server - the request fd, reply fd and wait fd (used for blocking).

8.5. Structured Exception Handling

Structured Exception Handling (or SEH) is an implementation of exceptions inside the Windows core. It allows code written in different languages to throw exceptions across DLL boundaries, and Windows reports various errors like access violations by throwing them. This section looks at how it works, and how it's implemented in Wine.

8.5.1. How SEH works

SEH is based on embedding EXCEPTION_REGISTRATION_RECORD structures in the stack. Together they form a linked list rooted at offset zero in the TEB (see the threading section if you don't know what this is). A registration record points to a handler function, and when an exception is thrown the handlers are executed in turn. Each handler returns a code, and they can elect to either continue through the handler chain or it can handle the exception and then restart the program. This is referred to as unwinding the stack. After each handler is called it's popped off the chain.

Before the system begins unwinding the stack, it runs vectored handlers. This is an extension to SEH available in Windows XP, and allows registered functions to get a first chance to watch or deal with any exceptions thrown in the entire program, from any thread.

A thrown exception is represented by an EXCEPTION_RECORD structure. It consists of a code, flags, an address and an arbitrary number of DWORD parameters. Language runtimes can use these parameters to associate language-specific information with the exception.

Exceptions can be triggered by many things. They can be thrown explicitly by using the RaiseException API, or they can be triggered by a crash (i.e. translated from a signal). They may be used internally by a language runtime to implement language-specific exceptions. They can also be thrown across DCOM connections.

Visual C++ has various extensions to SEH which it uses to implement, e.g. object destruction on stack unwind as well as the ability to throw arbitrary types. The code for this is in dlls/msvcrt/except.c

8.5.2. Translating signals to exceptions

In Windows, compilers are expected to use the system exception interface, and the kernel itself uses the same interface to dynamically insert exceptions into a running program. By contrast on Linux the exception ABI is implemented at the compiler level (inside GCC and the linker) and the kernel tells a thread of exceptional events by sending signals. The language runtime may or may not translate these signals into native exceptions, but whatever happens the kernel does not care.

You may think that if an app crashes, it's game over and it really shouldn't matter how Wine handles this. It's what you might intuitively guess, but you'd be wrong. In fact some Windows programs expect to be able to crash themselves and recover later without the user noticing, some contain buggy binary-only components from third parties and use SEH to swallow crashes, and still others execute privileged (kernel-level) instructions and expect it to work. In fact, at least one set of APIs (the IsBad*Ptr() series) can only be implemented by performing an operation that may crash and returning TRUE if it does, and FALSE if it doesn't! So, Wine needs to not only implement the SEH infrastructure but also translate Unix signals into SEH exceptions.

The code to translate signals into exceptions is a part of NTDLL, and can be found in dlls/ntdll/signal_i386.c. This file sets up handlers for various signals during Wine startup, and for the ones that indicate exceptional conditions translates them into exceptions. Some signals are used by Wine internally and have nothing to do with SEH.

Signal handlers in Wine run on their own stack. Each thread has its own signal stack which resides 4k after the TEB. This is important for a couple of reasons. Firstly, because there's no guarantee that the app thread which triggered the signal has enough stack space for the Wine signal handling code. In Windows, if a thread hits the limits of its stack it triggers a fault on the stack guard page. The language runtime can use this to grow the stack if it wants to. However, because a guard page violation is just a regular segfault to the kernel, that would lead to a nested signal handler and that gets messy really quick so we disallow that in Wine. Secondly, setting up the exception to throw requires modifying the stack of the thread which triggered it, which is quite hard to do when you're still running on it.

Windows exceptions typically contain more information than the Unix standard APIs provide. For instance, a STATUS_ACCESS_VIOLATION exception (0xC0000005) structure contains the faulting address, whereas a standard Unix SIGSEGV just tells the app that it crashed. Usually this information is passed as an extra parameter to the signal handler, however its location and contents vary between kernels (BSD, Solaris, etc). This data is provided in a SIGCONTEXT structure, and on entry to the signal handler it contains the register state of the CPU before the signal was sent. Modifying it will cause the kernel to adjust the context before restarting the thread.

8.6. File management

With time, Windows API comes closer to the old Unix paradigm "Everything is a file". Therefore, this whole section dedicated to file management will cover firstly the file management, but also some other objects like directories, and even devices, which are manipulated in Windows in a rather coherent way. We'll see later on some other objects fitting (more or less) in this picture (pipes or consoles to name a few).

First of all, Wine, while implementing the file interface from Windows, needs to maps a file name (expressed in the Windows world) onto a file name in the Unix world. This encompasses several aspects: how to map the file names, how to map access rights (both on files and directories), how to map physical devices (hardisks, but also other devices - like serial or parallel interfaces - and even VxDs).

8.6.1. Various Windows formats for file names

Let's first review a bit the various forms Windows uses when it comes to file names. The DOS inheritance

At the beginning was DOS, where each file has to sit on a drive, called from a single letter. For separating device names from directory or file names, a ':' was appended to this single letter, hence giving the (in)-famous C: drive designations. Another great invention was to use some fixed names for accessing devices: not only where these named fixed, in a way you couldn't change the name if you'd wish to, but also, they were insensible to the location where you were using them. For example, it's well known that COM1 designates the first serial port, but it's also true that c:\foo\bar\com1 also designates the first serial port. It's still true today: on XP, you still cannot name a file COM1, whatever the directory!!!

Well later on (with Windows 95), Microsoft decided to overcome some little details in file names: this included being able to get out of the 8+3 format (8 letters for the name, 3 letters for the extension), and so being able to use "long names" (that's the "official" naming; as you can guess, the 8+3 format is a short name), and also to use very strange characters in a file name (like a space, or even a '.'). You could then name a file My File V0.1.txt, instead of myfile01.txt. Just to keep on the fun side of things, for many years the format used on the disk itself for storing the names has been the short name as the real one and to use some tricky aliasing techniques to store the long name. When some newer disk file systems have been introduced (NTFS with NT), in replacement of the old FAT system (which had little evolved since the first days of DOS), the long name became the real name while the short name took the alias role.

Windows also started to support mounting network shares, and see them as they were a local disk (through a specific drive letter). The way it has been done changed along the years, so we won't go into all the details (especially on the DOS and Win9x side). The NT way

The introduction of NT allowed a deep change in the ways DOS had been handling devices:

  • There's no longer a forest of DOS drive letters (even if the assign was a way to create symbolic links in the forest), but a single hierarchical space.

  • This hierarchy includes several distinct elements. For example, \Device\Hardisk0\Partition0 refers to the first partition on the first physical hard disk of the system.

  • This hierarchy covers way more than just the files and drives related objects, but most of the objects in the system. We'll only cover here the file related part.

  • This hierarchy is not directly accessible for the Win32 API, but only the NTDLL API. The Win32 API only allows to manipulate part of this hierarchy (the rest being hidden from the Win32 API). Of course, the part you see from Win32 API looks very similar to the one that DOS provided.

  • Mounting a disk is performed by creating a symbol link in this hierarchy from \Global??\C: (the name seen from the Win32 API) to \Device\Harddiskvolume1 which determines the partition on a physical disk where C: is going to be seen.

  • Network shares are also accessible through a symbol link. However in this case, a symbol link is created from \Global??\UNC\host\share\ for the share share on the machine host) to what's called a network redirector, and which will take care of 1/ the connection to the remote share, 2/ handling with that remote share the rest of the path (after the name of the server, and the name of the share on that server).

    Note: In NT naming convention, \Global?? can also be called \?? to shorten the access.

All of these things, make the NT system pretty much more flexible (you can add new types of filesystems if you want), you provide a unique name space for all objects, and most operations boil down to creating relationship between different objects. Wrap up

Let's end this chapter about files in Windows with a review of the different formats used for file names:

  • c:\foo\bar is a full path.

  • \foo\bar is an absolute path; the full path is created by appending the default drive (i.e. the drive of the current directory).

  • bar is a relative path; the full path is created by adding the current directory.

  • c:bar is a drive relative path. Note that the case where c: is the drive of the current directory is rather easy; it's implemented the same way as the case just below (relative path). In the rest of this chapter, drive relative path will only cover the case where the drive in the path isn't the drive of the default directory. The resolution of this to a full pathname defers according to the version of Windows, and some parameters. Let's take some time browsing through these issues. On Windows 9x (as well as on DOS), the system maintains a process wide set of default directories per drive. Hence, in this case, it will resolve c:bar to the default directory on drive c: plus file bar. Of course, the default per drive directory is updated each time a new current directory is set (only the current directory of the drive specified is modified). On Windows NT, things differ a bit. Since NT implements a namespace for file closer to a single tree (instead of 26 drives), having a current directory per drive is a bit awkward. Hence, Windows NT default behavior is to have only one current directory across all drives (in fact, a current directory expressed in the global tree) - this directory is of course related to a given process -, c:bar is resolved this way:

    • If c: is the drive of the default directory, the final path is the current directory plus bar.

    • Otherwise it's resolved into c:\bar.

    • In order to bridge the gap between the two implementations (Windows 9x and NT), NT adds a bit of complexity on the second case. If the =C: environment variable is defined, then it's value is used as a default directory for drive C:. This is handy, for example, when writing a DOS shell, where having a current drive per drive is still implemented, even on NT. This mechanism (through environment variables) is implemented on CMD.EXE, where those variables are set when you change directories with cd. Since environment variables are inherited at process creation, the current directories settings are inherited by child processes, hence mimicking the behavior of the old DOS shell. There's no mechanism (in NTDLL or KERNEL32) to set up, when current directory changes, the relevant environment variables. This behavior is clearly band-aid, not a full featured extension of current directory behavior.

    Wine fully implements all those behaviors (the Windows 9x vs NT ones are triggered by the version flag in Wine).

  • \\host\share is UNC (Universal Naming Convention) path, i.e. represents a file on a remote share.

  • \\.\device denotes a physical device installed in the system (as seen from the Win32 subsystem). A standard NT system will map it to the \??\device NT path. Then, as a standard configuration, \??\device is likely to be a link to in a physical device described and hooked into the \Device\ tree. For example, COM1 is a link to \Device\Serial0.

  • On some versions of Windows, paths were limited to MAX_PATH characters. To circumvent this, Microsoft allowed paths to be 32,767 characters long, under the conditions that the path is expressed in Unicode (no Ansi version), and that the path is prefixed with \\?\. This convention is applicable to any of the cases described above.

To summarize, what we've discussed so, let's put everything into a single table...

Table 8-1. DOS, Win32 and NT paths equivalences

Type of pathWin32 exampleNT equivalentRule to construct
Full pathc:\foo\bar.txt\Global??\C:\foo\bar.txtSimple concatenation
Absolute path\foo\bar.txt\Global??\J:\foo\bar.txt Simple concatenation using the drive of the default directory (here J:)
Relative pathgee\bar.txt \Global??\J:\mydir\mysubdir\gee\bar.txt Simple concatenation using the default directory (here J:\mydir\mysubdir)
Drive relative pathj:gee\bar.txt

  • On Windows 9x (and DOS), J:\toto\gee\bar.txt.

  • On Windows NT, J:\gee\bar.txt.

  • On Windows NT, J:\tata\titi\bar.txt.

  • On Windows 9x (and DOS), \toto is the default directory on drive J:.

  • On Windows NT, if =J: isn't set.

  • On Windows NT, if =J: is set to J:\tata\titi.

UNC (Uniform Naming Convention) path\\host\share\foo\bar.txt \Global??\UNC\host\share\foo\bar.txt Simple concatenation.
Device path\\.\device\Global??\deviceSimple concatenation
Long paths\\?\...  With this prefix, paths can take up to 32,767 characters, instead of MAX_PATH for all the others). Once the prefix stripped, to be handled like one of the previous ones, just providing internal buffers large enough).

8.6.2. Wine implementation

We'll mainly cover in this section the way Wine opens a file (in the Unix sense) when given a Windows file name. This will include mapping the Windows path onto a Unix path (including the devices case), handling the access rights, the sharing attribute if any... Mapping a Windows path into an absolute Windows path

First of all, we described in previous section the way to convert any path in an absolute path. Wine implements all the previous algorithms in order to achieve this. Note also, that this transformation is done with information local to the process (default directory, environment variables...). We'll assume in the rest of this section that all paths have now been transformed into absolute from. Mapping a Windows (absolute) path onto a Unix path

When Wine is requested to map a path name (in DOS form, with a drive letter, e.g. c:\foo\bar\myfile.txt), Wine converts this into the following Unix path $(WINEPREFIX)/dosdevices/c:/foo/bar/myfile.txt. The Wine configuration process is responsible for setting $(WINEPREFIX)/dosdevices/c: to be a symbolic link pointing to the directory in Unix hierarchy the user wants to expose as the C: drive in the DOS forest of drives.

This scheme allows:

  • a very simple algorithm to map a DOS path name into a Unix one (no need of Wine server calls)

  • a very configurable implementation: it's very easy to change a drive mapping

  • a rather readable configuration: no need of sophisticated tools to read a drive mapping, a ls -l $(WINEPREFIX)/dosdevices says it all.

This scheme is also used to implement UNC path names. For example, Wine maps \\host\share\foo\bar\MyRemoteFile.txt into $(WINEPREFIX)/dosdevices/unc/host/share/foo/bar/MyRemoteFile.txt. It's then up to the user to decide where $(WINEPREFIX)/dosdevices/unc/host/share shall point to (or be). For example, it can either be a symbolic link to a directory inside the local machine (just for emulation purpose), or a symbolic link to the mount point of a remote disk (done through Samba or NFS), or even the real mount point. Wine will not do any checking here, nor will help in actually mounting the remote drive.

We've seen how Wine maps a drive letter or a UNC path onto the Unix hierarchy, we now have to look at how the filename is searched within this hierarchy. The main issue is about case sensitivity. Here's a reminder of the various properties for the file systems in the field.

Table 8-2. File systems' properties

FS NameLength of elementsCase sensitivity (on disk)Case sensitivity for lookup
FAT, FAT16 or FAT32 Short name (8+3)Names are always stored in upper-caseCase insensitive
VFATShort name (8+3) + alias on long name Short names are always stored in upper-case. Long names are stored with case preservation. Case insensitive
NTFSLong name + alias on short name (8+3). Long names are stored with case preservation. Short names are always stored in upper-case. Case insensitivity
Linux FS (ext2fs, ext3fs, reiserfs...) Long nameCase preservingCase sensitive

Case sensitivity vs. preservation: When we say that most systems in NT are case insensitive, this has to be understood for looking up for a file, where the matches are made in a case insensitive mode. This is different from VFAT or NTFS "case preservation" mechanism, which stores the file names as they are given when creating the file, while doing case insensitive matches.

Since most file systems used in NT are case insensitive and since most Unix file systems are case sensitive, Wine undergoes a case insensitive search when it has found the Unix path is has to look for. This means, for example, that for opening the $(WINEPREFIX)/dosdevices/c:/foo/bar/myfile.txt, Wine will recursively open all directories in the path, and check, in this order, for the existence of the directory entry in the form given in the file name (i.e. case sensitive), and if it's not found, in a case insensitive form. This allows to also pass, in most Win32 file API, a Unix path (instead of a DOS or NT path), but we'll come back to this later. This also means that the algorithm described doesn't correctly handle the case of two files in the same directory, whose names only differ on the case of the letters. This means that if there are two files in the same directory whose names match in a case sensitive comparison, Wine will pick up the right one if the filename given matches one of the names (in a case sensitive way), but will pickup one of the two (without defining the one it's going to pickup) if the filename given matches none of the two names in a case sensitive way (but in a case insensitive way). For example, if the two filenames are my_neat_file.txt and My_Neat_File.txt, Wine behavior when opening MY_neat_FILE.txt is undefined.

As Windows, at the early days, didn't support the notion of symbolic links on directories, lots of applications (and some old native DLLs) are not ready for this feature. Mainly, they imply that the directory structure is a tree, which has lots of consequences on navigating in the forest of directories (i.e. there cannot be two ways for going from directory to another, there cannot be cycles...). In order to prevent some bad behavior for such applications, Wine sets up an option. By default, symbolic links on directories are not followed by Wine. There's an option to follow them (see the Wine User Guide), but this could be harmful.

Wine considers that Unix file names are long filename. This seems a reasonable approach; this is also the approach followed by most of the Unix OSes while mounting Windows partitions (with filesystems like FAT, FAT32 or NTFS). Therefore, Wine tries to support short names the best it can. Basically, they are two options:

  • The filesystem on which the inspected directory lies in a real Windows FS (like FAT, or FAT32, or NTFS) and the OS has support to access the short filename (for example, Linux does this on FAT, FAT32 or VFAT). In this case, Wine makes full use of this information and really mimics the Windows behavior: the short filename used for any file is the same than on Windows.

  • If conditions listed above are not met (either, FS has no physical short name support, or OS doesn't provide the access access to the short name), Wine decides and computes on its own the short filename for a given long filename. We cannot ensure that the generated short name is the same than on Windows (because the algorithm on Windows takes into account the order of creation of files, which cannot be implemented in Wine: Wine would have to cache the short names of every directory it uses!). The short name is made up of part of the long name (first characters) and the rest with a hashed value. This has several advantages:

    • The algorithm is rather simple and low cost.

    • The algorithm is stateless (doesn't depend of the other files in the directory).

    But, it also has the drawbacks (of the advantages):

    • The algorithm isn't the same as on Windows, which means a program cannot use short names generated on Windows. This could happen when copying an existing installed program from Windows (for example, on a dual boot machine).

    • Two long file names can end up with the same short name (Windows handles the collision in this case, while Wine doesn't). We rely on our hash algorithm to lower at most this possibility (even if it exists).

Wine also allows in most file API to give as a parameter a full Unix path name. This is handy when running a Wine (or Winelib) program from the command line, and one doesn't need to convert the path into the Windows form. However, Wine checks that the Unix path given can be accessed from one of the defined drives, insuring that only part of the Unix / hierarchy can be accessed.

As a side note, as Unix doesn't widely provide a Unicode interface to the filenames, and Windows implements filenames as Unicode strings (even on the physical layer with NTFS, the FATs variant are ANSI), we need to properly map between the two. At startup, Wine defines what's called the Unix Code Page, that's is the code page the Unix kernel uses as a reference for the strings. Then Wine uses this code page for all the mappings it has to do between a Unicode path (on the Windows side) and a Ansi path to be used in a Unix path API. Note, that this will work as long as a disk isn't mounted with a different code page than the one the kernel uses as a default.

We describe below how Windows devices are mapped to Unix devices. Before that, let's finish the pure file round-up with some basic operations. Access rights and file attributes

Now that we have looked how Wine converts a Windows pathname into a Unix one, we need to cover the various meta-data attached to a file or a directory.

In Windows, access rights are simplistic: a file can be read-only or read-write. Wine sets the read-only flag if the file doesn't have the Unix user-write flag set. As a matter of fact, there's no way Wine can return that a file cannot be read (that doesn't exist under Windows). The file will be seen, but trying to open it will return an error. The Unix exec-flag is never reported. Wine doesn't use this information to allow/forbid running a new process (as Unix does with the exec-flag). Last but not least: hidden files. This exists on Windows but not really on Unix! To be exact, in Windows, the hidden flag is a metadata associated to any file or directory; in Unix, it's a convention based on the syntax of the file name (whether it starts with a '.' or not). Wine implements two behaviors (chosen by configuration). This impacts file names and directory names starting by a '.'. In first mode (ShowDotFile is FALSE), every file or directory starting by '.' is returned with the hidden flag turned on. This is the natural behavior on Unix (for ls or even file explorer). In the second mode (ShowDotFile is TRUE), Wine never sets the hidden flag, hence every file will be seen.

Last but not least, before opening a file, Windows makes use of sharing attributes in order to check whether the file can be opened; for example, a process, being the first in the system to open a given file, could forbid, while it maintains the file opened, that another process opens it for write access, whereas open for read access would be granted. This is fully supported in Wine by moving all those checks in the Wine server for a global view on the system. Note also that what's moved in the Wine server is the check, when the file is opened, to implement the Windows sharing semantics. Further operation on the file (like reading and writing) will not require heavy support from the server.

The other good reason for putting the code for actually opening a file in the server is that an opened file in Windows is managed through a handle, and handles can only be created in Wine server!

Just a note about attributes on directories: while we can easily map the meaning of Windows FILE_ATTRIBUTE_READONLY on a file, we cannot do it for a directory. Windows semantics (when this flag is set) means do not delete the directory, while the w attribute in Unix means don't write nor delete it. Therefore, Wine uses an asymmetric mapping here: if the directory (in Unix) isn't writable, then Wine reports the FILE_ATTRIBUTE_READONLY attribute; on the other way around, when asked to set a directory with FILE_ATTRIBUTE_READONLY attribute, Wine simply does nothing. Operations on file Reading and writing

Reading and writing are the basic operations on files. Wine of course implements this, and bases the implementation on client side calls to Unix equivalents (like read() or write()). Note, that the Wine server is involved in any read or write operation, as Wine needs to transform the Windows-handle to the file into a Unix file descriptor it can pass to any Unix file function. Getting a Unix fd

This is major operation in any file related operation. Basically, each file opened (at the Windows level), is first opened in the Wine server, where the fd is stored. Then, Wine (on client side) uses recvmsg() to pass the fd from the wine server process to the client process. Since this operation could be lengthy, Wine implement some kind of cache mechanism to send it only once, but getting a fd from a handle on a file (or any other Unix object which can be manipulated through a file descriptor) still requires a round trip to the Wine server. Locking

Windows provides file locking capabilities. When a lock is set (and a lock can be set on any contiguous range in a file), it controls how other processes in the system will have access to the range in the file. Since locking range on a file are defined on a system wide manner, its implementation resides in wineserver. It tries to make use of Unix file locking (if the underlying OS and the mounted disk where the file sits support this feature) with fcntl() and the F_SETLK command. If this isn't supported, then wineserver just pretends it works. I/O control

There's no need (so far) to implement support (for files and directories) for DeviceIoControl(), even if this is supported by Windows, but for very specific needs (like compression management, or file system related information). This isn't the case for devices (including disks), but we'll cover this in the hereafter section related to devices. Buffering

Wine doesn't do any buffering on file accesses but rely on the underlying Unix kernel for that (when possible). This scheme is needed because it's easier to implement multiple accesses on the same file at the kernel level, rather than at Wine levels. Doing lots of small reads on the same file can turn into a performance hog, because each read operation needs a round trip to the server in order to get a file descriptor (see above). Overlapped I/O

Windows introduced the notion of overlapped I/O. Basically, it just means that an I/O operation (think read/write to start with) will not wait until it's completed, but rather return to the caller as soon as possible, and let the caller handle the wait operation and determine when the data is ready (for a read operation) or has been sent (for a write operation). Note that the overlapped operation is linked to a specific thread.

There are several interests to this: a server can handle several clients without requiring multi-threading techniques; you can handle an event driven model more easily (i.e. how to kill properly a server while waiting in the lengthy read() operation).

Note that Microsoft's support for this feature evolved along the various versions of Windows. For example, Windows 95 or 98 only supports overlapped I/O for serial and parallel ports, while NT also supports files, disks, sockets, pipes, or mailslots.

Wine implements overlapped I/O operations. This is mainly done by queuing in the server a request that will be triggered when something the current state changes (like data available for a read operation). This readiness is signaled to the calling processing by queuing a specific APC, which will be called within the next waiting operation the thread will have. This specific APC will then do the hard work of the I/O operation. This scheme allows to put in place a wait mechanism, to attach a routine to be called (on the thread context) when the state changes, and to be done is a rather transparent manner (embedded any the generic wait operation). However, it isn't 100% perfect. As the heavy operations are done in the context of the calling threads, if those operations are lengthy, there will be an impact on the calling thread, especially its latency. In order to provide an effective support for this overlapped I/O operations, we would need to rely on Unix kernel features (AIO is a good example). Devices & volume management

We've covered so far the ways file names are mapped into Unix paths. There's still need to cover it for devices. As a regular file, devices are manipulated in Windows with both read / write operations, but also control mechanisms (speed or parity of a serial line, volume name of a hard disk, ...). Since this is also supported in Linux, there's also a need to open (in a Unix sense) a device when given a Windows device name. This section applies to DOS device names, which are seen in NT as nicknames to other devices.

Firstly, Wine implements the Win32 to NT mapping as described above, hence every device path (in NT sense) is of the following form: /??/devicename (or /DosDevices/devicename). As Windows device names are case insensitive, Wine also converts them to lower case before any operation. Then, the first operation Wine tries is to check whether $(WINEPREFIX)/dosdevices/devicename exists. If so, it's used as the final Unix path for the device. The configuration process is in charge of creating for example, a symbolic link between $(WINEPREFIX)/dosdevices/PhysicalDrive0 and /dev/hda0. If such a link cannot be found, and the device name looks like a DOS disk name (like C:), Wine first tries to get the Unix device from the path $(WINEPREFIX)/dosdevices/c: (i.e. the device which is mounted on the target of the symbol link); if this doesn't give a Unix device, Wine checks whether $(WINEPREFIX)/dosdevices/c:: exists. If so, it's assumed to be a link to the actual Unix device. For example, for a CD Rom, $(WINEPREFIX)/dosdevices/e:: would be a symbolic link to /dev/cdrom. If this doesn't exist (we're still handling a device name of the C: form), Wine tries to get the Unix device from the system information (/etc/mtab and /etc/fstab on Linux). We cannot apply this method in all the cases, because we have no insurance that the directory can actually be found. One could have, for example, a CD Rom which he/she want only to use as audio CD player (i.e. never mounted), thus not having any information of the device itself. If all of this doesn't work either, some basic operations are checked: if the devicename is NUL, then /dev/null is returned. If the device name is a default serial name (COM1 up to COM9) (resp. printer name LPT1 up to LPT9), then Wine tries to open the Nth serial (resp. printer) in the system. Otherwise, some basic old DOS name support is done AUX is transformed into COM1 and PRN into LPT1), and the whole process is retried with those new names.

To sum up:

Table 8-3. Mapping of Windows device names into Unix device names

Windows device nameNT device nameMapping to Unix device name
any_pathAUX\Global??\AUX Treated as an alias to COM1
any_pathPRN\Global??\PRNTreated as an alias to LPT1
any_pathCOMN\Global??\COMN $(WINEPREFIX)/dosdevices/comN (if the symbol link exists) or the Nth serial line in the system (on Linux, /dev/ttySN-1).
any_pathLPTN\Global??\LPTN $(WINEPREFIX)/dosdevices/lptN (if the symbol link exists) or the Nth printer in the system (on Linux, /dev/lpN-1).
\\.\E:\Global??\E: $(WINEPREFIX)/dosdevices/e:: (if the symbolic link exists) or guessing the device from /etc/mtab or /etc/fstab.
\\.\device_name \Global??\device_name $(WINEPREFIX)/dosdevices/device_name (if the symbol link exists).

Now that we know which Unix device to open for a given Windows device, let's cover the operation on it. Those operations can either be read/write, IO control (and even others).

Read and write operations are supported on real disks & CDROM devices, under several conditions:

  • Foremost, as the ReadFile() and WriteFile() calls are mapped onto the Unix read() and write() calls, the user (from the Unix perspective of the one running the Wine executable) must have read (resp. write) access to the device. It wouldn't be wise to let a user write directly to a hard disk!!!

  • Block sizes for read and write must be of the size of a physical block (generally 512 for a hard disk, depending on the type of CD used), and offsets must also be a multiple of the block size.

Wine also reads (if the first condition above about access rights is met) the volume information from a hard disk or a CD ROM to be displayed to a user.

Wine also recognizes VxDs as devices. But those VxDs must be the Wine builtin ones (Wine will never allow to load native VxDs). Those are configured with symbolic links in the $(WINEPREFIX)/dosdevices/ directory, and point to the actual builtin DLL. This DLL exports a single entry point, that Wine will use when a call to DeviceIoControl is made, with a handle opened to this VxD. This allows to provide some kind of compatibility for old Win9x apps, still talking directly to VxD. This is no longer supported on Windows NT, newest programs are less likely to make use of this feature, so we don't expect lots of development in this area, even though the framework is there and working. Note also that Wine doesn't provide support for native VxDs (as a game, report how many times this information is written in the documentation; as an advanced exercise, find how many more occurrences we need in order to stop questions whether it's possible or not).

8.7. NTDLL module

NTDLL provides most of the services you'd expect from a kernel. In lots of cases, KERNEL32 APIs are just wrappers to NTDLL APIs. There are however, some difference in the APIs (the NTDLL ones have quite often a bit wider semantics than their KERNEL32 counterparts). All the detailed functions we've described since the beginning of this chapter are in fact implemented in NTDLL, plus a great numbers of others we haven't written about yet.

8.8. KERNEL32 Module

As already explained, KERNEL32 maps quite a few of its APIs to NTDLL. There are however a couple of things which are handled directly in KERNEL32. Let's cover a few of them...

8.8.1. Console NT implementation

Windows implements console solely in the Win32 subsystem. Under NT, the real implementation uses a dedicated subsystem csrss.exe Client/Server Run-time SubSystem) which is in charge, among other things, of animating the consoles. Animating includes for example handling several processes on the same console (write operations must be atomic, but also a character keyed on the console must be read by a single process), or sending some information back to the processes (changing the size or attributes of the console, closing the console). Windows NT uses a dedicated (RPC based) protocol between each process being attached to a console and the csrss.exe subsystem, which is in charge of the UI of every console in the system. Wine implementation

Wine tries to integrate as much as possible into the Unix consoles, but the overall situation isn't perfect yet. Basically, Wine implements three kinds of consoles:

  • the first one is a direct mapping of the Unix console into the Windows environment. From the windows program point of view, it won't run in a Windows console, but it will see its standard input and output streams redirected to files; those files are hooked into the Unix console's output and input streams respectively. This is handy for running programs from a Unix command line (and use the result of the program as it was a Unix programs), but it lacks all the semantics of the Windows consoles.

  • the second and third ones are closer to the NT scheme, albeit different from what NT does. The wineserver plays the role of the csrss.exe subsystem (all requests are sent to it), and are then dispatched to a dedicated wine process, called (surprise!) wineconsole which manages the UI of the console. There is a running instance of wineconsole for every console in the system. Two flavors of this scheme are actually implemented: they vary on the backend for the wineconsole. The first one, dubbed user, creates a real GUI window (hence the USER name) and renders the console in this window. The second one uses the (n)curses library to take full control of an existing Unix console; of course, interaction with other Unix programs will not be as smooth as the first solution.

The following table describes the main implementation differences between the three approaches.

Table 8-4. Function consoles implementation comparison

FunctionBare streams Wineconsole with user backend Wineconsole with curses backend
Console as a Win32 Object (and associated handles) No specific Win32 object is used in this case. The handles manipulated for the standard Win32 streams are in fact "bare handles" to their corresponding Unix streams. The mode manipulation functions (GetConsoleMode() / SetConsoleMode()) are not supported. Implemented in server, and a specific Winelib program (wineconsole) is in charge of the rendering and user input. The mode manipulation functions behave as expected. Implemented in server, and a specific Winelib program (wineconsole) is in charge of the rendering and user input. The mode manipulation functions behave as expected.
Inheritance (including handling in CreateProcess() of CREATE_DETACHED, CREATE_NEW_CONSOLE flags). Not supported. Every process child of a process will inherit the Unix streams, so will also inherit the Win32 standard streams. Fully supported (each new console creation will be handled by the creation of a new USER32 window) Fully supported, except for the creation of a new console, which will be rendered on the same Unix terminal as the previous one, leading to unpredictable results.
ReadFile() / WriteFile() operations Fully supportedFully supportedFully supported
Screen-buffer manipulation (creation, deletion, resizing...) Not supportedFully supported Partly supported (this won't work too well as we don't control (so far) the size of underlying Unix terminal
APIs for reading/writing screen-buffer content, cursor position Not supportedFully supportedFully supported
APIs for manipulating the rendering window sizeNot supportedFully supported Partly supported (this won't work too well as we don't control (so far) the size of underlying Unix terminal
Signaling (in particular, Ctrl-C handling) Nothing is done, which means that Ctrl-C will generate (as usual) a SIGINT which will terminate the program. Partly supported (Ctrl-C behaves as expected, however the other Win32 CUI signaling isn't properly implemented). Partly supported (Ctrl-C behaves as expected, however the other Win32 CUI signaling isn't properly implemented).

The Win32 objects behind a console can be created in several occasions:

  • When the program is started from wineconsole, a new console object is created and will be used (inherited) by the process launched from wineconsole.

  • When a program, which isn't attached to a console, calls AllocConsole(), Wine then launches wineconsole, and attaches the current program to this console. In this mode, the USER32 mode is always selected as Wine cannot tell the current state of the Unix console.

Please also note, that starting a child process with the CREATE_NEW_CONSOLE flag, will end-up calling AllocConsole() in the child process, hence creating a wineconsole with the USER32 backend.

Another interesting point to note is that Windows implements handles to console objects (input and screen buffers) only in the KERNEL32 DLL, and those are not sent nor seen from the NTDLL level, albeit, for example, console are waitable on input. How is this possible? Well, Windows NT is a bit tricky here. Regular handles have an interesting property: their integral value is always a multiple of four (they are likely to be offsets from the beginning of a table). Console handles, on the other hand, are not multiple of four, but have the two lower bit set (being a multiple of four means having the two lower bits reset). When KERNEL32 sees a handle with the two lower bits set, it then knows it's a console handle and takes appropriate decisions. For example, in the various kernel32!WaitFor*() functions, it transforms any console handle (input and output - strangely enough handles to console screen buffers are waitable) into a dedicated wait event for the targetted console. There's an (undocumented) KERNEL32 function GetConsoleInputWaitHandle() which returns the handle to this event in case you need it. Another interesting handling of those console handles is in ReadFile() (resp. WriteFile()), which behavior, for console handles, is transferred to ReadConsole() (resp. WriteConsole()). Note that's always the ANSI version of ReadConsole() / WriteConsole() which is called, hence using the default console's code page. There are some other spots affected, but you can look in dlls/kernel32 to find them all. All of this is implemented in Wine.

Wine also implements the same layout of the registry for storing the preferences of the console as Windows does. Those settings can either be defined globally, or on a per process name basis. wineconsole provides the choice to the user to pick you which registry part (global, current running program) it wishes to modify the settings for.

Table 8-5. Console registry settings

NameDefault valuePurpose
CursorSize25 Percentage of cell height to which the cursor extents
CursorVisible1Whether the cursor is visible or not
EditionMode0 The way the edition takes place in the console: 0 is insertion mode, 1 is overwrite mode.
ExitOnDie1 Whether the console should close itself when last running program attached to it dies
FaceNameNo default Name of the font to be used for display. When none is given, wineconsole tries its best to pick up a decent font
FontSize0x0C08 The high word in the font cell height, and the low word is the font cell width. The default value is 12 pixels in height and 8 pixels in width.
FontWeight0 Weigth of the font. If none is given (or 0) wineconsole picks up a decent font size
HistoryBufferSize50 Number of entries in history buffer (not actually used)
HistoryNoDup0 Whether the history should store twice the same entry
MenuMask0 This mask only exists for Wine console handling. It allows to know which combination of extra keys is needed to open the configuration window on right click. The mask can include MK_CONTROL or MK_SHIFT bits. This can be needed when programs actually need the right click to be passed to them instead of being intercepted by wineconsole.
QuickEdit0 If null, mouse events are sent to the application. If non null, mouse events are used to select text on the window. This setting must really be set on a application per application basis, because it deals with the fact the CUI application will use or not the mouse events.
ScreenBufferSize0x1950 The high word is the number of font cells in the height of the screen buffer, while the low word is the number of font cells in the width of the screen buffer.
ScreenColors0x000F Default color attribute for the screen buffer (low char is the foreground color, and high char is the background color)
WindowSize0x1950 The high word is the number of font cells in the height of the window, while the low word is the number of font cells in the width of the window. This window is the visible part of the screen buffer: this implies that a screen buffer must always be bigger than its window, and that the screen buffer can be scrolled so that every cell of the screen buffer can be seen in the window.

8.8.2. Win16 processes support Starting a NE (Win16) process

Wine is also able to run 16-bit processes, but this feature is only supported on Intel IA-32 architectures.

When Wine is requested to run a NE (Win 16 process), it will in fact hand over the execution of it to a specific executable winevdm. VDM stands for Virtual DOS Machine. This winevdm is a Winelib application, but will in fact set up the correct 16-bit environment to run the executable. We will get back later on in details to what this means.

Any new 16-bit process created by this executable (or its children) will run into the same winevdm instance. Among one instance, several functionalities will be provided to those 16-bit processes, including the cooperative multitasking, sharing the same address space, managing the selectors for the 16-bit segments needed for code, data and stack.

Note that several winevdm instances can run in the same Wine session, but the functionalities described above are only shared among a given instance, not among all the instances. winevdm is built as Winelib application, and hence has access to any facility a 32-bit application has.

Each Win16 application is implemented in winevdm as a Win32 thread. winevdm then implements its own scheduling facilities (in fact, the code for this feature is in the krnl386.exe DLL). Since the required Win16 scheduling is non pre-emptive, this doesn't require any underlying OS kernel support. SysLevels

SysLevels are an undocumented Windows-internal thread-safety system dedicated to 16-bit applications (or 32-bit applications that call - directly or indirectly - 16-bit code). They are basically critical sections which must be taken in a particular order. The mechanism is generic but there are always three syslevels:

  • level 1 is the Win16 mutex,

  • level 2 is the USER mutex,

  • level 3 is the GDI mutex.

When entering a syslevel, the code (in dlls/krnl386.exe16/syslevel.c) will check that a higher syslevel is not already held and produce an error if so. This is because it's not legal to enter level 2 while holding level 3 - first, you must leave level 3.

Throughout the code you may see calls to _ConfirmSysLevel() and _CheckNotSysLevel(). These functions are essentially assertions about the syslevel states and can be used to check that the rules have not been accidentally violated. In particular, _CheckNotSysLevel() will break probably into the debugger) if the check fails. If this happens the solution is to get a backtrace and find out, by reading the source of the wine functions called along the way, how Wine got into the invalid state. Memory management

Every Win16 address is expressed in the form of selector:offset. The selector is an entry in the LDT, but a 16-bit entry, limiting each offset to 64 KB. Hence, the maximum available memory to a Win16 process is 512 MB. Note, that the processor runs in protected mode, but using 16-bit selectors.

Windows, for a 16-bit process, defines a few selectors to access the "real" memory (the one provided) by DOS. Basically, Wine also provides this area of memory.

8.8.3. DOS processes support

The behaviour we just described also applies to DOS executables, which are handled the same way by winevdm. This is only supported on Intel IA-32 architectures.

Wine implements also most of the DOS support in a Wine specific DLL (winedos). This DLL is called under certain conditions, like:

  • In winevdm, when trying to launch a DOS application (.EXE or .COM, .PIF).

  • In kernel32, when an attempt is made in the binary code to call some DOS or BIOS interrupts (like Int 21h for example).

When winevdm runs a DOS program, this one runs in real mode (in fact in V86 mode from the IA-32 point of view).

Wine also supports part of the DPMI (DOS Protected Mode Interface).

Wine, when running a DOS programs, needs to map the 1 MB of virtual memory to the real memory (as seen by the DOS program). When this is not possible (like when someone else is already using this area), the DOS support is not possible. Note also that by doing so, access to linear address 0 is enabled (as it's also real mode address 0 which is valid). Hence, NULL pointer dereference faults are no longer caught.

Chapter 9. Windowing system

9.1. USER32 Module

USER32 implements windowing and messaging subsystems. It also contains code for common controls and for other miscellaneous stuff (rectangles, clipboard, WNet, etc). Wine USER32 code is located in the dlls/user32/ directory.

9.1.1. Windowing subsystem



Windows are arranged into parent/child hierarchy with one common ancestor for all windows (desktop window). Each window structure contains a pointer to the immediate ancestor (parent window if WS_CHILD style bit is set), a pointer to the sibling (returned by GetWindow(..., GW_NEXT)), a pointer to the owner window (set only for popup window if it was created with valid hwndParent parameter), and a pointer to the first child window (GetWindow(.., GW_CHILD)). All popup and non-child windows are therefore placed in the first level of this hierarchy and their ancestor link (wnd->parent) points to the desktop window.

Desktop window                       - root window
 |     \      `-.
 |      \        `-.
popup -> wnd1  ->  wnd2i             - top level windows
 |        \  `-.      `-.
 |         \     `-.      `-.
child1   child2 -> child3  child4    - child windows

Horizontal arrows denote sibling relationship, vertical lines - ancestor/child. To summarize, all windows with the same immediate ancestor are sibling windows, all windows which do not have desktop as their immediate ancestor are child windows. Popup windows behave as topmost top-level windows unless they are owned. In this case the only requirement is that they must precede their owners in the top-level sibling list (they are not topmost). Child windows are confined to the client area of their parent windows (client area is where window gets to do its own drawing, non-client area consists of caption, menu, borders, intrinsic scrollbars, and minimize/maximize/close/help buttons).

Another fairly important concept is z-order. It is derived from the ancestor/child hierarchy and is used to determine "above/below" relationship. For instance, in the example above, z-order is


Current active window ("foreground window" in Win32) is moved to the front of z-order unless its top-level ancestor owns popup windows.

All these issues are dealt with (or supposed to be) in dlls/user32/winpos.c with SetWindowPos() being the primary interface to the window manager.

Wine specifics: in default and managed mode each top-level window gets its own X counterpart with desktop window being basically a fake stub. In desktop mode, however, only desktop window has an X window associated with it. Also, SetWindowPos() should eventually be implemented via Begin/End/DeferWindowPos() calls and not the other way around. Visible region, clipping region and update region



|_________               |  A and B are child windows of C
|    A    |______        |
|         |      |       |
|---------'      |       |
|   |      B     |       |
|   |            |       |
|   `------------'       |
|                   C    |

Visible region determines which part of the window is not obscured by other windows. If a window has the WS_CLIPCHILDREN style then all areas below its children are considered invisible. Similarly, if the WS_CLIPSIBLINGS bit is in effect then all areas obscured by its siblings are invisible. Child windows are always clipped by the boundaries of their parent windows.

B has a WS_CLIPSIBLINGS style:

 .          ______
 :         |      |
 |   ,-----'      |
 |   |      B     | - visible region of B
 |   |            |
 :   `------------'

When the program requests a display context (DC) for a window it can specify an optional clipping region that further restricts the area where the graphics output can appear. This area is calculated as an intersection of the visible region and a clipping region.

Program asked for a DC with a clipping region:

   ,--|--.   |     .    ,--.
,--+--'  |   |     :   _:  |
|  |   B |   |  => |  |    | - DC region where the painting will
|  |     |   |     |  |    |   be visible
`--|-----|---'     :  `----'

When the window manager detects that some part of the window became visible it adds this area to the update region of this window and then generates WM_ERASEBKGND and WM_PAINT messages. In addition, WM_NCPAINT message is sent when the uncovered area intersects a nonclient part of the window. Application must reply to the WM_PAINT message by calling the BeginPaint()/EndPaint() pair of functions. BeginPaint() returns a DC that uses accumulated update region as a clipping region. This operation cleans up invalidated area and the window will not receive another WM_PAINT until the window manager creates a new update region.

A was moved to the left:

 ________________________       ...          / C update region
|______                  |     :      .___ /
| A    |_________        |  => |   ...|___|..
|      |         |       |     |   :  |   |
|------'         |       |     |   :  '---'
|   |      B     |       |     |   :      \
|   |            |       |     :            \
|   `------------'       |                    B update region
|                   C    |

Windows maintains a display context cache consisting of entries that include the DC itself, the window to which it belongs, and an optional clipping region (visible region is stored in the DC itself). When an API call changes the state of the window tree, window manager has to go through the DC cache to recalculate visible regions for entries whose windows were involved in the operation. DC entries (DCE) can be either private to the window, or private to the window class, or shared between all windows (Windows 3.1 limits the number of shared DCEs to 5).

9.1.2. Messaging subsystem


Each Windows task/thread has its own message queue - this is where it gets messages from. Messages can be:

  1. generated on the fly (WM_PAINT, WM_NCPAINT, WM_TIMER)

  2. created by the system (hardware messages)

  3. posted by other tasks/threads (PostMessage)

  4. sent by other tasks/threads (SendMessage)

Message priority:

First the system looks for sent messages, then for posted messages, then for hardware messages, then it checks if the queue has the "dirty window" bit set, and, finally, it checks for expired timers. See dlls/user32/message.c.

From all these different types of messages, only posted messages go directly into the private message queue. System messages (even in Win95) are first collected in the system message queue and then they either sit there until Get/PeekMessage gets to process them or, as in Win95, if system queue is getting clobbered, a special thread ("raw input thread") assigns them to the private queues. Sent messages are queued separately and the sender sleeps until it gets a reply. Special messages are generated on the fly depending on the window/queue state. If the window update region is not empty, the system sets the QS_PAINT bit in the owning queue and eventually this window receives a WM_PAINT message (WM_NCPAINT too if the update region intersects with the non-client area). A timer event is raised when one of the queue timers expire. Depending on the timer parameters DispatchMessage either calls the callback function or the window procedure. If there are no messages pending the task/thread sleeps until messages appear.

There are several tricky moments (open for discussion):

  • System message order has to be honored and messages should be processed within correct task/thread context. Therefore when Get/PeekMessage encounters unassigned system message and this message appears not to be for the current task/thread it should either skip it (or get rid of it by moving it into the private message queue of the target task/thread - Win95, AFAIK) and look further or roll back and then yield until this message gets processed when system switches to the correct context (Win16). In the first case we lose correct message ordering, in the second case we have the infamous synchronous system message queue. Here is a post to one of the OS/2 newsgroup I found to be relevant:


    Here's the problem in a nutshell, and there is no good solution. Every possible solution creates a different problem.

    With a windowing system, events can go to many different windows. Most are sent by applications or by the OS when things relating to that window happen (like repainting, timers, etc.)

    Mouse input events go to the window you click on (unless some window captures the mouse).

    So far, no problem. Whenever an event happens, you put a message on the target window's message queue. Every process has a message queue. If the process queue fills up, the messages back up onto the system queue.

    This is the first cause of apps hanging the GUI. If an app doesn't handle messages and they back up into the system queue, other apps can't get any more messages. The reason is that the next message in line can't go anywhere, and the system won't skip over it.

    This can be fixed by making apps have bigger private message queues. The SIQ fix does this. PMQSIZE does this for systems without the SIQ fix. Applications can also request large queues on their own.

    Another source of the problem, however, happens when you include keyboard events. When you press a key, there's no easy way to know what window the keystroke message should be delivered to.

    Most windowing systems use a concept known as "focus". The window with focus gets all incoming keyboard messages. Focus can be changed from window to window by apps or by users clicking on windows.

    This is the second source of the problem. Suppose window A has focus. You click on window B and start typing before the window gets focus. Where should the keystrokes go? On the one hand, they should go to A until the focus actually changes to B. On the other hand, you probably want the keystrokes to go to B, since you clicked there first.

    OS/2's solution is that when a focus-changing event happens (like clicking on a window), OS/2 holds all messages in the system queue until the focus change actually happens. This way, subsequent keystrokes go to the window you clicked on, even if it takes a while for that window to get focus.

    The downside is that if the window takes a real long time to get focus (maybe it's not handling events, or maybe the window losing focus isn't handling events), everything backs up in the system queue and the system appears hung.

    There are a few solutions to this problem.

    One is to make focus policy asynchronous. That is, focus changing has absolutely nothing to do with the keyboard. If you click on a window and start typing before the focus actually changes, the keystrokes go to the first window until focus changes, then they go to the second. This is what X-windows does.

    Another is what NT does. When focus changes, keyboard events are held in the system message queue, but other events are allowed through. This is "asynchronous" because the messages in the system queue are delivered to the application queues in a different order from that with which they were posted. If a bad app won't handle the "lose focus" message, it's of no consequence - the app receiving focus will get its "gain focus" message, and the keystrokes will go to it.

    The NT solution also takes care of the application queue filling up problem. Since the system delivers messages asynchronously, messages waiting in the system queue will just sit there and the rest of the messages will be delivered to their apps.

    The OS/2 SIQ solution is this: When a focus-changing event happens, in addition to blocking further messages from the application queues, a timer is started. When the timer goes off, if the focus change has not yet happened, the bad app has its focus taken away and all messages targeted at that window are skipped. When the bad app finally handles the focus change message, OS/2 will detect this and stop skipping its messages.

    As for the pros and cons:

    The X-windows solution is probably the easiest. The problem is that users generally don't like having to wait for the focus to change before they start typing. On many occasions, you can type and the characters end up in the wrong window because something (usually heavy system load) is preventing the focus change from happening in a timely manner.

    The NT solution seems pretty nice, but making the system message queue asynchronous can cause similar problems to the X-windows problem. Since messages can be delivered out of order, programs must not assume that two messages posted in a particular order will be delivered in that same order. This can break legacy apps, but since Win32 always had an asynchronous queue, it is fair to simply tell app designers "don't do that". It's harder to tell app designers something like that on OS/2 - they'll complain "you changed the rules and our apps are breaking."

    The OS/2 solution's problem is that nothing happens until you try to change window focus, and then wait for the timeout. Until then, the bad app is not detected and nothing is done.

    --David Charlap 
  • Intertask/interthread SendMessage. The system has to inform the target queue about the forthcoming message, then it has to carry out the context switch and wait until the result is available. Win16 stores necessary parameters in the queue structure and then calls DirectedYield() function. However, in Win32 there could be several messages pending sent by preemptively executing threads, and in this case SendMessage has to build some sort of message queue for sent messages. Another issue is what to do with messages sent to the sender when it is blocked inside its own SendMessage.

9.1.3. Accelerators

There are three differently sized accelerator structures exposed to the user:

  1. Accelerators in NE resources. This is also the internal layout of the global handle HACCEL (16 and 32) in Windows 95 and Wine. Exposed to the user as Win16 global handles HACCEL16 and HACCEL32 by the Win16/Win32 API. These are 5 bytes long, with no padding:

    BYTE   fVirt;
    WORD   key;
    WORD   cmd;

  2. Accelerators in PE resources. They are exposed to the user only by direct accessing PE resources. These have a size of 8 bytes:

    BYTE   fVirt;
    BYTE   pad0;
    WORD   key;
    WORD   cmd;
    WORD   pad1;
  3. Accelerators in the Win32 API. These are exposed to the user by the CopyAcceleratorTable and CreateAcceleratorTable functions in the Win32 API. These have a size of 6 bytes:

    BYTE   fVirt;
    BYTE   pad0;
    WORD   key;
    WORD   cmd;

Why two types of accelerators in the Win32 API? We can only guess, but my best bet is that the Win32 resource compiler can/does not handle struct packing. Win32 ACCEL is defined using #pragma(2) for the compiler but without any packing for RC, so it will assume #pragma(4).

9.2. X Windows System interface

9.2.1. Keyboard mapping

Wine now needs to know about your keyboard layout. This requirement comes from a need from many apps to have the correct scancodes available, since they read these directly, instead of just taking the characters returned by the X server. This means that Wine now needs to have a mapping from X keys to the scancodes these programs expect.

On startup, Wine will try to recognize the active X layout by seeing if it matches any of the defined tables. If it does, everything is alright. If not, you need to define it.

To do this, open the file dlls/winex11.drv/keyboard.c and take a look at the existing tables.

What you really would need to do, is find out which scancode each key needs to generate. Find it in the main_key_scan table, which looks like this:

static const int main_key_scan[MAIN_LEN] =
/* this is my (102-key) keyboard layout, sorry if it doesn't quite match yours */
0x56 /* the 102nd key (actually to the right of l-shift) */

Next, assign each scancode the characters imprinted on the keycaps. This was done (sort of) for the US 101-key keyboard, which you can find near the top in keyboard.c. It also shows that if there is no 102nd key, you can skip that.

However, for most international 102-key keyboards, we have done it easy for you. The scancode layout for these already pretty much matches the physical layout in the main_key_scan, so all you need to do is to go through all the keys that generate characters on your main keyboard (except spacebar), and stuff those into an appropriate table. The only exception is that the 102nd key, which is usually to the left of the first key of the last line (usually Z), must be placed on a separate line after the last line.

After you have written such a table, you need to add it to the main_key_tab[] layout index table. This will look like this:

static struct {
WORD lang, ansi_codepage, oem_codepage;
const char (*key)[MAIN_LEN][4];
} main_key_tab[]={

After you have added your table, recompile Wine and test that it works. If it fails to detect your table, try running

WINEDEBUG=+key,+keyboard wine > key.log 2>&1

and look in the resulting key.log file to find the error messages it gives for your layout.

Note that the LANG_* and SUBLANG_* definitions are in include/winnls.h, which you might need to know to find out which numbers your language is assigned, and find it in the WINEDEBUG output. The numbers will be (SUBLANG * 0x400 + LANG), so, for example the combination LANG_NORWEGIAN (0x14) and SUBLANG_DEFAULT (0x1) will be (in hex) 14 + 1*400 = 414, so since I'm Norwegian, I could look for 0414 in the WINEDEBUG output to find out why my keyboard won't detect.

Chapter 10. COM in Wine

10.1. Writing COM Components for Wine

This section describes how to create your own natively compiled COM components.

10.1.1. Macros to define a COM interface

The goal of the following set of definitions is to provide a way to use the same header file definitions to provide both a C interface and a C++ object oriented interface to COM interfaces. The type of interface is selected automatically depending on the language but it is always possible to get the C interface in C++ by defining CINTERFACE.

It is based on the following assumptions:

  • all COM interfaces derive from IUnknown, this should not be a problem.

  • the header file only defines the interface, the actual fields are defined separately in the C file implementing the interface.

The natural approach to this problem would be to make sure we get a C++ class and virtual methods in C++ and a structure with a table of pointer to functions in C. Unfortunately the layout of the virtual table is compiler specific, the layout of g++ virtual tables is not the same as that of an egcs virtual table which is not the same as that generated by Visual C++. There are workarounds to make the virtual tables compatible via padding but unfortunately the one which is imposed to the Wine emulator by the Windows binaries, i.e. the Visual C++ one, is the most compact of all.

So the solution I finally adopted does not use virtual tables. Instead I use in-line non virtual methods that dereference the method pointer themselves and perform the call.

Let's take Direct3D as an example:

#define ICOM_INTERFACE IDirect3D
#define IDirect3D_METHODS \
    ICOM_METHOD1(HRESULT,Initialize,    REFIID,) \
    ICOM_METHOD2(HRESULT,CreateLight,   LPDIRECT3DLIGHT*,, IUnknown*,) \
#define IDirect3D_IMETHODS \
    IUnknown_IMETHODS \

// *** IUnknown methods *** //
#define IDirect3D_QueryInterface(p,a,b) ICOM_CALL2(QueryInterface,p,a,b)
#define IDirect3D_AddRef(p)             ICOM_CALL (AddRef,p)
#define IDirect3D_Release(p)            ICOM_CALL (Release,p)
// *** IDirect3D methods *** //
#define IDirect3D_Initialize(p,a)       ICOM_CALL1(Initialize,p,a)
#define IDirect3D_EnumDevices(p,a,b)    ICOM_CALL2(EnumDevice,p,a,b)
#define IDirect3D_CreateLight(p,a,b)    ICOM_CALL2(CreateLight,p,a,b)
#define IDirect3D_CreateMaterial(p,a,b) ICOM_CALL2(CreateMaterial,p,a,b)
#define IDirect3D_CreateViewport(p,a,b) ICOM_CALL2(CreateViewport,p,a,b)
#define IDirect3D_FindDevice(p,a,b)     ICOM_CALL2(FindDevice,p,a,b)


The ICOM_INTERFACE macro is used in the ICOM_METHOD macros to define the type of the this pointer. Defining this macro here saves us the trouble of having to repeat the interface name everywhere. Note however that because of the way macros work, a macro like ICOM_METHOD1 cannot use ICOM_INTERFACE##_VTABLE because this would give ICOM_INTERFACE_VTABLE and not IDirect3D_VTABLE.

ICOM_METHODS defines the methods specific to this interface. It is then aggregated with the inherited methods to form ICOM_IMETHODS.

ICOM_IMETHODS defines the list of methods that are inheritable from this interface. It must be written manually (rather than using a macro to generate the equivalent code) to avoid macro recursion (which compilers don't like).

The ICOM_DEFINE finally declares all the structures necessary for the interface. We have to explicitly use the interface name for macro expansion reasons again. Inherited methods are inherited in C by using the IDirect3D_METHODS macro and the parent's Xxx_IMETHODS macro. In C++ we need only use the IDirect3D_METHODS since method inheritance is taken care of by the language.

In C++ the ICOM_METHOD macros generate a function prototype and a call to a function pointer method. This means using once t1 p1, t2 p2, ... and once p1, p2 without the types. The only way I found to handle this is to have one ICOM_METHOD macro per number of parameters and to have it take only the type information (with const if necessary) as parameters. The undef ICOM_INTERFACE is here to remind you that using ICOM_INTERFACE in the following macros will not work. This time it's because the ICOM_CALL macro expansion is done only once the IDirect3D_Xxx macro is expanded. And by that time ICOM_INTERFACE will be long gone anyway.

You may have noticed the double commas after each parameter type. This allows you to put the name of that parameter which I think is good for documentation. It is not required and since I did not know what to put there for this example (I could only find doc about IDirect3D2), I left them blank.

Finally the set of IDirect3D_Xxx macros is a standard set of macros defined to ease access to the interface methods in C. Unfortunately I don't see any way to avoid having to duplicate the inherited method definitions there. This time I could have used a trick to use only one macro whatever the number of parameters but I preferred to have it work the same way as above.

You probably have noticed that we don't define the fields we need to actually implement this interface: reference count, pointer to other resources and miscellaneous fields. That's because these interfaces are just that: interfaces. They may be implemented more than once, in different contexts and sometimes not even in Wine. Thus it would not make sense to impose that the interface contains some specific fields.

10.1.2. Bindings in C

In C this gives:

typedef struct IDirect3DVtbl IDirect3DVtbl;
struct IDirect3D {
    IDirect3DVtbl* lpVtbl;
struct IDirect3DVtbl {
    HRESULT (*fnQueryInterface)(IDirect3D* me, REFIID riid, LPVOID* ppvObj);
    ULONG (*fnAddRef)(IDirect3D* me);
    ULONG (*fnRelease)(IDirect3D* me);
    HRESULT (*fnInitialize)(IDirect3D* me, REFIID a);
    HRESULT (*fnEnumDevices)(IDirect3D* me, LPD3DENUMDEVICESCALLBACK a, LPVOID b);
    HRESULT (*fnCreateLight)(IDirect3D* me, LPDIRECT3DLIGHT* a, IUnknown* b);
    HRESULT (*fnCreateMaterial)(IDirect3D* me, LPDIRECT3DMATERIAL* a, IUnknown* b);
    HRESULT (*fnCreateViewport)(IDirect3D* me, LPDIRECT3DVIEWPORT* a, IUnknown* b);

// *** IUnknown methods *** //
#define IDirect3D_QueryInterface(p,a,b) (p)->lpVtbl->fnQueryInterface(p,a,b)
#define IDirect3D_AddRef(p)             (p)->lpVtbl->fnAddRef(p)
#define IDirect3D_Release(p)            (p)->lpVtbl->fnRelease(p)
// *** IDirect3D methods *** //
#define IDirect3D_Initialize(p,a)       (p)->lpVtbl->fnInitialize(p,a)
#define IDirect3D_EnumDevices(p,a,b)    (p)->lpVtbl->fnEnumDevice(p,a,b)
#define IDirect3D_CreateLight(p,a,b)    (p)->lpVtbl->fnCreateLight(p,a,b)
#define IDirect3D_CreateMaterial(p,a,b) (p)->lpVtbl->fnCreateMaterial(p,a,b)
#define IDirect3D_CreateViewport(p,a,b) (p)->lpVtbl->fnCreateViewport(p,a,b)
#define IDirect3D_FindDevice(p,a,b)     (p)->lpVtbl->fnFindDevice(p,a,b)


IDirect3D only contains a pointer to the virtual/jump table. This is the only thing the user needs to know to use the interface. Of course the structure we will define to implement this interface will have more fields but the first one will match this pointer.

The code generated by ICOM_DEFINE defines both the structure representing the interface and the structure for the jump table. ICOM_DEFINE uses the parent Xxx_IMETHODS macro to automatically repeat the prototypes of all the inherited methods and then uses IDirect3D_METHODS to define the IDirect3D methods.

Each method is declared as a pointer to function field in the jump table. The implementation will fill this jump table with appropriate values, probably using a static variable, and initialize the lpVtbl field to point to this variable.

The IDirect3D_Xxx macros then just dereference the lpVtbl pointer and use the function pointer corresponding to the macro name. This emulates the behavior of a virtual table and should be just as fast.

This C code should be quite compatible with the Windows headers both for code that uses COM interfaces and for code implementing a COM interface.

10.1.3. Bindings in C++

And in C++ (with g++):

typedef struct IDirect3D: public IUnknown {
    private: HRESULT (*fnInitialize)(IDirect3D* me, REFIID a);
    public: inline HRESULT Initialize(REFIID a) { return ((IDirect3D*)t.lpVtbl)->fnInitialize(this,a); };
    private: HRESULT (*fnEnumDevices)(IDirect3D* me, LPD3DENUMDEVICESCALLBACK a, LPVOID b);
    public: inline HRESULT EnumDevices(LPD3DENUMDEVICESCALLBACK a, LPVOID b)
        { return ((IDirect3D*)t.lpVtbl)->fnEnumDevices(this,a,b); };
    private: HRESULT (*fnCreateLight)(IDirect3D* me, LPDIRECT3DLIGHT* a, IUnknown* b);
    public: inline HRESULT CreateLight(LPDIRECT3DLIGHT* a, IUnknown* b)
        { return ((IDirect3D*)t.lpVtbl)->fnCreateLight(this,a,b); };
    private: HRESULT (*fnCreateMaterial)(IDirect3D* me, LPDIRECT3DMATERIAL* a, IUnknown* b);
    public: inline HRESULT CreateMaterial(LPDIRECT3DMATERIAL* a, IUnknown* b)
        { return ((IDirect3D*)t.lpVtbl)->fnCreateMaterial(this,a,b); };
    private: HRESULT (*fnCreateViewport)(IDirect3D* me, LPDIRECT3DVIEWPORT* a, IUnknown* b);
    public: inline HRESULT CreateViewport(LPDIRECT3DVIEWPORT* a, IUnknown* b)
        { return ((IDirect3D*)t.lpVtbl)->fnCreateViewport(this,a,b); };
        { return ((IDirect3D*)t.lpVtbl)->fnFindDevice(this,a,b); };


In C++ IDirect3D does double duty as both the virtual/jump table and as the interface definition. The reason for this is to avoid having to duplicate the method definitions: once to have the function pointers in the jump table and once to have the methods in the interface class. Here one macro can generate both. This means though that the first pointer, t.lpVtbl defined in IUnknown, must be interpreted as the jump table pointer if we interpret the structure as the interface class, and as the function pointer to the QueryInterface method, t.fnQueryInterface, if we interpret the structure as the jump table. Fortunately this gymnastic is entirely taken care of in the header of IUnknown.

Of course in C++ we use inheritance so that we don't have to duplicate the method definitions.

Since IDirect3D does double duty, each ICOM_METHOD macro defines both a function pointer and a non-virtual inline method which dereferences it and calls it. This way this method behaves just like a virtual method but does not create a true C++ virtual table which would break the structure layout. If you look at the implementation of these methods you'll notice that they would not work for void functions. We have to return something and fortunately this seems to be what all the COM methods do (otherwise we would need another set of macros).

Note how the ICOM_METHOD generates both function prototypes mixing types and formal parameter names and the method invocation using only the formal parameter name. This is the reason why we need different macros to handle different numbers of parameters.

Finally there is no IDirect3D_Xxx macro. These are not needed in C++ unless the CINTERFACE macro is defined in which case we would not be here.

This C++ code works well for code that just uses COM interfaces. But it will not work with C++ code implement a COM interface. That's because such code assumes the interface methods are declared as virtual C++ methods which is not the case here.

10.1.4. Implementing a COM interface.

This continues the above example. This example assumes that the implementation is in C.

typedef struct _IDirect3D {
    void* lpVtbl;
    // ...
 } _IDirect3D;

static ICOM_VTABLE(IDirect3D) d3dvt;

// implement the IDirect3D methods here

int IDirect3D_fnQueryInterface(IDirect3D* me)
    // ...

// ...

static ICOM_VTABLE(IDirect3D) d3dvt = {


We first define what the interface really contains. This is the _IDirect3D structure. The first field must of course be the virtual table pointer. Everything else is free.

Then we predeclare our static virtual table variable, we will need its address in some methods to initialize the virtual table pointer of the returned interface objects.

Then we implement the interface methods. To match what has been declared in the header file they must take a pointer to an IDirect3D structure and we must cast it to an _IDirect3D so that we can manipulate the fields. This is performed by the ICOM_THIS macro.

Finally we initialize the virtual table.

10.2. A brief introduction to DCOM in Wine

This section explains the basic principles behind DCOM remoting as used by InstallShield and others.

10.2.1. Basics

The basic idea behind DCOM is to take a COM object and make it location transparent. That means you can use it from other threads, processes and machines without having to worry about the fact that you can't just dereference the interface vtable pointer to call methods on it.

You might be wondering about putting threads next to processes and machines in that last paragraph. You can access thread safe objects from multiple threads without DCOM normally, right? Why would you need RPC magic to do that?

The answer is of course that COM doesn't assume that objects actually are thread-safe. Most real-world objects aren't, in fact, for various reasons. What these reasons are isn't too important here, though; it's just important to realize that the problem of thread-unsafe objects is what COM tries hard to solve with its apartment model. There are also ways to tell COM that your object is truly thread-safe (namely the free-threaded marshaller). In general, no object is truly thread-safe if it could potentially use another not so thread-safe object, though, so the free-threaded marshaller is less used than you'd think.

For now, suffice it to say that COM lets you "marshal" interfaces into other apartments. An apartment (you may see it referred to as a context in modern versions of COM) can be thought of as a location, and contains objects.

Every thread in a program that uses COM exists in an apartment. If a thread wishes to use an object from another apartment, marshalling and the whole DCOM infrastructure gets involved to make that happen behind the scenes.

So. Each COM object resides in an apartment, and each apartment resides in a process, and each process resides in a machine, and each machine resides in a network. Allowing those objects to be used from any of these different places is what DCOM is all about.

The process of marshalling refers to taking a function call in an apartment and actually performing it in another apartment. Let's say you have two machines, A and B, and on machine B there is an object sitting in a DLL on the hard disk. You want to create an instance of that object (activate it) and use it as if you had compiled it into your own program. This is hard, because the remote object is expecting to be called by code in its own address space - it may do things like accept pointers to linked lists and even return other objects.

Very basic marshalling is easy enough to understand. You take a method on a remote interface (that is a COM interface that is implemented on the remote computer), copy each of its parameters into a buffer, and send it to the remote computer. On the other end, the remote server reads each parameter from the buffer, calls the method, writes the result into another buffer and sends it back.

The tricky part is exactly how to encode those parameters in the buffer, and how to convert standard stdcall/cdecl method calls to network packets and back again. This is the job of the RPCRT4.DLL file: the Remote Procedure Call Runtime.

The backbone of DCOM is this RPC runtime, which is an implementation of DCE RPC. DCE RPC is not naturally object oriented, so this protocol is extended with some new constructs and by assigning new meanings to some of the packet fields, to produce ORPC or Object RPC. You might see it called MS-RPC as well.

RPC packets contain a buffer containing marshalled data in NDR format. NDR is short for "Network Data Representation" and is similar to the XDR format used in SunRPC (the closest native equivalent on Linux to DCE RPC). NDR/XDR are all based on the idea of graph serialization and were worked out during the 80s, meaning they are very powerful and can do things like marshal doubly linked lists and other rather tricky structures.

In Wine, our DCOM implementation is not currently based on the RPC runtime, as while few programs use DCOM even fewer use RPC directly so it was developed some time after OLE32/OLEAUT32 were. Eventually this will have to be fixed, otherwise our DCOM will never be compatible with Microsoft's. Bear this in mind as you read through the code however.

10.2.2. Proxies and Stubs

Manually marshalling and unmarshalling each method call using the NDR APIs (NdrConformantArrayMarshall etc) is very tedious work, so the Platform SDK ships with a tool called midl which is an IDL compiler. IDL or the "Interface Definition Language" is a language designed specifically for describing interfaces in a reasonably language neutral fashion, though in reality it bears a close resemblance to C++.

By describing the functions you want to expose via RPC in IDL therefore, it becomes possible to pass this file to MIDL which spits out a huge amount of C source code. That code defines functions which have the same prototype as the functions described in your IDL but which internally take each argument, marshal it using NDR, send the packet, and unmarshal the return.

Because this code proxies the code from the client to the server, the functions are called proxies. Easy, right?

Of course, in the RPC server process at the other end, you need some way to unmarshal the RPCs, so you have functions also generated by MIDL which are the inverse of the proxies; they accept an NDR buffer, extract the parameters, call the real function and then marshal the result back. They are called stubs, and stand in for the real calling code in the client process.

The sort of marshalling/unmarshalling code that MIDL spits out can be seen in dlls/oleaut32/oleaut32_oaidl_p.c - it's not exactly what it would look like as that file contains DCOM proxies/stubs which are different, but you get the idea. Proxy functions take the arguments and feed them to the NDR marshallers (or picklers), invoke an NdrProxySendReceive and then convert the out parameters and return code. There's a ton of goop in there for dealing with buffer allocation, exceptions and so on - it's really ugly code. But, this is the basic concept behind DCE RPC.

10.2.3. Interface Marshalling

Standard NDR only knows about C style function calls - they can accept and even return structures, but it has no concept of COM interfaces. Confusingly DCE RPC does have a concept of RPC interfaces which are just convenient ways to bundle function calls together into namespaces, but let's ignore that for now as it just muddies the water. The primary extension made by Microsoft to NDR then was the ability to take a COM interface pointer and marshal that into the NDR stream.

The basic theory of proxies and stubs and IDL is still here, but it's been modified slightly. Whereas before you could define a bunch of functions in IDL, now a new object keyword has appeared. This tells MIDL that you're describing a COM interface, and as a result the proxies/stubs it generates are also COM objects.

That's a very important distinction. When you make a call to a remote COM object you do it via a proxy object that COM has constructed on the fly. Likewise, a stub object on the remote end unpacks the RPC packet and makes the call.

Because this is object-oriented RPC, there are a few complications: for instance, a call that goes via the same proxies/stubs may end up at a different object instance, so the RPC runtime keeps track of "this" and "that" in the RPC packets.

This leads naturally onto the question of how we got those proxy/stub objects in the first place, and where they came from. You can use the CoCreateInstanceEx API to activate COM objects on a remote machine, this works like CoCreateInstance API. Behind the scenes, a lot of stuff is involved to do this (like IRemoteActivation, IOXIDResolver and so on) but let's gloss over that for now.

When DCOM creates an object on a remote machine, the DCOM runtime on that machine activates the object in the usual way (by looking it up in the registry etc) and then marshalls the requested interface back to the client. Marshalling an interface takes a pointer, and produces a buffer containing all the information DCOM needs to construct a proxy object in the client, a stub object in the server and link the two together.

The structure of a marshalled interface pointer is somewhat complex. Let's ignore that too. The important thing is how COM proxies/stubs are loaded.

10.2.4. COM Proxy/Stub System

COM proxies are objects that implement both the interfaces needing to be proxied and also IRpcProxyBuffer. Likewise, COM stubs implement IRpcStubBuffer and understand how to invoke the methods of the requested interface.

You may be wondering what the word "buffer" is doing in those interface names. I'm not sure either, except that a running theme in DCOM is that interfaces which have nothing to do with buffers have the word "Buffer" appended to them, seemingly at random. Ignore it and don't let it confuse you :) This stuff is convoluted enough ...

The IRpcProxyBuffer and IRpcStubBuffer interfaces are used to control the proxy/stub objects and are one of the many semi-public interfaces used in DCOM.

DCOM is theoretically an internet RFC and is specced out, but in reality the only implementation of it apart from ours is Microsoft's, and as a result there are lots of interfaces which can be used if you want to customize or control DCOM but in practice are badly documented or not documented at all, or exist mostly as interfaces between MIDL generated code and COM itself. Don't pay too much attention to the MSDN definitions of these interfaces and APIs.

COM proxies and stubs are like any other normal COM object - they are registered in the registry, they can be loaded with CoCreateInstance and so on. They have to be in process (in DLLs) however. They aren't activated directly by COM however, instead the process goes something like this:

  • COM receives a marshalled interface packet, and retrieves the IID of the marshalled interface from it

  • COM looks in HKEY_CLASSES_ROOT/Interface/whatever-iid/ProxyStubClsId32 to retrieve the CLSID of another COM object, which implements IPSFactoryBuffer.

  • IPSFactoryBuffer has only two methods, CreateProxy and CreateStub. COM calls whichever is appropriate: CreateStub for the server, CreateProxy for the client. MIDL will normally provide an implementation of this object for you in the code it generates.

Once CreateProxy has been called, the resultant object is QueryInterfaced to IRpcProxyBuffer, which only has one method, IRpcProxyBuffer::Connect. This method only takes one parameter, the IRpcChannelBuffer object which encapsulates the "RPC Channel" between the client and server.

On the server side, a similar process is performed: the PSFactoryBuffer is created, CreateStub is called, result is QueryInterfaced to IRpcStubBuffer, and IRpcStubBuffer::Connect is used to link it to the RPC channel.

10.2.5. RPC Channels

Remember the RPC runtime? Well, that's not just responsible for marshalling stuff, it also controls the connection and protocols between the client and server. We can ignore the details of this for now, suffice it to say that an RPC Channel is a COM object that implements IRpcChannelBuffer, and it's basically an abstraction of different RPC methods. For instance, in the case of inter-thread marshalling (not covered here) the RPC connection code isn't used, only the NDR marshallers are, so IRpcChannelBuffer in that case isn't actually implemented by RPCRT4 but rather just by the COM/OLE DLLS.

On this topic, Ove Kåven says:

It depends on the Windows version, I think. Windows 95 and Windows NT 4 certainly had very different models when I looked. I'm pretty sure the Windows 98 version of RPCRT4 was able to dispatch messages directly to individual apartments. I'd be surprised if some similar functionality was not added to Windows 2000. After all, if an object on machine A wanted to use an object on machine B in an apartment C, wouldn't it be most efficient if the RPC system knew about apartments and could dispatch the message directly to it? And if RPC does know how to efficiently dispatch to apartments, why should COM duplicate this functionality? There were, however, no unified way to tell RPC about them across Windows versions, so in that old patch of mine, I let the COM/OLE dlls do the apartment dispatch, but even then, the RPC runtime was always involved. After all, it could be quite tricky to tell whether the call is merely interthread, without involving the RPC runtime...

RPC channels are constructed on the fly by DCOM as part of the marshalling process. So, when you make a call on a COM proxy, it goes like this:

Your code -> COM proxy object -> RPC Channel -> COM stub object -> Their code

10.2.6. How this actually works in Wine

Right now, Wine does not use the NDR marshallers or RPC to implement its DCOM. When you marshal an interface in Wine, in the server process a _StubMgrThread thread is started. I haven't gone into the stub manager here. The important thing is that eventually a _StubReaderThread is started which accepts marshalled DCOM RPCs, and then passes them to IRpcStubBuffer::Invoke on the correct stub object which in turn demarshalls the packet and performs the call. The threads started by our implementation of DCOM are never terminated, they just hang around until the process dies.

Remember that I said our DCOM doesn't use RPC? Well, you might be thinking "but we use IRpcStubBuffer like we're supposed to... isn't that provided by MIDL which generates code that uses the NDR APIs?". If so pat yourself on the back, you're still with me. Go get a cup of coffee.

10.2.7. Typelib Marshaller

In fact, the reason for the PSFactoryBuffer layer of indirection is because not all interfaces are marshalled using MIDL generated code. Why not? Well, to understand that you have to see that one of the driving forces behind OLE and by extension DCOM was the development of Visual Basic. Microsoft wanted VB developers to be first class citizens in the COM world, but things like writing IDL and compiling them with a C compiler into DLLs wasn't easy enough.

So, type libraries were invented. Actually they were invented as part of a parallel line of COM development known as "OLE Automation", but let's not get into that here. Type libraries are basically binary IDL files, except that despite there being two type library formats neither of them can fully express everything expressible in IDL. Anyway, with a type library (which can be embedded as a resource into a DLL) you have another option beyond compiling MIDL output - you can set the ProxyStubClsId32 registry entry for your interfaces to the CLSID of the "type library marshaller" or "universal marshaller". Both terms are used, but in the Wine source it's called the typelib marshaller.

The type library marshaller constructs proxy and stub objects on the fly. It does so by having generic marshalling glue which reads the information from the type libraries, and takes the parameters directly off the stack. The CreateProxy method actually builds a vtable out of blocks of assembly stitched together which pass control to _xCall, which then does the marshalling. You can see all this magic in dlls/oleaut32/tmarshal.c

In the case of InstallShield, it actually comes with typelibs for all the interfaces it needs to marshal (fixme: is this right?), but they actually use a mix of MIDL and typelib marshalling. In order to cover up for the fact that we don't really use RPC they're all forced to go via the typelib marshaller - that's what the 1 || hack is for and what the "Registering non-automation type library!" warning is about (I think).

10.2.8. Apartments

Before a thread can use COM it must enter an apartment. Apartments are an abstraction of a COM objects thread safety level. There are many types of apartments but the only two we care about right now are single threaded apartments (STAs) and the multi-threaded apartment (MTA).

Any given process may contain at most one MTA and potentially many STAs. This is because all objects in MTAs never care where they are invoked from and hence can all be treated the same. Since objects in STAs do care, they cannot be treated the same.

You enter an apartment by calling CoInitializeEx() and passing the desired thread model in as a parameter. The default if you use the deprecated CoInitialize() is a STA, and this is the most common type of apartment used in COM.

An object in the multi-threaded apartment may be accessed concurrently by multiple threads, e.g. it's supposed to be entirely thread safe. It must also not care about thread-affinity, the object should react the same way no matter which thread is calling it.

An object inside a STA does not have to be thread safe, and all calls upon it should come from the same thread - the thread that entered the apartment in the first place.

The apartment system was originally designed to deal with the disparity between the Windows NT/C++ world in which threading was given a strong emphasis, and the Visual Basic world in which threading was barely supported and even if it had been fully supported most developers would not have used it. Visual Basic code is not truly multi-threaded, instead if you start a new thread you get an entirely new VM, with separate sets of global variables. Changes made in one thread do not reflect in another, which pretty much violates the expected semantics of multi-threading entirely but this is Visual Basic, so what did you expect? If you access a VB object concurrently from multiple threads, behind the scenes each VM runs in a STA and the calls are marshaled between the threads using DCOM.

In the Windows 2000 release of COM, several new types of apartment were added, the most important of which are RTAs (the rental threaded apartment) in which concurrent access are serialised by COM using an apartment-wide lock but thread affinity is not guaranteed.

10.2.9. Structure of a marshaled interface pointer

When an interface is marshaled using CoMarshalInterface(), the result is a serialized OBJREF structure. An OBJREF actually contains a union, but we'll be assuming the variant that embeds a STDOBJREF here which is what's used by the system provided standard marshaling. A STDOBJREF (standard object reference) consists of the magic signature MEOW, then some flags, then the IID of the marshaled interface. Quite what MEOW stands for is a mystery, but it's definitely not "Microsoft Extended Object Wire". Next comes the STDOBJREF flags, identified by their SORF_ prefix. Most of these are reserved, and their purpose (if any) is unknown, but a few are defined.

After the SORF flags comes a count of the references represented by this marshaled interface. Typically this will be 5 in the case of a normal marshal, but may be 0 for table-strong and table-weak marshals (the difference between these is explained below). The reasoning is this: in the general case, we want to know exactly when an object is unmarshaled and released, so we can accurately control the lifetime of the stub object. This is what happens when cPublicRefs is zero. However, in many cases, we only want to unmarshal an object once. Therefore, if we strengthen the rules to say when marshaling that we will only unmarshal once, then we no longer have to know when it is unmarshaled. Therefore, we can give out an arbitrary number of references when marshaling and basically say "don't call me, except when you die".

The most interesting part of a STDOBJREF is the OXID, OID, IPID triple. This triple identifies any given marshaled interface pointer in the network. OXIDs are apartment identifiers, and are supposed to be unique network-wide. How this is guaranteed is currently unknown: the original algorithm Windows used was something like the current UNIX time and a local counter.

OXIDs are generated and registered with the OXID resolver by performing local RPCs to the RPC subsystem (rpcss.exe). In a fully security-patched Windows system they appear to be randomly generated. This registration is done using the ILocalOxidResolver interface, however the exact structure of this interface is currently unknown.

OIDs are object identifiers, and identify a stub manager. The stub manager manages interface stubs. For each exported COM object there are multiple interfaces and therefore multiple interface stubs (IRpcStubBuffer implementations). OIDs are apartment scoped. Each interface stub is identified by an IPID, which identifies a marshaled interface pointer. IPIDs are apartment scoped.

Unmarshaling one of these streams therefore means setting up a connection to the object exporter (the apartment holding the marshaled interface pointer) and being able to send RPCs to the right interface stub. Each apartment has its own RPC endpoint and calls can be routed to the correct interface pointer by embedding the IPID into the call using RpcBindingSetObject. IRemUnknown, discussed below, uses a reserved IPID. Please note that this is true only in the current implementation. The native version generates an IPID as per any other object and simply notifies the SCM of this IPID.

Both standard and handler marshaled OBJREFs contains an OXID resolver endpoint which is an RPC string binding in a DUALSTRINGARRAY. This is necessary because an OXID alone is not enough to contact the host, as it doesn't contain any network address data. Instead, the combination of the remote OXID resolver RPC endpoint and the OXID itself are passed to the local OXID resolver. It then returns the apartment string binding.

This step is an optimisation: technically the OBJREF itself could contain the string binding of the apartment endpoint and the OXID resolver could be bypassed, but by using this DCOM can optimise out a server round-trip by having the local OXID resolver cache the query results. The OXID resolver is a service in the RPC subsystem (rpcss.exe) which implements a raw (non object-oriented) RPC interface called IOXIDResolver. Despite the identical naming convention this is not a COM interface.

Unmarshaling an interface pointer stream therefore consists of reading the OXID , OID and IPID from the STDOBJREF, then reading one or more RPC string bindings for the remote OXID resolver. Then RpcBindingFromStringBinding is used to convert this remote string binding into an RPC binding handle which can be passed to the local IOXIDResolver::ResolveOxid implementation along with the OXID. The local OXID resolver consults its list of same-machine OXIDs, then its cache of remote OXIDs, and if not found does an RPC to the remote OXID resolver using the binding handle passed in earlier. The result of the query is stored for future reference in the cache, and finally the unmarshaling application gets back the apartment string binding, the IPID of that apartments IRemUnknown implementation, and a security hint (let's ignore this for now).

Once the remote apartments string binding has been located the unmarshalling process constructs an RPC Channel Buffer implementation with the connection handle and the IPID of the needed interface, loads and constructs the IRpcProxyBuffer implementation for that IID and connects it to the channel. Finally the proxy is passed back to the application.

10.2.10. Handling IUnknown

There are some subtleties here with respect to IUnknown. IUnknown itself is never marshaled directly: instead a version of it optimised for network usage is used. IRemUnknown is similar in concept to IUnknown except that it allows you to add and release arbitrary numbers of references at once, and it also allows you to query for multiple interfaces at once.

IRemUnknown is used for lifecycle management, and for marshaling new interfaces on an object back to the client. Its definition can be seen in dcom.idl - basically the IRemUnknown::RemQueryInterface method takes an IPID and a list of IIDs, then returns STDOBJREFs of each new marshaled interface pointer.

There is one IRemUnknown implementation per apartment, not per stub manager as you might expect. This is OK because IPIDs are apartment not object scoped (in fact, according to the DCOM draft spec, they are machine-scoped, but this implies apartment-scoped).

10.2.11. Table marshaling

Normally once you have unmarshaled a marshaled interface pointer that stream is dead, you can't unmarshal it again. Sometimes this isn't what you want. In this case, table marshaling can be used. There are two types: strong and weak. In table-strong marshaling, selected by a specific flag to CoMarshalInterface(), a stream can be unmarshaled as many times as you like. Even if all the proxies are released, the marshaled object reference is still valid. Effectively the stream itself holds a ref on the object. To release the object entirely so its server can shut down, you must use CoReleaseMarshalData() on the stream.

In table-weak marshaling the stream can be unmarshaled many times, however the stream does not hold a ref. If you unmarshal the stream twice, once those two proxies have been released remote object will also be released. Attempting to unmarshal the stream at this point will yield CO_E_DISCONNECTED.

10.2.12. RPC dispatch

Exactly how RPC dispatch occurs depends on whether the exported object is in a STA or the MTA. If it's in the MTA then all is simple: the RPC dispatch thread can temporarily enter the MTA, perform the remote call, and then leave it again. If it's in a STA things get more complex, because of the requirement that only one thread can ever access the object.

Instead, when entering a STA a hidden window is created implicitly by COM, and the user must manually pump the message loop in order to service incoming RPCs. The RPC dispatch thread performs the context switch into the STA by sending a message to the apartments window, which then proceeds to invoke the remote call in the right thread.

RPC dispatch threads are pooled by the RPC runtime. When an incoming RPC needs to be serviced, a thread is pulled from the pool and invokes the call. The main RPC thread then goes back to listening for new calls. It's quite likely for objects in the MTA to therefore be servicing more than one call at once.

10.2.13. Message filtering and re-entrancy

When an outgoing call is made from a STA, it's possible that the remote server will re-enter the client, for instance to perform a callback. Because of this potential re-entrancy, when waiting for the reply to an RPC made inside a STA, COM will pump the message loop. That's because while this thread is blocked, the incoming callback will be dispatched by a thread from the RPC dispatch pool, so it must be processing messages.

While COM is pumping the message loop, all incoming messages from the operating system are filtered through one or more message filters. These filters are themselves COM objects which can choose to discard, hold or forward window messages. The default message filter drops all input messages and forwards the rest. This is so that if the user chooses a menu option which triggers an RPC, they then cannot choose that menu option *again* and restart the function from the beginning. That type of unexpected re-entrancy is extremely difficult to debug, so it's disallowed.

Unfortunately other window messages are allowed through, meaning that it's possible your UI will be required to repaint itself during an outgoing RPC. This makes programming with STAs more complex than it may appear, as you must be prepared to run all kinds of code any time an outgoing call is made. In turn this breaks the idea that COM should abstract object location from the programmer, because an object that was originally free-threaded and is then run from a STA could trigger new and untested codepaths in a program.

10.2.14. Wrapup

There are still a lot of topics that have not been covered:

  • Format strings/MOPs

  • IRemoteActivation

  • Complex/simple pings, distributed garbage collection

  • Marshalling IDispatch

  • ICallFrame

  • Interface pointer swizzling

  • Runtime class object registration (CoRegisterClassObject), ROT

  • Exactly how InstallShield uses DCOM

10.2.15. Further Reading

Most of these documents assume you have knowledge only contained in other documents. You may have to reread them a few times for it all to make sense. Don't feel you need to read these to understand DCOM, you don't, you only need to look at them if you're planning to help implement it.

Chapter 11. Wine and OpenGL

11.1. What is needed to have OpenGL support in Wine

Basically, if you have a Linux OpenGL ABI compliant libGL ( installed on your computer, you should have everything that is needed.

To be more clear, I will detail one step after another what the configure script checks.

If, after Wine compiles, OpenGL support is not compiled in, you can always check config.log to see which of the following points failed.

11.1.1. Header files

The needed header files to build OpenGL support in Wine are:


the definition of all OpenGL core functions, types and enumerants


how OpenGL integrates in the X Window environment


the list of all registered OpenGL extensions

The latter file (glext.h) is, as of now, not necessary to build Wine. But as this file can be easily obtained from SGI (, and that all OpenGL should provide one, I decided to keep it here.

11.1.2. OpenGL library itself

To check for the presence of libGL on the system, the script checks if it defines the glXCreateContext function.

11.1.3. glXGetProcAddressARB function

The core of Wine OpenGL implementation (at least for all extensions) is the glXGetProcAddressARB function. Your OpenGL library needs to have this function defined for Wine to be able to support OpenGL.

11.2. How it all works

The core OpenGL function calls are the same between Windows and Linux. So what is the difficulty to support it in Wine? Well, there are two different problems:

  1. the interface to the windowing system is different for each OS. It's called "GLX" for Linux (well, for X Window) and "wgl" for Windows. Thus, one need first to emulate one (wgl) with the other (GLX).

  2. the calling convention between Windows (the Pascal convention or stdcall) is different from the one used on Linux (the C convention or cdecl). This means that each call to an OpenGL function must be "translated" and cannot be used directly by the Windows program.

Add to this some brain-dead programs (using GL calls without setting-up a context or deleting three times the same context) and you have still some work to do :-)

11.2.1. The Windowing system integration

This integration is done at two levels:

  1. At GDI level for all pixel format selection routines (i.e. choosing if one wants a depth / alpha buffer, the size of these buffers, ...) and to do the "page flipping" in double buffer mode. This is implemented in dlls/winex11.drv/opengl.c.

  2. In the OpenGL32.DLL itself for all other functionalities (context creation / deletion, querying of extension functions, ...). This is done in dlls/opengl32/wgl.c.

11.2.2. The thunks

The thunks are the Wine code that does the calling convention translation and they are auto-generated by a Perl script. In Wine Git tree, these thunks are already generated for you. Now, if you want to do it yourself, here is how it all works....

The script is located in dlls/opengl32 and is called make_opengl. It requires Perl5 to work and takes an OpenGL version as argument. It's the version to "simulate". This fixes the list of functions that the Windows application can link directly to without having to query them from the OpenGL driver.

This option can be 1.0 to 1.5. The default is 1.1.

This script generates three files:

  1. opengl32.spec gives Wine's linker the signature of all function in the OpenGL32.DLL library so that the application can link them. Only 'core' functions are listed here.

  2. opengl_norm.c contains all the thunks for the 'core' functions. Your OpenGL library must provide ALL the functions used in this file as these are not queried at run time.

  3. opengl_ext.c contains all the functions that are not part of the 'core' functions. Contrary to the thunks in opengl_norm.c, these functions do not depend at all on what your libGL provides.

    In fact, before using one of these thunks, the Windows program first needs to "query" the function pointer. At this point, the corresponding thunk is useless. But as we first query the same function in libGL and store the returned function pointer in the thunk, the latter becomes functional.

11.3. Known problems

11.3.1. Unknown extension error message:

Extension defined in the OpenGL library but NOT in opengl_ext.c...
Please report ( !

This means that the extension requested by the application is found in the libGL used by Linux (i.e. the call to glXGetProcAddressARB returns a non-NULL pointer) but that this string was NOT found in Wine extension registry.

This can come from two causes:

  1. The opengl_ext.c file is too old and needs to be generated again.

  2. Use of obsolete extensions that are not supported anymore by SGI or of "private" extensions that are not registered. An example of the former are glMTexCoord2fSGIS and glSelectTextureSGIS as used by Quake 2 (and apparently also by old versions of Half Life). If documentation can be found on these functions, they can be added to Wine extension set.

If you have this, run with WINEDEBUG=+opengl and send me the TRACE.

11.3.2. is built but it is still not working

This may be caused by some missing functions required by opengl_norm.c but that your Linux OpenGL library does not provide.

To check for this, do the following steps:

  1. create a dummy .c file:

    int main(void)
        return 0;
  2. try to compile it by linking both libwine and opengl32 (this command line supposes that you installed the Wine libraries in /usr/local/lib, YMMV):

    gcc dummy.c -L/usr/local/lib -L/usr/local/lib/wine -lwine -lopengl32
  3. if it works, the problem is somewhere else (and you can send me an email). If not, you could re-generate the thunk files for OpenGL 1.1 for example (and send me your OpenGL version so that this problem can be detected at configure time).

Chapter 12. Outline of DirectDraw Architecture

This is an outline of the architecture. Many details are skipped, but hopefully this is useful.

12.1. DirectDraw inheritance tree

XVidMode    DGA2

Most of the DirectDraw functionality is implemented in a common base class. Derived classes are responsible for providing display mode methods (EnumDisplayModes, SetDisplayMode, RestoreDisplayMode), GetCaps, GetDeviceIdentifier and internal functions called to create primary and backbuffer surfaces.

User provides for DirectDraw capabilities based on drawing to a Wine window. It uses the User DirectDrawSurface implementation for primary and backbuffer surfaces.

XVidMode attempt to use the XFree86 VidMode extension to set the display resolution to match the parameters to SetDisplayMode.

DGA2 attempt to use the XFree86 DGA 2.x extension to set the display resolution and direct access to the framebuffer, if the full-screen-exclusive cooperative level is used. If not, it just uses the User implementation.

12.2. DirectDrawSurface inheritance tree

 |              |
DIB        Fake Z-Buffer
 |      |         |
User   DGA2   DIBTexture

Main provides a very simple base class that does not implement any of the image-related functions. Therefore it does not place any constraints on how the surface data is stored.

DIB stores the surface data in a DIB section. It is used by the Main DirectDraw driver to create off-screen surfaces.

User implements primary and backbuffer surfaces for the User DirectDraw driver. If it is a primary surface, it will attempt to keep itself synchronized to the window.

DGA2 surfaces claims an appropriate section of framebuffer space and lets DIB build its DIB section on top of it.

Fake Z-Buffer surfaces are used by Direct3D to indicate that a primary surface has an associated z-buffer. For a first implementation, it doesn't need to store any image data since it is just a placeholder.

(Actually 3D programs will rarely use Lock or GetDC on primary surfaces, backbuffers or z-buffers so we may want to arrange for lazy allocation of the DIB sections.)

12.3. Interface Thunks

Only the most recent version of an interface needs to be implemented. Other versions are handled by having thunks convert their parameters and call the root version.

Not all interface versions have thunks. Some versions could be combined because their parameters were compatible. For example if a structure changes but the structure has a dwSize field, methods using that structure are compatible, as long as the implementation remembers to take the dwSize into account.

Interface thunks for Direct3D are more complicated since the paradigm changed between versions.

12.4. Logical Object Layout

The objects are split into the generic part (essentially the fields for Main) and a private part. This is necessary because some objects can be created with CoCreateInstance, then initialized later. Only at initialization time do we know which class to use. Each class except Main declares a Part structure and adds that to its Impl.

For example, the DIBTexture DirectDrawSurface implementation looks like this:

struct DIBTexture_DirectDrawSurfaceImpl_Part
        union DIBTexture_data data; /*declared in the real header*/

typedef struct
        struct DIB_DirectDrawSurfaceImpl_Part dib;
        struct DIBTexture_DirectDrawSurfaceImpl_Part dibtexture;
} DIBTexture_DirectDrawSurfaceImpl;

So the DIBTexture surface class is derived from the DIB surface class and it adds one piece of data, a union.

Main does not have a Part structure. Its fields are stored in IDirectDrawImpl/IDirectDrawSurfaceImpl.

To access private data, one says

DIBTexture_DirectDrawSurfaceImpl* priv = This->private;

12.5. Creating Objects

Classes have two methods relevant to object creation, Create and Construct. To create a new object, the Create method of the class is called. It allocates enough memory for IDirectDrawImpl or IDirectDrawSurfaceImpl as well as the private data for derived classes and then calls Construct.

Each class Construct method calls the base class Construct, then does the necessary initialization.

For example, creating a primary surface with the user ddraw driver calls User_DirectDrawSurface_Create which allocates memory for the object and calls User_DirectDrawSurface_Construct to initialize it. This calls DIB_DirectDrawSurface_Construct which calls Main_DirectDrawSurface_Construct.

Chapter 13. Wine and Multimedia

This file contains information about the implementation of the multimedia layer of Wine.

The implementation can be mainly found in a number of directories:

  • dlls/mmdevapi, dlls/dsound and dlls/winmm for the main sound APIs.

  • dlls/msacm32 (for the audio compression/decompression manager)

  • dlls/msvfw32 (for the video compression/decompression manager).

13.1. Overview

  • the high level abstraction layer: mmdevapi, dsound and winmm DLLs.

  • the mid level layer: handles MCI commands, providing coarser grain operations (like playing a MIDI file, or playing a video stream).

  • the low level layer: device drivers, currently implemented in winealsa.drv, winecoreaudio.drv, and wineoss.drv DLLs. It also has some helper DLLs (like msacm32 and msvfw32), which can make use of external DLLs like mciavi32 or application-provided DLLs.

The low level layer may depend on current hardware and OS services (like OSS on Unix). It provides the core of playback/record using fine grain objects (audio/midi streams...).

The dsound and winmm legacy APIs are implemented on top of mmdevapi, the new Windows audio system introduced with Windows Vista (Multimedia Device API).

13.2. Multimedia architecture

13.2.1. Windows 95 multimedia architecture

Kernel space |                    Client applications
             |           | |         ^ ^       | |          | |
             |        16>| |<32   16>| |<32 16>| |<32    16>| |<32
             |           | v         | |       | v          | v
             |      +----|-----------|---------|------------|-------+
             |      |    |           |         |            |       |  WinMM.dll
             |      |    |           |         |            |       |   32-bit
             |      +----|-----------|---------|------------|-------+
             |           | |         | ^       | |          |
             |  +------+ | |<16      | |       | |<16       |
             |  |   16>| | |         | |       | |          |
             |  |      v v v         | |       v v          v
             |  |   +---------------+---+-------------+-------------+
             |  |   | waveInXXX     |   | mciXXX      | *playSound* |
             |  |   | waveOutXXX    |   |             | mmioXXX     |
             |  |   | midiInXXX     |   |             | timeXXX     |
             |  |   | midiOutXXX    |   |             | driverXXX   |
             |  |   | midiStreamXXX |   |             |             |  MMSystem.dll
             |  |   | mixerXXX      |   |             |             |     16-bit
 +--------+  |  |   | auxXXX    +---+   +---+ mmThread|             |
 |MMDEVLDR|<------->| joyXXX    | Call back | mmTask  |             |
 +--------+  |  |   +-----------+-----------+---------+-------------+
     ^       |  |          |      ^    ^       | ^
     |       |  |       16>|      |<16>|    16>| |<16
     v       |  |          v      |    |       v |
 +--------+  |  |   +-------------+    +----------+
 |  VxD   |<------->|    *.drv    |    | mci*.drv |
 +--------+  |  |   +--------------+   +-----------+
             |  |    |  msacm.drv  |    | mciwave  |
             |  |    +--------------+   +-----------+
             |  |     | midimap.drv |    | mcimidi  |
             |  |     +-------------+    +-----------+
             |  |    Low-level drivers    |    ...   | MCI drivers
             |  |                         +----------+
             |  |                               |
             |  |                               |<16
             |  +-------------------------------+

The important points to notice are:

  • all drivers (and most of the core code) is 16-bit

  • all hardware (or most of it) dependent code reside in the kernel space (which is not surprising)

13.2.2. Windows NT multimedia architecture

Note that Win 98 has mixed 95/NT architecture, so when speaking about Windows 95 (resp. NT) architecture, it refers to the type of architecture, not what's actually implemented. For example, Windows 98 implements both types of architectures.

The important points to notice (compared to the Windows 95 architecture) are:

  • drivers (low level, MCIs...) are 32-bit and Unicode

  • the interfaces between kernel and user drivers has changed, but it doesn't impact much Wine. Those changes allow some good things (like kernel mixing, where different apps share the audio hardware) and of course bad things (like kernel mixing, which adds latency).

13.2.3. Wine multimedia architecture

Kernel space |                    Client applications
             |           | |         ^ ^       | |          | |
             |        16>| |<32   16>| |<32 16>| |<32    16>| |<32
             |           | |         | |       | |          | |
             |  +------+ | |         | |       | |          | |
             |  |32/16>| | |         | |       | |          | |
             |  |      v v v         | |       v v          v v
             |  |   +---------------+---+-------------+-------------+
             |  |   | waveInXXX     |   | mciXXX      | *playSound* |
             |  |   | waveOutXXX    |   |             | mmioXXX     | winmm.dll
             |  |   | midiInXXX     |   |             | timeXXX     |   32-bit
             |  |   | midiOutXXX    |   |             | driverXXX   |
             |  |   | midiStreamXXX |   |             |             | mmsystem.dll16
             |  |   | mixerXXX      |   |             |             |   16-bit
             |  |   | auxXXX    +---+   +---+ mmThread|             |
             |  |   | joyXXX    | Call back | mmTask  |             |
             |  |   +-----------+-----------+---------+-------------+
             |  |         ||      ^    ^       | ^
             |  |      16>||<32   |<16>|    16>| |<16
             |  |         vv      |<32>|    32>v |<32
+---------+  |  |   +-------------+    +----------+
|HW driver|<------->|    *.drv    |    | mci*.drv |
+---------+  |  |   +--------------+   +-----------+
             |  |    | msacm32.drv |    | mciwave  |
             |  |    +--------------+   +-----------+
             |  |     | midimap.drv |    | mcimidi  |
             |  |     +-------------+    +-----------+
             |  |    Low-level drivers    |    ...   | MCI drivers
             |  |                         +----------+
             |  |                               |
             |  |                               |<32/16
             |  +-------------------------------+

From the previous drawings, the most noticeable differences are:

  • low-level drivers can either be 16- or 32-bit (in fact, Wine supports only native wave and audio mappers).

  • MCI drivers can either be 16- or 32-bit

  • all built-in drivers (low-level and MCI) will be written as 32-bit drivers

Wine WinMM automatically adapts the messages to be sent to a driver so that it can convert it to 16 or 32-bit interfaces.

13.3. Low level layers

The low level drivers abstract the hardware specific features from the rest of the multimedia code. Those are implemented on top of mmdevapi, as new versions of Windows do.

Please note that native low level drivers are not currently supported in Wine, because they either access hardware components or require VxDs to be loaded; Wine does not correctly supports those two so far.

There are two specific low level drivers (msacm32.drv for wave input/output, midimap.drv for MIDI output only). These drivers (also present in Windows):

  • allow to choose one low level driver between many (we'll discuss how the choice is made later on)

  • add the possibility to convert stream format (i.e. ADPCM => PCM) (this is useful if the format required by the application for playback isn't supported by the soundcard).

  • add the possibility to filter a stream (adding echo, equalizer... to a wave stream, or modify the instruments that have to be played for a MIDI stream).

13.3.1. Hardware-bound low level drivers

Wine currently supports the following (kernel) multimedia interfaces.

  • Open Sound System (OSS) as supplied in the Linux and FreeBSD kernels by 4Front Technologies. The presence of this driver is checked by configure (depends on the <sys/soundcard.h> file). Source code resides in dlls/wineoss.drv.

  • Advanced Linux Sound Architecture (ALSA) as supplied in the Linux kernel. Source code resides in dlls/winealsa.drv.

  • Core Audio as introduced in Mac OS X 10.3. Source code resides in dlls/winecoreaudio.drv.

13.3.2. Wave mapper (msacm32.drv)

The Wave mapper device allows to load on-demand audio codecs in order to perform software conversion for the types the actual low level driver (hardware). Those codecs are provided through the standard ACM drivers in msacm32 DLL.

Wave mapper driver implementation can be found in dlls/msacm32.drv directory. This driver heavily relies on msacm DLL found in dlls/msacm32. This DLL loads ACM drivers which provide the conversion to PCM format (which is normally supported by low level drivers). A-Law, μLaw, ADPCM, MP3... fit into the category of non PCM formats.

13.3.3. MIDI mapper (midimap)

Midi mapper allows to map each one of 16 MIDI channels to a specific instrument on an installed sound card. This allows for example to support different MIDI instrument definitions (XM, GM...). It also permits to output on a per channel basis to different MIDI renderers.

A built-in MIDI mapper can be found in dlls/midimap. It partly provides the same functionality as the Windows one. It allows to pick up destination channels: you can map a given channel to a specific playback device channel (see the configuration bits for more details).

13.4. Mid level drivers (MCI)

The mid level drivers are represented by some common API functions, mostly mciSendCommand and mciSendString. Wine implements several MCI mid level drivers.

Table 13-1. Wine MCI drivers

MCI NameDLL NameRoleLocationComments
CdAudiomcicdaMCI interface to a CD audio playerdlls/mcicda Relies on NTDLL CD-ROM raw interface (through DeviceIoControl).
WaveAudiomciwave MCI interface for wave playback and record dlls/mciwaveIt uses the low level audio API.
SequencermciseqMidi Sequencer (playback)dlls/mciseqIt uses the low level midi APIs
AviVideomciavi32AVI playback and recorddlls/mciavi32 It rather heavily relies on msvfw32 DLL to work.

The MCI Name column is the name of the MCI driver, as it is searched in configuration. The DLL Name column is the name of the DLL the configuration provides as a value. The name listed here is the default one (see the configuration section for the details).

Adding a new MCI driver is just a matter of writing the corresponding DLL with the correct interface (see existing MCI drivers for the details), and to provide the relevant setup information for wine.inf.

13.5. High level layers

13.5.1. winmm

The high level layers encompass basically the winmm DLL exported APIs. It also provides the skeleton for the core functionality for multimedia playback and recording. Note that native WINMM and MMSYSTEM do not currently work under Wine and there is no plan to support them (it would require to also fully support VxD, which is not done yet).

winmm and mmsystem.dll16 in Wine can handle both 32- and 16-bit drivers (for low level and MCI drivers). Wine will handle all the conversions transparently for all the calls to winmm and mmsystem.dll16, as it knows what the driver interface is (32- or 16-bit) and it manages the information sent accordingly.

MCI drivers are seen as regular Wine modules, and can be loaded (with a correct load order between builtin, native), as any other DLL.

Multimedia timers are implemented with a dedicated thread, run in the context of the calling process, which should correctly mimic Windows behavior. The only drawback is that the thread will appear the calling process if it enumerates the running processes.

13.5.2. dsound

Wine also provide a dsound (DirectX) DLL with the proper COM implementation.

13.6. MS ACM DLLs

13.6.1. Contents

The msacm32 DLL provides a way to map a given wave format to another format. It also provides filtering capabilities. Those DLLs only implement the proper switch between a caller and a driver providing the implementation of the requested format change or filter operation.

There's nothing specific in Wine implementation compared to Windows one. Here's however a list of the builtin format change drivers (there's no filter driver yet):

Table 13-2. Wine ACM drivers

imaadp32IMA ADPCM (adaptative PCM)
msadp32Microsoft ADPCM (adaptative PCM)
msg711Microsoft G.711 (A-Law and μ-Law)
msgsm32Microsoft GSM 06.10
winemp3 Wine MP3 (MPEG Layer 3), based on mpglib library

Note that Wine also supports native audio codecs as well.

All builtin ACM drivers are 32-bit Unicode DLLs.

13.6.2. Caching

msacm32 keeps some data cached for all known ACM drivers. Under the key

are kept for values:

  • aFormatTagCache which contains an array of DWORD. There are two DWORDs per cFormatTags entry. The first DWORD contains a format tag value, and the second the associated maximum size for a WAVEFORMATEX structure. (Fields dwFormatTag and cbFormatSize from ACMFORMATDETAILS)

  • cFilterTags contains the number of tags supported by the driver for filtering.

  • cFormatTags contains the number of tags support by the driver for conversions.

  • fdwSupport (the same as the one returned from acmDriverDetails).

The cFilterTags, cFormatTags, fdwSupport are the same values as the ones returned from acmDriverDetails function.

13.7. MS Video DLLs

13.7.1. Contents

msvfw32 provides encode/decode video streams. Those DLLs only implement the proper switch between a caller and a driver providing the implementation of the requested format coding/decoding operation.

There's nothing specific in Wine implementation compared to Windows one. Here's however a list of the builtin decoding drivers:

Table 13-3. Wine VIDC drivers

msrle32Microsoft RLE (Run-Length encoded)
msvidc32Microsoft Video-1
iccvidRadius Cinepak Video Decoder

Note that Wine also supports native video codecs as well.

All builtin VIDC drivers are 32-bit Unicode DLLs.

13.8. Multimedia configuration

Unfortunately, multimedia configuration evolved over time:

  • In the early days on Windows 3.x, configuration was stored in system.ini file, under various sections ([drivers] for low level drivers, [mci] (resp. [mci32]) for 16-bit (resp. 32-bit) MCI drivers...).

  • With the apparition of the registry, in Windows 95, configuration as been duplicated there, under the key


  • Windows NT also adopted the registry, but decided to store the configuration information under another key than Windows 9x did.

    HKLM\Software\Microsoft\Windows NT\CurrentVersion
    And with a different layout of keys and values beneath this key.

Currently, Wine tries to load first a driver (low-level or MCI) from the NT registry settings. If it fails, it will try the system.ini configuration.

An out-of-the-box configuration is provided in wine.inf, and shall be stored in registry and system.ini at Wine installation time. It will setup correctly the MCI drivers configuration (as well as the wave and MIDI mappers). As the low-level drivers depend on hardware, their setup will be handled by winecfg.

Table 13-4. Wine multimedia configuration scheme

DriverRead from NT registryRead from system.iniSetup by wine.infSetup by winecfg
MCI driversYes (1)Yes (2)YesNo
Wave and MIDI mappersYesNoYesNo
Hardware-bound low level driversYesNoNoYes
ACM and VIDC drivers (audio & video codecs)NoYesYesNo

This will allow most settings to be correctly loaded and handled. However, it won't if an app tries to search directly the registry for the actual configuration, as the three potential configuration places may not be in sync.

It still lacks a correct installation scheme (as any multimedia device under Windows), so that all the correct keys are created in the registry. This requires an advanced model since, for example, the number of wave out devices can only be known on the destination system (depends on the sound card driven by the OSS interface).

The following sections describe which type of information (depending on the location) Wine multimedia DLLs understand.

13.8.1. NT configuration

Under the

HKLM\Software\Microsoft\Windows NT\CurrentVersion
key, are stored the names of the DLLs to be loaded for each MCI driver name:

13.8.2. system.ini

Wine will read the MCI drivers from the [mci] or [mci32] section. Wine won't make any difference between the two.

Here's a sample configuration:


ACM drivers configuration is read (only so far) from the system.ini (and setup at Wine installation from the wine.inf file).


Video (aka vidc) drivers configuration is read (only so far) from the system.ini (and setup at Wine installation from the wine.inf file).


See also the configuration part of the User's Guide for other information on low level drivers.

13.8.3. Per driver/DLL configuration Midi mapper

The Midi mapper configuration is the same as on Windows 9x. Under the key:

if the UseScheme value is not set, or is set to a null value, the MIDI mapper will always use the driver identified by the CurrentInstrument value. Note: Wine (for simplicity while installing) allows to define CurrentInstrument as #n (where n is a number), whereas Windows only allows the real device name here. If UseScheme is set to a non null value, CurrentScheme defines the name of the scheme to map the different channels. All the schemes are available with keys like

For every scheme, under this key, will be a sub-key (which name is usually a two digit index, starting at 00). Its default value is the name of the output driver, and the value Channels lists all channels (of the 16 standard MIDI ones) which have to be copied to this driver.

To provide enhanced configuration and mapping capabilities, each driver can define under the key

a link to and .IDF file which allows to remap channels internally (for example 9 -> 16), to change instruments identification, event controllers values. See dlls/midimap/midimap.c for the details (this isn't implemented yet).