%bochsdefs; ]> KevinLawton BryceDenney ChristopheBothamy MichaelCalabrese The development guide describes resources that are intended for developers in particular. Many Bochs resources are also covered in the User Guide, including compile instructions, bochsrc options, how to find the mailing lists, etc.

If you are an official SourceForge developer, then you can use SVN with write access. The SVN contains the most recent copy of the source code, and with write access you can upload any changes you make to the SVN server for others to use. The SVN checkout command is identical to the one for normal users, but you might want to get the whole tree to work with branches and tags. svn co https://svn.code.sf.net/p/bochs/code bochs-svn Depending on your network connection this may take a long time, since it downloads all files from all branches and tags that exist in the repository at the current revision.

Once you have a Bochs directory, you can compile the files, edit them, test them, etc. See the documentation section, Tracking the source code with SVN for more info on SVN, in the User Manual. But what's new and different is that you can now do SVN commits. When a file is all fixed and ready to share with the rest of the world, you run a commit command to upload your version to the server. First, it's good to do a SVN update to make sure nobody else has changed it since you downloaded it last. At the first commit you'll always have to specify your SF username and type your password. $ svn update file.cc $ svn commit --username sfusername file.cc [editor opens. type log message, save, and exit.] Login area: <https://svn.code.sf.net:443> SourceForge Subversion area Username: sfusername Password for 'sfusername': <--type your password Sending file.cc Transmitting file data . Committed revision 10. When SVN starts an editor, The default is usually vi. If you want a different editor, set the EDITOR environment variable to the name of your preferred editor. When you're done, just save the file and quit the editor. Unless there's some problem, you will see a message that says what the new SVN revision number is, and then "done". If while you're editing the log message, you decide that you don't want to commit after all, don't save the file. Quit the editor, and when it asks where the log message went, tell it to abort. Here is an example of a successful checkin: $ svn commit misc.txt [edit log msg] Sending misc.txt Transmitting file data . Committed revision 6. And here is an aborted one: $ svn commit misc.txt [quit editor without saving] Log message unchanged or not specified a)bort, c)ontinue, e)dit: a

Backups of the SVN repository can be made with the rsync utility. In case of data corruption or other problems on the server, the repository with all revisions, branches and tags can be restored easily. It is recommended to update this backup frequently. The following example creates a folder called bochs-svn-rsync that contains the repository. rsync -av svn.code.sf.net::p/bochs/code bochs-svn-rsync

The Bochs SVN repository is set up to send a notification email to the "bochs-cvs" mailing list after each successful commit. This email contains the log message, a list of the modified files and a diff against the previous revision. The diff of large commits will be truncated at 96 kByte. After each commit the SVN server runs the script post-commit located in the hooks folder. On SourceForge, this script forces a refresh of the Allura code browser and it can call a script post-commit-user for addition operations if it exists. For Bochs we have set up this script and call svnnotify from it to create the notification email. #!/bin/sh svnnotify --repos-path $1 --revision $2 -O -C -d -e 98304 -t bochs-cvs@lists.sourceforge.net

Ideas: - how to browse code with the Allura code browser - how to find an identifier, variable, or specific text in the code - how to make patches with SVN

The initial versions of some sections in this chapter are based on a document written by Peter "Firefly" Lund. It was added and updated in January 2006. The Bochs virtual PC consists of many pieces of hardware. At a bare minimum there are always a CPU, a PIT (Programmable Interval Timer), a PIC (Programmable Interrupt Controller), a DMA controller, some memory (this includes both RAM and BIOS ROMs), a video card (usually VGA), a keyboard port (also handles the mouse), an RTC with battery backed NVRAM, and some extra motherboard circuitry. There might also be an ethernet card, a PCI controller, a soundcard, an IDE controller (+ harddisks/CDROM), a SCSI controller (+ harddisks), a floppy controller, an APIC ... There may also be more than one CPU. Most of these pieces of hardware have their own C++ class - and if Bochs is configured to have more than one piece of a type of hardware, each will have its own object. The pieces of hardware communicates over a couple of buses with each other - some of the things that the buses carry are reads and writes in memory space, reads and writes in I/O space, interrupt requests, interrupt acknowledges, DMA requests, DMA acknowledges, and NMI request/acknowledge. How that is simulated is explained later.&FIXME; In addition to the simulator itself, some other components are required for the communication with the user. The most important parts are these: the window that simulates the monitor and receives keyboard / mouse events the configuration interface that allows to adjust simulation settings the simulator interface for the communication between the other componnents the parameter tree (for configuration settings and save/restore) the logfunctions class (handle and configure panic/error/info/debug) These componnents of Bochs are optional: the plugin interface the builtin debugger the instrumentation feature The simulation window is handled by the GUI object (other terms used in the sources are "display library", "VGAW"). There are many different but compatible implementations of the GUI object, depending on whether you compile for X (Unix/Linux), Win32, Macintosh (two versions: one for Mac OS X and one for older OS's), Amiga, etc. The cross-platform libraries SDL and wxWidgets are also supported. For the configuration interface there are also some different implementations: textconfig (text menus only), wxdialog (wxWidgets port), win32dialog/win32paramdlg (Windows port).

Location Meaning biosSystem and VGA BIOS images, system BIOS sources and makefilebuildadditional stuff required for building Bochs on different platformsbx_debugthe builtin Bochs debuggercputhe cpu emulation sourcescpu/avxsources for emulating AVX instructionscpu/cpudbsources for emulating different cpu modelscpu/fputhe fpu emulation sourcesdisasmthe disassembler for the Bochs debuggerdoc/docbookthe Bochs documentation in DocBook formatdoc/manBochs manual pagesdocs-htmlold Bochs documentation in HTML (will be replaced by DocBook)guidisplay libraries (guis), the simulator interface, the parameter tree and text mode config interfacegui/bitmapsbitmaps for the headerbargui/fontthe default VGA font used by most of the display librariesgui/keymapskeymaps for the keyboard mapping featurehosthost specific drivers (currently only used by the pcidev kernel module for Linux)instrumentdirectory tree for the instrumentation featureiodevstandard PC devices, PCI core devicesiodev/displaydisplay adapters (vga, cirrus, voodoo)iodev/hdimagesupport for different disk image types and lowlevel cdrom accessiodev/networkingnetworking devices and lowlevel modulesiodev/soundsound devices and lowlevel modulesiodev/usbUSB HCs and pluggable devicesmemorymemory management and ROM loadermiscuseful utilities (e.g. bximage, niclist)misc/sb16tool to control the SB16 emulation from the guest sidepatchespending patches

Bochs has many macros with inscrutable names. One might even go as far as to say that Bochs is macro infested. Some of them are gross speed hacks, to cover up the slow speed that C++ causes. Others paper over differences between the simulated PC configurations. Many of the macros exhibit the same problem as C++ does: too much stuff happens behind the programmer's back. More explicitness would be a big win.

C++ methods have an invisible parameter called the this pointer - otherwise the method wouldn't know which object to operate on. In many cases in Bochs, there will only ever be one object - so this flexibility is unnecessary. There is a hack that can be enabled by defining e.g. BX_USE_PIC_SMF to 1 in config.h that makes most methods static, which means they have a "special relationship" with the class they are declared in but apart from that are normal C functions with no hidden parameters. Of course they still need access t/o the internals of an object, so the single object of their class has a globally visible name that these functions use. It is all hidden with macros. Declaration of a class, from iodev/pic.h: ... #if BX_USE_PIC_SMF # define BX_PIC_SMF static # define BX_PIC_THIS thePic-> #else # define BX_PIC_SMF # define BX_PIC_THIS this-> #endif ... class bx_pic_c : public bx_pic_stub_c { public: bx_pic_c(); ~bx_pic_c(); ... BX_PIC_SMF void service_master_pic(void); BX_PIC_SMF void service_slave_pic(void); BX_PIC_SMF void clear_highest_interrupt(bx_pic_t *pic); }; And iodev/pic.cc: ... #define LOG_THIS thePic-> bx_pic_c *thePic = NULL; ... void bx_pic_c::service_master_pic(void) { Bit8u unmasked_requests; int irq; Bit8u isr, max_irq; Bit8u highest_priority = BX_PIC_THIS s.master_pic.lowest_priority + 1; if(highest_priority > 7) highest_priority = 0; if (BX_PIC_THIS s.master_pic.INT) { /* last interrupt still not acknowleged */ return; } isr = BX_PIC_THIS s.master_pic.isr; if (BX_PIC_THIS s.master_pic.special_mask) { /* all priorities may be enabled. check all IRR bits except ones * which have corresponding ISR bits set */ max_irq = highest_priority; } else { /* normal mode */ /* Find the highest priority IRQ that is enabled due to current ISR */ max_irq = highest_priority; ... } ... Ugly, isn't it? If we use static methods, methods prefixed with BX_PIC_SMF are declared static and references to fields inside the object, which are prefixed with BX_PIC_THIS, will use the globally visible object, thePic->. If we don't use static methods, BX_PIC_SMF evaluates to nothing and BX_PIC_THIS becomes this->. Making it evaluate to nothing would be a lot cleaner, but then the scoping rules would change slightly between the two Bochs configurations, which would be a load of bugs just waiting to happen. Some classes use BX_SMF, others have their own version of the macro, like BX_PIC_SMF above.

The CPU class is a special case of the above: if Bochs is simulating a uni- processor machine then there is obviously only one bx_cpu_c object and the static methods trick can be used. If, on the other hand, Bochs is simulating an SMP machine then we can't use the trick. In a UP configuration, the CPU object is declared as bx_cpu. In an SMP configuration it will be an array of pointers to CPU objects (bx_cpu_array[]). Access of a CPU object often goes through the BX_CPU(x) macro, which either ignores the parameter and evaluates to &bx_cpu, or evaluates to bx_cpu_array [n], so the result will always be a pointer. If static methods are used then BX_CPU_THIS_PTR evaluates to BX_CPU(0)->. Ugly, isn't it?

The central object for the communication between several Bochs componnents is the simulator interface (AKA "siminterface"). It especially makes the separation between the configuration interface and the simulator possible. These jobs are done using the SIM object: event handling parsing / writing config file accessing the parameter tree begin and quit simulation execute save and restore code adjust log options access to the debugger core The event handling supports both synchronous and asynchronous types. In case of a synchronous event, the sender waits until the receiver has processed the event (finally setting return code). Asynchronous events are simply added to the event queue and fetched from it later when the receiver processes events. The event handling is important if a multithreaded implementation of gui and config interface like wxWidgets is used. An example for the event handling is the procedure after pressing the "snapshot" tool bar button: wxmain.cc: MyFrame::OnToolbarClick() adds BX_ASYNC_EVT_TOOBAR to the queue wx.cc: bx_wx_gui_c::handle_events() fetches event and calls gui button handler gui.cc: bx_gui_c::snapshot_handler() calls siminterface to ask for a file name siminterface.cc: bx_real_sim_c::ask_filename() creates event BX_SYNC_EVT_ASK_PARAM and waits for completion siminterface.cc: bx_real_sim_c::sim_to_ci_event() calls registered config interface function wxmain.cc: SimThread::SiminterfaceCallback2() creates a wxCommandEvent wxmain.cc: MyFrame::OnSim2CIEvent() calls handler for this event wxmain.cc: MyFrame::HandleAskParamString() calls wxFileDialog() wxmain.cc: MyFrame::OnSim2CIEvent() calls SimThread::SendSyncResponse to set return code() wxmain.cc: SimThread::SiminterfaceCallback2() returns event result siminterface.cc: bx_real_sim_c::ask_filename() returns file name if successful gui.cc: bx_gui_c::snapshot_handler() saves simulation screenshot to file

Starting with version 1.3, the Bochs configuration parameters are stored in parameter objects. These objects have get/set methods with min/max checks and it is possible to define parameter handlers to perform side effects and to override settings. Each parameter type has it's own object type with specific features (numeric, boolean, enum, string and file name). A special object type containing a list of parameters is designed for building and managing configuration menus or dialogs automatically. In the original implementation the parameters could be accessed only with their unique id from a static list or a special structure containing pointers to all parameters. Starting with version 2.3, the Bochs parameter object handling has been rewritten to a parameter tree. There is now a root list containing child lists, and these lists can contain lists or parameters and so on. The parameters are now accessed by a name build from all the list names in the path and finally the parameter name separated by periods. Bit32u megs = SIM->get_param_num("memory.standard.ram.size")->get(); The example above shows how to get the memory size in megabytes from the simulator interface. In the root list (".") there is child list named "memory" containing a child list "standard". It's child list "ram" contains the numeric parameter type "size". The SIM->get_param_num() methods returns the object pointer and the get() method returns the parameter value. The table below shows all parameter types used by the Bochs configuration interface. Type Description bx_object_c Base class for all the other parameter types. It contains the unique parameter id and the object type value. bx_param_c Generic parameter class. It contains the name, label, description and the input/output formats. bx_param_num_c Numerical (decimal/hex) config settings are stored in this parameter type. bx_param_bool_c This parameter type is based on bx_param_num_c, but it is designed for boolean values. A dependency list can be defined to enable/disable other parameters depending on the value change. bx_param_enum_c Based on bx_param_num_c this parameter type contains a list of valid values. bx_param_string_c Configuration strings are stored in this type of parameter. bx_param_filename_c Based on bx_param_string_c this parameter type is used for file names. bx_list_c Contains a list of pointers to parameters (bx_param_*_c and bx_list_c). In the config interface it is used for menus/dialogs.

The save/restore feature is based on an extension to the parameter tree concept. A subtree (list) called "bochs" appears in the root of the parameter tree and some new "shadow" parameter types store pointers to values instead of the values itself. All the hardware objects have register_state() methods to register pointers to the device registers and switches that need to be saved. The simulator interface saves the registered data in text format to the specified folder (usually one file per item in the save/restore list). Large binary arrays are registered with a special parameter type, so they are saved as separate files. The filename is then created from the full parameter path without the prefix "bochs.". The table below shows the additional parameter types for save/restore. Type Description bx_shadow_num_c Based on bx_param_num_c this type stores a pointer to a numerical variable. bx_shadow_bool_c This parameter type stores a pointer to a boolean variable (bit #0 only) or a numerical one (only one selected bit). bx_shadow_data_c This special parameter type stores pointer size of a binary array. The data is saved in a separate file and the text file uses the file name as the value. bx_shadow_filedata_c This special parameter type stores the descriptor of an open file (added in Bochs 2.5). It is also possible to use the bx_param_num_c object with parameter save/restore handlers. With this special way several device settings can be saved to and restored from one single parameter. The disk image state is also handled this way (see below). All devices can uses these two save/restore specific methods: register_state() is called after the device init() to register the device members for save/restore after_restore_state() is an optional method to do things directly after restore (e.g. vga: force a display update) To implement save/restore for hard drive images, the new method register_state() has been added to the base class of the disk image objects. It creates a bx_param_bool_c object called "image" and installs static save and restore handlers. The save operation finally sets the parameter's value (1 = success) and the save/restore handlers are doing the main job. The static handlers call the class-specific code. Depending on the image "mode" they copy either the whole image file or the file containing changes (journal). The files are saved similar to binary arrays with the same naming convention. The restore methods are doing some format or coherency checks, close the open image, copy the file(s) and finally re-open the image.

Like many other open source projects, Bochs uses a configure script created with autoconf. The configure script generates all makefiles and a set of header and support files from templates. This example shows how to add an option to the template file configure.in. The resulting configure script sets up symbols like BX_SUPPORT_BUSMOUSE in the output file config.h and replaces @BUSM_OBJS@ entries in the makefile output. BUSM_OBJS='' AC_MSG_CHECKING(for Busmouse support) AC_ARG_ENABLE(busmouse, AS_HELP_STRING([--enable-busmouse], [enable Busmouse support (InPort)]), [if test "$enableval" = yes; then AC_MSG_RESULT(yes) AC_DEFINE(BX_SUPPORT_BUSMOUSE, 1) BUSM_OBJS='busmouse.o' else AC_MSG_RESULT(no) AC_DEFINE(BX_SUPPORT_BUSMOUSE, 0) fi], [ AC_DEFINE(BX_SUPPORT_BUSMOUSE, 0) AC_MSG_RESULT(no)] ) AC_SUBST(BUSM_OBJS) These output files are generated by the configure script in addition to the makefiles. config.h - the main header file ltdlconf.h - header file required for compiling with libtool bxversion.h - header file containing version strings bxversion.rc - resource file for Windows with version information build/linux/bochs-dlx - DLX Linux shortcut script (Linux only) build/macosx/Info.plist - property list file for MacOSX build/win32/nsis/bochs.nsi - NSIS script for creating Windows installer package

The logfunctions class is one of the base classes of Bochs. It supports 4 log levels (debug, info, error, panic) and 4 possible "actions" that can be done when a log event occurs. Most of the higher level C++ classes of Bochs inherit this class to make the logging configuration per object (here called "module") possible. In the Bochs sources the log events appear as macros (BX_DEBUG, BX_INFO, BX_ERROR, BX_PANIC) and they call the related logfunction methods, unless the symbol BX_NO_LOGGING is set to 1. This is the definition in bochs.h: typedef class BOCHSAPI logfunctions { char *name; char *prefix; int onoff[N_LOGLEV]; class iofunctions *logio; // default log actions for all devices, declared and initialized // in logio.cc. BOCHSAPI_CYGONLY static int default_onoff[N_LOGLEV]; public: logfunctions(void); logfunctions(class iofunctions *); ~logfunctions(void); void info(const char *fmt, ...) BX_CPP_AttrPrintf(2, 3); void error(const char *fmt, ...) BX_CPP_AttrPrintf(2, 3); void panic(const char *fmt, ...) BX_CPP_AttrPrintf(2, 3); void ldebug(const char *fmt, ...) BX_CPP_AttrPrintf(2, 3); void fatal (const char *prefix, const char *fmt, va_list ap, int exit_status); void ask (int level, const char *prefix, const char *fmt, va_list ap); void put(const char *p); void put(const char *n, const char *p); void setio(class iofunctions *); void setonoff(int loglev, int value) { assert (loglev >= 0 && loglev < N_LOGLEV); onoff[loglev] = value; } const char *get_name() const { return name; } const char *getprefix() const { return prefix; } int getonoff(int level) const { assert (level>=0 && level<N_LOGLEV); return onoff[level]; } static void set_default_action(int loglev, int action) { assert (loglev >= 0 && loglev < N_LOGLEV); assert (action >= 0 && action < N_ACT); default_onoff[loglev] = action; } static int get_default_action(int loglev) { assert (loglev >= 0 && loglev < N_LOGLEV); return default_onoff[loglev]; } } logfunc_t;

Here is a short description of some logfunctions methods. The constructor registers a new log module with default values. The module's log prefix is empty and the log levels are set up with default actions. The destructor removes the log module from the table. The info(), error(), panic() and ldebug() methods are called via macros to create a log event of the related level. The fatal() method is called if a log event occurs and it's action is set to "fatal". It is used to shut down the Bochs simulation. The ask() method is called if a log event occurs and it's action is set to "ask". It sends an event to the simulator interface and depending on the return value the simulation continues or it is terminated by calling fatal(). The simulator interface either prompts the user on the console or calls some platform / gui specific code to handle the ask request. The put() methods are used to set up the log module prefix in that appears in the log file and the log module name that appears in the config interface. If the name is not specified, the prefix is used instead. The setio() method sets up the iofunctions class for the log file output. This method is only used by the logfunctions constructors. The getonoff() and setonoff() methods are used by the config interface to display and change the log actions for a Bochs facility. The get_default_action() and set_default_action() methods are also used by the config interface to set up the default action for a log level. The get_name() and getprefix() methods return the strings set up with the put() method. The config interface is also using them to build the menu / dialog to set up the log functions.

The Bochs internal timers are required to provide timer features in the device emulation and for the interaction between simulator and gui. They are implemented in the bx_pc_system_c class and driven by the cpu. When programming a timer the interval is specified in useconds and the timer code translates the value to cpu ticks using the IPS value. In the original implementation the cpu object calls a timer method to increment the system time by one tick after completing one instruction. If a timer has expired, the related timer handler function is called. Now it is also possible to execute a number of cpu instructions, finally update the timer subsystem with this number and possibly call several timer handlers. Here are some examples for timers in the devices and gui code: the PIT (i82C54) system timer at 18.2 Hz the CMOS RTC one-second-timer the display update timer (set up with "vga: update_freq=X") the devices timer (polls keyboard/mouse events from the gui every 1 emulated msecond) the LED auto-off timer (indicating data transfer for min 0.5 seconds) the synchronization timers (realtime/slowdown) are also based on the standard timers These are the capabilities of the Bochs internal timers: register / unregister at runtime activate / deactivate at runtime timer period changeable one-shot or continuous mode

Here are the timer-related definitions and members in pc_system.h: #define BX_MAX_TIMERS 64 #define BX_NULL_TIMER_HANDLE 10000 typedef void (*bx_timer_handler_t)(void *); struct { bool inUse; // Timer slot is in-use (currently registered). Bit64u period; // Timer periodocity in cpu ticks. Bit64u timeToFire; // Time to fire next (in absolute ticks). bool active; // 0=inactive, 1=active. bool continuous; // 0=one-shot timer, 1=continuous periodicity. bx_timer_handler_t funct; // A callback function for when the // timer fires. void *this_ptr; // The this-> pointer for C++ callbacks // has to be stored as well. #define BxMaxTimerIDLen 32 char id[BxMaxTimerIDLen]; // String ID of timer. Bit32u param; // Device-specific value assigned to timer (optional) } timer[BX_MAX_TIMERS]; unsigned numTimers; // Number of currently allocated timers. unsigned triggeredTimer; // ID of the actually triggered timer. Bit32u currCountdown; // Current countdown ticks value (decrements to 0). Bit32u currCountdownPeriod; // Length of current countdown period. Bit64u ticksTotal; // Num ticks total since start of emulator execution. Bit64u lastTimeUsec; // Last sequentially read time in usec. Bit64u usecSinceLast; // Number of useconds claimed since then. // A special null timer is always inserted in the timer[0] slot. This // make sure that at least one timer is always active, and that the // duration is always less than a maximum 32-bit integer, so a 32-bit // counter can be used for the current countdown. static const Bit64u NullTimerInterval; static void nullTimer(void* this_ptr); These are the public timer-related methods for timer control, driving the timers with the cpu and retrieving the internal time implemented in the bx_pc_system_c class: void initialize(Bit32u ips); int register_timer(void *this_ptr, bx_timer_handler_t, Bit32u useconds, bool continuous, bool active, const char *id); bool unregisterTimer(unsigned timerID); void setTimerParam(unsigned timerID, Bit32u param); void start_timers(void); void activate_timer(unsigned timer_index, Bit32u useconds, bool continuous); void deactivate_timer(unsigned timer_index); unsigned triggeredTimerID(void) { return triggeredTimer; } Bit32u triggeredTimerParam(void) { return timer[triggeredTimer].param; } static BX_CPP_INLINE void tick1(void) { if (--bx_pc_system.currCountdown == 0) { bx_pc_system.countdownEvent(); } } static BX_CPP_INLINE void tickn(Bit32u n) { while (n >= bx_pc_system.currCountdown) { n -= bx_pc_system.currCountdown; bx_pc_system.currCountdown = 0; bx_pc_system.countdownEvent(); // bx_pc_system.currCountdown is adjusted to new value by countdownevent(). } // 'n' is not (or no longer) >= the countdown size. We can just decrement // the remaining requested ticks and continue. bx_pc_system.currCountdown -= n; } int register_timer_ticks(void* this_ptr, bx_timer_handler_t, Bit64u ticks, bool continuous, bool active, const char *id); void activate_timer_ticks(unsigned index, Bit64u instructions, bool continuous); Bit64u time_usec(); Bit64u time_usec_sequential(); static BX_CPP_INLINE Bit64u time_ticks() { return bx_pc_system.ticksTotal + Bit64u(bx_pc_system.currCountdownPeriod - bx_pc_system.currCountdown); } static BX_CPP_INLINE Bit32u getNumCpuTicksLeftNextEvent(void) { return bx_pc_system.currCountdown; } This private method is called when the function handling the clock ticks finds that an event has occurred: void countdownEvent(void);

The Bochs timer implementation requires at least one timer to be active. That's why there is a so-called nullTimer to make it work. It is initialized in the constructor on the first timer slot with the highest possible timer interval and it's handler is an empty function. The most important variables of the timer subsystem are initialized on startup with the nullTimer values and updated after each timer modification (register / unregister / activate / deactivate / processing handler). ticksTotal: number of ticks total from emulator startup to the last update of timer subsystem currCountdownPeriod: length of the period from ticksTotal to the next timer event currCountdown: number of ticks remaining until the next timer event occurs The number if ticks since emulator startup is calculated with the formula ticksTotal + currCountdownPeriod - currCountdown and returned with the time_ticks() method. The number of useconds since emulator startup is returned with the time_usec() method computed from the return value of time_ticks() and the IPS value. &FIXME; To be continued

All devices models located on or connected to the Bochs "mainboard" are based on the bx_devmodel_c class. It is defined in iodev.h and it has been introduced along with the plugin interface. That's why the device registration code is still located in plugin.cc, but it is present independent from the --enable-plugins switch. Some code in devices.cc uses the device registration to initialize or reset all devices. It also calls required methods of the save/restore feature. If the debugger is present, a device can resister a method to dump it's status. This is the definition of this class: class BOCHSAPI bx_devmodel_c : public logfunctions { public: virtual ~bx_devmodel_c() {} virtual void init(void) {} virtual void reset(unsigned type) {} virtual void register_state(void) {} virtual void after_restore_state(void) {} #if BX_DEBUGGER virtual void debug_dump(int argc, char **argv) {} #endif }; A device registers it registers itself in the module / plugin init code. In the plugins case the unregister function is called before unloading the plugin. These are the macro definitions that point to functions for registering / unregistering a device and for checking whether or not a device is registered: #define BX_REGISTER_DEVICE_DEVMODEL(a,b,c,d) pluginRegisterDeviceDevmodel(a,b,c,d) #define BX_UNREGISTER_DEVICE_DEVMODEL(a,b) pluginUnregisterDeviceDevmodel(a,b) #define PLUG_device_present(a) pluginDevicePresent(a)

In addition to the default CMOS RAM layout, the Bochs BIOS uses some additional registers for harddisk parameters and the boot sequence. The following table shows all CMOS registers and their meaning. Legend: S - set by the emulator (Bochs) Q - set by the emulator (Qemu) B - set by the bios U - unused by the bios LOC NOTES MEANING 0x00 S rtc seconds 0x01 B second alarm 0x02 S rtc minutes 0x03 B minute alarm 0x04 S rtc hours 0x05 B hour alarm 0x06 S,U day of week 0x07 S,B date of month 0x08 S,B month 0x09 S,B year 0x0a S,B status register A 0x0b S,B status register B 0x0c S status register C 0x0d S status register D 0x0f S shutdown status values: 0x00: normal startup 0x09: normal 0x0d+: normal 0x05: eoi ? else: unimpl 0x10 S fd drive type (2 nibbles: high=fd0, low=fd1) values: 1: 360K 5.25" 2: 1.2MB 5.25" 3: 720K 3.5" 4: 1.44MB 3.5" 5: 2.88MB 3.5" !0x11 configuration bits!! 0x12 S how many disks first (hd type) !0x13 advanced configuration bits!! 0x14 S,U equipment byte (?) bits where what 7-6 floppy.cc 5-4 vga.cc 0 = vga 2 keyboard.cc 1 = enabled 0 floppy.cc 0x15 S,U base memory - low 0x16 S,U base memory - high 0x17 S,U extended memory in k - low 0x18 S,U extended memory in k - high 0x19 S hd0: extended type 0x1a S hd1: extended type 0x1b S,U hd0:cylinders - low 0x1c S,U hd0:cylinders - high 0x1d S,U hd0:heads 0x1e S,U hd0:write pre-comp - low 0x1f S,U hd0:write pre-comp - high 0x20 S,U hd0:retries/bad_map/heads>8 0x21 S,U hd0:landing zone - low 0x22 S,U hd0:landing zone - high 0x23 S,U hd0:sectors per track 0x24 S,U hd1:cylinders - low 0x25 S,U hd1:cylinders - high 0x26 S,U hd1:heads 0x27 S,U hd1:write pre-comp - low 0x28 S,U hd1:write pre-comp - high 0x29 S,U hd1:retries/bad_map/heads>8 0x2a S,U hd1:landing zone - low 0x2b S,U hd1:landing zone - high 0x2c S,U hd1:sectors per track 0x2d S boot from (bit5: 0:hd, 1:fd) 0x2e S,U standard cmos checksum (0x10->0x2d) - high 0x2f S,U standard cmos checksum (0x10->0x2d) - low 0x30 S extended memory in k - low 0x31 S extended memory in k - high 0x32 S rtc century 0x34 S extended memory in 64k - low 0x35 S extended memory in 64k - high 0x37 S ps/2 rtc century (copy of 0x32, needed for winxp) 0x38 S eltorito boot sequence + boot signature check bits 0 floppy boot signature check (1: disabled, 0: enabled) 7-4 boot drive #3 (0: unused, 1: fd, 2: hd, 3:cd, else: fd) 0x39 S ata translation policy - ata0 + ata1 bits 1-0 ata0-master (0: none, 1: LBA, 2: LARGE, 3: R-ECHS) 3-2 ata0-slave 5-4 ata1-master 7-6 ata1-slave 0x3a S ata translation policy - ata2 + ata3 (see above) 0x3b S ata biosdetect flags - ata0 + ata1 (unimplemented) bits 1-0 ata0-master (0: auto, 1: cmos, 2: none) 3-2 ata0-slave 5-4 ata1-master 7-6 ata1-slave 0x3c S ata biosdetect flags - ata2 + ata3 (unimplemented) 0x3d S eltorito boot sequence (see above) bits 3-0 boot drive #1 7-4 boot drive #2 0x3f S BIOS options bits 0 fastboot (skip boot menu delay) 7-1 reserved 0x5b S extra memory above 4GB 0x5c S extra memory above 4GB 0x5d S extra memory above 4GB 0x5f Q number of processors

This section is a detailed description for configuring Sound Blaster 16 from source. If you have a binary and all you want to know is what to put in your bochsrc file, see the sb16 bochsrc option in the user guide. The original version of the Sound Blaster 16 (SB16) emulation for Bochs was written and donated by Josef Drexler. The entire set of his SB16 patches have been integrated into Bochs, however, so you can find everything you need here.

Right now, MPU401 emulation is next to perfect. It supports UART and SBMIDI mode, because the SB16's MPU401 ports can't do anything else as well. The digital audio basically works, but the emulation is too slow for fluent output unless the application doesn't do much in the background (or the foreground, really). The sound tends to looping or crackle on slower computer, but the emulation appears to be correct. Even a MOD player works, although only for lower sampling speeds. The OPL3 chip now also produces output. The source code has been ported from DOSBox and the output data is polled from the mixer thread. Also, the MIDI data running through the MPU401 ports can be written into a SMF, that is the standard midi file. The wave output can be written into a VOC file, which has a format defined by Creative Labs. Output to a WAV file and dual output (device and file at the same time) is now also supported.

Output to the host sound system is supported on Windows, Linux, FreeBSD, MacOS 9, MacOSX and platforms supported by SDL. On Linux using OSS, the output goes to any file or device. If you have a wavetable synthesizer, midi can go to /dev/midi00, otherwise you may need a midi interpreter. For example, the midid program from the DosEmu project would work. Wave output should go to /dev/dsp. These devices are assumed to be OSS devices, if they're not some of the ioctl's might fail. If ALSA is present on Linux and the sound driver is set to alsa, Bochs uses it's default PCM output device and MIDI sequencer. On Windows, midi and wave output go to the midi mapper and the wave mapper, respectively. The device ID for the midi is now selectable. A future version might also have selectable wave output devices. See the next section for more information about the sound lowlevel interface.

You need to configure Bochs using the --enable-sb16 option. There are a few values in config.h that are relevant to the sound functions. Editing config.h after running configure is usually not necessary, since it detects the available drivers and enables them for the compilation. BX_USE_SB16_SMF should be 1 unless you intend to have several sound cards running at the same time. BX_SOUND_LOWLEVEL_NAME is the name of the driver used as the "default" one for all features. The default value of this setting is the dummy driver with no output. The configure script usually changes this value. The following are supported at the moment: alsa Output for Linux with ALSA PCM and sequencer interface oss Output for Linux, to /dev/dsp and /dev/midi00 osx Output for MacOSX midi and wave device sdl Wave output with SDL/SDL2 win Output for Windows midi and wave mappers file Wave and midi output to file dummy Dummy functions, no output Setup the SB16 emulation in your bochsrc, according to instructions in that file (see sb16 option in the user guide).

The source and the DOS executable for the SB16CTRL program that is used to modify the runtime behaviour of the SB16 emulator is included in misc/sb16. See the section SB16CTRL in the user documentation for information about the commands of SB16CTRL.

This file is intended for programmers who would like to port the sound output routines to their platform. It gives a short outline what services have to be provided. You should also have a look at the exisiting files, SOUNDLOW.CC, SOUNDMOD.CC and e.g. SOUNDLNX.CC for Linux or SOUNDWIN.CC for Windows and their respective header files to get an idea about how these things really work.

The main include file for a lowlevel sound driver is iodev.h. It has all definitions for the system-independent functions that a sound driver uses. The sound driver also needs to include soundlow.h for the definitions of the base classes bx_sound_lowlevel_c, bx_soundlow_waveout_c, bx_soundlow_wavein_c and bx_soundlow_midiout_c. Additionally, every output driver will have an include file, which should be included on top of soundmod.cc to allow the emulator to use that driver. The code to initialize the object for the selected drivers can be found in that file, so a soundcard emulation does not need to include the specific driver headers. To actually make the emulator use any specific driver as the default, BX_SOUND_LOWLEVEL_NAME has to be set to the name of the respective driver. Note that if your class contains any system-specific statements, include-files and so on, you should enclose both the include-file and the CC-file in an #if defined (OS-define) construct. Also don't forget to add your file to the list of lowlevel sound object files (SOUNDLOW_OBJS) in the file configure.in and to regenerate the configure script,

#define BX_SOUNDLOW_WAVEPACKETSIZE 19200 #define BX_SOUNDLOW_OK 0 #define BX_SOUNDLOW_ERR 1 typedef struct { Bit16u samplerate; Bit8u bits; Bit8u channels; Bit8u format; Bit16u volume; } bx_pcm_param_t; const bx_pcm_param_t default_pcm_param = {44100, 16, 2, 1}; The maximum size of a wave data packet, the return values of the lowlevel functions, the structure for the PCM parameters and the default parameter set are also important for the sound driver development. They can be found in the main include file soundlow.h. All lowlevel sound methods called from the device code have to return either BX_SOUNDLOW_OK if the function was successful, or BX_SOUNDLOW_ERR if not. If any of the initialization functions fail, the device emulation should disable the affected feature.

The following classes are involved with the sound lowlevel interface: bx_soundmod_ctl_c is a pseudo device that is used to initialize the sound drivers depending on the configuration. bx_sound_lowlevel_c is the base class of the lowlevel sound support. It has methods to return pointers to the objects for the available services waveout, wavein and midiout. The base class returns NULL for all services. bx_sound_dummy_c is derived from bx_sound_lowlevel_c. It returns vaild pointers for all services, but the output classes are only implemented as stubs and the wavein service returns silence. This "dummy" driver is used whenever a OS specific driver does not implement all services. bx_soundlow_waveout_c, bx_soundlow_wavein_c and bx_soundlow_midiout_c are the base classes for the services provided by the Bochs lowlevel sound support. Some methods are stubs and used by the "dummy" sound driver, others are helper methods and used by the OS specific implementations derived from these base classes. bx_sound_OS_c is derived from bx_sound_lowlevel_c. It returns vaild pointers for all services it implements for the selected OS (operating system / library) or NULL for services it does not implement. In the second case the Bochs sound init code falls back to the "dummy" driver.

class bx_sound_lowlevel_c : public logfunctions { public: bx_sound_lowlevel_c(); virtual ~bx_sound_lowlevel_c(); virtual bx_soundlow_waveout_c* get_waveout() {return NULL;} virtual bx_soundlow_wavein_c* get_wavein() {return NULL;} virtual bx_soundlow_midiout_c* get_midiout() {return NULL;} protected: bx_soundlow_waveout_c *waveout; bx_soundlow_wavein_c *wavein; bx_soundlow_midiout_c *midiout; }; The base class for sound lowlevel support is derived from the logfunctions class to make the Bochs logging capabilities available in the sound driver code. The constructor of this base class only initializes all pointers to NULL and the destructor deletes the objects if necessary.

class bx_soundlow_waveout_c : public logfunctions { public: bx_soundlow_waveout_c(); virtual ~bx_soundlow_waveout_c(); virtual int openwaveoutput(const char *wavedev); virtual int set_pcm_params(bx_pcm_param_t *param); virtual int sendwavepacket(int length, Bit8u data[], bx_pcm_param_t *src_param); virtual int get_packetsize(); virtual int output(int length, Bit8u data[]); virtual int closewaveoutput(); virtual int register_wave_callback(void *, get_wave_cb_t wd_cb); virtual void unregister_wave_callback(int callback_id); virtual bool mixer_common(Bit8u *buffer, int len); protected: void convert_pcm_data(Bit8u *src, int srcsize, Bit8u *dst, int dstsize, bx_pcm_param_t *param); void start_mixer_thread(void); bx_pcm_param_t emu_pcm_param, real_pcm_param; int cvt_mult; int cb_count; struct { void *device; get_wave_cb_t cb; } get_wave[BX_MAX_WAVE_CALLBACKS]; int pcm_callback_id; }; The base class for wave output support is also derived from the logfunctions class. In addition to wave output methods used from sound devices, it contains everything required for the mixer thread feature (register PCM sources, convert data formats, start mixer). The constructor should not allocate the output devices. This should be done in openwaveoutput(). This table shows the waveout class methods, where are they called from and if a platform / library specific implementation is required. Method Called from Platform code openwaveoutput()Sound init codeRequiredset_pcm_params()openwaveoutput() and sendwavepacket()Requiredsendwavepacket()Sound device emulationOptionalget_packetsize()Mixer threadOptionaloutput()Mixer threadRequiredclosewaveoutput()Sound device emulationOptionalregister_wave_callback()openwaveoutput() and sound device emulationOptionalunregister_wave_callback()class destructor and sound device emulationOptionalmixer_common()Mixer threadOptionalconvert_pcm_data()InternalNostart_mixer_thread()InternalNo

openwaveoutput() is called when the sound output subsystem initializes. It should do the following: Set up the default PCM parameters for output. Open the given device, and prepare it for wave output. Register the callback function for the PCM buffer queue (sendwavepacket() adds the output to the queue and the mixer thread gets it from there). Start the mixer thread, unless the sound library has it's own one (e.g. SDL). openwaveoutput() will only be called once, whereas set_pcm_params() is called whenever the PCM samplerate has been changed. The parameters are the following: wavedev is the wave output device selected by the user. It is strictly system-dependent. Some sound libraries currently ignore this value and use the default one instead. The value is that of the waveout=device configuration parameter of the sound bochsrc option. Note that only one wave output device will be used at any one time. wavedev may not have the same value throughout one session, but it will be closed before it is changed.

This function should called from openwaveoutput() to initialize the output device with the default parameters and from sendwavepacket() whenever the samplerate has been changed in the emulated sound device. It should do the following: Open the wave output device, unless openwaveoutput() did that already. Prepare the device for data and set the device parameters to those given in the function call. The parameters are the following: param is a pointer to a structure containing the set of parameters required to set up a sound device for PCM output. The members of the structure bx_pcm_param_t are these: samplerate is the desired frequency of the output. Because of the capabities of the soundcards, it can have any value between 5000 and 48,000. bits is either 8 or 16, denoting the resolution of one sample. channels is the number of channels (2 for stereo output, or 1 for mono output. format is a bit-coded value (see below). volume is the output volume to be used by the mixer code. The 16 bit value consists of two 8 bit values for each channel. Bit number Meaning 0 (LSB) 0: unsigned data 1: signed data 1..6 Type of codec (see below) 7 0: no reference byte 1: with reference byte 8..x reserved (0) Value Meaning 0 PCM (raw data) 1 reserved 2 2-bit ADPCM (Creative Labs format) 3 2.4-bit (3-bit) ADPCM (Creative Labs format) 4 4-bit ADPCM (Creative Labs format) Other codecs are not supported by the SB hardware. In fact, most applications will translate their data into raw data, so that in most cases the codec will be zero. The number of bytes per sample can be calculated from this as (bits / 8) * channels.

This function is called whenever a data packet of at most BX_SOUNDLOW_WAVEPACKETSIZE is ready at the soundcard emulation. It should then do the following: Add this wave packet to the waveout buffer chain after converting to 16 bit signed little endian. If the samplerate has been changed set_pcm_params() should be called to update the sound hardware settings. Parameters: length is the number of data bytes in the data stream. It will never be larger than BX_SOUNDLOW_WAVEPACKETSIZE. data is the array of data bytes. src_param is a pointer to a structure containing the PCM parameters (see above). The order of bytes in the data stream is the same as that in the Wave file format: Output type Sequence of data bytes 8 bit mono Sample 1; Sample 2; Sample 3; etc. 8 bit stereo Sample 1, Channel 0; Sample 1, Channel 1; Sample 2, Channel 0; Sample 2, Channel 1; etc. 16 bit mono Sample 1, LSB; Sample 1, MSB; Sample 2, LSB; Sample 2, MSB; etc. 16 bit stereo Sample 1, LSB, Channel 0; Sample 1, MSB, Channel 0; Sample 1, LSB, Channel 1; Sample 1, MSB, Channel 1; etc. Typically 8 bit data will be unsigned with values from 0 to 255, and 16 bit data will be signed with values from -32768 to 32767, although the soundcard emulations are not limited to this. site.

This function is called from the mixer thread to retrieve the size of a wave data packet based on the current samplerate. By default the packet size is big enough to send output for 0.1 seconds. If the host sound driver / library uses a different value, this value should be returned with this method.

This function is called from the mixer thread to send the mixed PCM output to the host sound hardware. Parameters: length is the number of data bytes in the data stream. It will never be larger than the value returned from get_packetsize. data is the array of data bytes.

This function is currently only called from the soundcard emulation if the "file" driver is used. This makes the runtime change of the output file possible. By default this method does nothing and the wave output device is closed in the destructor of the specific class.

This function is called from openwaveoutput() to register the function to retrieve data from the PCM output buffer chain. Other sound emulation devices (e.g. OPL3, PC speaker) can register a function to poll the data from the device emulation. The return value is the ID of the registered function and it is usually used to unregister the source. Parameters: arg is the pointer to the device emulation object. wd_cb is the pointer to a static function that returns wave data from the device emulation. This function is usually called from the mixer_common() method.

This function is usually called from the destructor of the sound emulation device to unregister it's registered function to poll PCM data. If the driver / library doesn't use the default mixer thread, a specific implementation of this method my be required. Parameter: callback_id is the ID of the function to unregister.

This is the main wave output mixing function. It is called from the mixer thread, it polls the wave data from all registered sources and it mixes the data using a simple algorithm (addition and clipping). The return value indicates whether or not wave data is available for output. Parameters: buffer is the output buffer for the wave data. len is the maximum length of the output buffer.

This function converts the PCM data sent from the sound device emulation to the 16 bit stereo signed little endian format. It should be called in sendwavepacket() after allocating the output buffer in the buffer queue. Future versions might also perform resampling here. Parameters: src is the buffer containing data sent from the sound emulation. srcsize is the amount of wave data to be converted. dst is the buffer for the converted wave data. dstsize is the size of the destination buffer. param is a pointer to the struture containing the format parameters of the source data.

This function starts the mixer thread and it should be called in openwaveoutput() unless the sound driver / library has it's own way to do this (e.g. SDL). This function also initializes the mutex required for locking the mixer thread when adding data to the buffer chain or unregistering a source.

class bx_soundlow_wavein_c : public logfunctions { public: bx_soundlow_wavein_c(); virtual ~bx_soundlow_wavein_c(); virtual int openwaveinput(const char *wavedev, sound_record_handler_t rh); virtual int startwaverecord(bx_pcm_param_t *param); virtual int getwavepacket(int length, Bit8u data[]); virtual int stopwaverecord(); static void record_timer_handler(void *); void record_timer(void); protected: int record_timer_index; int record_packet_size; sound_record_handler_t record_handler; }; The base class for wave input support is also derived from the logfunctions class. It contains the framework for wave input (recording) support. The base class is used by the "dummy" sound driver and returns silence to let the input mechanism of the soundcard emulation work. The soundcard emulator object needs to implement a callback function to notifies the emulation about available data. This function usually calls the driver method to get the wave data packet. The driver objects has a periodic timer with an interval of 0.1 emulated seconds that is active during recording. The timer handler processes the wave data recorded with platform or library specific function and finally notifies the emulator. The constructor of the base class only initializes the timer ID. OS specific implementations should initialize other required members here. The destructor of the base class only calls stopwaverecord(). OS specific implementations should close the input device here if necessary.

openwaveinput() is called when the sound emulation first receives a sound recording command. It should do the following: Open the given device, and prepare it for wave input or Store the device name so that the device can be opened in startwaverecord(). In addition to this the record handler value should be stored and the record timer should be registered. This is the definition of record handler callback function: typedef Bit32u (*sound_record_handler_t)(void *arg, Bit32u len); openwaveinput() will only be called once, whereas startwaverecord() is called for every new wave input command to the soundcard emulation. If feasible, it could be useful to open and/or lock the input device in startwaverecord() as opposed to openwaveinput() to ensure that it can be used by other applications while Bochs doesn't need it. The parameters are the following: device is the wave device selected by the user. It is strictly system-dependent. The value is that of the wavein=device configuration parameter of the sound bochsrc option. rh is a pointer to the record handler method of the sound emulation. When sound recording is active, this handler is called periodicly to notify the sound emulation about newly available data. Note that only one wave input device will be used at any one time. device may not have the same value throughout one session, but it will be closed before it is changed.

This method receives a pointer to the required PCM parameters (samplerate, data format) as the argument and it should set up the input device for recording, calculate the size of the recording packet for 0.1 second and start the record timer.

This method is called from the record handler method of the sound emulation device to retrieve the recorded wave data packet.

This method is called to stop the wave recording. It deactivates the timer that calls the method to perform the recording.

class bx_soundlow_midiout_c : public logfunctions { public: bx_soundlow_midiout_c(); virtual ~bx_soundlow_midiout_c(); virtual int openmidioutput(const char *mididev); virtual int midiready(); virtual int sendmidicommand(int delta, int command, int length, Bit8u data[]); virtual int closemidioutput(); }; The base class for MIDI output support is also derived from the logfunctions class. OS specific implementations should initialize required members in the constructor. The destructor of the base class only calls closemidioutput(). OS specific implementations should close the input device here if necessary.

openmidioutput() is called when the first midi output starts. It is only called if the midi output to the driver is active (midimode 1). It should prepare the given MIDI hardware for receiving midi commands. Description of the parameters: mididev is a system-dependent variable. The value is that of the midiout=device configuration parameter of the sound bochsrc option. Note that only one midi output device will be used at any one time. device may not have the same value throughout one session, but it will be closed before it is changed.

midiready() is called whenever the applications asks if the midi queue can accept more data. Return values: BX_SOUNDLOW_OK if the midi output device is ready. BX_SOUNDLOW_ERR if it isn't ready. Note: midiready() will be called a few times before the device is opened. If this is the case, it should always report that it is ready, otherwise the application (not Bochs) will hang.

sendmidicommand()is called whenever a complete midi command has been written to the emulator. It should then send the given midi command to the midi hardware. It will only be called after the midi output has been opened. Note that if at all possible it should not wait for the completion of the command and instead indicate that the device is not ready during the execution of the command. This is to avoid delays in the program while it is generating midi output. Description of the parameters: delta is the number of delta ticks that have passed since the last command has been issued. It is always zero for the first command. There are 24 delta ticks per quarter, and 120 quarters per minute, thus 48 delta ticks per second. command is the midi command byte (sometimes called status byte), in the usual range of 0x80..0xff. For more information please see the midi standard specification. length is the number of data bytes that are contained in the data structure. This does not include the status byte which is not replicated in the data array. It can only be greater than 3 for SysEx messages (commands 0xF0 and 0xF7) data[] is the array of these data bytes, in the order they have in the standard MIDI specification. Note, it might be NULL if length==0.

closemidioutput() is called before shutting down Bochs or when the emulator gets the stop_output command through the emulator port. After this, no more output will be necessary until openmidioutput() is called again, but midiready() might still be called. It should do the following: Wait for all remaining messages to be completed Reset and close the midi output device

This section describes how the three new disk images "undoable", "growing", and "volatile" are implemented in Bochs 2.1. It also applies to the write support the "vvfat" disk image mode in Bochs 2.4.6. undoable -> base r/o file, plus growing, commitable, rollbackable redolog file growing -> growing files, all previously unwritten sectors go to the end of file volatile -> base r/o file, plus hidden growing redolog vvfat -> virtual VFAT disk created from directory, plus hidden growing redolog

The idea behind volatile and undoable disk images is to have a read-only base file, associated with one redolog file. In case of vvfat, a directory is associated with the redolog file. Reading a sector is done from the redolog file if it contains the sector, or from the base file / vvfat directory otherwise. Sectors written go to the redolog, so base image files are opened in read only mode in this configuration. The redolog is designed in a way so it starts as a small file and grows with every new sectors written to it. Previously written sectors are done in place. Redolog files can not shrink. The redolog is a growing file that can be created on the fly. Now, it turns out that if you only use a redolog without any base image file, you get a "growing" disk image. So "undoable", "volatile", "growing" and "vvfat" harddisk images classes are implemented on top of a redolog class.

At the start of a redolog file, there is a header, so Bochs can check whether a file is consistent. This header is also checked when the automatic type and size detection is selected. The generic part of the header contains values like type of image, and spec version number. The header also has a specific part. For redologs, the number of entries of the catalog, the extent, bitmap and disk size are stored. In a redolog, the disk image is divided in a number of equal size "extents". Each extent is a collection of successive 512-bytes sectors of the disk image, preceeded by a n*512bytes bitmap. the n*512bytes bitmap defines the presence (data has been written to it) of a specific sector in the extent, one bit for each sector. Therefore with a 512bytes bitmap, each extent can hold up to 4k blocks Typically the catalog can have 256k entries. With a 256k entries catalog and 512bytes bitmaps, the redolog can hold up to 512GiB All data is stored on images as little-endian values

At the start of a redolog file, there is a header. This header is designed to be reusable by other disk image types. The header length is 512 bytes. It contains : Start position in bytes Length in bytes Data type Description Possible values 0 32 string magical value Bochs Virtual HD Image 32 16 string type of file Redolog 48 16 string subtype of file Undoable, Volatile, Growing 64 4 Bit32u version of used specification 0x00010000, 0x00020000 68 4 Bit32u header size 512 The current version of the header is 0x00020000 (2.0) - see below for details. Start position in bytes Length in bytes Data type Description 72 4 Bit32u number of entries in the catalog 76 4 Bit32u bitmap size in bytes 80 4 Bit32u extent size in bytes 84 4 Bit32u timestamp in FAT format ("undoable" mode only - otherwise reserved) 88 8 Bit64u disk size in bytes The reserved field between "extent" and "disk" has been added in redolog version 2.0 to fix an alignment bug on some platforms. It is now used for consistency check of the "undoable" mode. When creating the redolog file, the timestamp of the read-only file is stored there (in FAT format). After that, the "undoable" mode init code compares the timestamp of the r/o file with the one stored in the redolog.

Immediately following the header, there is a catalog containing the position number (in extents) where each extent is located in the file. Each position is a Bit32u entity.

Each extent starts with a bitmap block of n*512 bytes size. Each byte of the bitmap stores the write status of 8 coresponding disk sectors in the extent (1 = data written).

This is a collection of successive 512-bytes sectors of the disk image. The bitmap preceeding this data block contains the write status of each sector.

The following tables shows what parameters are used when creating redologs or creating "growing" images : Catalog entries Catalog size(KiB) Bitmap size (B) Extent size (KiB) Disk Max Size 512 2 1 4 2MiB 512 2 2 8 4MiB 1k 4 2 8 8MiB 1k 4 4 16 16MiB 2k 8 4 16 32MiB 2k 8 8 32 64MiB 4k 16 8 32 128MiB 4k 16 16 64 256MiB 8k 32 16 64 512MiB 8k 32 32 128 1GiB 16k 64 32 128 2GiB 16k 64 64 256 4GiB 32k 128 64 256 8GiB 32k 128 128 512 16GiB 64k 256 128 512 32GiB 64k 256 256 1024 64GiB 128k 512 256 1024 128GiB 128k 512 512 2048 256GiB 256k 1024 512 2048 512GiB 256k 1024 1024 4096 1TiB 512k 2048 1024 4096 2TiB 512k 2048 2048 8192 4TiB 1024k 4096 2048 8192 8TiB 1024k 4096 4096 16384 16TiB 2048k 8192 4096 16384 32TiB

The class redolog_t(); implements the necessary methods to create, open, close, read and write data to a redolog. It also contains methods for the subtype and consistency check and for the save/restore support. Managment of header catalog and sector bitmaps is done internally by the class.

#define STANDARD_HEADER_MAGIC "Bochs Virtual HD Image" #define STANDARD_HEADER_VERSION (0x00020000) #define STANDARD_HEADER_SIZE (512) These constants are used in the generic part of the header. #define REDOLOG_TYPE "Redolog" #define REDOLOG_SUBTYPE_UNDOABLE "Undoable" #define REDOLOG_SUBTYPE_VOLATILE "Volatile" #define REDOLOG_SUBTYPE_GROWING "Growing" These constants are used in the specific part of the header. #define REDOLOG_PAGE_NOT_ALLOCATED (0xffffffff) This constant is used in the catalog for an unwritten extent.

redolog_t(); instanciates a new redolog. int make_header(const char* type, Bit64u size); creates a header structure in memory, and sets its type and parameters based on the disk image size. Returns 0. int create(const char* filename, const char* type, Bit64u size); creates a new empty redolog file, with header and catalog, named filename of type type for a size bytes image. Returns 0 for OK or -1 if a problem occurred. int create(int filedes, const char* type, Bit64u size); creates a new empty redolog file, with header and catalog, in a previously opened file described by filedes, of type type for a size bytes image. Returns 0 for OK or -1 if a problem occurred. int open(const char* filename, const char* type, Bit64u size); opens a redolog file named filename, and checks for consistency of header values against a type and size. Returns 0 for OK or -1 if a problem occurred. int open(const char* filename, const char* type, Bit64u size, int flags); opens a redolog file with flags applied. This allows to open a redolog in read-only mode. All other parameters and the return value are similar to the default open() method above. void close(); closes a redolog file. off_t lseek(off_t offset, int whence); seeks at logical data offset offset in a redolog. offset must be a multiple of 512. Only SEEK_SET and SEEK_CUR are supported for whence. Returns -1 if a problem occurred, or the current logical offset in the redolog. ssize_t read(void* buf, size_t count); reads count bytes of data of the redolog, from current logical offset, and copies it into buf. count must be 512. Returns the number of bytes read, that can be 0 if the data has not previously be written to the redolog. ssize_t write(const void* buf, size_t count); writes count bytes of data from buf to the redolog, at current logical offset. count must be 512. Returns the number of bytes written. Bit64u get_size(); returns the size stored in the "disk" field in the header. This is used for size autodetection feature ("growing" mode) and the consistency check ("undoable" mode). Bit32u get_timestamp(); returns the value of the "timestamp" field in the header (only used by the "undoable" mode). bool set_timestamp(Bit32u timestamp); writes the timestamp to the header. This is done by the "undoable" mode init code if get_timestamp() returns 0 or the redolog is newly created. static int check_format(int fd, const char *subtype); checks the format of the file with descriptor fd. Returns HDIMAGE_FORMAT_OK if the subtype matches the requested one. This is used for for the image mode autodetection feature. bool save_state(const char *backup_fname); copies the redolog file to a new file backup_fname. This is used by the hdimage save/restore feature.

"volatile" and "undoable" disk images are easily implemented by instanciating a device_image_t object (base image) and a redolog_t object (redolog). "growing" disk images only instanciates a redolog_t object. Class names are undoable_image_t, volatile_image_t and growing_image_t. When using these disk images, the underlying data structure and layout is completely hidden to the caller. Then, all offset and size values are "logical" values, as if the disk was a flat file.

#define UNDOABLE_REDOLOG_EXTENSION ".redolog" #define UNDOABLE_REDOLOG_EXTENSION_LENGTH (strlen(UNDOABLE_REDOLOG_EXTENSION)) #define VOLATILE_REDOLOG_EXTENSION ".XXXXXX" #define VOLATILE_REDOLOG_EXTENSION_LENGTH (strlen(VOLATILE_REDOLOG_EXTENSION)) These constants are used when building redolog file names

undoable_image_t(Bit64u size, const char* redolog_name); instanciates a new undoable_image_t object. This disk image logical length is size bytes and the redolog filename is redolog_name. int open(const char* pathname); opens the disk image pathname in read-only mode, as an undoable disk image. The image mode of this base image is auto-detected. All supported disk image modes can be used here. The associated redolog will be named pathname with a UNDOABLE_REDOLOG_EXTENSION suffix, unless set in the constructor. Returns 0 for OK or -1 if a problem occurred. void close(); closes the base image and its redolog. off_t lseek(off_t offset, int whence); seeks at logical data position offset in the undoable disk image. Only SEEK_SET and SEEK_CUR are supported for whence. Returns -1 if a problem occurred, or the current logical offset in the undoable disk image. ssize_t read(void* buf, size_t count); reads count bytes of data from the undoable disk image, from current logical offset, and copies it into buf. count must be 512. Returns the number of bytes read. Data will be read from the redolog if it has been previously written or from the base image otherwise. ssize_t write(const void* buf, size_t count); writes count bytes of data from buf to the undoable disk image, at current logical offset. count must be 512. Returns the number of bytes written. Data will always be written to the redolog. bool save_state(const char *backup_fname); calls the related redolog_t method to save the image state. void restore_state(const char *backup_fname); called by the hdimage restore code. Copies the backup file to the original location and overwrites the existing redolog file.

volatile_image_t(Bit64u size, const char* redolog_name); instanciates a new volatile_image_t object. This disk image logical length is size bytes and the redolog filename is redolog_name plus a random suffix. int open(const char* pathname); opens the disk image pathname in read-only mode, as a volatile disk image. The image mode is auto-detected. The associated redolog will be named pathname with a random suffix, unless set in the constructor. Returns 0 for OK or -1 if a problem occurred. void close(); closes the base image and its redolog. The redolog is deleted/lost after close is called. off_t lseek(off_t offset, int whence); seeks at logical data position offset in the volatile disk image. Only SEEK_SET and SEEK_CUR are supported for whence. Returns -1 if a problem occurred, or the current logical offset in the volatile disk image. ssize_t read(void* buf, size_t count); reads count bytes of data from the volatile disk image, from current logical offset, and copies it into buf. count must be 512. Returns the number of bytes read. Data will be read from the redolog if it has been previously written or from the base image otherwise. ssize_t write(const void* buf, size_t count); writes count bytes of data from buf to the volatile disk image, at current logical offset. count must be 512. Returns the number of bytes written. Data will always be written to the redolog. bool save_state(const char *backup_fname); calls the related redolog_t method to save the image state. void restore_state(const char *backup_fname); called by the hdimage restore code. Copies the backup file to the original location and overwrites the existing redolog file.

growing_image_t(Bit64u size); instanciates a new growing_image_t object. This disk image logical length is size bytes. int open(const char* pathname); opens the growing disk image pathname, Returns 0 for OK or -1 if a problem occurred. void close(); closes the growing disk image. off_t lseek(off_t offset, int whence); seeks at logical data position offset in the growable disk image. Only SEEK_SET and SEEK_CUR are supported for whence. Returns -1 if a problem occurred, or the current logical offset in the grwoing image. ssize_t read(void* buf, size_t count); reads count bytes of data from the growing disk image, from current logical offset, and copies it into buf. count must be 512. Returns the number of bytes read. The buffer will be filled with null bytes if data has not been previously written to the growing image. ssize_t write(const void* buf, size_t count); writes count bytes of data from buf to the growing disk image, at current logical offset. count must be 512. Returns the number of bytes written. static int check_format(int fd, Bit64u imgsize); checks the format of the file with descriptor fd. Returns HDIMAGE_FORMAT_OK if the file format matches the "growing" one. This is used for the image mode autodetection feature. bool save_state(const char *backup_fname); calls the related redolog_t method to save the image state. void restore_state(const char *backup_fname); called by the hdimage restore code. Copies the backup file to the original location and overwrites the existing redolog file.

Christophe Bothamy, wrote the keymapping code for Bochs, provided these instructions to help developers to add keymapping to a GUI. Bochs creates a bx_keymap_c object named bx_keymap. This object allows you to : - load the configuration specified keymap file - get the translated BX_KEY_* from your GUI key You have to provide a translation function from string to your Bit32u key constant. Casting will be necessary if your key constants are not Bit32u typed. The function must be "static Bit32u (*)(const char *)" typed, and must return BX_KEYMAP_UNKNOWN if it can not translate the parameter string. What you have to do is : - call once "void loadKeymap(Bit32u (*)(const char*))", providing your translation function, to load the keymap - call "Bit32u getBXKey(Bit32u)" that returns the BX_KEY_* constant, for each key you want to map. The file gui/x.cc implements this architecture, so you can refer to it as an example.

Writing a BIOS for a pc-compatible includes the task of embedding various checksums. At least there is the overall bios checksum stored in the very last byte of the program. Depending on the number and types of services the bios provides there are others, e.g. a checksum for the PCI BIOS extensions a checksum for the PCI interrupt routing table a checksum for the plug and play BIOS extensions All these checksums have one common point: using the usual assembler directives they are hard (if not impossible) to compute at compile time. You can either compute them by hand --- a tedious, error-prone task, where in addition you often have to make unreliable assumptions about the memory layout of the entire BIOS. Or you patch them directly into your compiled BIOS image. Apart from computing the checksums this is what biossums does for you.

With the exception of the overall BIOS checksum, in a modern PC BIOS checksums are not used to ensure data integrity. Instead they are used in conjunction with certain signatures to securely identify the entry points or the addresses of important data of some BIOS extensions. Because these services are often invoked from x86 protected mode the original method via interrupts is not applicable. Scanning (even only parts) of the BIOS for (short) signatures and solely relying on this is insecure though, cause the found signature might not refer to the sought service but rather be some obscure machine code resembling the signature by accident. Since signatures are usually part of a larger header or table the above mentioned problem is being circumvented by checksumming over this header and comparing the result to a checksum stored next to the signature. In practice the checksum is often part of the header, chosen in a way that the contents of the header add up to zero.

biossums is very simple and straightforward. The only (and mandatory) argument is the file name of the BIOS image. The file is being read, patched and written. So if you want to keep your original file for reference, use biossums on a copy of your BIOS image. For now, biossums can only rely on signatures to find the locations of the accompanying checksums. Therefore biossums refuses to set any checksums if it finds more than one signature of the same type.

Running upon the current BIOS-bochs-legacy biossums displays: PCI-Bios header at: 0x9610 Current checksum: 0x17 Calculated checksum: 0x17 $PIR header at: 0x99C0 Current checksum: 0x37 Calculated checksum: 0x37 $PnP header at: 0x9C80 Current checksum: 0xD1 Calculated checksum: 0x0F Setting checksum. Bios checksum at: 0xFFFF Current checksum: 0x00 Calculated checksum: 0x28 Setting checksum. If we patch in a second "_32_" signature at offset 0x9F00 and reset the $PIR checksum to 0x00 we get: PCI-Bios header at: 0x9610 Current checksum: 0x17 Calculated checksum: 0x17 PCI-Bios header at: 0x9F00 Current checksum: 0x00 Calculated checksum: 0x00 Multiple PCI headers! No checksum set. $PIR header at: 0x99C0 Current checksum: 0x00 Calculated checksum: 0x37 Setting checksum. $PnP header at: 0x9C80 Current checksum: 0x0F Calculated checksum: 0x0F Bios checksum at: 0xFFFF Current checksum: 0x28 Calculated checksum: 0x05 Setting checksum.

Although biossums takes care of all checksums being used by the BIOS of the Bochs project (as of version 2.02) there are more to cover, e.g. the checksums for "Plug and Play" BIOS extension. In addition it was planned to provide further information to biossums via map-/symbol-files to verify the locations of checksums apart from scanning for signatures. For now this seems not to be necessary; in practice no double signatures have been observed yet.

By default Bochs is compiled as one single executable file providing all features enabled at compile time. To add a new feature (device, gui, driver) several files need to be modified (configure script, makefile, some headers) and the complete build process must be performed. As an alternative, Bochs can be compiled as an executable file containing the core funtionality and a number of loadable modules providing config interface, device emulation, display library (gui) capabilities or drivers for some other features. On Windows platforms such a module is called DLL, other platforms like Linux call it shared library and use libtool to create it. Bochs uses the environment variable LTDL_LIBARY_PATH to search for plugins. To build Bochs in this mode, the configure option --enable-plugins must be used. In this mode it is possible to add externally developed extensions (AKA "user plugins") that can be detected at startup. Rebuilding the whole project is not necessary in that case. Basically these Bochs facilities are currently supported by the plugin interface: config interface i/o device pluggable USB device display library (gui) disk image format (mode) networking module lowlevel sound module

The plugin interface expects a special file name format that depends on the plugin type and platform. On Linux and other platforms using libtool-based compilation the plugin file for the CMOS device gets this name: libbx_cmos.so On Windows platforms (Cygwin, MinGW/MSYS, MSVC) the output file is a DLL and the name is slightly different: bx_cmos.dll The names of device plugins are created from the base name of source files like the example above. For other module types the naming is similar, but with some extensions. This table shows how the names of some existing modules are created: Type Module name Source file name Libtool file name DLL file name Display library sdl2 sdl2.cc libbx_sdl2_gui.so bx_sdl2_gui.dll Disk image module vbox vbox.cc libbx_vbox_img.so bx_vbox_img.dll Networking module slirp eth_slirp.cc libbx_eth_slirp.so bx_eth_slirp.dll Lowlevel sound driver file soundfile.cc libbx_soundfile.so bx_soundfile.dll

Device plugins are categorized into some types depending on their purpose. Some devices are mandatory for the x86 PC emulation or need to be initialized early to make other devices work correctly. A plugin can be loaded only one single time, but some types may be capable to create multiple objects. This is the full list of plugin types defined in extplugin.h with some description: Type Description Entry function name Base C++ class used Remarks PLUGTYPE_CORE Core device plugin, always required or depending on core option, highest priority in init order libmodule_plugin_entry bx_devmodel_c Single device only PLUGTYPE_STANDARD Device plugin, mostly required or depending on core plugins, lower priority in init order libmodule_plugin_entry bx_devmodel_c Single device only PLUGTYPE_OPTIONAL Device plugin depending on normal config option libmodule_plugin_entry bx_devmodel_c Network device plugins can create up to 4 instances PLUGTYPE_VGA VGA-compatible device plugin selected with vga: extension=X option libmodule_plugin_entry bx_vgacore_c One plugin of type required / supported PLUGTYPE_USB USB device plugin selected with the portX parameter of the host controller option libmodule_plugin_entry usb_device_c Number of instances not limited PLUGTYPE_CI Config interface plugin selected with the config_interface option libmodule_gui_plugin_entry - One plugin of type required / supported PLUGTYPE_GUI Display library (gui) plugin selected with the display_library option libmodule_gui_plugin_entry bx_gui_c One plugin of type required / supported PLUGTYPE_IMG Additional disk image format selected with the "mode" parameter when setting up a disk image libmodule_img_plugin_entry device_image_t Number of instances not limited PLUGTYPE_NET Networking driver / emulation module selected with the "ethmod" parameter of the NIC options libeth_module_plugin_entry eth_pktmover_c Some modules can create multiple instances PLUGTYPE_SND Sound driver selected with the sound option libsoundmodule_plugin_entry bx_sound_lowlevel_c One plugin of type required / max. 4 different plugins supported

Each plugin has an entry function that is called during plugin detection, after loading and before unloading the modules. For compatiblity with the "monolithic" Bochs compilation each plugin must have a unique name. When plugins are disabled, the macros / functions for loading / unloading the plugin directly call the plugin entry function. The entry function can be called with these mode arguments: PLUGIN_PROBE returns the plugin type as defined above (mandatory for all plugins) PLUGIN_FLAGS returns flags associated with the plugin PLUGIN_INIT is called after loading the plugin PLUGIN_FINI is called before unloading the plugin At Bochs startup, but before initializing the config options, it searches the plugin paths for modules with the correct file name format and the expected entry function name. Each valid plugin is temporarily loaded to read the plugin type and flags by calling the entry function with the corresponding mode. The plugin interface builds up a database with all information for loading detected modules later on demand. A plugin can return multiple types, but it can only be loaded with one of it (currently used by the "wx" gui and "voodoo" device plugin). The flags can be used to indicate specific capabilities of the plugin (currently on used for device plugins that can be connected to a PCI slot). For unimplemented calling modes the entry function must return 0. To simplify the naming of the plugin entry function some macros have been defined. This example shows how the entry function is defined for normal device plugins on platforms other than Windows: #define PLUGIN_ENTRY_FOR_MODULE(mod) \ extern "C" int CDECL lib##mod##_plugin_entry(plugin_t *plugin, Bit16u type, Bit8u mode) Please see plugin.h for all supported definitions of the entry function. The example below shows the plugin-related section of the "unmapped" device source file. // Define BX_PLUGGABLE in files that can be compiled into plugins. For // platforms that require a special tag on exported symbols, BX_PLUGGABLE // is used to know when we are exporting symbols and when we are importing. #define BX_PLUGGABLE #include "iodev.h" #include "unmapped.h" #define LOG_THIS theUnmappedDevice-> bx_unmapped_c *theUnmappedDevice = NULL; PLUGIN_ENTRY_FOR_MODULE(unmapped) { if (mode == PLUGIN_INIT) { theUnmappedDevice = new bx_unmapped_c(); BX_REGISTER_DEVICE_DEVMODEL(plugin, type, theUnmappedDevice, BX_PLUGIN_UNMAPPED); } else if (mode == PLUGIN_FINI) { delete theUnmappedDevice; } else if (mode == PLUGIN_PROBE) { return (int)PLUGTYPE_OPTIONAL; } return(0); // Success }

To ensure compatibility between both compilation modes a bunch of macros have been defined in plugin.h. If required the specific functions are implemented in plugin.cc. That's why the code for the modules that can be plugins doesn't need special cases for "plugin" and "non-plugin" mode. For the plugin types PLUGTYPE_CORE and PLUGTYPE_STANDARD the macros for loading / unloading plugin directly call the entry function. For the other types a static list is created at compile time using a modified version of the plugin_t structure. This is the counterpart to the dynamic list in plugin mode created at startup. The load / unload functions are similar in both modes, except that the "non-plugin" version of these functions finally just call the entry function. These macros are defined for both modes, but calling mode specific code: PLUG_load_plugin(name,type) PLUG_get_plugins_count(type) PLUG_get_plugin_name(type,index) PLUG_get_plugin_flags(type,index) PLUG_load_plugin_var(name,type) PLUG_load_opt_plugin(name) PLUG_unload_opt_plugin(name)

This device was added by Dave Poirier (eks@void-core.2y.net). Compiling Bochs with iodebug support ./configure --enable-iodebug make Other optional fields may be added to the ./configure line, see Bochs documentation for all the information. To enable the iodebug plugin at runtime, it must be loaded with the 'plugin_ctrl' bochsrc option. Using the I/O Interface to the debugger port range: 0x8A00 - 0x8A01 Port 0x8A00 servers as command register. You can use it to enable the i/o interface, change which data register is active, etc. Port 0x8A01 is used as data register for the memory monitoring.

0x8A00 Used to enable the device. Any I/O to the debug module before this command is sent is sent will simply be ignored. 0x8A01 Selects register 0: Memory monitoring range start address (inclusive) 0x8A02 Selects register 1: Memory monitoring range end address (exclusive) 0x8A80 Enable address range memory monitoring as indicated by register 0 and 1 and clears both registers 0x8AE0 - Return to Debugger Prompt If the debugger is enabled (via --enable-debugger), sending 0x8AE0 to port 0x8A00 after the device has been enabled will return the Bochs to the debugger prompt. Basically the same as doing CTRL+C. 0x8AE2 - Instruction Trace Disable If the debugger is enabled (via --enable-debugger), sending 0x8AE2 to port 0x8A00 after the device has been enabled will disable instruction tracing 0x8AE3 - Instruction Trace Enable If the debugger is enabled (via --enable-debugger), sending 0x8AE3 to port 0x8A00 after the device has been enabled will enable instruction tracing 0x8AE4 - Register Trace Disable If the debugger is enabled (via --enable-debugger), sending 0x8AE4 to port 0x8A00 after the device has been enabled will disable register tracing. 0x8AE5 - Register Trace Enable If the debugger is enabled (via --enable-debugger), sending 0x8AE5 to port 0x8A00 after the device has been enabled will enable register tracing. This currently output the value of all the registers for each instruction traced. Note: instruction tracing must be enabled to view the register tracing 0x8AFF Disable the I/O interface to the debugger and the memory monitoring functions. all accesses must be done using word reading this register will return 0x8A00 if currently activated, otherwise 0

All accesses to this port must be done using words. Writing to this port will shift to the left by 16 the current value of the register and add the provided value to it. Sample: reg0 = 0x01234567 out port: 0x8A01 data: 0xABCD reg0 = 0x4567ABCD

Enable memory monitoring on first page of text screen (0xb8000-0xb8fa0): add in bochrc file: optromimage1: file="asmio.rom", address=0xd0000 /* * Make asmio ROM file: * gcc -c asmio.S * objcopy -O binary asmio.o asmio.rom */ .text .global start .code16 /* ROM Header */ .byte 0x55 .byte 0xAA .byte 1 /* 512 bytes long */ start: /* Monitor memory access on first page of text screen */ mov $0x8A00,%dx /* Enable iodebug (0x8A00->0x8A00) */ mov %dx,%ax out %ax,%dx mov $0x8A01,%ax /* Select register 0 start addr (0x8A01->0x8A00) */ out %ax,%dx mov $0x8A01,%dx /* Write start addr 0xB8000 (high word first) */ mov $0xB,%ax out %ax,%dx mov $0x8000,%ax /* Write start addr (low word) */ out %ax,%dx mov $0x8A02,%ax /* Select register 1 end addr (0x8A02->0x8A00) */ mov $0x8A00,%dx out %ax,%dx mov $0x8A01,%dx /* Write end addr 0xB8FA0 (high word first) */ mov $0xB,%ax out %ax,%dx mov $0x8FA0,%ax /* Write end addr (low word) */ out %ax,%dx mov $0x8A00,%dx /* Enable addr range memory monitoring (0x8A80->0x8A00) */ mov $0x8A80,%ax out %ax,%dx mov $0x8A00,%dx /* Return to Bochs Debugger Prompt (0x8AE0->0x8A00) */ mov $0x8AE0,%ax out %ax,%dx lret .byte 0x6b /* Checksum (code dependent!, update it as needed) */ .align 512 /* NOP follow */

&FIXME; Not written yet.

&FIXME; Not written yet (take stuff from bxdebugger.html).

Don't make use of any external C++ classes. They are not offered on all platforms and this would make Bochs non-portable. There is use of such classes in the optional debugger. I plan on removing this use. Don't use fancy C++ features. Bochs is incredibly performance sensitive, and will be increasingly so as more speed enhancements are added. There's a time and place for most everything and this is not it. Some advanced features create overhead in the generated code that you don't see. They also convolute the code, and sometimes occlude that is really going on. Don't use templates Don't use virtual functions if not strictly required Don't use C++ exceptions Use soft tabs. At least when you submit code, convert all hard tabs to spaces. There is no uniform way to handle tabs properly. Please do compile with all warnings turned on. It's really difficult to spot interesting warnings when a compile is littered with non-interesting ones. Don't use signed ints where unsigned will do. Make sure that contributed code / patches are LGPL compatible.

Update version number and strings in configure.in. VERSION="2.6.9" VER_MAJOR=2 VER_MINOR=6 VER_REVISION=9 VER_SVN=0 REL_STRING="Build from SVN snapshot on April 9, 2017" In the README file you have to update version number and date. Add some information about new features if necessary. Bochs x86 Pentium+ Emulator Updated: Sun Apr 9 08:45:00 CEST 2017 Version: 2.6.9 In the file bochs.manifest you have to update the version number for the Windows build. version="2.6.9.0" Check date, update/sumup info in CHANGES. Run autoconf to regenerate configure and check them in. Create an SVN tag that contains all files of the revision that was used in the release. For prereleases I create a normal SVN tag like this: svn mkdir tags/REL_2_5_pre1_FINAL svn copy trunk/bochs tags/REL_2_5_pre1_FINAL/bochs svn commit But for a real release, I make an SVN branch tag AND a normal tag. svn mkdir tags/REL_2_5_FINAL svn copy trunk/bochs tags/REL_2_5_FINAL/bochs svn mkdir branches/REL_2_5 svn copy trunk/bochs branches/REL_2_5/bochs svn commit The tag marks where the branch split off of the main trunk. This is very useful in maintaining the branch since you can do diffs against it. svn diff tags/REL_2_5_FINAL/bochs trunk/bochs svn diff tags/REL_2_5_FINAL/bochs branches/REL_2_5 etc. All bugfix-only releases after the final release should be created from the REL_2_5 branch. Now you can start building packages with the sources from the created release tag.

An anonymous SVN checkout from the release tag is the base for all official release packages. Do this checkout from the release tag and specify a not yet existing directory name with the version number as the destination. Then create the source package from this new directory. These steps can be done both on Linux and Windows (Cygwin). svn co http://svn.code.sf.net/p/bochs/code/tags/REL_2_5_FINAL/bochs bochs-2.5 tar czvf bochs-2.5.tar.gz --exclude=.svn bochs-2.5 The source TAR file bochs-2.5.tar.gz is ready for upload.

The RPM will be building using the configuration in .conf.linux with a few parameters from build/redhat/make-rpm. Make any last minute changes to .conf.linux. Any changes will go into the source RPM. The DLX Linux demo package will be downloaded from the Bochs website to the Bochs root directory if it is not already present there. ./build/redhat/make-rpm | tee ../build.txt This produces two rpm files in the current directory. Test and upload.

These instructions require cygwin or mingw32/mingw64 to prepare the sources and MSVC++ to build the binaries. Use the Bochs sources from the SVN checkout or unpack the TAR file. In Cygwin or mingw32/mingw64: sh .conf.win32-vcpp # runs configure make win32_snap # unzip workspace, make a win32 source ZIP The source ZIP is present in the parent directory of the Bochs root and now ready for upload. To build the binary package, copy it to a Windows machine, if necessary. Open up Visual C++ and load the workspace file Bochs.sln. Check the Build:Set Active Project Configuration is set the way you want it. For releases I use "Win32 Release". To create "bochsdbg.exe" with Bochs debugger support, manually change these lines in config.h to turn on the debugger and the enhanced debugger gui. #define BX_DEBUGGER 1 #define BX_DEBUGGER_GUI 1 One of the optimization features must be turned off, since it is currently not compatible with the debugger. #define BX_SUPPORT_HANDLERS_CHAINING_SPEEDUPS 0 VC++ will rebuild Bochs with debugger and overwrite bochs.exe. To avoid trashing the non-debug version, move it out of the way while the debugger version is being built. Then rename the debugger version to bochsdbg.exe. cd obj-release mv bochs.exe bochs-normal.exe (build again with BX_DEBUGGER=1 this time) mv bochs.exe bochsdbg.exe mv bochs-normal.exe bochs.exe Do make install_win32 into the NSIS folder in the Bochs source tree: make install_win32 INSTDIR=./build/win32/nsis/bochs-2.5 This downloads and unpacks both the DLX Linux demo and the HTML docs from the Bochs website, copies all the files into ./build/win32/nsis/bochs-2.5 and then creates a binary ZIP file in the NSIS folder. Now make the NSIS installer package (the current script is known to work with NSIS 3.04) cd build/win32/nsis make That gives an installer called Bochs-2.5.exe. Test and upload it.

When you are ready with creating release packages you have to upload them using the SF file manager feature. Create a subdirectory with the version number in the bochs directory. Point the download destination to the new directory and start uploading packages. The top of the CHANGES file should be used as the release notes. After setting up the file properties the new release is ready for download. After having all files set up in the download area, don't forget to post an announcement containing a brief summary of changes to the bochs-announce mailing list and the "Project News" section on SF.

The Bochs project webspace is stored under the SF directory /home/project-web/bochs. It can be accessed from the SF shell using SSH or with the commands sftp, scp and rsync. Some parts of the directory structure must be updated from the local CVS repository, others from Bochs SVN (directories bochs and sfsite). The online documentation, disk images and screenshots must be uploaded manually. Location Meaning cgi-binCGI scripts for the websitehtdocsroot directory of the websitehtdocs/doc/docbookBochs online documentationhtdocs/docs-htmlold Bochs documentationhtdocs/guestosdisk images directly stored on the Bochs websitehtdocs/screenshotscreenshots of Bochs running several guest operating systemshtdocs/svn-snapshotlink to current snapshothtdocs/techspectechnical specifications of several hardware componentslxrBochs source browsersfsite-cvsrootlocal CVS repositorysitebinshell scripts (e.g. for snapshot generation)sitemanwebsite manual pagessnapshotSVN snapshot storage areatmptemp directory for shell scripts

The main HTML content of the Bochs website (except online documentation) is stored in the sfsite directory of the Bochs SVN repository. Unlike other SF projects you don't need to upload these files to the Bochs project webspace. Running a simple SVN update on the SF shell is enough after the files have been updated in the repository. Please see Setting up SVN write access for general instructions. The only difference is the directory name sfsite instead of bochs. The example below shows how to start the SF shell with SSH and to update the HTML files. ssh -t vruppert,bochs@shell.sourceforge.net create vruppert,bochs@shell.sourceforge.net's password: Requesting a new shell for "vruppert" and waiting for it to start. queued... starting... This is an interactive shell created for user vruppert,bochs. Use the "timeleft" command to see how much time remains before shutdown. Use the "shutdown" command to destroy the shell before the time limit. For path information and login help, type "sf-help". [vruppert@shell-24002 ~]$ svnserve -d -r /home/svn/p/bochs/code [vruppert@shell-24002 ~]$ cd /home/project-web/bochs/htdocs/ [vruppert@shell-24002 htdocs]$ svn update U index.html Updated to revision 10752 [vruppert@shell-24002 htdocs]$ shutdown Requesting that your shell be shut down. This request will be processed soon. [vruppert@shell-24002 htdocs]$ Broadcast message from root (Mon Oct 31 09:45:04 2011): The system is going down for system halt NOW! Connection to shell-24002 closed by remote host. Connection to shell-24002 closed. Connection to shell.sourceforge.net closed.

The SVN snapshot The SVN snapshot link can be found on the bottom of the page getcurrent.html. can be updated with SF shell access using SSH. There is a script called update-svn-snapshot.sh that can do all the required steps (checking out SVN, packing the source tree into one archive, updating the website link). See previous section how to create a shell. cd /home/project-web/bochs/sitebin/ ./update-svn-snapshot.sh

To update the online documentation, a file called bochsdoc.tar.gz must be generated with the make. This file must be uploaded to the location of the online documentation on SF using scp. cd doc/docbook make bochsdoc.tar.gz scp bochsdoc.tar.gz vruppert,bochs@web.sf.net:htdocs/doc/docbook After a successful upload, the HTML files must be unpacked from the SF shell. See section Updating the Bochs website content how to create a shell. cd /home/project-web/bochs/htdocs/doc/docbook tar xvzf bochsdoc.tar.gz The updated files can be accessed from the sidebar of the Bochs website.

&FIXME; sources, tmp

&FIXME; sources, tmp