smarter. If you use --enable-x86-64, it has the following side
effects on other options:
- cpu level defaults to 6 instead of 5
- ignore bad msr defaults to on (Peter T. requested this)
- if cpu level is 6 or greater, these options are enabled by default.
- 4meg pages (PSE)
- physical address extensions (PAE)
- global pages (PGE)
- apic support
- a few weeks ago I changed the gui code so that it always made a copy
of the clipboard data using new Bit8u[] and passed it into the keyboard
code. But I didn't get the keyboard code quite right, and Christophe
noticed the incorrect malloc that I had forgotten to remove. I changed
it to work as I intended:
1. gui code allocates paste buffer (new), copies clipboard data in
2. gui code passes the buffer to the keyboard code, and forgets about it
3. keyboard code uses the buffer, then frees it when finished (delete)
- modified: iodev/keyboard.h iodev/keyboard.cc
. 4 ata interfaces support
. devices auto-detection
. 16bits/32bits device access
. EDD3.0 support
. ata/atapi layer rewrite
. added verious DEBUG_XXX to debug only parts of the bios, because we're running low on data space in debug mode
. moved ata data at offset 0x121 in EBDA. Previously, data was being overwritten by keyboard.sys
. added jump to [0x40:0x67] after reset if shutdown status is 0x05. Fixes bug [ 601166 ] CMOS Problem @ "0x0F Index 0x05 data"
. DMA controller init on POST entry
configuring for a different machine. This is used to short-circuit
certain sanity checks, such as existence of certain required libraries.
- in --with-wx configuration, when $cross_configure=1, do not abort
configure if library is not found.
- if wxWindows version not found, print "not_found" instead of nothing
From his patch file text:
> This patch adds Promise DC2300 VLB IDE Support.
> You may find msdos and win95 drivers on the net. Look for
> P2300W95.ZIP and DC2300VLBIDEver260b.ZIP.
>
> The good news is that now win95 natively sees my cdrom,
> and that the disks are not in msdos compatibility mode any more.
>
> The bad news is that it works only for the first ata interface.
>
> I tested that patch on msdos and win95 only.
>
> Some info on VLB IDE can be found at http://ryston.cz/petr/vlb/
I got Win95 running with 32-bit paging/filesystem using the
recommended driver and these patches. Since the patches did
such a good job bracketing code modifications with a #define,
they might as well become part of the current CVS code.
use getB_CF() etc. getB_CF() and friends are only for a relatively
small number of cases where a true boolean/binary number (0 or 1) is required
rather than 0 or non-0 as is returned by get_CF().
loadSRegLMNominal() which should be used to load a segment register
in long-mode with nominal values which are compatible with existing
checks and expectations for descriptor cache values.
Fixed 64-bit iret to not do a descriptor fetch if SS selector is null.
Also load SS with loadSRegLMNorminal() in the same case.
was not correct (used == 0, rather than s&0xffc == 0). Also,
with a null SS selector, it was fetching the descriptor anyways.
Put more code inside the if (selector != NULL) clause.
For a temporary measure I added the local INIT_64_DESCRIPTOR
from segment_ctrl_pro.cc, and used it in the case that the
SS selector is null. We need to make a real function which
sets a descriptor in long-mode to nominal values. I'm going
to do that next... I can't stand seeing the current hacks. :^)
section title, and less like the grammar that you're supposed to use.
Someone on irc was confused because it was written in uppercase, but then
bochs did not accept the uppercase ATA0-MASTER parameter.
Fixed/updated/cleaned guest2host TLB speedups for Long mode.
I now can boot the Linux x86-64 kernel to the VFS mount message,
using all the accelerations.
these from interfering from a normal compile here's what I did.
In config.h.in (which will generate config.h after a configure),
I added a #define called KPL64Hacks:
#define KPL64Hacks
*After* running configure, you must set this by hand. It will
default to off, so you won't get my hacks in a normal compile.
This will go away soon. There is also a macro just after that
called BailBigRSP(). You don't need to enabled that, but you
can. In many of the instructions which seemed like they could
be hit by the fetchdecode64() process, but which also touched
EIP/ESP, I inserted a macro. Usually this macro expands to nothing.
If you like, you can enabled it, and it will panic if it finds
the upper bits of RIP/RSP set. This helped me find bugs.
Also, I cleaned up the emulation in ctrl_xfer{8,16,32}.cc.
There were some really old legacy code snippets which directly
accessed operands on the stack with access_linear. Lots of
ugly code instead of just pop_32() etc. Cleaning those up,
minimized the number of instructions which directly manipulate
the stack pointer, which should help in refining 64-bit support.
user can turn on/off use of native host specific inline asm
statements. By default, this option is enabled, so you only
need it to disable inline asms in your compile for now.
Currently only on x86+GCC environments, will inline asm()
statements be used. Eventually, other platforms could specify
some asm()s; probably for endian issues such as byte-swapping
and unaligned memory accesses. On x86, there are some inline
asm()s which do the arithmetic EFLAGS processing so that the
lazy flags handling is somewhat bypassed. Eventually, I'll
add more, at least for the more common instructions. This
adds a little extra performance.
giving it Bit32u pointers. On MacOSX for some reason Boolean is set to
unsigned char instead of unsigned int, so it actually cares that we get the
type right.
- return model=2 so that Linux recognizes the processor as having an APIC.
We don't really know what Hammer returns.
- in SetCR4, allow bits 9 and 10 to be written
- added jump to [0x40:0x67] after reset if shutdown status is 0x05. Fixes bug [ 601166 ] CMOS Problem @ "0x0F Index 0x05 data"
- changed various occurences of "mov ax, #0000" by "xor ax, ax"
- fixed serious bugs in EDD function 0x48
- included Volker's rombios recent changes to PCIBIOS
- initialization of DMA controller on POST entry
The function gets the real time in useconds and puts it in
a Bit64u. This function is defined when:
BX_HAVE_REALTIME_USEC is 1.
Right now, BX_HAVE_REALTIME_USEC is defined to be BX_HAVE_GETTIMEOFDAY
and bx_get_realtime64_usec is defined in terms of gettimeofday().
However, it could be defined in terms of any other method of obtaining
the current time accurate to the usecond. That is why I moved the
function to osdep and added the new define.
the icache pageStamp check too early, before it was known
that the TLB entry would produce a physical address in
range of the normal part of physical memory. PCI accesses
were causing seg faults because of this. I haven't tested
this for PCI.
- Features :
. number of active channels defined at boot-time config
. new options in bochsrc
. up to 8 devices support (disks or cdroms)
. up to 4 cdrom devices can be changed at runtime config
. wxwindows config interface
which says to paste getB_ with flag and then paste with (. It should
be "getB_##flag(void)". Some preprocessors are complaining about pasting
the symbol with the paren.
of (1 & (val32>>N)), and added a getB_?F() accessor for special
cases which need a strict binary value (exactly 0 or 1). Most
code only needed a value for logical comparison. I modified the
special cases which do need a binary number for shifting and
comparison between flags, to use the special getB_?F() accessor.
Cleaned up memory.cc functions a little, now that all accesses
are within a single page.
Fixed a (not very likely encountered) bug in fetchdecode.cc (and
fetchdecode64.cc) where a 2-byte opcode starting with a prefix
starts at the last offset on a page. There were no checks
on the segment overrides for a boundary condition. I added them.
The eflags enhancements added just a tiny bit of performance.
- don't allow MMX on cpu level < 5.
- require FPU support on cpu level >= 55
- don't allow MMX support without FPU support (moved this check from
cpu/i387.h to config.h)
- PacketSetReadTimeout: instead of calling with a very short timeout,
now we call it with -1 which means no timeout.
- Psyon found that Peter Tattam's changes, while necessary for win95, broke
WinNT/2000. So now on Windows NT/2000, we still call
WaitForSingleObject(lpAdapter->ReadEvent,0) but we ignore the return value.
parameter values associated with the dialog and updated the wxWindows
controls. At the time I didn't realize that I was overriding
wxWindow::Refresh() which repaints the window. Later, I renamed the method
to CopyParamToGui() to make it more clear, but many of the callers in
wxmain.cc continued to call Refresh(), which now reverted to the parent class
wxWindow::Refresh(). Since there was no compile error I didn't notice for a
while, but it caused the ParamDialogs to repaint themselves constantly but
never actually change their values. This is now fixed by changing those
method calls to CopyParamToGui().
started using a wxTimer to trigger the redraws. Now instead of calling
MyPanel::OnPaint directly, I call Refresh() instead. This makes the Windows
display work correctly.
I'm not sure why this wasn't needed before, but after switching to
wxWindows 2.3.3 it seems to be important.
- in a wxLogDebug() call I was trying to pass a struct instead of a string.
Fixed that.
- hardcode BX_HAVE_GETTIMEOFDAY=0 if cross compiling for VC++
(this part hasn't been tested)
- modified files: configure configure.in config.h.in iodev/pit_wrap.cc
so frequently.
Coded asm() statements for INC/DEC_ERX() instructions.
Cleaned up the iCache a litle including a bug fix. The
generation ID was decrementing the whole field including
some high meta bits. That could roll over after 1 Billion
cycles. I know only decrement if the field is valid, to
save the write.
I implemented inline functions which can serve the value of
the arithmetic flags if they are cached, and redirect to
the lazy_flags.cc routines if not.
Most of this was just prep work for adding more asm() statements
for native eflags processing when on x86.
'char_changed'. These variables are currently only used by the win32 gui for
the update of the font bitmaps. SDL and wxWindows do not use the variables
since they are using the charmap data directly
- free text snapshot memory when the user has cancelled the dialog
- write text snapshot file in binary mode (disables the LF -> CRLF conversion
in cygwin)
- pci2isa prefix changed to "P2I"
- device name changed to "PIIX3 PCI-to-ISA bridge"
- added more default values to the reset() function
- added stubs for i/o mapped registers
- array pci_conf[] is now a part of the structure s
a precursor to making it a runtime option. I'm not
sure how to do this, but it should be trivial. All it
would take is adding the option and changing the
macro for BX_USE_REALTIME_PIT to point to that option.
also extended by the REX.B field on Hammer) is passed to instructions.
I rearranged the bxInstruction_c to free up a field to be used
to pass this info when mod-rm bytes are not used. This got rid
of the ugly ((i->b1 & 7) + i->rex_b) code.
Probably shaved just a very little run time off Hammer emulation,
and even less on x86-32. The resultant is a little cleaner anyways.
and I could never explain it. The problem was that in keyboard.cc it
included "math.h" with quotes instead of <math.h> with less than/greater
than, so gcc -MM didn't realize it was a system header.
- MyPanel::blankCursor
- MyFrame::panel
- AdvancedLogOptions::action (2d array of wxChoice *'s)
- ParamDialog::idHash, ParamDialog::paramHash, and ParamStructs
- file dialog in BrowseTextCtrl
- fix illegal use of a wxObject. I had been clearing the ParamStruct
with memset(), but I forgot that ParamStruct was a subclass of wxObject
so I was trashing the wxObject fields too! Instead I created a
ParamStruct constructor that clears the pointers to NULL.
- comment out debug output from AdvancedLogOptionsDialog::SetAction
- modified files: gui/wx.cc gui/wxdialog.cc gui/wxdialog.h
gui/wxmain.cc gui/wxmain.h
- put off some log related features until later
- advanced log options and debug log are done, so I removed them from the list
- added some wxWindows coding hints I found on usenet
interface) by using get/set_default_log_action. This has very minimal effect
on the behavior of the text config interface.
- modified: main.cc gui/control.cc
time, so I've tried to improve it. Now the logfunctions class has a
static field default_onoff[] which represents the default actions for
each kind of log message. Default_onoff[] is initialized with static
data so it should be valid by the time it's used. This can be reached by
static accesors logfunctions::set_default_action and
logfunctions::get_default_action. It seemed appropriate to put the defaults
inside the logfunctions class, rather than in bx_options or siminterface.
However, to make them accessible to the config interface, I added similar
methods in siminterface that map to the accessors in logfunctions.
- logio.cc had two different definitions of LOG_THIS for different halves
of the file, which was ugly and confusing. (LOG_THIS is needed for BX_INFO,
BX_DEBUG, etc.) I removed the first definition and fixed the minor compile
problems that ensued. In the initializers for iofunctions, everything
is complicated because of the unpredictable order that constructors get
called. They must use this->log to print things, because genlog hasn't
been initialized yet.
- now if SIM->set_log_action(int mod, int level, int action) is called
with mod<0, it sets the action for all modules/devices instead of just one.
- modified: bochs.h logio.cc main.cc gui/siminterface.cc gui/siminterface.h
- LogOptionsDialog: show the "no change" option. When options other than
"no change" are selected, set both the default log action and the
specific log action for all devices.
- AdvancedLogOptionsDialog: store the wxChoice*'s in a 2d array since I
have to refer to them later.
- since both the log options and advanced log options dialogs needed to
create many wxChoices using similar rules, I moved the creation code
into a utility function makeLogOptionChoiceBox.
- I finally discovered the wxADJUST_MINSIZE option to wxSizer::Add(),
which solves some layout problems that I was having with wxChoices.
With this setting, the wxChoice will automatically grow to the size of
the largest string that can be selected.
- improve sizing of the scrollWin according to the desired size of the panel
that contains all the wxChoices.
- add sketches for proposed dialog that combines both LogOptionsDialog and
AdvancedLogOptionsDialog using tabs
From sf patch page :
This patch adds remote GDB stub capability to Bochs.
It's an updated version of the patched bochs 1.3pre1 in
found in this URL:
http://www.rtmk.org/bochs-gdb.html
more details in the same site. thanks Johan Rydberg
in cpu.cc out of the main loop, and into the asynchronous
events handling. I went through all the code paths, and
there doesn't seem to be any reason for that code to be
in the hot loop.
Added another accessor for getting instruction data, called
modC0(). A lot of instructions test whether the mod field
of mod-nnn-rm is 0xc0 or not, ie., it's a register operation
and not memory. So I flag this in fetchdecode{,64}.cc.
This added on the order of 1% performance improvement for
a Win95 boot.
Macroized a few leftover calls to Write_RMV_virtual_xyz()
that didn't get modified in the x86-64 merge. Really, they
just call the real function for now, but I want to have them
available to do direct writes with the guest2host TLB pointers.
but if you hand edit cpu/cpu.h, and change BxICacheEntries,
you can try different sizes. I'll make this more flexible
with configure. For now, use "--enable-icache" with no parameters.
- Modified fetchdecode.cc/fetchdecode64.cc just enough so that
instructions which encode a direct address now use a memory
resolution function which just sticks the immediate address
into rm_addr. With cached instructions we need this.
set_text_charbyte()
- vga: store the address of the active charmap in the new variable
charmap_address
- vga: text mode hack removed. The write modes, operations and masks must be
used in text mode too.
- sdl: clear_screen() is not necessary when the charmap has changed
- win32: update only the changed font bitmaps before drawing the text
- All mouse events in the VGA window go to MyPanel::OnMouse. Middle mouse
button and F12 both toggle mouse capture. OnMouse queues an event
for the simulation thread to process. The simulation thread calls
bx_devices.keyboard->mouse_motion() when it sees the event on the queue.
- add IFDBG_VGA around some display debug code. All wx mouse debug code
is controlled by IFDBG_MOUSE.
- modified: gui/wx.cc gui/wxmain.cc gui/wxmain.h
the wxwindows thread, it just sets a boolean flag needRefresh. Meanwhile, a
wxTimer running in wall clock time (as opposed to sim time) triggers a
repaint only if the the boolean is true.
- now MyRefresh is implemented by just setting needRefresh=true.
- make graphics_tile_update call MyRefresh.
- make dimension_update call MyRefresh.
to bitfields. bxInstruction_c is now 24 bytes, including 4 for
the memory addr resolution function pointer, and 4 for the
execution function pointer (16 + 4 + 4).
Coded more accessors, to abstract access from most code.
with accessors. Had to touch a number of files to update the
access using the new accessors.
Moved rm_addr to the CPU structure, to slim down bxInstruction_c
and to prevent future instruction caching from getting sprayed
with writes to individual rm_addr fields. There only needs to
be one. Though need to deal with instructions which have
static non-modrm addresses, but which are using rm_addr since
that will change.
bxInstruction_c is down to about 40 bytes now. Trying to
get down to 24 bytes.
read all param values from CPU #0. The only solution I can come up with
is to change the siminterface handler function interface to pass a void*
to the callback function. I'll take care of it eventually.
you can switch on and off using the $TEST_* varables at top.
- when running with --parallel, use "-geom +x+y" argument to xterm
to make the windows appear in a regular pattern.
use accessors. This lets me work on compressing the
size of fetch-decode structure (now called bxInstruction_c).
I've reduced it down to about 76 bytes. We should be able
to do much better soon. I needed the abstraction of the
accessors, so I have a lot of freedom to re-arrange things
without making massive future changes.
Lost a few percent of performance in these mods, but my
main focus was to get the abstraction.
I'm keeping a separate patch in case Christophe wants to make revisions
to his patch without my changes getting in the way.
I have just been working on the configuration interface part, for
example making it so that when you disable the ata[0123] channel, all the
devices controlled by that channel will be disabled as well. I haven't
gotten around to the wxWindows part, but these changes will make the
wxWindows work much easier. I will update this patch as my work progresses.
instead of bx_param_bool_c. There are several cases where I need an
enum to enable/disable some fields, and this change should allow
that.
- modified: gui/siminterface.h gui/siminterface.cc
no longer used. Also rearranged that struct a little
to be more compressed. Over time, I'm going to reduce
it further, for use with future accelerations.
enhancement to bochs. You can now configure with
--enable-guest2host-tlb.
Force the support of big pages (PSE) when x86-64 is configured.
Reverted back to only one kind of TLB entry style, since everything
is ported.
Fixed one bug in io.cc with as_64 and the index registers.
There are others, as noticed by Peter.
wxWindows without debugger, I needed to add some more cases of
#if BX_DEBUGGER to make it work. It is certainly possible to
handle such problems by always compiling in the debug dialogs but
only instantiating them if debug support is compiled in, but I have
chosen (for now) to put #if BX_DEBUGGER around things like this.
- modified: gui/wxdialog.h gui/wxdialog.cc
wxWindows without debugger, I needed to add some more cases of #if
BX_DEBUGGER to make it work. All the problems I found were things like
referencing a debug structure which was never initialized (NULL). It is
possible to handle such problems by always compiling in the debug dialogs but
only instantiating them if debug support is compiled in, but I have chosen
(for now) to put #if BX_DEBUGGER around things like this.
class declaration, for example:
static const unsigned os_64=0, as_64=0;
After reading some suggestions on usenet, I changed these into
enums instead, like this:
enum { os_64=0, as_64=0 };
you have AC_DEFINE(var, value) executed more than once, autoconf is
supposed to use the last AC_DEFINE value, but sometimes it doesn't.
I haven't been able to figure out why, so I've just worked around it
instead.
- for BX_SupportGlobalPages, which could call AC_DEFINE twice, I kept the
value in a shell variable $support_global_pages until the end and then
did one AC_DEFINE after it had reached its final value. Same thing
with BX_USE_CONFIG_INTERFACE.
- also I realized that we are using AC_SUBST() in many cases where it is
not needed. AC_SUBST(name) substitutes the string @name@ with a
value from the configure script. For preprocessor symbols like
BX_SupportPAE, we only need the AC_DEFINE; the AC_SUBST has no effect.
be used at all, and Peter didn't want it. "extdb.o" is compiled
into libcpu.a, if configured for it.
Removed a few #warnings for x86-64 compile, based on Peter's
line-item comments regarding the warnings I inserted during
the port/merge.
- try to find the configure script and set the path correctly, in case
the script is run from the main directory, build, etc.
- name the build directories build-$name so they don't overwrite anything.
Now that I have "rm -rf" in my script I wanted to be a little more careful.
printed to stderr in the text debugger. Also allows the user to
type (text) debugger commands directly, which also appear in the log.
- all text output in the debugger now passes through dbg_printf()
(used to be fprintf to stderr) so that in wxWindows I can redirect
it all to the wxWindows debug log screen. Added debug_fputs to
siminterface which actually sends the text to the GUI by creating
a BX_ASYNC_EVT_DBG_MSG event.
- changed prefix and msg fields of BxLogMsgEvent to const char *,
and also in args of logmsg method of siminterface.
- don't trap SIGINT in wxWindows. There are other ways to stop execution.
Also, signal handling with multiple threads is very strange and different
on different platforms.
- minor changes to fix gcc -Wall warnings in dbg_main.cc
- add a new boolean parameter BXP_DEBUG_RUNNING that tells if the debugger is
running freely or not. This is used by the wxWindows GUI to enable or
disable certain choices.
- CpuRegistersDialog has continue,stop,step buttons. When the sim is running
freely, I disable continue and step, and enable stop. When the sim stops
to wait for the user, I disable stop and enable continue and step. The
change of enables used to be triggered by actually pressing the button,
but then if you started/stopped the simulation in some other way (typing
in debug log window) the enables were never changed. Now the enables are
controlled by the value of BXP_DEBUG_RUNNING, which is set by the debug code
itself, and the buttons are enabled at the right time.
- ParamDialog::Refresh() is now virtual so that child classes can redefine
its refresh behavior.
- in safeWxStrcpy, force the last element of the array to be a 0, since
I noticed that strncpy is not guaranteed to terminate the string!
- modified: debug/dbg_main.cc debug/debug.h gui/siminterface.cc
gui/siminterface.h gui/wxdialog.cc gui/wxdialog.h gui/wxmain.cc
gui/wxmain.h
printing a message when a reserved bit was set, but not causing
a #GP(0). As well, I force a new PAE support option to 1 when
Hammer support is enabled.
circular dependencies between 3 cpu related libs that I need
as part of this transition. I changed the "ar rv" to "ld -i -o"
to do an incremental load instead of an archive. Hope this
doesn't break any platforms. We can reset this later.
at the appropriate times. For example, you can't stop when you're
stopped. You can't step when it's running.
- modified: gui/wxdialog.cc gui/wxdialog.h
called cpu_mode. Now there is one for cpu32, but it is declared:
static const unsigned cpu_mode=BX_MODE_IA32;
This way the compiler can compile-out if-then-else clauses based
on it, allowing for easier code sharing.
Bochs debugger. The Bochs debugger calls SIM->debug_get_next_command() which
does not return until a debugger command is found. The siminterface sends an
synchronous event to the wxWindows thread with a blank to be filled in with a
debugger command. wxWindows fills in the blank and sends the synchronous
event back, and the Bochs debugger interprets it as if it was typed on
the command line. For the long term I haven't decided whether to stick with
sending text strings vs. some other method.
- so far the wxWindows debugger consists of one big dialog box that shows
all the standard registers, and a working Continue, Stop, and Step button.
- modify ParamDialog so that it is more useful as a base class, by moving
some things to protected members&fields, separating out functionality
that is most likely to be replaced into virtual functions, and making it
generally more flexible. The new CpuRegistersDialog is based on
ParamDialog.
- in wxdialog.cc, continue the convention of using wxID_HELP, wxID_OK,
wxID_CANCEL, etc. for the id's of buttons, instead of wxHELP, wxOK, etc.
which are intended to be ORred together in a bit field.
- cpu/init.cc: put ifdefs around DEFPARAMs for flags in configurations
where they don't exist. Add an eflags shadow parameter that represents all
of the bits of eflags at once. There are also boolean shadow params for
each bit.
- modified files: cpu/init.cc debug/dbg_main.cc debug/debug.h
gui/siminterface.cc gui/siminterface.h gui/wxdialog.cc gui/wxdialog.h
gui/wxmain.cc gui/wxmain.h
member functions are turned on, BX_CPU_C_PREFIX expands to nothing, and any
method that uses BX_CPU_C_PREFIX instead of explictly writing "BX_CPU_C::"
will not be a member function at all. This makes it impossible for code
outside the BX_CPU_C object to call the accessor because sometimes the method
is at ptr_to_cpu->get_EIP() and other times you'd have to do just get_EIP().
The only way I've found to solve this is to remove the BX_CPU_C_PREFIX
and write BX_CPU_C:: instead.
- in debug/dbg_main.cc I removed the EBP, EIP, ESP, SP shortcuts. Now
the accessors are used everywhere. Also I replaced a reference to
the short-lived get_erx() accessor with ones that work: get_EAX(), etc.
- with these changes the current cvs compiles with any combination of
debugger enabled/disabled, SMP enabled/disabled, and x86-64 enabled/disabled.
BX_READ_8BIT_REG() --> BX_READ_8BIT_REGx()
BX_WRITE_8BIT_REG() --> BX_WRITE_8BIT_REGx()
They use an extra parameter "extended". I coded this
as the macro without the "x" for cpu32 compiles. This
allows for ease of merging and code sharing.
to incrementally merge files. For a test, shift16.cc is always
compiled in the cpu/ directory regardless of 32/64-bit configure.
Ultimately, all files will migrate from cpu64 to cpu.
- add get_erx() method to bx_gen_reg_t which returns the erx field of the
structure (which is has a different name in cpu and cpu64). Providing
an accessor is one strategy for avoiding igly "#ifdef BX_SUPPORT_X86_64"
statements in the rest of the code.
- cpu64/init.cc: the "eflags" before get_flag and set_flag is no longer
correct. removed.
- modified files: load32bitOShack.cc logio.cc cpu/cpu.h cpu64/apic.cc
cpu64/cpu.h cpu64/init.cc cpu64/proc_ctrl.cc debug/dbg_main.cc
cpu64 directories. Instead of using the macros introduced in cpu.h rev 1.37
such as GetEFlagsDFLogical and SetEFlagsDF and ClearEFlagsDF, I made inline
methods on the BX_CPU_C object that access the eflags fields. The problem
with the macros is that they cannot be used outside the BX_CPU_C object. The
macros have now been removed, and all references to eflags now use these new
accessors.
- I debated whether to put the accessors as members of the BX_CPU_C object
or members of the bx_flags_reg_t struct. I chose to make them members
of BX_CPU_C for two reasons: 1. the lazy flags are implemented as
members of BX_CPU_C, and 2. the eflags are referenced in many many places
and it is more compact without having to put eflags in front of each. (The
real problem with compactness is having to write BX_CPU_THIS_PTR in front of
everything, but that's another story.)
- Kevin pointed out a major bug in my set accessor code. What a difference a
little tilde can make! That is fixed now.
- modified: load32bitOShack.cc debug/dbg_main.cc
and in both cpu and cpu64 directories:
cpu.cc cpu.h ctrl_xfer_pro.cc debugstuff.cc exception.cc flag_ctrl.cc
flag_ctrl_pro.cc init.cc io.cc io_pro.cc proc_ctrl.cc soft_int.cc
string.cc vm8086.cc
This adds a whole new directory cpu64 with the new emulation code.
Very few changes were necessary outside cpu64. To try it, configure
with --enable-x86-64 and make.
- also this adds Peter Tattam's external debugger interface.
- modified files: Makefile.in bochs.h config.h.in configure.in
load32bitOShack.cc logio.cc cpu/Makefile.in cpu/cpu.cc debug/dbg_main.cc
- added files: cpu/extdb.cc cpu/extdb.h and cpu64/*
This adds a whole new directory cpu64 with the new emulation code.
Very few changes were necessary outside cpu64. To try it, configure
with --enable-x86-64 and make.
- also this adds Peter Tattam's external debugger interface.
- modified files: Makefile.in bochs.h config.h.in configure.in
load32bitOShack.cc logio.cc cpu/Makefile.in cpu/cpu.cc debug/dbg_main.cc
- added files: cpu/extdb.cc cpu/extdb.h and cpu64/*
performance of the ne2K driver in win9x."
- receive timer set to 10000 instructions instead of 1000000.
- placed if (WaitForSingleObject(lpAdapter->ReadEvent,0) == WAIT_OBJECT_0) {}
around the read code.
> This is the bug fix to make the reset button work properly when the cpu
> is in the halt state. There is another patch in init.cc as well to clear
> async_event. If you don't do this, if a cpu goes into HLT, the only thing
> which will fix it is another interrupt. The reset button won't work.
a consistent way of accessing these flags that works both inside and
outside the BX_CPU class, I added inline accessor methods for each
flag: assert_FLAG(), clear_FLAG(), set_FLAG(value), and get_FLAG ()
that returns its value. I use assert to mean "set the value to one"
to avoid confusion, since there's also a set method that takes a value.
- the eflags access macros (e.g. GetEFlagsDFLogical, ClearEFlagsTF) are
now defined in terms of the inline accessors. In most cases it will
result in the same code anyway. The major advantage of the accesors
is that they can be used from inside or outside the BX_CPU object, while
the macros can only be used from inside.
- since almost all eflags were stored in val32 now, I went ahead and
removed the if_, rf, and vm fields. Now the val32 bit is the
"official" value for these flags, and they have accessors just like
everything else.
- init.cc: move the registration of registers until after they have been
initialized so that the initial value of each parameter is correct.
Modified files:
debug/dbg_main.cc cpu/cpu.h cpu/debugstuff.cc cpu/flag_ctrl.cc
cpu/flag_ctrl_pro.cc cpu/init.cc
length. (The first guinea pig is the 2-bit IOPL field in eflags.)
Also it can have a pointer to a Bit8u, Bit16u, Bit32u and signed
equivalents and do the right thing.
- add lots more CPU fields as parameters: EBP ESI EDI ESP, all segment regs,
LDTR, GDTR, eflags, DR*, TR*, CR*. These are all visible on a
ridiculously tall dialog box that will one day become the debugger.
wxWindows thread. However, when the wxWindows thread calls
Bochs code, for example SIM->some_action() that triggers a
BX_PANIC(), then the Sim2CI event is created in the wxWindows
thread. This used to cause thread deadlock, but now it is
recognized and handled safely.
You need to use '--enable-global-pages' to configure in support.
If you have something to boot that uses them, give them a
spin. Really the were introduced for PPro and above, but
I haven't put in any limits. CPUID and CR4 report the proper
bits when configured, regardless of --enable-cpu-level at the
moment.
if off, we were still reading CR3 from the TSS and reloading
it! This was causing problems with a DOS extender. When
paging is turned back on, CR3 would be incorrect.
with GCC) align them with the GCC special alignment attribute.
Since there was then one available field, I split the protection
attributes and native host pointers into their own fields.
Before, with 3 dwords per TLB entry, some entries (about 3/8)
were spanning two processor cache lines (assuming a 32-byte
cache line). Now, they all fit within one cache line.
Knocked about 1.4% off Win95 boot time, probably more off normal
software runs.
BX_READ not 0. BX_READ was 10. While I was at it, I did
change BX_{READ,WRITE,RW} to {0,1,2} rather than {10,11,12}
in case that helps optimize code.
There may be more paging checks we should do before changing
any state, to avoid receiving a page fault in the middle.
I put some extra comments in there.
to request bulk IO operations to IO devices which are bulk IO aware.
Currently, I modified only harddrv.cc to be aware. I added some
fields to the bx_devices_c class for the IO instructions to
place requests and receive responses from the IO device emulation.
Devices except the hard drive, don't monitor these fields so they
respond as normal. The hard drive now monitors these fields for
bulk requests, and if enabled, it memcpy()'s data straight from
the disk buffer to memory. This eliminates numerous inp/outp calling
sequences per disk sector.
I used the fields in bx_devices_c so that I would not have to
disrupt most IO device modules. Enhancements can be made to
other devices if they use high-bandwidth IO via in/out instructions.
vga_charmap
- the SDL gui uses the charmap data for the vga text display
* TODO: implement this feature for other guis
- removed unused variables in sdl.cc and gui.cc
- fixed a warning in vga.cc
All the EFLAGS bits used to be cached in separate fields. I left
a few of them in separate fields for now - might remove them
at some point also. When the arithmetic fields are known
(ie they're not in lazy mode), they are all cached in a
32-bit EFLAGS image, just like the x86 EFLAGS register expects.
All other eflags are store in the 32-bit register also, with
a few also mirrored in separate fields for now.
The reason I did this, was so that on x86 hosts, asm() statements
can be #ifdef'd in to do the calculation and get the native
eflags results very cheaply. Just to test that it works, I
coded ADD_EdId() and ADD_EwIw() with some conditionally compiled
asm()s for accelerated eflags processing and it works.
-Kevin
it can decide how to proceed. Some of those bits are necessary
to make TLB invalidation decisions. INVLPG doesn't cause
a whole TLB flush anymore, just one page. Some of the
current CPU behaviours model the P6, especially on CR0
reloads. Earlier processors kept some pre-change pre-fetched
instructions until a branch. We could probably model that
by setting a flag, and letting the revalidate_prefetch_q
function cause serialization.
The TLB flush code only invalidates entries which are not
already invalidated for the case where the TLB invalidation
ID trick is not in use.
Read-Modify-Write instructions. The first read phase stores
the host pointer in the "pages" field if a direct use pointer
is available. The Write phase first checks if a pointer was
issued and uses it for a direct write if available.
I chose the "pages" field since it needs to be checked by the
write_RMW_virtual variants anyways and thus needs to be
cached anyways.
Mostly the mods where to access.cc, but I did also macro-ize
the calls to write_RMW_virtual...() in files which use it
and cpu.h. Right now, the macro is just a straight pass-through.
I tried expanding it to a quick initial check for the pointer
availability to do the write in-place, with a function call
as a fall-back. That didn't seemed to matter at all.
Booting is not helped by this really. The upper bound of
the gain is 5 or 6%, and that's only if you have a loop that
looks like:
label:
add [eax], ebx ;; mega read-modify-write instruction
jmp label ;; intensive loop.
Kevin Lawton says he doesn't get a performance benefit.
I'm not sure if I do. Either way, the difference isn't
very large.
This code may get removed if it turns out to be useless.
- modified files: config.h.in cpu/init.cc debug/dbg_main.cc gui/control.cc
gui/siminterface.cc gui/siminterface.h gui/wxdialog.cc gui/wxdialog.h
gui/wxmain.cc gui/wxmain.h iodev/keyboard.cc
----------------------------------------------------------------------
Patch name: patch.wx-show-cpu2
Author: Bryce Denney
Date: Fri Sep 6 12:13:28 EDT 2002
Description:
Second try at implementing the "Debug:Show Cpu" and "Debug:Show
Keyboard" dialog with values that change as the simulation proceeds.
(Nobody gets to see the first try.) This is the first step toward
making something resembling a wxWindows debugger.
First, variables which are going to be visible in the CI must be
registered as parameters. For some variables, it might be acceptable
to change them from Bit32u into bx_param_num_c and access them only
with set/get methods, but for most variables it would be a horrible
pain and wreck performance.
To deal with this, I introduced the concept of a shadow parameter. A
normal parameter has its value stored inside the struct, but a shadow
parameter has only a pointer to the value. Shadow params allow you to
treat any variable as if it was a parameter, without having to change
its type and access it using get/set methods. Of course, a shadow
param's value is controlled by someone else, so it can change at any
time.
To demonstrate and test the registration of shadow parameters, I
added code in cpu/init.cc to register a few CPU registers and
code in iodev/keyboard.cc to register a few keyboard state values.
Now these parameters are visible in the Debug:Show CPU and
Debug:Show Keyboard dialog boxes.
The Debug:Show* dialog boxes are created by the ParamDialog class,
which already understands how to display each type of parameter,
including the new shadow parameters (because they are just a subclass
of a normal parameter class). I have added a ParamDialog::Refresh()
method, which rereads the value from every parameter that it is
displaying and changes the displayed value. At the moment, in the
Debug:Show CPU dialog, changing the values has no effect. However
this is trivial to add when it's time (just call CommitChanges!). It
wouldn't really make sense to change the values unless you have paused
the simulation, for example when single stepping with the debugger.
The Refresh() method must be called periodically or else the dialog
will show the initial values forever. At the moment, Refresh() is
called when the simulator sends an async event called
BX_ASYNC_EVT_REFRESH, created by a call to SIM->refresh_ci ().
Details:
- implement shadow parameter class for Bit32s, called bx_shadow_num_c.
implement shadow parameter class for Boolean, called bx_shadow_bool_c.
more to follow (I need one for every type!)
- now the simulator thread can request that the config interface refresh
its display. For now, the refresh event causes the CI to check every
parameter it is watching and change the display value. Later, it may
be worth the trouble to keep track of which parameters have actually
changed. Code in the simulator thread calls SIM->refresh_ci(), which
creates an async event called BX_ASYNC_EVT_REFRESH and sends it to
the config interface. When it arrives in the wxWindows gui thread,
it calls RefreshDialogs(), which calls the Refresh() method on any
dialogs that might need it.
- in the debugger, SIM->refresh_ci() is called before every prompt
is printed. Otherwise, the refresh would wait until the next
SIM->periodic(), which might be thousands of cycles. This way,
when you're single stepping, the dialogs update with every step.
- To improve performance, the CI has a flag (MyFrame::WantRefresh())
which tells whether it has any need for refresh events. If no
dialogs are showing that need refresh events, then no event is sent
between threads.
- add a few defaults to the param classes that affect the settings of
newly created parameters. When declaring a lot of params with
similar settings it's more compact to set the default for new params
rather than to change each one separately. default_text_format is
the printf format string for displaying numbers. default_base is
the default base for displaying numbers (0, 16, 2, etc.)
- I added to ParamDialog to make it able to display modeless dialog
boxes such as "Debug:Show CPU". The new Refresh() method queries
all the parameters for their current value and changes the value in
the wxWindows control. The ParamDialog class still needs a little
work; for example, if it's modal it should have Cancel/Ok buttons,
but if it's going to be modeless it should maybe have Apply (commit
any changes) and Close.
- bx_gen_reg cannot be declared with BX_SMF or it can't read gen_reg
when static member functions are turned on.
- use "BX_CPU_C_PREFIX" instead of "BX_CPU_C::" for get_segment_base.
- the SMF (static member function) tricks are just plain wierd. The only way to
really be sure that you're not breaking something is to try compiling it with
SMF on and with SMF off. e.g. "configure && make" and
"configure --enable-processors=2 && make".
work (control-C kills process instead of returning to debugger prompt), give
a configure-time error if they are enabled together. If we can fix the
signal problem in readline, this check can be removed.
problems with control-C handling if you enable readline, because readline()
installs its own signal handlers every time you call it. I'm having
good luck with "--with-wx --enable-debugger --disable-readline" now.
X windows, wxWindows. Each platform has its own way of returning
a variable length string, and its own rules about how you're supposed
to dispose of the string. Now all platforms do the same thing: they
allocate a Bit8u buffer with C++ "new" and copy the clipboard data in,
then release the clipboard data in the platform-specific correct way.
The Bit8u buffer is sent to the keyboard code, which frees it with
delete [] when finished.
- modified: gui/wxmain.cc gui/wxmain.h
X windows, wxWindows. Each platform has its own way of returning
a variable length string, and its own rules about how you're supposed
to dispose of the string. Now all platforms do the same thing: they
allocate a Bit8u buffer with C++ "new" and copy the clipboard data in,
then release the clipboard data in the platform-specific correct way.
The Bit8u buffer is sent to the keyboard code, which frees it with
delete [] when finished.
- modified: gui/wx.cc gui/x.cc gui/win32.cc iodev/keyboard.cc
files that need them. This is more in line with the other gui libraries,
and the compile line is easier to read.
- modified: Makefile.in configure.in configure gui/Makefile.in
delivered to both the GUI and simulator thread, so they both call
bx_signal_handler. This can lead to deadlock as multiple threads enter
BX_PANIC and try to show a dialog box at once. To solve the problem, I made
a function isSimThread() which can be called from anywhere. If the
bx_signal_handler is called from any thread OTHER THAN the simulation thread,
it returns without doing anything. As a result, only one thread is allowed
to enter the signal handler code, and now control-C works correctly.
Bochs to immediately read the bochsrc and start simulating immediately (as
opposed to going into the configuration interface first). Now -q does
the right thing in the wxWindows interface. It behaves as if you selected
Read Configuration and then Simulate:Start.
- modified: main.cc gui/siminterface.cc gui/siminterface.h gui/wxmain.cc
direct reads/writes from native variables to the x86 (guest)
memory image. Look at the end of bochs.h. Don't know if that's
the right place to put them, but here you can extend these
macros to platform-specific asm() code if you like, or just
use the generic C code I supplied. Some platforms have special
instructions for byte-order swapping etc. Also, you can't
make any assumptions about the alignment of the pointers
passed.
mode uses the notion of the guest-to-host TLB. This has the
benefit of allowing more uniform and streamlined acceleration
code in access.cc which does not have to check if CR0.PG
is set, eliminating a few instructions per guest access.
Shaved just a little off execution time, as expected.
Also, access_linear now breaks accesses which span two pages,
into two calls the the physical memory routines, when paging
is off, just like it always has for paging on. Besides
being more uniform, this allows the physical memory access
routines to known the complete data item is contained
within a single physical page, and stop reapplying the
A20ADDR() macro to pointers as it increments them.
Perhaps things can be optimized a little more now there too...
I renamed the routines to {read,write}PhysicalPage() as
a reminder that these routines now operate on data
solely within one page.
I also added a little code so that the paging module is
notified when the A20 line is tweaked, so it can dump
whatever mappings it wants to.
I have not tested these functions, but they model the format and
acceleration principals of the byte/word/dword functions. Give them
a try on both little/big endian machines.
to diff configure...you'll have to run autoconf yourself.
- in config.h.in set BX_SUPPORT_MMX to 0, not 1. The configure script
will replace it with a 1 if appropriate. That's the way all other
options are done.
- you need to use patch -p1 for this
wxHAS_RAW_KEY_CODES is not available in the wxWindows library.
my patch "patch.wx-raw-keycodes" adds wxHAS_RAW_KEY_CODES to the wxWindows
library, and it will also be in wxWindows 2.3.3 and beyond.
correct name for floppy and cdrom devices:on windows, A:, B:, etc;
on Linux /dev/fd0 and /dev/cdrom.
- It's perfectly ok to type device names into the blank on those dialog
boxes, but the label "Disk Image File:" sort of implies that it has to
be a file. To try to avoid confusion change the label to simply
"Disk Image:".
on win32. I believe this is the same problem that scarlip referred to as
"it doesn't start reliably. sometimes it works, sometimes it won't".
See comments in gui/wx.cc (dimension_update) for details.
- since dimension_update is called from the simulator thread, it needs to
get the GUI mutex before calling any wxWindows functions. Now it does.
- add lots of IFDBG_VGA(...) lines which I use to diagnose various problems.
They are compiled away by default (for performance). See definition of
IFDBG_VGA() in gui/wxmain.h.
values. All wxWindows dialogs return wxID_OK or wxID_CANCEL. The
wxOK and wxCANCEL symbols are ONLY used in input arguments that determine
whether to display the ok and cancel buttons. Now I'm doing the same.
Return values are wxID_OK if they pressed the ok button, and wxID_CANCEL
if they pressed cancel.
so that a compare of the current access could be done more
efficiently against the cached values, both in the normal
paging routines, and in the accelerated code in access.cc.
This cut down the amount of code path needed to get to
direct use of a host address nicely, and speed definitely
got a boost as a result, especially if you use the
--enable-guest2host-tlb option.
The CR0.WP flag was a real pain, because it imparts
a complication on the way protections work. Fortunately
it's not a high-change flag, so I just base the new
cached info on the current CR0.WP value, and dump
the TLB cache when it changes.
an allocation scheme that is guaranteed to return a block that is
aligned correctly.
- Kevin asked me to go ahead and align the memory to 4k page boundaries,
so I did. If we need to change this, just change BX_MEM_VECTOR_ALIGN
in memory/memory.h (now 4096).
checks were honoring the EFLAGS.DF bit, but assuming it was always
equal to 0 (increment upward). Plus some general cleanup of the
acceleration code.
I left the default of '--enable-repeat-speedups' to disabled, but
it seems pretty solid. Definitely adds performance for disk
heavy workloads.
