This card has 2 ata channels, so this would be better
than the promise DC2300. It also has drivers for win95,
available on tekram website.
Many thanks to Volker for fixing the interrupt bug!
It solved the problem I had with this patch...
This is an interim update to allow others to test.
We have userland code running!!! (up to a point)
Able to start executing "sash" as /sbin/init in userland from linux 64 bit
kernel until it crashes trying to access a null pointer. No kernel panics
though, just a segfault loop.
specific to X windows. I changed a few names of variables and methods to
sound less X windows specific, for example in the BXKeyEntry struct
the "xwinKey" field is now called "hostKey".
- remove the hacks in bx_keymap_c::loadKeymap that checked that every
keysym started with "XK_". Now I can still make that check, but it
is done in x.cc's convertStringToXKeysym() instead.
- in sdl.cc, load the keymap in specific_init function. I had to
create a conversion function from SDL key names to SDLkey values.
I created a table of key names and associated values, and I just
do a strcmp on each name. It's not a fast algorithm obviously but this
conversion only has to be done for maybe 200 keys at startup time.
- added a key map file for SDL called sdl-pc-us.map. Since this has
basically all the SDLK_* symbols defined, it is a good starting point
if you need to make any other keymaps.
- modified: gui/keymap.cc gui/keymap.h gui/sdl.cc gui/x.cc iodev/keyboard.cc
- added: gui/sdlkeys.h gui/keymaps/sdl-pc-us.map
into inline functions with asm() statements in cpu.h. This cleans
up the *.cc code (which now doesn't have any asm()s in it), and
centralizes the asm() code so constraints can be modified in one
place. This also makes it easier to cover more instructions
with asm()s for more efficient eflags handling.
wxLog::SetActiveTarget (new wxLogStderr ());
which used to be inside #ifdef __WXDEBUG__ ... #endif. This fixes the
problem of tons of little debug message dialogs appearing in wxwindows
libraries built without debug support.
- This reverts revision 1.61. I cannot figure out why I had trouble back on
Sept 27, because now both my debug and release libraries work ok with the
SetActiveTarget line.
the toolbar button and the mouse capture shortcuts (F12 and middle button)
call ToggleMouse(). The only difference is that the first time you
press the toolbar button it gives a little message about mouse capture
and how to turn it off.
- modified: gui/wx.cc gui/wxmain.cc gui/wxmain.h
Quit while the simulation was running. I used to just do
wxFrame::Close(TRUE) in the wxwindows thread, but the simulation
thread would keep on running even after the window had disappeared.
If the simulation thread sent any messages to the gui, it could
segfault. Now, when you click on Quit it asks the simulator to
quit, but does not close the gui. When the simulator thread finally
quits, THEN it calls wxFrame::Close(TRUE) from the sim_thread (with
appropriate gui mutex). Such tricks are only needed when the
simulation is running. I hope this will fix bug #616142: "wx: crash
from events after gui shutdown".
- move MyFrame::closing flag out into a global variable wxBochsClosing
so that it will be valid even when the MyFrame is deleted. See
comments on in wxmain.h.
- when wxBochsClosing is set, don't create any new windows. If any
errors appear, just print them to stderr because the gui is just
about to be closed.
- fix comment in SimThread::Entry to talk about longjmp instead of
kill_bochs_request
- (minor) move contents of DefaultCallback2 into DefaultCallback. It
doesn't refer to the wxApp at all. This is a bit safer in case it
gets called when the wxApp is in the process of being deleted.
- modified: gui/wxmain.cc gui/wxmain.h
hack with longjmp() back to cpu.cc main decode loop, and added a
check in there to return control when bx_guard.special_unwind_stack
is set (compiling with debugger enabled only).
If in the debugger you try to execute further instructions
(which you shouldn't), other fields need to be reset I would
think, such as EXT and errorno, and have to make sure ESP/EIP
are corrected properly. Basically, this hack is only good
for examining the current situation of a nasty fault.
works under FreeBSD.
- Fix some formatting problems with long lines.
- Cleanup wait_for_connect(). Only set SO_REUSEADDR once. Reuse sockaddr for
bind() and accept().
- Turn on TCP_NODELAY to speed up GDB remote protocol transactions.
- In bx_gdbstub_init(), remove `Communicating...' message. Add a `Waiting...'
message just before calling wait_for_connect() to give the user some idea
that the emulator is waiting for something.
if init() is called a second time. This allows me to restart a
simulation (wxwindows interface only) without restarting the whole
application.
- modified: iodev/*.cc
in pc_system.h to flag each timer slot as being allocated or not.
register_timer*() functions will claim a free slot if one
exists before using one at the end of the list. This will allow
for this function to be called repeatedly and not have to run
out of timer slots.
This must be done carefully though, because the call to SDL_WarpPointer
generates a mouse motion event of its own. To work around this, I have
added code that should ignore motion events generated by SDL_WarpPointer.
This should fix bug [ 619283 ] SDL: os mouse pointer leaves the window.
On Linux, SDL 1.2.3 it works fine.
wants to free up that resource.
unsigned unregisterTimer(int timerID);
Pass in the timer index received from the register function. A return
of 1 means success. NOTE: you must make sure the timer is deactivated
first. Call deactivate_timer() to be sure. A return of 0 means
failure, though a panic is really generated.
For now, this function does not really free up the slot, but will soon.
to give the compiler some hints:
BX_CPP_AttrPrintf(formatArg, firstArg)
BX_CPP_AttrNoReturn()
The first is to tell the compiler that a function receives printf-like
arguments so it can do some smart argument checking w.r.t. the
format string. The 2nd tells the compiler that the function does
not ever return; it's not used yet, but I'd like to use it on
exception() after we fix the situation of it returning for debugging.
I fixed one parameter mismatch in cpu/ by deleting a deprecated
debug print statement. There are several other mismatches in
other code modules.
bx_list_c can now be displayed as either a wxStaticBox with the
child parameters inside, or as a wxNotebook with each child
parameter in a separate tab. (The children can also be lists of
course.) The default display is the wxStaticBox type, but if you
set the option bit bx_list_c::USE_TAB_WINDOW in the list,
ParamDialog will use the wxNotebook display instead.
- to get the param trees working, I created a new struct
AddParamContext, which is passed to AddParam(). This struct is
critical when AddParam calls itself recursively to display lists
within lists.
- use the wxNotebook display feature for the ATA0,1,2,3 controller
dialog box. Now instead of being hundreds of pixels tall, it is
reasonable height with three different tabs. This fixed bug
#619074: "wx: ATA interface editor too tall" and was the whole
reason I started messing with this at all.
plus some minor cleanups
- when I added the enum constant bx_list_c::USE_TAB_WINDOW, I also
removed the BX_ prefix from all the other enum constants that are
used in parameter options in siminterface.cc. Since these constants
are enums within a class, there is no possibility of namespace
conflicts so the prefix is not needed.
- added wxADJUST_MINSIZE to all wxChoice controls, since that tells
wxWindows to adjust its size to the length of the longest string.
- instead of calling SetSize or SetSizeHints on every textcontrol with
a hardcoded width, I am using just two wxSize specifications for
everything: either normalTextSize or longTextSize.
- edit names of a few menus and params. For example now instead of
the tab saying "Master ATA device on channel 0" it will say
"First HD/CD on channel 0".
Modified Files:
main.cc gui/control.cc gui/gui.cc gui/siminterface.cc gui/siminterface.h
gui/wxdialog.cc gui/wxdialog.h gui/wxmain.cc
to be signed instead of unsigned, it was looking up color values
like wxBochsPalette[-126] instead of wxBochsPalette[130]. This affected
any colors in the palette between 128 and 255. I simply hanged newBits
to signed, and it fixed bug #618462: wx: 256 color modes not working properly
- I am still not satisfied with the way that wxWindows maps colors on an
8-bit display, but that is a different issue.
after the palette changes. This makes the iodev/vga.cc code do the
right thing. This also fixes the toolstar splash screen, but at a
fraction of the cost of the solution in rev 1.37.
need to be updated. This was impossible before, because we only stored the
RGB values in wxScreen. I added a new array called wxPixel that stores the
8-bit color index that should be displayed. Now if the VGA palette changes,
I use wxPixel data to reconstruct wxScreen using the new palette.
- this fixes the display of the splash screen of the Toolstar disk, since
it relied on drawing the picture and then fading the colormap from back
up to the normal values. I was hoping it would fix the splash screen of
win95 too, but it does not.
exit out of cpu_loop() and back to the caller can be honored.
Previously, code in this function was a part of cpu_loop so
a "return;" would already do that. Now, a value is passed
back to cpu_loop() to denote such a request, and then a return
is executed from cpu_loop().
I haven't tested this yet, but previously I must have broke
certain debugging requests by moving the code to a separate
function and not fixing the "return;" statements.
Symptom: Linux kernel 2.4.19 would hang in random places. CPU still
running, but in dle loop.
Cause: if APIC interrupt occurred while a PIC interrupt was pending, the
PIC interrupt would be lost. This is because either an APIC or PIC
interrupt would trash any pending interrupt event because INTR is only a state,
not an event queue.
Temporary fix: reworked apic.cc to have it's own copy of INTR state. cpu.cc now
checks for both cpu.INTR and local_apic.INTR.
Need to do further research to see if local_apic and pic can be integrated in such
a way as properly manage the combined effects of both devices accessing INTR state.
value and a change-mask, rather than passing all the boolean
change flags as arguments.
Recoded the POPF instruction in flag_ctrl.cc to use the
new writeEFlags() function, and to make it more sane.
Also, the old write_flags() and write_eflags() functions
redirect to writeEFlags() for now. Later, when we get
back in a development mode, it would be better to make
all calls use the new function and get rid of the old ones.
been using the Boolean type for a number of multi-bit fields on the
assumption that it is actually many bits wide. However, this assumption is
unsafe and has caused some bugs that are hard to track down.
- in the Carbon library on MacOS X, Boolean is defined to be an unsigned char.
This has been causing some of the EFLAGS accessors to fail (bits 8-31)
because they depended on Boolean being 32 bits wide. I changed these
accessors to return Bit32u instead. I believe that this will finally fix
[ 618388 ] Unable to boot under MacOS X.
- It would be possible to create a bochs specific type for booleans (bx_bool),
but it's cleaner to simply use "Boolean" when we actually mean a 1-bit true
or false field, and Bit8u/Bit32u when it is a multibit field.
32-bits rather than 64. This is possible, because there is
always an active null (heartbeat) timer, with periodicity
of less than or equal to the maximum 32-bit int value.
This generates a little less code in the hot part of cpu_loop,
and saved about 3% execution time on a Win95 boot.
Moved the asynchronous handling code from cpu_loop() to its
own function since it's a long path. This neatened up the
code a little (less gotos and all), and made it more clear
to use a "while (1)" around the iterative code in cpu_loop().
