* added save state support to the SHARC CPU core
* added save state support to the PowerPC recompiler
* added save state support to the virtual TLB system
* added save state support to the RF5C400 sound core
* added save state support to konppc module
* added save state support to K056800 host controller
* added save state support to the Konami hornet driver
Fixed poor default CLUT handling in the voodoo driver
* added save state support to the MIPS3 recompiler
* added save state support to CAGE audio system
* added save state support to the voodoo emulator
* added save state support to the smc91c9x emulator
* added save state support to the kinst, seattle, and vegas drivers
* fixed core video handling of save states with dynamic screen resolutions
SMC91C9x:
* converted to proper device
* updated seattle and vegas drivers to allocate devices
* added separate 91C96 device for eventual 2049 use
* cleaned up code
* changed SPU receive model to a push model; updated drivers accordingly
* added macros for setting the SPU transmit handler and sending bytes
* cleaned up ppc.h
* changed from tracking "live" registers to tracking "necessary" registers
* genericized register tracking to be more flexible
* added previous instruction pointer to opcode descriptions
PowerPC frontend/DRC:
* cleaned up register tracking implementation
* fixed numerous errors and shortcomings in the tracking
* added support for removing unnecessary XER CA and CR0 computations
* updated UML logging to output new frontend statistics
MIPS3 frontend/DRC:
* tweaked register tracking to match new DRC frontend system
* updated UML logging to output new frontend statistics
* Removed explicit flag requests from the shorthand opcodes
* Added optimization function to drcuml which is called at block end
* Added logic to compute the necessary flags based on upcoming opcodes
and only select those flags which are required
* Updated the PPC and MIPS3 DRCs to no longer explicitly specify flags
- rewrote PowerPC implementation as a dynamic recompiler on top
of the universal recompiler engine
- wrote a front-end to analyze PowerPC code paths and register usage
- wrote a common shared module with C implementations of tricky
CPU behaviors
- added separate CPU types for the variants supported, instead of
relying on a hidden model enum
- rewrote the serial port emulation for the 4xx series to be more
accurate and not rely on separate DMA handlers
- rewrote the MMU handling to implement a software TLB that faults
in pages and handles changed bits appropriately
- implemented emulation of the PowerPC 603's software TLB, which
allows the model 3 games to run without a hack to disable the MMU
Updated the PowerPC disassembler to share constants with the rest of
the core, and to more aggressively use simplified mnemonics, especially
for branches. [Aaron Giles]
Universal recompiler:
- fixed frontend to handle opcode widths different from bus width
- added several new opcodes:
* (D)GETFLGS - copies the UML flags to a destination operand
* FDRNDS - rounds a double precision value to single precision
- renamed several opcodes:
* SETC -> CARRY
* XTRACT -> ROLAND
* INSERT -> ROLINS
- consolidated the following opcodes:
* LOAD?U -> LOAD
* LOAD?S -> LOADS
* STORE? -> STORE
* READ?U -> READ
* READ?M -> READM
* WRITE? -> WRITE
* WRITM? -> WRITEM
* SEXT? -> SEXT
* FTOI?? -> FTOINT
* FFRI? -> FFRINT
* FFRF? -> FFRFLT
- removed some opcodes:
* FLAGS - can be done with GETFLGS/LOAD4/ROLINS
* ZEXT - can be achieved with AND
* READ?S - can be achieved with READ/SEXT
- updated C, x86, and x64 back-ends to support these opcode changes
- updated disassembler to support these opcode changes
MIPS3 dynamic recompiler:
- updated to use new/changed opcode forms
- changed context switch so that it only swaps a single pointer
Konami Hornet changes: [Aaron Giles]
- updated to new PowerPC configurations
- updated some memory handlers to be native 8-bit handlers
- cleaned up JVS implementation to work with new serial code
- added fast RAM for the work RAM to give a small speed boost
Konami GTI Club changes: [Aaron Giles]
- updated to new PowerPC configurations
- updated some memory handlers to be native 8-bit handlers
Konami Viper/ZR107 changes: [Aaron Giles]
- updated to new PowerPC configurations
Sega Model 3 changes: [Aaron Giles]
- updated to new PowerPC configurations
- reimplemented/centralized interrupt handling
- these games are broken for the moment
Fixed crasher due to some Konami games using 8 layers in
the K056832 implementation, even though it was only written
for 4. [Aaron Giles]
Added fisttp opcode to i386 disassembler. [Aaron Giles]
1. In the MIPS core:
- renamed struct mips3_config -> mips3_config
- updated all drivers to the new names
- removed MIPS3DRC_STRICT_COP0 flag, which is no longer used
- a few minor cleanups
2. In the CPU interface:
- added new 'intention' parameter to the translate callback to
indicate read/write/fetch access, user/supervisor mode, and
a flag for debugging
- updated all call sites to pass an appropriate value
- updated all CPU cores to the new prototype
3. In the UML:
- added new opcode SETC to set the carry flag from a source bit
- added new opcode BSWAP to swap bytes within a value
- updated C, x86, x64 back-ends to support the new opcodes
- updated disassembler to support the new opcodes
4. In the DRC frontend:
- fixed bug in handling edge case with the PC near the 0 or ~0
Subject: [patch] More Machine->machine changes, add machine to irq
callbacks
Hi mamedev,
Here are two more patches to eliminate Machine globals. The first
patch was autogenerated by the attached fixup script. That script has
been updated to catch additional cases which it previously missed
(when Machine is the last parameter to a function or Machine is used
in an assignment). This makes ~50 more files deprecat.h free.
A sizable chunk (~20%) of the remaining uses of the Machine global in
the drivers are due to irq callbacks for sound and machine updates.
Typically such callbacks need to call cpunum_set_input_line, which
requires a machine parameter, so if the callbacks don't pass the
machine parameter, these routines have no choice but to reference the
global variable.
The second patch attempts to address most cases of this by adding the
machine parameter to the callback interfaces. This allows us to
remove #include "deprecat.h" from ~150 files, at the cost of having to
fix up hundreds of callbacks.
In total, these patches reduced the number of files with deprecat.h
from 783 to 575.
~aa
* Added new opcode LZCNT which returns the number of leading zeros in
a parameter.
* Added new opcode XTRACT which is a combined rotate/mask (basically
rlwinm from PowerPC)
* Added new opcode INSERT which is a combined rotate/mask/blend
(basically rlwimi from PowerPC). Best. Opcode. Ever.
* Updated all back-ends to support these new opcodes.
* Fixed several bugs relating to shifts/rotates and optimizing out
cases incorrectly.
MIPS3 DRC changes:
* Updated to use INSERT and XTRACT where appropriate
* Cleaned up register usage to enable an additional direct mapping;
this means Linux gets 1 now and Windows gets 3
machine registers if the machine supports it. Currently only x64 on
Windows has enough free registers to do so, though PowerPC will almost
certainly be able to take advantage of this. Gives a minor speedup.
* Fixed front-end so that virtual no-op instructions are still targeted
as branch targets.
* Fixed front-end to mark the beginning of each sequence as needing TLB
validation, since any sequence can be jumped to from anywhere.
* Redid the MIPS3 TLB implementation. Fixed the exception vector and
type handling. Changed the bitfields to directly map from the MIPS TLB
format. Added distinction between TLB fill and TLB valid/modified
exceptions.
* Added separate modes for user, supervisor, and kernel modes. Each mode
does proper verification of addresses now and generates address errors
for invalid accesses.
* Fixed several bugs in the TLB implementation; not everything works
yet but it's a lot closer.
* Made COP0 access checking mandatory in non-kernel modes.
* Fixed several crashes when recompiling virtual no-ops.
* Fixed TLB bug where entries for virtual address 0 were present by
default.
* Fixed bug in the map variable implementation that would sometimes
result in incorrectly recovered values.
This will be expanded in the future.
Added two new opcodes: SAVE and RESTORE to save and restore the entire
virtual machine state for examination/setup.
Added new back-end function get_info() which returns information from
the back-end about how many actual registers will be mapped.
Fixed a bug that mapped the high a low parts of registers to the same
address. This should help the C back-end run better on big-endian
architectures.
is drcuml.c, which defines a universal machine language
syntax that can be generated by a frontend recompiler and
then retargeted via a generic backend interface to any of
a number of different architectures. A disassembler for the
UML is also included to allow examination of the generated
UML code.
Currently supported backend architectures include 32-bit x86,
64-bit x86, and a platform-neutral interpreted C backend that
can be used as a fallback for platforms without native
support. The C backend also performs additional validation
to ensure assumptions are met.
Along with the new architecture is a new MIPS III/IV
recompiler frontend. This frontend has been rewritten from
the old x64-specific recompiler to generate UML opcodes
instead. This means that the single recompiler can be used
to target multiple backend architectures and should in
theory produce identical results across all of them.
The old 32-bit and 64-bit MIPS recompilers are now officially
retired. The new system provides similar performance (within
5% generally) to the old system and has similar compatibility.
The only currently known issues are some problems with the
two Gauntlet 3D games.
Subject: [patch] Remove more Machine globals, #include "deprecat.h"
Hi mamedev,
The attached patch goes through and converts a number of Machine
globals to machine locals, and then removes #include "deprecat.h" if
appropriate. The script that generated it is included, since the
patch itself is rather large and would have been time consuming to
produce otherwise.
The script doesn't convert cases of Machine that aren't in common
macros. I'll try to tackle those later if someone doesn't beat me to
it.
~aa
direct byte, word, dword, and qword accessors for all bus sizes,
there are now masked word, dword, and qword accessors for all
bus sizes.
IMPORTANT: masks that are passed to the _masked_* functions are
NOT inverted. Although inverted masks are still passed to callback
functions, when you request a masked read or write the masks should
represent the bits you want.
Updated the various MIPS cores that use these functions to invert
their masks.
accordingly.
Added new functions for dynamically installing device memory
read/write handlers.
Updated install_memory_XXX_handler() functions to take a machine
parameter. Updated all drivers accordingly.
Merged installation of read and write handlers where appropriate.
Simplified memory.c code for dynamic installation so that a single
function handles all the work; a NULL read or write handler
indicates not to install anything for reads or writes.
only remaining form is the one that takes a pointer parameter.
Added macros for STATE_PRESAVE and STATE_POSTLOAD to define common
functions. Added machine parameter to these functions.
Updated all drivers and CPU/sound cores to use the new macros
and consolidate on the single function type. As a result pushed
the machine parameter through a few initialization stacks.
Removed unnecessary postload callbacks which only marked all tiles
dirty, since this is done automatically by the tilemap engine.
trace through in a debug build, yet should operate the same as before.
Created a complete set of functions for all databus sizes (8,16,32,64) and
all endiannesses. A few functions are redundant, but it is now very clear
which functions to use in which scenarios. It is also now possible to rely
on being able to access values of 8, 16, 32 or 64 bits via the built-in
accessors without fear of crashing.
Updated all cores using 8-bit handlers to explicitly call the 8-bit handlers
with the appropriate endianness.
Fixed a few games which were calling n-bit handlers directly to use the
generic forms. In the future, this is all the access drivers will have.
fixed reading from SXYP
fixed reading from IRGB
fixed writing to LZCR
fixed sign extension of GTE control registers
fixed writing to FLAG
New games added or promoted from NOT_WORKING status
---------------------------------------------------
1 on 1 Government (JAPAN)
Added ability to test the instruction/data cache ram. The scratchpad and BIU register are now handled internally to the CPU.
All writes are performed with masks. SWL/SWR used to be implemented with two writes ( one byte and one word ) when writing three bytes, now it only ever performs one. Byte and Word writes use masks as they leave the rest of the register on the bus, which can be picked up by larger registers.
The read/write functions to use are cached when the SR bits are updated, as are the bad address masks.
Added coprocessor 1 & 3 support, though they don't do anything useful.
All loads now go through the delay pipeline, a lwl/lwr will grab the value out of the pipeline if it's updating the same register.
Added undocumented behaviour of BLEZ/BGTZ. The comparison for zero can be changed by specifying an alternate register in the RT field ( the documentation says you should always use register 0 ).
Restricted to 16 COP0 registers & generate an exception if any of the 5 for the MMU are used.
Added BCF/BCT instructions, although I have found no conditions that affect them yet.
Generates an exception if any MMU instructions are executed.
Sets the CE instruction for all exceptions, not just those involving a coprocessor. The bits of the opcode that specify the coprocessor are grabbed no matter what the instruction.
Added TAR register and BT bit in SR. When an exception occurs during a branch, BT determines whether it was taken or not. The TAR register gets set to the destination of the branch.
Fixed the BD bit when you are in a branch delay slot and you didn't take the branch, this shows up in the pipeline as !pc.
Fixed branches within a branch delay slot.
Multiply & divide instructions can be aborted if you write to HI/LO before reading the result.
Added data breakpoints, you don't appear to be able to set breakpoints on any of the addresses internal to the CPU.
Multiply/divide/GTE instructions can execute when an exception is taken, although the EPC indicates that it hasn't. The BIOS avoids rerunning GTE instructions as they are destructive, so you have to make sure they run.
Added bus error handling, PSXCPU is limited to 8mb of ram & any access outside this range will trigger an exception. I believe this is to be an internal limit.
Added CXD8611R as a specific CPU type, System 12 appears to allow more than 8mb of ram & it's possible that this is different.
Mapped out all instructions to either generate an exception or ignore bits.
Updated the disassembler to match the decoding.
Fixed disassembling of branch instructions in a branch delay slot.
Lui checks for a ori/addiu following and will show you the result.
Added step over/out support.
Fixed standalone disassembler.
All occurrences of ACCESSING_LSB, ACCESSING_MSB, ACCESSING_LSB16, ACCESSING_MSB16, ACCESSING_LSB32, ACCESSING_MSB32, ACCESSING_LSW32, ACCESSING_MSW32 & simple mem_mask checks have been replace with the new macros.
The old macros are gone.
Converted address maps to tokens. Changed the address_map structure
to house global map-wide information and hung a list of entries off
of it corresponding to each address range. Introduced new functions
address_map_alloc() and address_map_free() to build/destroy these
structures. Updated all code as necessary.
Fixed several instances of porttagtohandler*() in the address maps.
Drivers should use AM_READ_PORT() macros instead.
ADDRESS_MAP_EXTERN() now is required to specify the number of
databits, just like ADDRESS_MAP_START.
Removed ADDRESS_MAP_FLAGS() grossness. There are now three new macros
which replace its former usage. ADDRESS_MAP_GLOBAL_MASK(mask)
specifies a global address-space-wide mask on all addresses. Useful
for cases where one or more address lines simply are not used at
all. And ADDRESS_MAP_UNMAP_LOW/HIGH specifies the behavior of
unmapped reads (do they come back as 0 or ~0).
Changed internal memory mapping behavior to keep only a single
address map and store the byte-adjusted values next in the address
map entries rather than maintaining two separate maps. Many other
small internal changes/cleanups.
Updated all call-through handlers appropriately. Renamed read8_handler to
read8_machine_func, replicating this pattern throughout.
Defined new set of memory handler functions which are similar but which
pass a const device_config * in place of the running_machine *. These are
called read8_device_func, etc. Added macros READ8_DEVICE_HANDLER() for
specifying functions of this type. Note that some plumbing still needs to
happen in memory.c before this will work.
This check-in should remove the need for the global Machine and in turn
"deprecat.h" for a lot of drivers, but that work has not been done. On
the flip side, some new accesses to the global Machine were added in the
emu/ files. These should be addressed over time, but are smaller in
number than the references in the driver.
suffixed with _func. Did this throughout the core and
drivers I was familiar with.
Fixed gcc compiler error with recent render.c changes.
gcc does not like explicit (int) casts on float or
double functions. This is fracking annoying and stupid,
but there you have it.
a new compile-time define (ENABLE_DEBUGGER). This means that MAME_DEBUG no longer means
"enable debugger", it simply enables debugging features such as assertions and debug code
in drivers.
Also removed the various levels of opbase protection in memory.h and always just turned
on full bounds checking.
Fixed build break due to missing ampoker.lay -> ampoker2.lay renaming.
The idea is to create extra work if a driver wants to use these and hopefully
gives an incentive to look for an alternate solution
- Added #include of deprecat.h that rely on these contructs
- Removed a bunch of unneccassary #include's from these files
The system 10 & 12 clock speeds have been raised as they run on an upgraded chipset.
All clocks are currently set to divide by 2 externally, I don't know if this is correct.
The clock is also divided internally as we have no wait states, incorrect dma timing, no gpu timing, no dma bus stealing and no gte timing.
Updated all CPU cores to return a CPUINFO_INT_CLOCK_MULTIPLIER of 1.
Changed the core to actually respect both CPUINFO_INT_CLOCK_MULTIPLIER and CPUINFO_INT_CLOCK_DIVIDER.
Updated a number of drivers to use cpunum_get_clock() instead of Machine->drv->cpu[x].clock.
***** Raw input clock speeds should now be specified for all CPUs in the MACHINE_DRIVER. *****
Removed explicit divisors from all drivers using the following CPU types,
which were already specifying non-1 values for CPUINFO_INT_CLOCK_DIVIDER:
* COP4x0
* I8039/8048 families
* M68(7)05, HD63705
* M6809E
* PIC16C5X
* TMS32010
* TMS340x0
In a few cases, it appears that the divisor was not being used, so I guessed in those cases whether or not
the specified clock speed was raw.
Cleaned up jaguar driver:
* proper video timing, configured by the chipset
* 32-bit rendering, removing 16bpp hacks
* support for borders
* proper object processor timing, including multiple passes per line
* added R3041 as a clone of the R3000
* fixed XTALs based on documentation
Removed hack in setting the timer in the MIPS core, which caused missed timers on the aleck64 games.
Fixed icount management in the RSP core which caused it to report negative cycle counts.
Fixes aleck64_0120red and mtetrisc0115u1red.
This fixes the winwork.c and poly(new).c meory leaks in the viper.c
driver. I didn't look at the x86drc.c, because Aaron said it will be
changed soon and there is no need to investigate those leaks at all.
As I wanted to update my own personal TODO about that fix I
recognised it only happens in viper.c as well, so I took a stab at
it. It was just anothe rmissing cleanup function and I also cleaned
up the *_exit() potions/usage of the cores supporting DRC a bit.
- removed years from copyright notices
- removed redundant (c) from copyright notices
- updated "the MAME Team" to be "Nicola Salmoria and the MAME Team"
This is an updated version of my earlier ATTR_PRINTF patch. It was
reviewed by Atari Ace to use ATTR_PRINTF properly and fixes even more
format errors. I also reviewed the whole source again and it is now
used in all possible places.
Removed ui_popup(). Drivers should always be using popmessage() instead (has been this way for a while).
Augmented popmessage() so that you can pass NULL to immediately dismiss any messages.
