to eliminate device().machine().scheduler() chains all over the place (nw)
This change has been sitting in my local tree for months, from when I was
(mostly unsuccessfully) working on c64 performance. It gives a small speedup
to drivers with many tightly-interleaved executing devices (i.e. Commodore
drivers) I didn't think it was worth making everyone do a clean build for
such tiny gains but I didn't have the heart to chuck it. Since I'm already
making another core-touching, clean-build-needed commit tonight, in it goes.
Moved delegates into /src/lib/util to enable usage of delegates in other project parts
Moved mame_printf_* calls into /src/osd/osdcore.c and renamed them to osd_printf_*
Changed mess.mak to display compilation of ymmu100.ppm nicely
- updated device_execute_interface::interface_post_reset so screen is searched
from same context
- removed tilelgcy.h since no more users exist
- made bfm_adr2 a device
owner device, not the device itself, to make it consistent with devcb2
patterns. This change is mostly hidden since most drivers use
MCFG_DEVICE_VBLANK_INT_DRIVER and MCFG_DEVICE_PERIODIC_INT_DRIVER. But
for those few cases where a device was explicitly specified, it reduces
the awkwardness.
We should probably standardize this when using device_delegates for
consistency.
A full rebuild is necessary after taking this change, since the macros
mentioned above have changed their parameters.
sound devices for each of the Midway 8-bit sound
boards. This will also aid in eventually hooking them
up to pinballs.
Enhanced the mixer interface support to allow for
more than one output line. To use this you need to
use the MCFG_MIXER_ROUTE macro instead of
MCFG_SOUND_ROUTE so that the mixer output index can
be specified. See midway_ssio_device for an example.
make its purpose clearer.
Added delegate support for interrupt callbacks, and made driver_device
variants of the generic interrupt helper callbacks. Did not convert
existing callers yet due to logistical issues.
and paths consistently for devices, I/O ports, memory
regions, memory banks, and memory shares. [Aaron Giles]
NOTE: there are likely regressions lurking here, mostly
due to devices not being properly found. I have temporarily
added more logging to -verbose to help understand what's
going on. Please let me know ASAP if anything that is being
actively worked on got broken.
As before, the driver device is the root device and all
other devices are owned by it. Previously all devices
were kept in a single master list, and the hierarchy was
purely logical. With this change, each device owns its
own list of subdevices, and the hierarchy is explicitly
manifest. This means when a device is removed, all of its
subdevices are automatically removed as well.
A side effect of this is that walking the device list is
no longer simple. To address this, a new set of iterator
classes is provided, which walks the device tree in a depth
first manner. There is a general device_iterator class for
walking all devices, plus templates for a device_type_iterator
and a device_interface_iterator which are used to build
iterators for identifying only devices of a given type or
with a given interface. Typedefs for commonly-used cases
(e.g., screen_device_iterator, memory_interface_iterator)
are provided. Iterators can also provide counts, and can
perform indexed lookups.
All device name lookups are now done relative to another
device. The maching_config and running_machine classes now
have a root_device() method to get the root of the hierarchy.
The existing machine->device("name") is now equivalent to
machine->root_device().subdevice("name").
A proper and normalized device path structure is now
supported. Device names that start with a colon are
treated as absolute paths from the root device. Device
names can also use a caret (^) to refer to the owning
device. Querying the device's tag() returns the device's
full path from the root. A new method basetag() returns
just the final tag.
The new pathing system is built on top of the
device_t::subtag() method, so anyone using that will
automatically support the new pathing rules. Each device
has its own internal map to cache successful lookups so
that subsequent lookups should be very fast.
Updated every place I could find that referenced devices,
memory regions, I/O ports, memory banks and memory shares
to leverage subtag/subdevice (or siblingtag/siblingdevice
which are built on top).
Removed the device_list class, as it doesn't apply any
more. Moved some of its methods into running_machine
instead.
Simplified the device callback system since the new
pathing can describe all of the special-case devices that
were previously handled manually.
Changed the core output function callbacks to be delegates.
Completely rewrote the validity checking mechanism. The
validity checker is now a proper C++ class, and temporarily
takes over the error and warning outputs. All errors and
warnings are collected during a session, and then output in
a consistent manner, with an explicit driver and source file
listed for each one, as well as additional device and/or
I/O port contexts where appropriate. Validity checkers
should no longer explicitly output this information, just
the error, assuming that the context is provided.
Rewrote the software_list_device as a modern device, getting
rid of the software_list_config abstraction and simplifying
things.
Changed the way FLAC compiles so that it works like other
external libraries, and also compiles successfully for MSVC
builds.
a few CPU's to give octal output by default.
Note that for now just some functions things return octal, will add more, but since this change required clean compile better to put it sooner then later (no whatsnew)
using one cothread for each executable device. This functionality is
strictly internal and should result in no changes in scheduling
behavior. However, it now becomes possible to exit out of a device's
execution in the midst of an instruction by calling
machine().scheduler().make_active(). Note that this is somewhat dangerous
for cores that have logic at the start of their execute loop to check for
interrupts or other conditions, as those checks will be skipped upon
resume, so consider this sort of early exiting to be highly experimental
for now. [Aaron Giles, Miodrag Milanovic, byuu]
- non-device timer callbacks
- machine state changing callbacks
- configuration callbacks
- per-screen VBLANK callbacks
- DRC backend callbacks
For the timer case only, I added wrappers for the old-style functions.
Over time, drivers should switch to device timers instead, reducing the
number of timers that are directly allocated through the scheduler.
existing modern devices and the legacy wrappers to work in this
environment. This in general greatly simplifies writing a modern
device. [Aaron Giles]
General notes:
* some more cleanup probably needs to happen behind this change,
but I needed to get it in before the next device modernization
or import from MESS :)
* new template function device_creator which automatically defines
the static function that creates the device; use this instead of
creating a static_alloc_device_config function
* added device_stop() method which is called at around the time
the previous device_t's destructor was called; if you auto_free
anything, do it here because the machine is gone when the
destructor is called
* changed the static_set_* calls to pass a device_t & instead of
a device_config *
* for many devices, the static config structure member names over-
lapped the device's names for devcb_* functions; in these cases
the members in the interface were renamed to have a _cb suffix
* changed the driver_enumerator to only cache 100 machine_configs
because caching them all took a ton of memory; fortunately this
implementation detail is completely hidden behind the
driver_enumerator interface
* got rid of the macros for creating derived classes; doing it
manually is now clean enough that it isn't worth hiding the
details in a macro
Remove redundant machine items from address_space and device_t.
Neither machine nor m_machine are directly accessible anymore.
Instead a new getter machine() is available which returns a
machine reference. So:
space->machine->xxx ==> space->machine().xxx
device->machine->yyy ==> device->machine().yyy
Globally changed all running_machine pointers to running_machine
references. Any function/method that takes a running_machine takes
it as a required parameter (1 or 2 exceptions). Being consistent
here gets rid of a lot of odd &machine or *machine, but it does
mean a very large bulk change across the project.
Structs which have a running_machine * now have that variable
renamed to m_machine, and now have a shiny new machine() method
that works like the space and device methods above. Since most of
these are things that should eventually be devices anyway, consider
this a step in that direction.
98% of the update was done with regex searches. The changes are
architected such that the compiler will catch the remaining
errors:
// find things that use an embedded machine directly and replace
// with a machine() getter call
S: ->machine->
R: ->machine\(\)\.
// do the same if via a reference
S: \.machine->
R: \.machine\(\)\.
// convert function parameters to running_machine &
S: running_machine \*machine([^;])
R: running_machine \&machine\1
// replace machine-> with machine.
S: machine->
R: machine\.
// replace &machine() with machine()
S: \&([()->a-z0-9_]+machine\(\))
R: \1
// sanity check: look for this used as a cast
(running_machine &)
// and change to this:
*(running_machine *)
functionality in favor of alternate mechanisms. Errors are
now reported via an astring rather than via callbacks. Every
option must now specify a type (command, integer, float, string,
boolean, etc). Command behavior has changed so that only one
command is permitted. [Aaron Giles]
Changed fileio system to accept just a raw searchpath instead of
an options/option name combination. [Aaron Giles]
Created emu_options class dervied from core_options which wraps
core emulator options. Added mechanisms to cleanly change the
system name and add/remove system-specific options, versus the
old way using callbacks. Also added read accessors for all the
options, to ensure consistency in how parameters are handled.
Changed most core systems to access emu_options instead of
core_options. Also changed machine->options() to return emu_options.
[Aaron Giles]
Created cli_options class derived from emu_options which adds the
command-line specific options. Updated clifront code to leverage
the new class and the new core behaviors. cli_execute() now accepts
a cli_options object when called. [Aaron Giles]
Updated both SDL and Windows to have their own options classes,
derived from cli_options, which add the OSD-specific options on
top of everything else. Added accessors for all the options so
that queries are strongly typed and simplified. [Aaron Giles]
Out of whatsnew: I've surely screwed up some stuff, though I have
smoke tested a bunch of things. Let me know if you hit anything odd.
Also I know this change will impact the WINUI stuff, please let me
know if there are issues. All the functionality necessary should
still be present. If it's not obvious, please talk to me before
adding stuff to the core_options class.
to pass a core_options object to the constructor, along with
a search path. This required pushing either a running_machine
or a core_options through some code that wasn't previously
ready to handle it. emu_files can be reused over multiple
open/close sessions, and a lot of core code cleaned up
nicely as things were converted to them.
Also created a file_enumerator class for iterating over files
in a searchpath. This replaces the old mame_openpath functions.
Changed machine->options() to return a reference.
Removed public nvram_open() and fixed jchan/kaneko16 to
stop directly saving NVRAM.
Removed most of the mame_options() calls; this will soon go
away entirely, so don't add any more.
Added core_options to device_validity_check() so they can be
used to validate things.
are still intact. The new state_manager class has templatized methods
for saving the various types, and through template specialization can
save more complex system types cleanly (like bitmaps and attotimes).
Added new mechanism to detect proper state save types. This is much
more strict and there will likely be some games/devices that fatalerror
at startup until they are remedied. Spot checking has caught the more
common situations.
The new state_manager is embedded directly in the running_machine,
allowing objects to register state saving in their constructors now.
Added NAME() macro which is a generalization of FUNC() and can be
used to wrap variables that are registered when directly using the
new methods as opposed to the previous macros. For example:
machine->state().save_item(NAME(global_item))
Added methods in the device_t class that implicitly register state
against the current device, making for a cleaner interface.
Just a couple of required regexes for now:
state_save_register_postload( *)\(( *)([^,;]+), *
\3->state().register_postload\1\(\2
state_save_register_presave( *)\(( *)([^,;]+), *
\3->state().register_presave\1\(\2
timers into the scheduler. Retain TIMER devices as a separate wrapper
in timer.c/.h. Inline wrappers are currently provided for all timer
operations; a future update will bulk clean these up.
Rather than using macros which hide generation of a string-ified name
for callback functions, the new methods require passing both a function
pointer plus a name string. A new macro FUNC() can be used to output
both, and another macro MFUNC() can be used to output a stub-wrapped
class member as a callback.
Also added a time() method on the machine, so that machine->time() gives
the current emulated time. A wrapper for timer_get_time is currently
provided but will be bulk replaced in the future.
For this update, convert all classic timer_alloc, timer_set,
timer_pulse, and timer_call_after_resynch calls into method calls on
the scheduler.
For new device timers, added methods to the device_t class that make
creating and managing these much simpler. Modern devices were updated
to use these.
Here are the regexes used; some manual cleanup (compiler-caught) will
be needed since regex doesn't handle nested parentheses cleanly
1. Convert timer_call_after_resynch calls
timer_call_after_resynch( *)\(( *)([^,;]+), *([^,;]+), *([^,;]+), *([^);]+)\)
\3->scheduler().synchronize\1\(\2FUNC(\6), \5, \4\)
2. Clean up trailing 0, NULL parameters
(synchronize[^;]+), 0, NULL\)
\1)
3. Clean up trailing NULL parameters
(synchronize[^;]+), NULL\)
\1)
4. Clean up completely empty parameter lists
synchronize\(FUNC\(NULL\)\)
synchronize()
5. Convert timer_set calls
timer_set( *)\(( *)([^,;]+), *([^,;]+), *([^,;]+), *([^,;]+), *([^);]+)\)
\3->scheduler().timer_set\1\(\2\4, FUNC(\7), \6, \5\)
6. Clean up trailing 0, NULL parameters
(timer_set[^;]+), 0, NULL\)
\1)
7. Clean up trailing NULL parameters
(timer_set[^;]+), NULL\)
\1)
8. Convert timer_set calls
timer_pulse( *)\(( *)([^,;]+), *([^,;]+), *([^,;]+), *([^,;]+), *([^);]+)\)
\3->scheduler().timer_pulse\1\(\2\4, FUNC(\7), \6, \5\)
9. Clean up trailing 0, NULL parameters
(timer_pulse[^;]+), 0, NULL\)
\1)
10. Clean up trailing NULL parameters
(timer_pulse[^;]+), NULL\)
\1)
11. Convert timer_alloc calls
timer_alloc( *)\(( *)([^,;]+), *([^,;]+), *([^);]+)\)
\3->scheduler().timer_alloc\1\(\2FUNC(\4), \5\)
12. Clean up trailing NULL parameters
(timer_alloc[^;]+), NULL\)
\1)
13. Clean up trailing 0 parameters
(timer_alloc[^;]+), 0\)
\1)
14. Fix oddities introduced
\&m_machine->scheduler()
m_machine.scheduler()
global functions which are now superceded by the operators and
methods on the class. [Aaron Giles]
Required mappings are:
attotime_make(a,b) => attotime(a,b)
attotime_to_double(t) => t.as_double()
double_to_attotime(d) => attotime::from_double(d)
attotime_to_attoseconds(t) => t.as_attoseconds()
attotime_to_ticks(t,f) => t.as_ticks(f)
ticks_to_attotime(t,f) => attotime::from_ticks(t,f)
attotime_add(a,b) => a + b
attotime_add_attoseconds(a,b) => a + attotime(0, b)
attotime_sub(a,b) => a - b
attotime_sub_attoseconds(a,b) => a - attotime(0, b)
attotime_compare(a,b) == 0 => a == b
attotime_compare(a,b) != 0 => a != b
attotime_compare(a,b) < 0 => a < b
attotime_compare(a,b) <= 0 => a <= b
attotime_compare(a,b) > 0 => a > b
attotime_compare(a,b) >= 0 => a >= b
attotime_mul(a,f) => a * f
attotime_div(a,f) => a / f
attotime_min(a,b) => min(a,b)
attotime_max(a,b) => max(a,b)
attotime_is_never(t) => t.is_never()
attotime_string(t,p) => t.as_string(p)
In addition, some existing #defines still exist but will go away:
attotime_zero => attotime::zero
attotime_never => attotime::never
ATTOTIME_IN_SEC(s) => attotime::from_seconds(s)
ATTOTIME_IN_MSEC(m) => attotime::from_msec(m)
ATTOTIME_IN_USEC(u) => attotime::from_usec(u)
ATTOTIME_IN_NSEC(n) => attotime::from_nsec(n)
ATTOTIME_IN_HZ(h) => attotime::from_hz(h)
Also removed redundant m_machine from the state and execute
interfaces to fix ambiguity when using m_machine from within
a device that inherits from these.
