Unify internal code generation to make better use of GHC calling convention.
Ideally, it would just work on Windows as well, but some random bug appeared.
This bug was causing freezes on SPU LLVM compilation.
This commit desperately attempts to workaround it.
Move source files to Emu/GDB.cpp, GDB.h
Remove "WITH_GDB" option, enable GDB Server by default.
Change class name to gdb_thread.
Alias for external access gdb_server.
Change config option name to "GDB Server"
Bind on 127.0.0.1 by default.
- Multiple header files where missing #includes to other headers that
where used in the header. Correct header was included in correct
order in source files which caused everything to compile.
- Added missing #includes so header files correctly include all their
dependencies and fixes problems with IDEs being unable to parse
headers correctly due to missing symbols
- Add vm_locking.h and vm_reservation.h and move relevant functions
and types to these headers.
- Change include order and make vm_ptr.h, vm_var.h and vm_ref.h headers
usable invidually and them including vm.h instead of other way around
- Because usage of vm::ptr now requires including vm_ptr.h instead of
vm.h updated multiple #includes
- Added additional #includes to vm_reservation.h and vm_locking to
where vm::reservation_* and locking related functions are used
Misc: fix EH frame registration (LLVM, non-Windows).
Misc: constant-folding bitcast (cpu_translator).
Misc: add syntax for LLVM arrays (cpu_translator).
Misc: use function names for proper linkage (SPU LLVM).
Changed function search and verification in Giga mode.
Basic stack frame layout analysis.
Function detection in Giga mode.
Basic use of new information in SPU LLVM.
Fixed jump table compilation in SPU LLVM.
Disable broken optimization in Accurate xfloat mode.
Make compiled SPU modules position-independent in SPU LLVM.
Optimizations include but not limited to:
* Compiling SPU functions as native functions when eligible
* Avoiding register context write-out
* Aligned stack assumption (CWD alike instruction)
Are made composable in expressions similar to arithmetic ops.
Implement noncast in addition to bitcast (no-op case).
Implement bitcast constant folding.
Fixed some misuse of sext<>.
Properly forward value categories in expression structs.
Simplify SFINAE tests (is_llvm_expr, llvm_common_t) in global operators.
Add llvm_const_int and remove llvm_add_const, llvm_sub_const, etc.
Add llvm_ord and llvm_uno for FP comparison via >=< operators.
Replace cpu_translator::fcmp with fcmp_ord and fcmp_uno.
* Fix SPU LR event setting in atomic commands according to hw test
* MFC: increment timestamp for PUT cmd in non-tsx path
* MFC: fix reservation lost test on non-tsx path in regard to the lock bit
* Reservation notification moved out of writer_lock scope to reduce its lifetime
* Use passive_lock/unlock in ppu atomic inctrustions to reduce redundancy
* Lock only once for dma transfers (non-TSX)
* Don't use RDTSC in reservation update logic
* Remove MFC cmd args passing to process_mfc_cmd
* Reorder check_state cpu_flag::memory check for faster unlocking
* Specialization for 128-byte data copy in SPU dma transfers
* Implement memory range locks and isolate PPU and SPU passive lock logic
Remove "atomic operator" classes
Remove test, test_and_set, test_and_reset, test_and_complement global functions
Simplify atomic_t<> with constexpr if, remove some garbage
Redesign bs_t<> to use class, mark its methods constexpr
Implement atomic_bs_t<> for optimizations
Remove unused __bitwise_ops concept (should be in other header anyway)
Bitsets can now be tested via safe bool conversion
* Made GDB debugger working with IDA
* Added async interrupts support
* Report proper thread after pausing
* Support attaching debugger before running app
`unveil<>` renamed to `fmt_unveil<>`, now packs args to u64 imitating va_args
`bijective...` removed, `cfg::enum_entry` now uses formatting system
`fmt_class_string<>` added, providing type-specific "%s" handler function
Added `fmt::append`, removed `fmt::narrow` (too obscure)
Utilities/cfmt.h: C-style format template function (WIP)
Minor formatting fixes and cleanup
* Optimizations
1) Some headers simplified for better compilation time
2) Some templates simplified for smaller executable size
3) Eliminate std::future to fix compilation for mingw64
4) PKG installation can be cancelled now
5) cellGame fixes
6) XAudio2 fix for mingw64
7) PPUInterpreter bug fixed (Clang)
* any_pod<> implemented
Aliases: any16, any32, any64
rsx::make_command fixed
This used to crash so it was removed. However I think this is a cruicial feature.
This is the only way to observe the stack-frames, registers and disasm around the PC
in a user friendly way.
The reason this caused crashes was that we got a reference to a std::vector and iterated
over it in one thread while another thread could modify it (and thereby invalidating the
iterators). The way I "fixed" it is to just copy the `std::vector<std::shared_ptr<...>>`
and iterate over the copy. This obviously has some overhead (locking the shared_ptr
counters and incrementing them). It also allows entities other than the Thread manager
to keep the Thread objects frome being deleted but that should not be a problem.
A copy constructor is necessarry for `class InstrBase`, as the implicit
copy constructor simply copies the pointer `m_args`. This results in a
double delete of the same memory region, causing a segmentation fault
when rpcs3 exited.
1) MakeFromLE replaced with make(), MakeFromBE removed. Compiler seems
to be able to optimize it anyway.
2) be_t<> conversion operator temporarily replaced with conversion
through LE because it can't work with all possible numeric conversions
(float<>int for example)
Reasons: 1) It's unsafe - we cannot guarantee that address is still good
while it is used. 2) It's wrong in many modules which are usually just
compare pointer with zero.
* use one central unified log with channels/priorities ad-hoc listener registration and de-registration
* disable buffering by default
* add multi-threaded ringbuffer implementation
* use buffered listener for the gui (using the ringbuffer)
move module initialization into a module manager, still has some issues like stopping not working and debug crashing
add #idef 0 to modules that aren't in the windows project
don't double initialize and don't de-initialize for now, since many modules don't expect it and it leads to many errors
remove duplicate module lists for empty modules and implemented ones, make Module non-copyable but movable
add secondary project, no real use for it now
add some memleak config to the emucore and add asmjit path to rpcs3
small rebase error fixed to get it to compile again
add filters for emucore
re-add the module manager and static file
WIP commit, linker errors abound
some more abstraction layer stuff
fix the remaining linker errors, re-enable platform specific mouse, pad and keyboard handlers
rebasing
fix memset undefined and re() usage of se_t before declaration
Add wxGUI define by default for cmake builds
fix copy constructors of Datetime header
fix copy constructors of other wx interface classes
remove static declarations of global variables
make wxGLCanvas constructor non-ambiguous even with wx2.8. compat mode, fix wrong std::exception constructor calls
remove duplicate definition for FromUTF8 and ToUTF8
temp changes
make precompiled header slimmer under Linux to increase CI and dev-machine build-times
make sure unused modules don't compile
add unused modules to the VS project to easier keep track of them
BUGFIX: Add break after NV4097_SET_TEXTURE_BORDER_COLOR in RSXThread.cpp
BUGFIX: Fix parameters passed to RSXTexture::SetControl3 (they were being
passed in reverse order)
BUGFIX: Remove invalid, non-sensical call to glPixelStorei in GLGSRender.h
BUGFIX: Fix signed/unsigned comparison compiler warnings in GLGSRender.h
CHANGE: Make GLFragmentProgram::Decompiler synchronous by default
CHANGE: Update wxWidgets submodule to latest commit
BUGFIX: Fix several memory leaks
ADDED: Created a new MSVC debug configuration to output locations
of allocations that end up leaking after the program
is closed.
BUGFIX: Fix the stupid PadHandler crash due to the lack of a virtual d'tor
* replace GetThreadID with std::this_thread.getId()
* name all anonymous structs and unions that contain non-trivially constructable objects
* made default constructor for big endian type noexcept to make it work with std::atomic
* move instantiated specialized template function members ouside of the class definition to comply with the standard
* added default instantiation for template parameter "=nullptr"
* used the C++11 standardized thread_local instead of the __declspec(thread)
* added transitional definitions to bridge the microsoft specific calls (compare and exchange and aligned alloc)
* removed cyclic dependency between Emulator->CPUThreadManager->CPUThread->SMutex->Emulator->...
* fixed some instances of indentation by space instead of tabs
* surrounded some unused code with an #if 0 block to make sure it doesn't compile
