remove xe_kernel_export_shim_fn field of Export function_data, trampoline is now the only way exports get invoked
Remove kernelstate argument from string functions in order to conform to the trampoline signature (the argument was unused anyway)
Constant-evaluated initialization of ppc_opcode_disasm_table, removal of unused std::vector fields
Constant-evaluated initialization of export tables
name field on export is just a const char* now, only immutable static strings are ever passed to it
Remove unused callcount field of export.
PM4 compare op function extracted
Globally apply /Oy, /GS-, /Gw on msvc windows
Remove imgui testwindow code call, it took up like 300 kb
But for normal msvc builds i would put it at around 30-50%
Added per-xexmodule caching of information per instruction, can be used to remember what code needs compiling at start up
Record what guest addresses wrote mmio and backpropagate that to future runs, eliminating dependence on exception trapping. this makes many games like h3 actually tolerable to run under a debugger
fixed a number of errors where temporaries were being passed by reference/pointer
Can now be compiled with clang-cl 14.0.1, requires -Werror off though and some other solution/project changes.
Added macros wrapping compiler extensions like noinline, forceinline, __expect, and cold.
Removed the "global lock" in guest code completely. It does not properly emulate the behavior of mfmsrd/mtmsr and it seriously cripples amd cpus. Removing this yielded around a 3x speedup in Halo Reach for me.
Disabled the microprofiler for now. The microprofiler has a huge performance cost associated with it. Developers can re-enable it in the base/profiling header if they really need it
Disable the trace writer in release builds. despite just returning after checking if the file was open the trace functions were consuming about 0.60% cpu time total
Add IsValidReg, GetRegisterInfo is a huge (about 45k) branching function and using that to check if a register was valid consumed a significant chunk of time
Optimized RingBuffer::ReadAndSwap and RingBuffer::read_count. This gave us the largest overall boost in performance. The memcpies were unnecessary and one of them was always a no-op
Added simplification rules for multiplicative patterns like (x+x), (x<<1)+x
For the most frequently called win32 functions i added code to call their underlying NT implementations, which lets us skip a lot of MS code we don't care about/isnt relevant to our usecases
^this can be toggled off in the platform_win header
handle indirect call true with constant function pointer, was occurring in h3
lookup host format swizzle in denser array
by default, don't check if a gpu register is unknown, instead just check if its out of range. controlled by a cvar
^looking up whether its known or not took approx 0.3% cpu time
Changed some things in /cpu to make the project UNITYBUILD friendly
The timer thread was spinning way too much and consuming a ton of cpu, changed it to use a blocking wait instead
tagged some conditions as XE_UNLIKELY/LIKELY based on profiler feedback (will only affect clang builds)
Shifted around some code in CommandProcessor::WriteRegister based on how frequently it was executed
added support for docdecaduple precision floating point so that we can represent our performance gains numerically
tons of other stuff im probably forgetting
- Changed name of config option to apply_title_update to better reflect what that option does
- Mount TU package to UPDATE: partition
- Simplified UserModule::title_id()
- Splitted loading module into two parts to allow applying TUs and custom patches
C++17ification!
- Filesystem interaction now uses std::filesystem::path.
- Usage of const char*, std::string have been changed to
std::string_view where appropriate.
- Usage of printf-style functions changed to use fmt.
This was a headache to work out, big thanks to the lack of documentation on .xexp files... a ton of guesswork was involved here but luckily it turned out well.
I did have to make some pretty major changes to the way XEX files are loaded though.
Previously it'd just load everything in one go: XEX headers -> decrypt/decompress data -> load imports/symbols -> set loader data table entries, etc...
Now it's changed to something like this:
- Load base XEX headers + decrypted/decompressed image data, return X_STATUS_PENDING
- In the LoadFromFile call used to load the XEX, search for XEXP patch file (only .xexp in same folder atm)
- If patch exists: load XEXP, decrypt headers/data, apply patch to base XEX, dispose of XEXP
- Finish XEX load via LoadXexContinue() (handles imports/symbols/loader data...)
This saves us from needing to reset the imports/function/symbol stuff after patching (since all the XEX code will be a lot different), but I'm not really sure if I went about it the best way.
Code is mainly just copy/pasted from kernel/util/xex2.cc, I've tried fixing it up to work better in a class, but there's probably some things I missed.
Also includes some minor improvements to the XEX loader, like being able to try both XEX keys (retail/devkit) automatically, and some fixes to how the base address is determined.
(Previously there was code that would get base address from optional header, code that'd get it from xex_security_info, code that'd use a stored base address value...
Now everything reads it from a single stored value instead, which is set either from the xex_security_info, or if it exists from the optional header instead.
Maybe this can help improve compatibility with any weird XEX's that don't have a base address optional header?)
Compressed XEX loader also has some extra checks to make sure the compressed data hash matches what's expected.
Might increase loading times by a fraction, but could save reports from people unknowingly using corrupt XEXs.
(still no checks for non-compressed data though, maybe need to compare data with xex_security_info->ImageHash?)
