The 3-bit field at +24bit in DWORD 1 of the XMA context was named
subframe_skip_count but never actually used for skipping subframes.
Testing across multiple titles (PGR4, Halo Reach) shows the field
might control extra output buffer blocks that must be reserved per decoded
frame. Renaming the field to output_buffer_padding to reflect this
observed behavior and adding padding to the minimum output space threshold
Pdding blocks are reserved from the output budget after each frame is
fully consumed, preventing the decoder from overrunning the space
that the game expects to remain free.
This is still likely not entirely correct but reduces some of the
observed noise in current implementation.
- Fake: No decoding
- Master: 1:1 from base version of Xenia
- Old: Master with improvements
- New: Completely new decoder with better stability (but not compability)
Co-authored-by: Herman S. <429230+has207@users.noreply.github.com>
- Added codepath for "consume only" pass. This should resolve constant clicking in Source engine games.
- Added smaller performance improvements to decoder
- Added some fixes introduced by RodoMa92 in PR198
- Lack of AVX2 extension (should be done differently in the future)
- Disable deprecated-volatile warning
- Added missing override in posix EventInfo, ImGui notification class and XContent class
- Removed not used XAudio2.h include in XMP
- Fixed missing switch-case in XObject
- Added fugly template in native_list.h
- Fixed multiple smaller issues
- Discovered new unknown kXNotificationID use by XNotifyBroadcast together and wanted to record them before I forget them.
- Changed kXNotificationID to kXNotification
- Added kXNotificationSystemPXLiveSystemUpdate
C++ relational operators are supposed to raise FE_INVALID if an argument is
NaN, use std::isless/greater[equal] instead where they were easy to locate
(though there are other places possibly, mostly min/max and clamp usage was
checked).
Also fixes a copy-paste error making the CPU shader interpreter execute
MINs as MAXs instead.
The "close window" keyboard hotkey (Guide-B) now toggles between loglevel -1 and the loglevel set in your config.
Added LoggerBatch class, which accumulates strings into the threads scratch buffer. This is only intended to be used for very high frequency debug logging. if it exhausts the thread buffer, it just silently stops.
Cleaned nearly 8 years of dust off of the pm4 packet disassembler code, now supports all packets that the command processor supports.
Added extremely verbose logging for gpu register writes. This is not compiled in outside of debug builds, requires LogLevel::Debug and log_guest_driven_gpu_register_written_values = true.
Added full logging of all PM4 packets in the cp. This is not compiled in outside of debug builds, requires LogLevel::Debug and disassemble_pm4.
Piggybacked an implementation of guest callstack backtraces using the stackpoints from enable_host_guest_stack_synchronization. If enable_host_guest_stack_synchronization = false, no backtraces can be obtained.
Added log_ringbuffer_kickoff_initiator_bts. when a thread updates the cp's read pointer, it dumps the backtrace of that thread
Changed the names of the gpu registers CALLBACK_ADDRESS and CALLBACK_CONTEXT to the correct names.
Added a note about CP_PROG_COUNTER
Added CP_RB_WPTR to the gpu register table
Added notes about CP_RB_CNTL and CP_RB_RPTR_ADDR. Both aren't necessary for HLE
Changed name of UNKNOWN_0E00 gpu register to TC_CNTL_STATUS. Games only seem to write 1 to it (L2 invalidate)
