- Implements render target data load (aka Read Color Buffer/Read Depth Buffer)
- Refactors vulkan surface barrier to be much cleaner.
- Removes redundant surface barrier invocations after doing a merged load
from surface cache.
- Adds explicit access modes when gathering surfaces from cache.
- Further improve aliased data preservation by unconditionally scanning.
Its is possible for cache aliasing to occur when doing memory split.
- Also sets up for RCB/RDB implementation
- use std::chrono to measure 20ms between XInputSetState calls, because clock() measures cpu time and not real time (so calls were made more often than intended)
- fixup data types for cached vibration values to prevent u8 values overflowing and giving false positives on m_dev->newVibrateData
- Added support for current default toolset, explitly specified c++17 standard and 8.1 win sdk.
- properties were defined via `common_default.props` and `common_default_macros.props`
Note: There are 2 files because some properties need to be imported before `Microsoft.Cpp.Default.props` , but other common properties can be imported after but they require marco defined in `Microsoft.Cpp.Default.props`.
- Added hidapi and libusb project files to repository to simplify msbuild changes.
- llvm and glslang projects modified to detect and use Visual Studio Version.
- Added `-Thost=x64` to llvm NMake CL (warning in cmake lists)
Note: `-A` option is used to set platform since rpcs3 requires cmake 3.8.3+ anyway.
Starting with QT 5.13, having both fonts causes stylesheets to use the fallback font even if the first font is present, but uses the first font when there's no fallback font.
The fallback font (Arial) can safely be removed as the default one is 'good enough' for the intended purpose of having a fallback.
With the option enabled GET commands are blocked until the current PUTLLC/PUTLLUC executer on that address finishes
Additional improvements:
- Minor race fix of sys_ppu_thread_exit (wait until the writer finishes)
- Max number of ppu threads bumped to 8
vkAcquireNextImageKHR can also return VK_SUBOPTIMAL_KHR and is non-fatal.
However, it's a good idea to still recreate the swap chain later to maintain
optimal presentation paths after temporary occlusion.
