When I dropped ARM from a generic target, this caused the vertexloader to try using the JIT path.
Instead of !_M_GENERIC, check for _M_X86 instead. Since it is only for the x86 target
The register cache can be filled to the point that when dumping them the FixupBranch goes over the maximum size of 0x80.
Force them to use the "5byte" variant of the jump.
If we were able to determine if the length we had to jump was <0x80 in the future this could be a slight optimization.
This has to be done in bcctrx, bclrx, and twx. It was already done in bcx before.
This fixes issue 7378.
This was some unused array just sitting in our global ppcState struct.
I've got no idea what its use was /supposed/ to be used for if ever implemented.
If a CPU string was incapable of being found we would return a null pointer, which would crash with strncpy.
Also if we couldn't get a CPU implementer we would call free() to a null pointer.
In addition, detect 64bit ARM running.
MemoryUtil.cpp was incorrectly using the old __x86_64__ define when it should be using _M_X86_64.
It was also using _ARCH_64 when it shouldn't have which was causing an errant PanicAlert to come up in my development.
Adds the ability to build our generic build option without manually enabling it in the cmake file.
If trying to build on a x86_32 host, throw a fatal error but say that a binary can be built as a generic build if one wants.
These were only compiled in on Windows and x86_32.
They provided "optimized" copies and compares based on blocksizes for the AMD Athlon and Duron CPU families.
The code was taken from something that AMD provides with a as-is license.
Just get rid of this crap.
The issue was that on memory exception we wouldn't call in to PPCAnalyst and our code_block would retain the previous blocks information.
This would cause us to compile the previous blocks instructions in prior to the exception exit.
Copies over the PP shaders to the APK's assets and installs them on run.
Exposes them via the video settings UI.
This is in anticipation of dropping the workaround for rotated blits on Adreno and instead forcing shader usage by the user.
We need to pull in function pointers for OpenGL 3.0 in order to use glAttribIPointer.
This isn't too big of an issue, and this code will be gone in the future when we change over to libepoxy.
Just need to push code upstream to libepoxy to support Android with GLES and GL first.
Interpreter::Helper_UpdateCR1 doesn't use the argument passed to UpdateCR1. It pulls its value from the FPSCR register.
Also there was a Interpreter::Helper_UpdateCR1(float) in addition to Helper_UpdateCR1(double) that hasn't ever existed. Remove the function
declaration.
The ARM insturction tables were missing addcox, addeox, and subfcox.
Implements support for addcox and addeox to the ArmJit, required little work since they are basically the same as addcx and addex.
Looking at the old code for the ButtonManager was a brainfsck. This fixes a ton of bugs I kept uncovering as I was moving along.
Fixes the gamepad configuration file being incorrect.
No longer treats touchscreen in a special way. Ends up as a regular device with a "Touchscreen" device name.
Was incorrectly converting a index from integer to ButtonType. Wouldn't work due to the addition of some unused(in JNI) enumerators in ButtonType.
Fixes an issue where a map had a key as an axis which was causing its binding to be overwritten for every axis that was used twice (eg main stick left and right);
Fixes Triggers not working at all.
Fixes DPad not working at all.
Fixes C-Stick only half working.
Removes touch screen specific nativelibrary types onTouchAxisEvent and onTouchEvent.
Adds a configuration version configuration option. Allows easy configuration overwriting if the options need to be changed during updating.
Supersedes github PR #291.
Adds a LTO option that isn't enabled by default.
Allows building release binaries with link time optimizations without generating a -dirty build.
Not enabled by default due to concerns of memory usage and increased build time.
Noticed this while messing with EFB to RAM.
We were having an implicit conversion from integer to float, GLSL ES doesn't allow implicit conversion.
Changes it to a explicit conversion to float.
This matches how ARM handles their naming in their drivers for different models.
Really it's that way because both Mali-T6xx and Mali-T7xx fall under Midgard.
While everything else (except Mali-55) fall under Utgard.
There are a few instructions in lXX that aren't currently fastmem capable due to using a register offset.
This implements fastmem for those few instructions.
Really I'll be changing how ARMv7 fastmem works in the future so this is really temporary code.
Just don't know how long it'll stay.
This relies on PR #257
They are similar enough that they will share bugs with their drivers, so make them fall under the same Mali-Txxx umbrella of bug issues.
If there is ever a need in the future for having separate bugs depending on family, we can support that then.
This is the only way we can determine the video driver version with mali.
Really it's a good thing that they only push driver updates once every two years, makes it easy to determine what driver anybody is running.
Fixes an issue in lXX where if the instruction was an update instruction and was a register address offset that we would add garbage values to the
writeback register.
Does some minor cleanups as well which save a few instructions here and there.
Fixes issue 6990.
This uses a bit of templating to remove the duplicate code that is the CodeBlocks in each emitter headers.
No actual functionality change in this.
When creating a Fixupbranch we were swapping the BL and B targets.
I think this was found by PPSSPP a while ago, but they never send PRs to merge their changes upstream.
GLSL ES 3.10 adds implicit support for the binding layout qualifier that we use.
Changes our GLSL version enums to bit values so we can check for both ES versions easily.
Ryan Houdek
Ryan Houdek