- Added build for Ghidra 11.1.2
- #36 - Fixed importing binaries with debug symbols
- #27 - Fixed processing of HI16/16 relocation pairs
- #39 - Fixed handling of
vtfmandvhtfminstructions
- Added build for Ghidra 11.1.1
- Updated internal dependencies
- #8 - switched to the official way of packaging the extension
- Warning: The installation method is different now, please see the updated README
- Note: If you're updating from version 18 in an already existing Ghidra installation you must delete the plugin first
- Fixed HI16 relocation calculation when relocating to a very high address
- Updated internal dependencies
- Updated to Ghidra 11.0
- Updated to Ghidra 10.4
- Updated to Ghidra 10.3
- #25 - added support for type A relocations only present in program headers
- Fixed debugger breakpoint placing
- Updated to Ghidra 10.2
- Added workaround for
Register must already existerrors during analysis
- Switched to the new version numbering scheme
- Added CI for automatic builds targeting multiple Ghidra versions
- Note: This version won't be released
- Updated to Ghidra 10.1.4
- Updated internal dependencies
- Updated to Ghidra 10.1
- Added missing implementation for
syncandwaitinstructions - #17 - Added workaround for missing symbol sections when applying relocations
- Ghidra Debugger can be used to debug games running in PPSSPP
- Fixed decompilation of
maxandmin - From now on, the release zip will contain precompiled Allegrex Sleigh spec
- Updated to Ghidra 10.0.4
- Added initial support for decompiling VFPU instructions
- Warning: For existing projects every VFPU branch instruction (
bvf,bvfl,bvt,bvtl) must be cleared and disassembled again to generate proper pcode
- Warning: For existing projects every VFPU branch instruction (
- Improved disassembly of VFPU instructions
- Warning: For existing projects you will need to clear the affected instructions and disassemble them again. Otherwise,
you will see wrong disassembly and decompilation
- Hint: You can search bookmarks for "Instruction pcode is unimplemented" to find VFPU instructions.
- Hint: If the project was disassembled before version 1.6 then you can easily spot wrongly disassembled load and store instruction as
they will have invalid references to
DAT(e.g.lv.q C000=>DAT_00000004,0x0(a0)). - Hint: To clear entire function: press
Select -> Functionthen pressCkey to clear andDto dissemble again - Note: It should be possible to write a script that automatically clears and dissembles every affected function, such script might be provided in future versions
- All VFPU load and store instructions: register name and offset is now shown correctly
- Operands of
vpfxs,vpfxtandvpfxdare now disassembled - Operands of all
vrotinstructions are now disassembled - Fixed wrong conditions and operands of
vmfvcandvmtvc - Fixed wrong vector size of first operand of
vcst - Fixed the mnemonic of
vwbn
- Warning: For existing projects you will need to clear the affected instructions and disassemble them again. Otherwise,
you will see wrong disassembly and decompilation
- Removed type suffixes for VFPU registers in disassembly
- Fixed decompilation of
ins - #7 - Added support for type B relocations (found in kernel modules)
- This also adds new option during import to use alternative relocation mapping (must be checked for some files)
- Fixed VFPU load and store instructions creating references to DAT instead of just showing register number
- #9 - Fixed issues with loading binaries with debug symbols
- Added
PpssppExportSymFilescript
- Updated to Ghidra 10.0.2
- #2 - Fixed wrong endianness for
long longreturn values
- Updated to Ghidra 9.2.1
- #6 - Rebasing image after importing can fail
- Note: This fix will only apply to new projects
- #5 - Added bitrev, mfic, mtic, and wsbw instructions
- #4 - Added Coprocessor 0 registers and transfers
- Initial release