Tags: Vector35/binaryninja-api
Tags
[ARMv7] Expand ARM and Thumb SIMD and VFP lifting Add lifting for vector comparisons, bitwise operations, widening and narrowing moves, multiply-long, negation, pairwise addition, shifts, swaps, transposes, tests, and interleaving operations. Use dedicated intrinsics for lane-wise operations and explicit IL for bitwise operations, register swaps, and scalar floating-point negation. Correct SIMD decoding, signedness, immediate types, D/Q operand widths, and reserved encodings. Support ARMv8 VMULL.P64 polynomial products. Add FLDMDBX/FLDMIAX and FSTMDBX/FSTMIAX lifting, accounting for the unused trailing word in addressing and writeback. Correct Thumb decoding of these forms, including SP-based transfers.
Eliminate polling latency in BinaryView search generators.
Unify the navigation paths between graph/linear views, and improve de… …rived string navigation
Unify the navigation paths between graph/linear views, and improve de… …rived string navigation
Rename the API struct conversion interface to use APIStruct names Value types convertible to their C API structures now use methods named `ToAPIStruct`, `FromAPIStruct`, and `FreeAPIStruct`. The `APIAble` trait is renamed to `APIStruct`. This frees up the `APIObject` naming to be used for representing non value type objects in the future.
Rename the API struct conversion interface to use APIStruct names Value types convertible to their C API structures now use methods named `ToAPIStruct`, `FromAPIStruct`, and `FreeAPIStruct`. The `APIAble` trait is renamed to `APIStruct`. This frees up the `APIObject` naming to be used for representing non value type objects in the future.
Add bn::base::DynamicLibrary, an abstraction over runtime loading of … …shared libraries `bn::base::DynamicLibrary` opens a shared library at runtime and unloads it when the object is destroyed. The platform handle can be handed off with `Release()` when the library should remain loaded indefinitely. `Open()` returns either the library or the platform's description of why it could not be loaded. `SymbolScope` chooses whether `GetSymbol()` also resolves symbols from the libraries that the loaded library links against. Each platform restricts this differently, so macOS passes `RTLD_FIRST` to `dlopen()`, other Unixes check which image defines each symbol, and Windows only ever searches the module itself.
[MIPS] Expand floating-point, coprocessor, and shuffle lifting - Add scalar and paired-single FP arithmetic and conditional moves - Lift TRUNC/ROUND/CEIL/FLOOR.W/L.S/D with fixed rounding semantics - Model paired-single arithmetic and per-lane conditional moves with intrinsics - Decode and lift ALIGN/DALIGN and BITSWAP/DBITSWAP - Lift JALX, coprocessor zero-test branches, and PREFX - Model COP2 operations, control/high-word transfers, and shadow GPR access with intrinsics - Preserve branch inputs across conditional and split-register delay-slot writes - Expand lifting regressions for endianness, register modes, aliasing, zero registers, and delay slots
[MIPS] Expand floating-point, coprocessor, and shuffle lifting - Add scalar and paired-single FP arithmetic and conditional moves - Lift TRUNC/ROUND/CEIL/FLOOR.W/L.S/D with fixed rounding semantics - Model paired-single arithmetic and per-lane conditional moves with intrinsics - Decode and lift ALIGN/DALIGN and BITSWAP/DBITSWAP - Lift JALX, coprocessor zero-test branches, and PREFX - Model COP2 operations, control/high-word transfers, and shadow GPR access with intrinsics - Preserve branch inputs across conditional and split-register delay-slot writes - Expand lifting regressions for endianness, register modes, aliasing, zero registers, and delay slots
[MIPS] Expand floating-point, coprocessor, and shuffle lifting - Add scalar and paired-single FP arithmetic and conditional moves - Lift TRUNC/ROUND/CEIL/FLOOR.W/L.S/D with fixed rounding semantics - Model paired-single arithmetic and per-lane conditional moves with intrinsics - Decode and lift ALIGN/DALIGN and BITSWAP/DBITSWAP - Lift JALX, coprocessor zero-test branches, and PREFX - Model COP2 operations, control/high-word transfers, and shadow GPR access with intrinsics - Preserve branch inputs across conditional and split-register delay-slot writes - Expand lifting regressions for endianness, register modes, aliasing, zero registers, and delay slots
PreviousNext