In order to reduce rounding problems from calculations, FPU stack
registers for x86 architectures contain values stored in an
80-bit extended precision format.
As a result, reading and writing to these registers requires
specific handling.
This update brings the Ruby bindings in line with the Python
bindings by supporting reading and writing the FPU stack registers
using 2-element arrays: [mantissa, exponent]
The mantissa array element contains the first 64 bits of the FPU
stack register.
The exponent array element contains the last 16 bits of the FPU
stack register.
Sometimes, the finalizer for an `UcRef` runs so late that the members of the
module have already been set to `None`. We need to make sure that we don't
depend on anything in the module, or we risk getting a Exception when we try
to access the `release_handle` method of `None` (`Uc`).
* Move assembly to S files
* more assembly files
* osx compilation change
* makefile mistake
* add objcopy from crosstool
* use gobjcopy on osx
* start cmocka install cleanup
* move wget to directory option
* move back to cd
* fix copy
* First cut
* free allocated memory
* bad idea
too much switching between python and c
* add debug
* cleanup bad size
* fix mem_unmap and query for Ruby bindings
* ruby bindings: fix issues with GC freeing callbacks while we still have references to them
* ruby bindings: add test for garbage collection of hooks
* ruby bindings: let the VM garbage collect hooks properly this time
* ruby bindings: update garbage collection test to make sure Proc is garbage collected after Uc is collected
* ruby bindings: fix m_uc_hook_add to return the ruby VALUE with proper memory management instead of making another one with bad memory management
* ruby bindings: fix cb_hook_intr signature
* add architecture query
* ruby bindings: only treat certain x86 registers specially if we're actually on x86
* only treat certain x86 registers specially if we're actually on x86 (uc_reg_read and uc_reg_write)
* ruby bindings: read and write ARM64's 128-bit NEON registers
* Update UC_HOOK_MEM_VALID
UC_HOOK_MEM_READ fires before handlers for invalid memory reads, so UC_HOOK_MEM_VALID would technically also be receiving invalid memory reads. Switching to UC_HOOK_MEM_READ_AFTER ensures that only actually valid reads are hooked
* Removed macro change, added comment
Removed the macro change, but added a clarifying comment. May submit a future PR with a new macro added
* Update unicorn.h
will -> may
* Update unicorn.h
* arm64eb: arm64 big endian also using little endian instructions.
* arm64: using another example that depends on endians.
example:
1. store a word: 0x12345678
2. load a byte:
* little endian : 0x78
* big endian : 0x12