Loading repository data…
maddiestone / repository
IDA Python Embedded Toolkit -- IDAPython scripts for automating analysis of firmware of embedded devices
The IDAPython Embedded Toolkit is a set of script to automate many of the steps associated with statically analyzing, or reverse engineering, the firmware of embedded devices in IDA Pro.
The IDAPython Embedded Toolkit has been presented at the following venues:
To understand how and why the IDAPython Embedded Toolkit was created, check out the slides and recording from the DerbyCon or RECON Presentations.
The IDAPython Embedded Toolkit is a set of IDAPython scripts written to be processor/architecture-agnostic and automate the triage, analysis, and annotation processes associated with reversing the firmware image of an embedded device. The currently available scripts:
Each script is written to be processor/architecture-agnostic, but in some scripts, this requires a regular expression to address each architecture's specific-syntax. Before running the scripts, verify that the architecture of the firmware image to be analyzed is supported in the script.Please see Architecture Agnostic Structure of Scripts for more details. The IDAPython Embedded Toolkit only becomes more powerful, the more processors that are supported, so please submit a pull request as you add new processors.
To run a script, you must have IDA Pro 6.95 installed. Open the IDA database on which you'd like to run a script and then select File > Script File... and select the script to run.
Currently, the IDAPython Embedded Toolkit has only been tested on IDA Pro 6.95. Testing on IDA Pro 7.0 is currently in process.
If you completed the default installation for IDA Pro, then IDAPython should be installed. You can verify by checking your IDA directory for a Python/ folder. If that is there, IDAPython is installed.
Otherwise, install IDAPython per: https://github.com/idapython/src
Once IDAPython is installed, the IDAPython Embedded Toolkit scripts may be run by opening an IDA database and selecting File > Script file... from the upper menu. Then, select the script to run. Each script is run individually by selecting it through this process.
The scripts in the IDAPython Embedded Toolkit are written to be architecture and processor-agnostic. This is done by finding the common structure and processes that are not dependent on architecture-specific syntax. For the scripts that require processor-specific syntax (for example: Special Function Register Names or Instruction Syntax), regular expressions are used for each architecture. For more information on how to write regular expressions in Python: https://docs.python.org/2/library/re.html
Thanks to the contribution by @tmr232, each script auto-identifies the architecture in use and selects the correct set of
regular expressions using the IDAPython function:
processor_name = idaapi.get_inf_structure().procName
If the processor-in-use does not have regular expressions defined within the script, then the script will exit with an "Unsupported Processor Type" error. To make the script work, you simply need to add the required regular expression. To do this:
idaapi.get_inf_structure().procName The command will output a string. That string is the processor name.
Example of the Regular Expressions for Processor-Specific Syntax in define_code_functions.py
################### USER DEFINED VALUES ###################
# Enter a regular expression for how this architecture usually
# begins and ends functions. If the architecture does not
# dictate how to start or end a function use r".*" to allow
# for any instruction.
#
processor_name = idaapi.get_inf_structure().procName
if processor_name == '8051': # 8051 Architecture Prologue and Epilogue smart_prolog = re.compile(r".*")
smart_epilog = re.compile(r"reti{0,1}")
elif processor_name == 'PIC18Cxx': # PIC18 Architecture Prologue and Epilogue
smart_prolog = re.compile(r".*")
smart_epilog = re.compile(r"return 0")
elif processor_name == 'm32r': # Mitsubishi M32R Architecutre Prologue and Epilogue
smart_prolog = re.compile(r"push +lr")
smart_epilog = re.compile(r"jmp +lr.*")
elif processor_name == 'TMS32028': # Texas Instruments TMS320C28x
smart_prolog = re.compile(r".*")
smart_epilog = re.compile(r"lretr")
elif processor_name == 'AVR': # AVR
smart_prolog = re.compile(r"push +r")
smart_epilog = re.compile(r"reti{0,1}")
else:
print "[define_code_functions.py] UNSUPPORTED PROCESSOR. Processor = %s is unsupported. Exiting." % processor_name
raise NotImplementedError('Unsupported Processor Type.')
For example, let's say we have firmware where fp = 0x808000 and the majority of memory accesses are as offsets from fp. This script will calculate that the instruction is reading 0x80C114, create a cross-reference to that location, and replace the operand in the instruction with this calculated value as shown below.
ld R1, @(0x4114, fp) --> ld R1, @[0x80C114]
add3 R10, fp, 0x4147 --> add3 R10, fp, 0x4147; @[0x80C147]
define_code_functions.py -- Define Code and Functions This script scans an area of the database from the user input "start address" to "end address" defining the bytes as code and attempting to define functions from that code. The script is architecture agnostic by having the user define a regular expression for the "function prologue" and the "function epilogue" for the architecture being analyzed.
define_data_as_types.py -- Define a Block as Data Defines a segment of addresses as the user-specified data type (byte, word, or double word). The byte length for each of these types is architecture dependent, but generally: 1 byte = Byte 2 bytes = Word 4 bytes = Double Word This script with undefine all bytes in the range first which means if you previously had code or strings defined in the area, they will be overwritten as data.
make_strings.py -- Define a Block as Strings This script is used to search for and declare blocks of "Unexplored" bytes as ASCII strings. The user inserts the starting and ending address of the areas to be analyzed. The script then checks if each byte is an ASCII character value and ends with a defined "ending string character." In this example, the ending string characters are 0xD, 0xA, and 0x00. The script only checks "undefined or unexplored" values in the database. For example, if a string is currently defined as code, it will not identify this string. This is to protect previously defined values.
label_funcs_with_no_xrefs.py -- Label All Functions without Cross-References/ Valid Code Paths This script identifies what could be "dead code". It checks each function for cross-references to the function in question. If there are none, it adds the prefix "noXrefs_" to the function name. This is very efficient for architectures that do not call functions indirectly.
identify_port_use_locations.py -- Find All CPU Port Usage Identifies all code using the CPU's ports and records the address and instruction in the identified file. There is the option to annotate each function that accesses a CPU port/pin with a prefix stating that it's using the specific port/pin.
find_mem_acceses.py -- Identify Memory Accesses Identifies the memory accesses used in the code. When a memory access is identified based on the user contributed regular expression, this script completes three different actions to help with the static analysis:
Copyright 2017 The Johns Hopkins University Applied Physics Laboratory LLC All rights reserved. Permission is hereby granted, free of charge, to any person obtain
Selected from shared topics, language and repository description—not editorial ratings.
nicoleshundrin /
I reverse engineered and patched ELF binaries using IDA and Python to modify program logic, fix corrupted files, and inject new functionality like executing embedded shell commands.
identify_operand_locations.py -- Identify Instructions that Reference an Operand Identifies the instructions in the range start_addr to end_addr that reference the input operand (regex_operand). The addresses of all instructions where the operand is found are printed to the IDA output window and saved to a file.