SMDA is a minimalist recursive disassembler library that is optimized for accurate Control Flow Graph (CFG) recovery from memory dumps. It is based on Capstone and currently supports x86/x64 Intel machine code. As input, arbitrary memory dumps (ideally with known base address) can be processed. The output is a collection of functions, basic blocks, and instructions with their respective edges between blocks and functions (in/out). Optionally, references to the Windows API can be inferred by using the ApiScout method.
With version 1.2.0, we have finally simplified things by moving to PyPI!
So installation now is as easy as:
$ pip install smda
A typical workflow using SMDA could like this:
>>> from smda.Disassembler import Disassembler >>> disassembler = Disassembler() >>> report = disassembler.disassembleFile("/bin/cat") >>> print(report) 0.777s -> (architecture: intel.64bit, base_addr: 0x00000000): 143 functions >>> for fn in report.getFunctions(): ... print(fn) ... for ins in fn.getInstructions(): ... print(ins) ... 0x00001720: (-> 1, 1->) 3 blocks, 7 instructions. 0x00001720: ( 4883ec08) - sub rsp, 8 0x00001724: (488b05bd682000) - mov rax, qword ptr [rip + 0x2068bd] 0x0000172b: ( 4885c0) - test rax, rax 0x0000172e: ( 7402) - je 0x1732 0x00001730: ( ffd0) - call rax 0x00001732: ( 4883c408) - add rsp, 8 0x00001736: ( c3) - ret 0x00001ad0: (-> 1, 4->) 1 blocks, 12 instructions. [...] >>> json_report = report.toDict()
There is also a demo script:
- analyze.py -- example usage: perform disassembly on a file or memory dump and optionally store results in JSON to a given output path.
The code should be fully compatible with Python 2 and 3. Further explanation on the innerworkings follow in separate publications but will be referenced here.
To take full advantage of SMDA's capabilities, make sure to (optionally) install:
- pdbparse (currently as fork from https://github.com/VPaulV/pdbparse to support Python3)
- 2022-08-12: v1.9.1 - Added support for parsing intel MachO files, including Go parsing.
- 2022-08-10: v1.8.5 - Fixed Go 64bit lavel parsing for v1.12 binaries.
- 2022-08-04: v1.8.4 - Dot export now uses hex formatted addresses in node names.
- 2022-08-03: v1.8.3 - Added support for producing a Dot export for SmdaFunction.
- 2022-08-01: v1.8.1 - Added support for parsing 32bit Go binaries as well.
- 2022-08-01: v1.8.0 - Added support for parsing Go function information (THX to @danielenders1!).
- 2022-07-22: v1.7.4 - Bugfix for marshalling of reports.
- 2022-07-08: v1.7.2 - Excluded overly aggressive tailcall recognition heuristics when processing Golang binaries.
- 2022-01-27: v1.7.0 - SmdaReports now contains a field
oep; SmdaFunctions now indicate
is_exportedand can provide CodeXrefs via
getCodeOutrefs(). (THX for the ideas: @mr-tz)
- 2021-08-20: v1.6.1 - Bugfix for alignment calculation of binary mappings. (THX: @williballenthin)
- 2021-08-19: v1.6.0 - Bugfix for truncation during ELF segment/section loading. API usage in ELF files is now resolved as well! (THX: @williballenthin)
- 2021-07-22: v1.5.19 - Now also parsing plt.sec structures to identify functions.
- 2021-06-07: v1.5.18 - Bugfix for struct.pack 8byte conversion using L instead Q (works on Linux, not on Windows).
- 2021-05-21: v1.5.17 - Bugfix for MemoryError when having LIEF try to process section data.
- 2021-05-20: v1.5.16 - Bugfix for formatting exceptions in report output (THX: @BonusPlay)
- 2021-05-18: v1.5.15 - Changed SHA256 in SmdaReports for unmapped files (was hash of memory-mapped image, not it's the input file's hash).
- 2021-04-07: v1.5.14 - Bugfix when processing Exception handler addresses as function entry point candidates (THX: capa team).
- 2021-01-20: v1.5.13 - Now using LIEF 0.11 and moved some print output to logging.
- 2021-01-15: v1.5.11 - Disassembler now offers
disassembleUnmappedBuffer(buffer)to load and process unmapped files directly from memory.
- 2020-12-11: v1.5.10 - Pinned LIEF to 0.10.1.
- 2020-12-01: v1.5.9 - Bugfix for section names. again. :)
- 2020-11-25: v1.5.6 - Now considering segments for content when ELF file has no sections (THX: @jcrussell).
- 2020-11-10: v1.5.5 - Unmarshalling setting default value for older reports.
- 2020-11-06: v1.5.4 - Minor fix on PE header parsing.
- 2020-11-05: v1.5.3 - Adjusted API thunk identification.
- 2020-10-30: v1.5.2 - One bugfix, also removed one print and reduced logging priority for the message in case the PDB parser module is missing.
- 2020-10-30: v1.5.1 - PE section table now contained in SmdaReport and added
- 2020-10-30: v1.4.12 - Bugfix in IndirectCallHandler (THX: @jcrussell).
- 2020-10-29: v1.4.11 - Populate exception handlers specified in PE64
.pdatasection as FEPs.
- 2020-10-29: v1.4.10 - Resolves 64bit API calls of style
call qword ptr [rip + offset]and more register-based API calls in general (THX: @jcrussell).
- 2020-10-29: v1.4.8 - Bugfixes. Verbose mode added (THX: @jcrussell).
- 2020-10-28: v1.4.6 - WinApiResolver now tries to resolve import by ordinal to their name if it is known - can be extended in the database of OrdinalHelper.
- 2020-10-28: v1.4.5 - Store the (mapped) buffer that was used to do disassembly along inside a SmdaReport - goal: enable to read strings/bytes at offsets at a later time.
- 2020-10-27: v1.4.4 - SmdaInstructions can now provide potential data references via
- 2020-10-27: v1.4.3 - SmdaInstructions can now on demand provide the detailed capstone instruction representation via
- 2020-10-27: v1.4.1 - 10-20% gain in processing speed by switching to
- 2020-10-26: v1.4.0 - Adding SmdaBasicBlock. Some convenience code to ease intgration with capa. (GeekWeek edition!)
- 2020-09-07: v1.3.11 - Summarizable DisassemblyStatistics.
- 2020-09-02: v1.3.10 - Fixed a bug where IDA Pro would crash when failing to demangle a function name while exporting a SMDA report.
- 2020-08-31: v1.3.9 - Adjusted Logging to avoid interference with other loggers configured outside of SMDA (THX: @BonusPlay).
- 2020-08-25: v1.3.6 - PicHash no longer stored as list.
- 2020-08-17: v1.3.5 - Bugfix for import parsing (ELF files).
- 2020-08-17: v1.3.4 - Recalculate PIC hash and nesting depth for older (v1.2.x) reports on import for compatibility.
- 2020-08-17: v1.3.3 - Added binary variation of
push ebp;mov ebp, espto list of default prologues and added exception handling for DominatorTrees (THX: @fxb).
- 2020-07-13: v1.3.2 - Use LIEF to parse Import Table for WinAPI usage data when processing unmapped files.
- 2020-07-13: v1.3.1 - Fixed
setup.pyto properly specify dependencies (THX: @BonusPlay).
- 2020-06-22: v1.3.0 - Added DominatorTree (Implementation by Armin Rigo) to calculate function nesting depth, shortened PIC hash to 8 byte, added some missing instructions for the InstructionEscaper, IdaInterface now demangles names.
- 2020-05-28: v1.2.15 - Bugfixes in IntelInstructionEscaper (handling of negative RIP-relative offsets), SmdaReport (datetime handling), PeFileParser (handling of empty pefile.sections); SCC calculation changed to iterative algorithm (using @bwesterb's implementation) and activated by default again.
- 2020-05-14: v1.2.10 - Bug in IdaInterface fixed.
- 2020-05-13: v1.2.9 - Bugfix in code gap identification in FunctionCandidateManager, SCC calculation is now optional.
- 2020-05-12: v1.2.7 - Added additional default metadata field "component" to SmdaReport.
- 2020-05-11: v1.2.6 - Export from IDA to SMDA data format is now supported (IDA 7.4).
- 2020-05-09: v1.2.5 - Fixed off-by-one that affected wildcarding of instructions (THX to Viviane Zwanger).
- 2020-05-04: v1.2.4 - Various minor fixes.
- 2020-04-29: v1.2.0 - Restructured config.py into smda/SmdaConfig.py to similfy usage and now available via PyPI! The smda/Disassembler.py now emits a report object (smda.common.SmdaReport) that allows direct (pythonic) interaction with the results - a JSON can still be easily generated by using toDict() on the report.
- 2020-04-28: v1.1.0 - Several improvements, including: x64 jump table handling, better data flow handling for calls using registers and tailcalls, extended list of common prologues based on much more groundtruth data, extended padding instruction list for gap function discovery, adjusted weights in candidate priority score, filtering code areas based on section tables, using exported symbols as candidates, new function output metadata: confidence score based on instruction mnemonic histogram, PIC hash based on escaped binary instruction sequence
- 2020-03-10: Various minor fixes and QoL improvements.
- 2019-08-20: IdaExporter is now handling failed instruction conversion via capstone properly.
- 2019-08-19: Minor fix for crashes caused by PDB parser.
- 2019-08-05: v1.0.3 - SMDA can now export reports from IDA Pro (requires capstone to be available for idapython).
- 2019-06-13: PDB symbols for functions are now resolved if given a PDB file using parameter "-d" (THX to @VPaulV).
- 2019-05-15: Fixed a bug in PE mapper where buffer would be shortened because of misinterpretation of section sizes.
- 2019-02-14: v1.0.2 - ELF symbols for functions are now resolved, if present in the file. Also "-m" parameter changed to "-p" to imply parsing instead of just mapping (THX: @VPaulV).
- 2018-12-12: all gcc jump table styles are now parsed correctly.
- 2018-11-26: Better handling of multibyte NOPs, ELF loader now provides base addr.
- 2018-09-28: We now have functional PE/ELF loaders.
- 2018-07-09: v1.0.1 - Performance improvements.
- 2018-07-01: v1.0.0 - Initial Release.
Thanks to Steffen Enders for his extensive contributions to this project! Thanks to Paul Hordiienko for adding symbol parsing support (ELF+PDB)! Thanks to Jonathan Crussell for helping me to beef up SMDA enough to make it a disassembler backend in capa! Thanks to Willi Ballenthin for improving handling of ELF files, including properly handling API usage! Thanks to Daniel Enders for his contributions to the parsing of the Golang function registry and label information! The project uses the implementation of Tarjan's Algorithm by Bas Westerbaan and the implementation of Lengauer-Tarjan's Algorithm for the DominatorTree by Armin Rigo.
Pull requests welcome! :)