Hashing the $#!+ out of firmware
summary
Project extracts firmware with binwalk and FMK, unpacks every file, computes cryptographic and fuzzy hashes, and builds a searchable database from 1.8 TB of firmware. Comparison of file hashes reveals unchanged core files, reused code across models, and surprising scope: a CVE listing of ~10 affected images yields 37 other affected firmwares in the dataset. Fuzzy-hash searching is useful but slow, a GUI for visualizing evolution is unfinished, and full real-world testing requires many physical devices.
AI-written summary · about
chapters (17)
- 0:00Intro
- 0:12Outline
- 1:12What is firmware?
- 4:43I think that is a fine change log
- 6:22Now we are talking!
- 8:34An example with cakes
- 12:52A better way to investigate firmware
- 13:09TP-Link MR3020
- 17:46Vulnerability info
- 19:04Number of CVEs listed per year
- 19:48Number of firmware related
- 20:52Some exploration
- 22:13CVE-2009-0322 (266)
- 24:12Go big!
- 25:29CVE-2018-9032
- 28:03Is this great?
- 30:16In summary
related talks
45mFirmware.RE: Firmware Unpacking, Analysis and Vulnerability-Discovery as a Service
01h 08mAutomated Dynamic Fireware Analysis At Scale: A Case Study on Embedded Web Interfaces
37mThe Firmware Supply-Chain Security Is Broken: Can We Fix It?
33mIdentifying Multi-Binary Vulnerabilities in Embedded Firmware at Scale
20mSoK: Insecurity of Cellular Basebands
30mWhy Open Source Firmware is Important
33mEmbedded Device Exploitation 101: An Introduction to Firmware Hacking
41mfirmWar: An Imminent Threat to the Foundation of Computing