diff options
Diffstat (limited to '2011/gsm-ensa2011/section-osmocombb.tex')
| -rw-r--r-- | 2011/gsm-ensa2011/section-osmocombb.tex | 296 |
1 files changed, 296 insertions, 0 deletions
diff --git a/2011/gsm-ensa2011/section-osmocombb.tex b/2011/gsm-ensa2011/section-osmocombb.tex new file mode 100644 index 0000000..a8f4cd1 --- /dev/null +++ b/2011/gsm-ensa2011/section-osmocombb.tex @@ -0,0 +1,296 @@ +\section{OsmocomBB Project} + +\begin{frame}{A GSM phone baseband processor} +\begin{itemize} + \item GSM protocol stack always runs in a so-called baseband processor (BP) + \item What is the baseband processor + \begin{itemize} + \item Typically ARM7 (2G/2.5G phones) or ARM9 (3G/3.5G phones) + \begin{itemize} + \item Runs some RTOS (often Nucleus, sometimes L4) + \item No memory protection between tasks + \end{itemize} + \item Some kind of DSP, model depends on vendor + \begin{itemize} + \item Runs the digital signal processing for the RF Layer 1 + \item Has hardware peripherals for A5 encryption + \end{itemize} + \end{itemize} + \item The software stack on the baseband processor + \begin{itemize} + \item is written in C and assembly + \item lacks any modern security features (stack protection, non-executable pages, address space randomization, ..) + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{A GSM Baseband Chipset} + \begin{figure}[h] + \centering + \includegraphics[width=100mm]{calypso-block.pdf} + \end{figure} + \url{http://laforge.gnumonks.org/papers/gsm_phone-anatomy-latest.pdf} +\end{frame} + +\begin{frame}{Requirements for GSM security analysis} +What do we need for protocol-level security analysis? +\begin{itemize} + \item A GSM MS-side baseband chipset under our control + \item A Layer1 that we can use to generate arbitrary L1 frames + \item A Layer2 protocol implementation that we can use + modify + \item A Layer3 protocol implementation that we can use + modify +\end{itemize} +None of those components existed, so we need to create them! +\end{frame} + +\begin{frame}{A GSM baseband under our control} +The two different DIY approaches +\begin{itemize} + \item Build something using generic components (DSP, CPU, ADC, FPGA) + \begin{itemize} + \item No reverse engineering required + \item A lot of work in hardware design + debugging + \item Hardware will be low-quantity and thus expensive + \end{itemize} + \item Build something using existing baseband chipset + \begin{itemize} + \item Reverse engineering or leaked documents required + \item Less work on the 'Layer 0' + \item Still, custom hardware in low quantity + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{A GSM baseband under our control} +Alternative 'lazy' approach +\begin{itemize} + \item Re-purpose existing mobile phone + \begin{itemize} + \item Hardware is known to be working + \item No prototyping, hardware revisions, etc. + \item Reverse engineering required + \item Hardware drivers need to be written + \item But: More time to focus on the actual job: Protocol software + \end{itemize} + \item Searching for suitable phones + \begin{itemize} + \item As cheap as possible + \item Readily available: Many people can play with it + \item As old/simple as possible to keep complexity low + \item Baseband chipset with lots of leaked information + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{Baseband chips with leaked information} +\begin{itemize} + \item Texas Instruments Calypso + \begin{itemize} + \item DBB Documentation on cryptome.org and other sites + \item ABB Documentation on Chinese phone developer websites + \item Source code of GSM stack / drivers was on sf.net (tsm30 project) + \item End of life, no new phones with Calypso since about 2008 + \item No cryptographic checks in bootloader + \end{itemize} + \item Mediatek MT622x chipsets + \begin{itemize} + \item Lots of Documentation on Chinese sites + \item SDK with binary-only GSM stack libraries on Chinese sites + \item 95 million produced/sold in Q1/2010 + \end{itemize} +\end{itemize} +Initial choice: TI Calypso (GSM stack source available) +\end{frame} + + +\subsection{OsmocomBB Introduction} + +\begin{frame}{OsmocomBB Introduction} +\begin{itemize} + \item Project was started only in January 2010 (9 months ago!) + \item Implementing a GSM baseband software from scratch + \item This includes + \begin{itemize} + \item GSM MS-side protocol stack from Layer 1 through Layer 3 + \item Hardware drivers for GSM Baseband chipset + \item Simple User Interface on the phone itself + \item Verbose User Interface on the PC + \end{itemize} + \item Note about the strange project name + \begin{itemize} + \item Osmocom = Open Source MObile COMmunication + \item BB = Base Band + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Software Architecture} +\begin{itemize} + \item Reuse code from OpenBSC where possible (libosmocore) + \begin{itemize} + \item We build libosmocore both for phone firmware and PC + \end{itemize} + \item Initially run as little software in the phone + \begin{itemize} + \item Debugging code on your host PC is so much easier + \item You have much more screen real-estate + \item Hardware drivers and Layer1 run in the phone + \item Layer2, 3 and actual phone application / MMI on PC + \item Later, L2 and L3 can me moved to the phone + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Software Interfaces} +\begin{itemize} + \item Interface between Layer1 and Layer2 called L1CTL + \begin{itemize} + \item Fully custom protocol as there is no standard + \item Implemented as message based protocol over Sercomm/HDLC/RS232 + \end{itemize} + \item Interface between Layer2 and Layer3 called RSLms + \begin{itemize} + \item In the GSM network, Um Layer2 terminates at the BTS but is controlled by the BSC + \item Reuse this GSM 08.58 Radio Signalling Link + \item Extend it where needed for the MS case + \end{itemize} +\end{itemize} +\end{frame} + +\subsection{OsmocomBB Software} + +\begin{frame}{OsmocomBB Target Firmware} +\begin{itemize} + \item Firmware includes software like + \begin{itemize} + \item Drivers for the Ti Calypso Digital Baseband (DBB) + \item Drivers for the Ti Iota TWL3025 Analog Baseband (ABB) + \item Drivers for the Ti Rita TRF6151 RF Transceiver + \item Drivers for the LCD/LCM of a number of phones + \item CFI flash driver for NOR flash + \item GSM Layer1 synchronous/asynchronous part + \item Sercomm - A HDLC based multiplexer for the RS232 to host PC + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Host Software} +\begin{itemize} + \item Current working name: layer23 + \item Includes + \begin{itemize} + \item Layer 1 Control (L1CTL) protocol API + \item GSM Layer2 implementation (LAPDm) + \item GSM Layer3 implementation (RR/MM/CC) + \item GSM Cell (re)selection + \item SIM Card emulation + \item Supports various 'apps' depending on purpose + \end{itemize} +\end{itemize} +\end{frame} + +\subsection{OsmocomBB Hardware Support} + +\begin{frame}{OsmocomBB Supported Hardware} +\begin{itemize} + \item Baseband Chipsets + \begin{itemize} + \item TI Calypso/Iota/Rita + \item Some early research being done on Mediatek (MTK) MT622x + \end{itemize} + \item Actual Phones + \begin{itemize} + \item Compal/Motorola C11x, C12x, C13x, C14x and C15x models + \item Most development/testing on C123 and C155 + \item GSM modem part of Openmoko Neo1973 and Freerunner + \end{itemize} + \item All those phones are simple feature phones built on a ARM7TDMI based DBB +\end{itemize} +\end{frame} + +\begin{frame}{The Motorola/Compal C123} + \begin{figure}[h] + \centering + \includegraphics[width=100mm]{c123_pcb.jpg} + \end{figure} +\end{frame} + + +\subsection{OsmocomBB Project Status} + +\begin{frame}{OsmocomBB Project Status: Working} +\begin{itemize} + \item Hardware Drivers for Calypso/Iota/Rita very complete + \item Drivers for Audio/Voice signal path + \item Layer1 + \begin{itemize} + \item Power measurements + \item Carrier/bit/TDMA synchronization + \item Receive and transmit of normal bursts on SDCCH + \item Transmit of RACH bursts + \item Automatic Rx gain control (AGC) + \item Frequency Hopping + \end{itemize} + \item Layer2 UI/SABM/UA frames and ABM mode + \item Layer3 Messages for RR / MM / CC + \item Cell (re)selection according GSM 03.22 +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Project Status: Working (2/2)} +OsmocomBB can now do GSM Voice calls (since 08/2010) +\begin{itemize} + \item Very Early Assignment + Late Assignment + \item A3/A8 Authentication of SIM + \item A5/1 + A5/2 Encryption + \item Full Rate (FR) and Enhanced Full Rate (EFR) codec +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Project Status: Not working} +\begin{itemize} + \item Layer1 + \begin{itemize} + \item Automatic Tx power control (APC) + \item Neighbor Cell Measurements (WIP) + \item In-call hand-over to other cells (WIP) + \end{itemize} + \item Actual UI on the phone + \item Circuit Switched Data (CSD) calls + \item GPRS (packet data) + \item No Type Approval for the stack! +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB Project Status: Executive Summary} +\begin{itemize} + \item We can establish control/signalling channels to both hopping and non-hopping GSM cells + \begin{itemize} + \item Control over synthesizer means we can even go to GSM-R band + \end{itemize} + \item We can send arbitrary data on those control channels + \begin{itemize} + \item RR messages to BSC + \item MM/CC messages to MSC + \item SMS messages to MSC/SMSC + \end{itemize} + \item TCH (Traffic Channel) support for voice calls + \begin{itemize} + \item Has been used on real networks for 30+ minute calls! + \end{itemize} +\end{itemize} +\end{frame} + +\begin{frame}{OsmocomBB use cases} +OsmocomBB can be used today for +\begin{itemize} + \item practical lab exercises in education on any level of GSM, +from the radio modem through the protocol stack + \item applied research in GSM protocols and GSM security + \item penetration testing of GSM operator equipment + \item measurement and exploration of real operator networks +\end{itemize} +With (your?) help, we can turn it into an actual mobile phone for +regular users, i.e. bringing the freedom of Free Software into one of +the most closed areas of computing. +\end{frame} |