Safe multi-rail control for sparsely encoded signals
Secure multi-rail control for sparsely encoded signals using multiple rails and FSMs with check modules addresses the inadequacy of existing access prevention methods, enhancing IC security against fault injection attacks without compromising performance or scalability.
Patent Information
- Application Number
- JP2024532818
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2021-12-03
- Filing Date
- 2022-12-02
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-12-02
AI Technical Summary
Existing approaches to prevent unauthorized access to sensitive information in electronic devices are inadequate against evolving software, hardware, and wireless attacks, particularly due to the reduction in effectiveness of security measures by modern circuit synthesis tools.
Implementing secure multi-rail control for sparsely encoded signals using multiple rails and finite state machines (FSMs) to transmit critical signals, ensuring a minimum Hamming distance and incorporating check modules to detect fault injection attacks.
Enhances the security of integrated circuits against fault injection attacks by making it difficult for attackers to modify critical signals, while maintaining compatibility with various ICs and avoiding the need to disable circuit synthesis tools or use custom circuits.
Smart Images

Figure 2024542751000001_ABST
Abstract
Description
[Background technology]
[0001] Electronic devices often contain at least one integrated circuit (IC) that enables various functions on the electronic device. To securely perform some of these functions, the electronic device may store or operate using secret information restricted to certain access levels within the IC. An attacker may attempt to unauthorizedly access this secret information of the electronic device to impair or modify the operation of the IC. However, existing approaches to prevent unauthorized access are often insufficient to counter evolving software, hardware, and wireless attacks employed by malicious actors attempting to access the secret information of electronic devices. Summary of the Invention
[0002] This specification discloses techniques, apparatus, and systems for secure multi-rail control for sparsely encoded signals. Integrated circuits (ICs) may transmit various signals to manage interactions between circuit components of the IC. These critical signals, when modified, can cause the IC to perform differently than intended and potentially bypass security measures, making them a common target for malicious attacks. While various strategies may be used to protect against these attacks, modern circuit synthesis tools reduce or eliminate the effectiveness of these security measures, making ICs vulnerable to manipulation. In contrast, the secure multi-rail control for sparsely encoded signals described herein utilizes multiple rails to transmit sparsely encoded critical signals. Each rail may be controlled by a separate finite state machine (FSM) to reduce vulnerabilities that may arise from circuit synthesis or provide a tunable solution that may be exploited in different ways based on the implementation, providing suitability for various ICs.
[0003] An IC is described that includes a first component and a second component. The first component may include a plurality of FSMs. In aspects, the first component and the second component are coupled to each other by a plurality of separate data lines. A sparsely encoded signal having a minimum Hamming distance is used to communicate between the first component and the second component. Each FSM of the plurality of FSMs may be coupled to the second component by a respective data line of the plurality of data lines. Each of the plurality of data lines may communicate a single bit of the sparsely encoded signal between the first component and the second component. In this manner, the IC may be protected from fault injection attacks and the IC may withstand undesired circuit synthesis. The second component may, for example, comprise a second plurality of FSMs. In embodiments, the first component may be any component of the IC and the second component may be any different component of the IC. The described systems and techniques may be utilized at any point where sensitive operations are occurring on a hardware device, for example, any process involving access control or privilege level operations (e.g., power, register access, system reset). In aspects, the described techniques may be utilized between any two components of an IC that perform such sensitive operations. In one embodiment, the described secure multi-rail control logic is implemented within an encryption IP block of the IC (e.g., between two components of the encryption IP block). As yet another example, the second component may be a multiplexer device or may include storage registers, or configuration and setting registers.
[0004] This Summary is provided to introduce simplified concepts for implementing safe multi-rail control for sparsely encoded signals. The simplified concepts are further described in the Detailed Description below. This Summary is not intended to identify essential features of the claimed invention, nor is it intended for use in determining the scope of the claimed invention.
[0005] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The details of one or more aspects of the described techniques, apparatus and systems for secure multi-rail control for sparsely encoded signals are set forth below. Where the same reference numbers are used in different instances in the description and figures, similar elements are indicated. [Brief description of the drawings]
[0006] [Figure 1] FIG. 1 illustrates an example operating environment including electronic devices that may implement safe multi-rail control for sparsely encoded signals. [Diagram 2] FIG. 1 illustrates an example implementation of an example security circuit including multiple circuit components that can be implemented to support aspects of secure multi-rail control for sparsely encoded signals. [Diagram 3] FIG. 1 illustrates an example cryptographic intellectual property (IP) block capable of implementing secure multi-rail control for sparsely encoded signals in accordance with one or more aspects. [Figure 4] FIG. 1 illustrates an example logic of an AND operation on sparsely encoded signals that is vulnerable to attacks by circuit composition. [Diagram 5] FIG. 1 illustrates example logic for an AND operation using safe multi-rail control for sparsely encoded signals, in accordance with one or more aspects. [Figure 6] 6A and 6B illustrate example logic for implementing secure multi-rail control for sparsely encoded signals in accordance with one or more aspects. [Figure 7] FIG. 1 illustrates an example method for transmitting a sparsely encoded signal using multiple rails. [Figure 8] FIG. 1 illustrates an example electronic device that can implement safe multi-rail control for sparsely encoded signals. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0007] In general, there are various methods that an attacker may use to access privileged information (e.g., keys) held by an integrated circuit (IC). For example, a fault injection attack may be used to inject faults into signals transmitted between different circuit components of an integrated circuit. Various methods may be used to inject faults into an IC. For example, faults may be injected into an IC by using a high-power laser, exposing the IC to electromagnetic interference, adjusting the IC's clock, changing the IC's power supply, or any combination of these methods. In some implementations, a high-power laser may be used to inject faults with high precision. For example, a laser may be shone on the surface of a transistor to flip the transistor and cause it to transmit a different value of a signal.
[0008] An IC may use various signals to control the operation of the circuit, and more specifically, to manage the interactions between various circuit components. In this way, these critical signals may directly affect or control which operations are performed by the IC and when the operations are performed. If a fault is injected along with (or in place of) these signals, the IC may perform unintended operations and possibly bypass security measures (e.g., secure boot / root of trust processing). For example, a fault may be injected into a handshake to cause the IC to prematurely release an intermediate value in a process that is subject to circuit operations. In this way, an attacker may obtain or access privileged information (e.g., a private key) about the circuit by bypassing operations that have not yet been performed when the value is prematurely released. In other implementations, the value itself may indicate privileged information held by the integrated circuit, for example, the value may represent a private key used by the IC.
[0009] Various techniques may be employed to reduce an IC's vulnerability to fault injection attacks; however, improvements in modern circuit synthesis tools may introduce complications in implementing these countermeasures. Specifically, modern circuit synthesis tools are designed to identify and reduce redundancies in circuit operation to improve IC performance. Some security countermeasures may exploit redundancies to increase the complexity of successfully implementing a malicious attack on an IC (e.g., successfully releasing privileged information). Thus, when employed within an encrypted IC, these circuit synthesis tools may optimize operation and produce more efficient circuits that do not contain redundancies intentionally designed for security purposes, thereby eliminating important security features. Overcoming these issues may require disabling the circuit synthesis tool in the encrypted IC, which may lead to reduced performance. Alternatively, the encrypted IC may require creating custom multi-bit gates, which may increase cost and manufacturing complexity, reduce scalability, and reduce adaptability.
[0010] In contrast to other solutions, the techniques described below utilize sparse encoding for signals (e.g., critical signals) transmitted between two circuit components. For example, each signal may be defined to have a predetermined minimum Hamming distance (e.g., the number of bits that differ between any two valid solutions). Specifically, if a value is representable by a 1-bit representation (e.g., 1'b0 for LOW (e.g., defined by transmitting a low voltage value) or 1'b1 for HIGH (e.g., defined by transmitting a high voltage value)) and the predetermined minimum Hamming distance is 3, then a LOW value may be represented as 3'b010 and a HIGH value may be represented as 3'b101. In this way, an attacker may need to instantly (or within a short time period) flip multiple bits to successfully inject a fault that creates a different valid solution.
[0011] Generally, in information theory, the Hamming distance between two data items of equal length (e.g., 32 bits) is the number of positions where corresponding symbols or bit positions in the data items differ. For example, given two integers, each represented by 16 bits, where each integer is a valid solution, the Hamming distance between these two integers could be the number of positions where corresponding bits differ between the two 16-bit data structures.
[0012] In embodiments, the HIGH and LOW values may be represented as a mixture of 1 and 0 bits to increase the difficulty of success in a fault injection attack. In general, it may be easier to invert all bits in a signal to a single value (HIGH or LOW, e.g., 1 or 0). Thus, at least one HIGH bit and at least one LOW bit may be used to represent each sparsely encoded signal. To transmit sparsely encoded signals, ICs described below may utilize multiple rails (e.g., dedicated data lines that can be used to transmit signals between a first circuit component and a second circuit component), each carrying a single bit of the sparsely encoded signal. Due to the mixture of values, the multiple rails may include both positive and negative rails. This is further described in connection with FIG. 3.
[0013] In some implementations, the multiple rails may transmit critical signals between two components of an IC. As a specific example, the multiple rails may communicatively couple a finite state machine (FSM) (e.g., in a first component of an IC) to another circuit component to transmit control signals (e.g., initialization, acknowledgement, termination) between the FSM and the circuit component. Some control signals are single-bit signals that may be represented as a single HIGH value or a single LOW value. In aspects, sparsely encoded signals (e.g., sparse encoding) may include signaling (e.g., transmitted logical values) having set encoded values with non-adjacent codewords and / or codewords having a Hamming distance of at least two. For example, a valid codeword transmitted as bits on separate data lines is a codeword that cannot be obtained from another codeword by glitching or flipping one bit value during an attack.
[0014] Due to the redundancy used to meet the minimum Hamming distance, circuit synthesis tools may reduce the effectiveness of the sparse encoding of these critical signals (discussed in more detail in connection with FIG. 4). Thus, to ensure that the described technique is resistant to circuit optimization, the first circuit component (described above) may include as many FSMs as there are rails. Thus, each FSM in the first circuit component may be communicatively coupled to a respective single rail of the multiple rails and used to drive a single bit of the sparsely encoded signal. In this manner, the ICs described below may be resistant to circuit optimization via multiple FSMs / rails without defeating circuit synthesis tools or requiring custom circuitry.
[0015] In yet another aspect, each of the multiple rails may be connected to a check module that may be used to compare the transmitted signal to multiple valid states, thereby allowing the IC to check for possible fault injection attempts. If the check module determines that the transmitted signal does not match any of the predefined valid states, the check module may signal an error and initiate an error handling process. Thus, the techniques, systems and apparatus for secure multi-rail control for sparsely encoded signals described below may provide an economical, scalable and adaptable solution for improving the security of ICs against fault injection attacks.
[0016] 1 illustrates an exemplary environment 100 including devices 102 capable of implementing aspects of secure multi-rail control for sparsely encoded signals. The devices 102 may be implemented as any suitable devices, some of which are shown as a smartphone 102-1, a tablet computer 102-2, a laptop computer 102-3, a game console 102-4, a desktop computer 102-5, a server computer 102-6, a wearable computing device 102-7 (e.g., a smart watch), and a broadband router 102-8 (e.g., a mobile hotspot). Although not shown, the devices 102 may be implemented as any of a mobile station (e.g., a fixed or mobile STA), a mobile communication device, a client device, a user equipment, a mobile phone, an entertainment device, a mobile game console, a personal media device, a media playback device, a health monitoring device, a drone, a camera, an Internet home appliance capable of wireless Internet access and browsing, an IoT device, and / or other types of electronic devices. Device 102 may provide other functionality or may include components or interfaces that have been omitted from FIG. 1 for clarity or visual simplicity.
[0017] The device 102 includes an integrated circuit (IC 104), which utilizes one or more processors 106 and a computer-readable medium 108, which may include a memory medium or a storage medium. The processor 106 may be implemented as a general-purpose processor (e.g., a multi-core central processing unit (CPU) or an application processor (AP)), an application specific integrated circuit (ASIC), a graphics processing unit (GPU), or a system-on-chip (SoC) integrated with other components of the device 102. The computer-readable medium 108 may include any suitable type of memory medium or storage medium, examples of which include read-only memory (ROM), programmable ROM (PROM), random access memory (RAM), dynamic RAM (DRAM), static RAM (SRAM), or flash memory. In the context of this description, the computer-readable medium 108 of the device 102 is implemented as at least one hardware-based or physical storage device that does not include a transitory signal or carrier wave. Applications, firmware, and / or operating systems (not shown) of device 102 may be embodied on computer-readable medium 108 as processor-executable instructions and executed by processor 106 to provide various functions described herein. Computer-readable medium 108 may also store device data 110, examples of which include user data or user media accessible through the applications, firmware, or operating system of device 102.
[0018] In this embodiment, the IC 104 has a security circuit 112. The device 102, the IC 104, or the security circuit 112 may implement a secure cryptographic processor. The security circuit 112 may be implemented using one or more circuit components 114, for example, circuit components 114-1 to 114-n. The circuit components 114 may be organized to perform any number of operations to enable the functionality of the device 102. An example of a circuit component includes a processor and multiple functional components, as described in FIG. 2. The security circuit 112 may be realized, for example, as a protected enclave, a trusted chip platform, a hardware-based root of trust (RoT) chip (e.g., a silicon RoT), etc. Regardless of how or where the security circuit 112 is incorporated into the electronic device, the security circuit 112 may resist many different types of attacks.
[0019] The security circuit 112 may include at least one encryption intellectual property (IP) block 116 that performs various encryption operations of the IC 104. In some implementations, the at least one encryption IP block 116 may be implemented in an Advanced Encryption Security (AES) unit. In aspects, the AES unit or encryption IP block 116 may implement one or more encryption operations, such as data encryption or data decryption. The data may be, for example, image data (e.g., captured by a camera or video camera) or audio data (e.g., captured by a microphone). In one embodiment, pseudorandom or truly random numerical values may be used to generate a private / public key pair. The public key is transmitted from the device 102 via the transceiver 126 and / or the I / O port 128. Data (e.g., image / audio data) that is encrypted using the public key and received by the device 102 via the transceiver 126 and / or the I / O port 128 may be decrypted using the private key. Alternatively or additionally, data such as image data or audio data (e.g., data acquired using sensor 130) may be hashed using the private key to generate a hash digest, and the data and hash digest may be transmitted from device 102 via transceiver 126 and / or I / O port 128 so that the hash digest acts as a signature of the data that can be verified by a third party using the public key.
[0020] In another embodiment, the encryption IP block 116 may perform Boolean masking and remasking operations on data operated on by the IC 104 based on pseudorandomness provided by a pseudorandom number generator or a true random number generator. The encryption IP block 116 may handle device secrets used to ensure proper use and performance of the IC 104, such as, for example, encryption keys, or other privileged information. To protect the privileged information handled by the encryption IP block 116, cryptographically secure operations may be performed to protect against various attacks that could release device secrets or information related to the IC 104.
[0021] In aspects, the encryption IP block 116 may be configured to perform encrypted and secure operations according to secure multi-rail control for sparsely encoded signals as described herein. For example, the encryption IP block 116 may have one or more finite state machine blocks (FSMs 118) that control the operation of the encryption processor within the IC 104. As further described in connection with FIG. 3, the encryption IP block 116 may have a master control FSM block that manages the interaction of the encryption IP block 116 and the processor 106. For example, the master control FSM may send signals that control when values are written to registers accessed by the processor 106. The FSMs 118 may also include a core control FSM block (shown in FIG. 3) that coordinates the internal operation of the encryption IP block 116.
[0022] In some implementations, the FSM 118 may be implemented using multiple copies of the same FSM. For example, the main control FSM or the core control FSM may be implemented as a set of identical (or similar) FSMs. In this manner, each bit of a sparsely encoded signal having a multi-bit representation may be handled by an individual FSM. Thus, each respective FSM of a first component (e.g., the main control FSM) may be communicatively coupled to a respective FSM of a second component (e.g., the core control FSM) and used to communicate a single bit of a sparsely encoded signal. In this manner, the encryption IP block 116, or the IC 104 in general, may withstand undesirable circuit synthesis that optimizes and skips security components.
[0023] The encryption IP block 116 may include multiple rails 120 between different components (e.g., FSM 118) of the encryption IP block 116. In aspects, the rails 120 may be used to transmit sparsely encoded signals between different components of the encryption IP block 116 or IC 104. When communicating between a first circuit component and an FSM 118 implemented through multiple copies of an FSM (e.g., main or core FSM of the encryption IP block 116), each rail of the rails 120 may couple a single FSM from the first component to a second component. As such, each FSM and each rail 120 may communicate a single bit of the sparsely encoded signal, thus overcoming issues associated with circuit synthesis tools.
[0024] As mentioned above, a sparsely encoded signal may be represented by a combination of HIGH and LOW bits. Thus, rail 120 may have both a positive rail and a negative rail. In some implementations, the positive rail may carry a HIGH signal and the negative rail may carry a LOW signal. In some other implementations, the signal carried by a particular rail (e.g., the negative rail) may be inverted to carry a desired signal state (e.g., a HIGH signal). In aspects, the negative rail, or a component (e.g., an FSM) transmitting one or more bits of a signal over the negative rail, may include an inverter that inverts the output of the value communicated by the first component. This also applies to the positive rail, or a component transmitting one or more bits of a signal over the positive rail. For example, if an FSM transmits a HIGH signal (e.g., a 1), the negative rail may transmit a LOW signal and an inverter of that FSM may invert that LOW signal to a HIGH signal. Thus, the negative rail may transmit a LOW signal as a HIGH signal and the positive rail may transmit a HIGH signal as a LOW signal. The control logic of the negative rail is described in more detail in conjunction with FIG. 6.
[0025] In some implementations, rail 120 may be coupled to at least one check module (not shown) to ensure that the sparsely encoded signals transmitted over rail 120 represent valid signal states (e.g., predetermined HIGH and LOW signal representations). The check module may compare the transmitted signals to a set of predetermined valid states of critical signals stored in the check module (or elsewhere in computer-readable medium 108). If the check module determines that the transmitted signals do not correspond to valid signal states (e.g., if a fault has been injected into the signal), it may pass an error up the chain and handle it appropriately (as further described in connection with FIG. 2).
[0026] As shown, the security circuitry 112 is coupled to an interconnect 122. The interconnect 122 may be implemented using, for example, a bus, switching fabric, or bus network that allows the various circuit components to communicate. Each circuit component 114 may be directly or indirectly coupled to the interconnect 122. Although described in conjunction with the encryption IP block 116, any portion of the interconnect 122 may utilize multiple rails (e.g., buses or wires) to enable the use of sparsely encoded signals. In other words, multi-rail control for sparsely encoded signals may be used to communicate signals between any two components of the IC 104, not just components within the encryption IP block 116. In some embodiments, the interconnect 122 may transmit all signals using multiple rails and sparse encoding. In other embodiments, multi-rail control for sparsely encoded signals may be implemented only for critical signals (e.g., signals that directly control the operation of the IC 104 or signals used to facilitate operations involving privileged data, such as private keys).
[0027] The device 102 may also include a display 124, a transceiver 126, input / output ports (I / O ports 128), and / or sensors 130. The display 124 may be operatively coupled to one of the processors 106 (e.g., a graphics processing unit (GPU)) and configured to graphically display the operating system or applications of the device 102. The transceiver 126 may be configured to enable wired or wireless communication of data (e.g., device data 110) across a wired or wireless network according to any suitable communication protocol. The I / O ports 128 of the device 102 may include universal serial bus (USB) ports, coaxial cable ports, and other serial or parallel connectors (including internal connectors) useful for coupling electronic devices to various components, peripherals, or accessories (such as a keyboard, microphone, or camera).
[0028] The device 102 also includes sensors 130 that enable the device 102 to sense various properties, changes, stimuli, or characteristics of the environment in which the device 102 operates. For example, the sensors 130 may include various motion sensors, ambient light sensors, acoustic sensors, capacitance sensors, infrared sensors, temperature sensors, radar sensors, or magnetic sensors. Alternatively, or in addition, the sensors 130 may enable the device 102 to interact with or accept input from a user, such as by touch or proximity sensing.
[0029] 2 illustrates an example implementation 200 of an example security circuit 112 that includes a number of circuit components (including a number of example circuit components 114) that can be implemented to support aspects of secure multi-rail control for sparsely encoded signals. As shown, the security circuit 112 includes a processor 106 coupled to an interconnect 122. Each processor 106, a number of memories, and a number of other circuit components 114 may be directly or indirectly coupled to the interconnect 122. In aspects, the components of FIG. 2 may be embodied as a secure computing platform or a secure system-on-chip that implements a RoT and / or other secure cryptographic functions.
[0030] In an exemplary embodiment, the memories may include a read only memory (ROM 202), a static random access memory (SRAM 204), and a flash memory 206. In aspects, the ROM 202, the SRAM 204, or the flash memory 206 may be implemented in the security circuit 112 or in the computer-readable medium 108. The components 114 may include an alert handler 208, an AES unit (not shown), an internal or external cryptographic IP block 116 of the AES unit, a hash-based message authentication code (HMAC) engine (HMAC engine 210), and a serial peripheral interface (SPI) device (SPI device 212). The components 114 may also include a universal asynchronous receiver / transmitter (UART) unit (UART unit 214), a general purpose input / output (GPIO) interface (GPIO interface 216), a pin multiplexer (pin MUX 218), and a pad controller 220. The plurality of components 114 may further include a random number generator (RNG 222) and a timer 224. Additionally, the components 114 may include any memory, as shown in Figure 2. Although particular examples of memory and other components 114 are shown in Figure 2 or described herein, a given implementation of the security circuit 112 may include more, fewer, and / or different instances of processors, controllers, memory, modules, or peripheral devices (including duplicates thereof).
[0031] The illustrated circuit components 114 may operate synchronously based on one or more clock signals. Although not shown in FIG. 2, the security circuit 112 may include at least one clock generator for generating clock signals or may include a reset circuit for resetting one or more individual components independently of one another, resetting multiple components jointly, or resetting the entire IC chip. Alternatively, the security circuit 112 may receive at least one clock or reset signal from a source external to the security circuit 112, which may or may not be on a separate chip. One or more separate components 114 may each operate in an individual clock domain. For example, the circuit components 114 may be synchronized to a clock local to the respective component. Components 114 in different clock domains may operate or communicate asynchronously with respect to one another.
[0032] Exemplary implementations of the illustrated components 114 are described below. The processor 106 may be realized as the "main," "central," or "core" processor of the security circuit 112. The processor 106 may be implemented with a 32-bit in-order reduced instruction set computing (RISC) core with a multi-stage pipeline, by way of example only. For example, using RISC-V features, the processor 106 may implement an M (machine) mode and a U (user) mode. Activating a reset pin (not shown) (e.g., by deassertion of an active-low reset pin) may cause the processor 106 to exit reset and begin executing code at a reset vector. The reset vector may begin in ROM 202, thereby enabling the code before jumping to an emulated embedded flash (eFlash). In other words, it is expected that the code will be instantiated in the eFlash before the reset is released. In some cases, the reset of the entire security circuit 112 may be made asynchronous active-low by compatibility specifications to aid in interoperability between various circuit components. A reset may be generated by the alert handler 208 as a security measure, may be generated by a watchdog timer, etc. A reset signal may also be sent to other circuit components (such as one of the memories or one of the other components 114).
[0033] A debug module 226 (DM) and an interrupt controller 228 (ItC) may be coupled to the processor 106 or may be compatible with either of them. The debug module 226 provides debug access to the processor 106. The logic of the debug module 226 allows the processor 106 to enter a debug mode by interfacing with specific pins of the IC, providing the ability to inject code into the device (e.g., by emulating instructions) or memory. The interrupt controller 228 may be located close to the processor 106. The interrupt controller 228 may receive a vector of interrupt sources from within the security circuit 112. The interrupt controller 228 may also assign a level and priority to the interrupt before forwarding the interrupt to the processor 106 for handling.
[0034] The processor 106 may provide any desired level of performance or include any internal circuit components. For example, the processor 106 may include at least one arithmetic logic unit (ALU) (e.g., including an "extra" ALU to calculate branch targets and eliminate latency cycles of taken conditional branches) and multiple pipeline stages. With multiple pipeline stages, the pipeline may perform register writeback to reduce latency cycles from loads and stores and prevent pipeline stalls where the response to a load or store is available the cycle after the request. The processor 106 may implement a single-cycle multiplier or generate imprecise exceptions in error response to a store, allowing the processor to continue execution beyond the store without waiting for a response. Although not shown, the processor 106 specifically, or the security circuitry 112 generally, may include an instruction cache to provide single-cycle access times for instructions.
[0035] In the illustrated example 200, the components of the security circuit 112 include or have access to three memory address spaces for instructions and data. The ROM 202 is the target of the processor 106 after the release of reset. The ROM 202 has hard-coded instructions to perform a subset of platform checks before checking the next stage code. The next stage code (e.g., a boot loader stored in eFlash memory) may be the first piece of code that is not hard-coded into the silicon of the device. This next stage code is therefore signature checked for integrity to improve security. The ROM 202 may perform this signature check by implementing a Rivest-Shamir-Adelman Check (RSA Check) algorithm on the entire contents of the boot loader.
[0036] The flash memory 206 can be implemented as an eFlash memory for code storage. The eFlash can house the boot loader mentioned herein, as well as the operating system and applications layered on top. The SPI device 212 can be used to bulk load the eFlash memory. Also, the debug module 226 can be used to load code. The SRAM 204 can operate as a scratchpad SRAM available for data storage by the processor 106 (e.g., stack and heap information). Also, the SRAM 204 can store code.
[0037] The security circuitry 112 may include circuit components 114, which may be subordinate execution units, coupled to the processor 106 via an interconnect 122. Each of these components 114 may adhere to an interface framework that ensures compatibility with each other and with the processor 106. The compatibility scheme may specify how the processor 106 communicates with a given circuit component 114 (e.g., using the interconnect 122), how the circuit component 114 communicates with the processor 106 (e.g., using interrupts), how the circuit component 114 communicates security events (e.g., using alert indications) to other circuit components 114 such as the alert handler 208, how the circuit component communicates with peripheral devices (e.g., by chip I / O), or a combination thereof. The illustrated components 114 may include circuit components associated with alert-related functions provided by the alert handler 208, associated with the processor 106, associated with one or more memories, associated with the chip I / O, etc. Thus, memory may also include components associated with each other, or other illustrated circuit components 114.
[0038] The circuit or chip I / O includes a pin MUX 218 and a pad controller 220. The pin MUX 218 provides signaling routes between at least some of the components 114 and available multiplexable I / O nodes of the security circuit 112 (e.g., pins of the chip on which the various components are integrated, or interfaces to other parts of the SoC). The pad controller 220 manages the control or pad attributes, such as drive strength, technology, pull-up vs. pull-down, etc., of each circuit component. The pin MUX 218 and the pad controller 220 themselves may be peripheral devices on the interconnect 122. Thus, each may have at least one register collection that provides software configurability or may be otherwise associated with them.
[0039] The UART unit 214 can implement UART functions such as single lane duplex UART functions. Its outputs and inputs can be configured to connect to any circuit I / O via pin MUX 218. The GPIO interface 216 creates G-bit bidirectional communication to external circuits via pin MUX 218, where G is a positive integer such as 16, 32, or 64. With respect to memory I / O, the SPI device 212 can implement a firmware mode, where the firmware mode can enable a function that provides the ability for an external driver to send firmware upgrade code to a bank of the flash memory 206 for in-field firmware updates. The firmware mode can include addressing of memory using SPI transactions. Although not shown, the security circuit 112 can include an Inter-Integrated Circuit (I2C) host to enable command of I2C devices. The command of I2C devices can include standard mode, full mode, and high speed mode.
[0040] Several "core security" components are also shown, including the encryption engine and alert handler 208. The encryption IP core 116, which may be implemented as described with reference to FIG. 1, may provide various symmetric encryption and decryption functions, such as by using one or more protocols and / or various key sizes, such as 128b, 192b, or 256b. In aspects, the encryption IP block 116 may be incorporated within an AES unit and may implement the methods described herein. The components may select encryption or decryption of data arriving in, for example, 16-byte amounts, to be encrypted or decrypted using different block cipher operating modes. The encryption IP block 116 may support Electronic Code Book (ECB) mode, Cipher Block Chaining (CBC) mode, Cipher Feedback (CFB) mode, Output Feedback (OFB) mode, Counter (CTR) mode, etc. Data transfer may be made available to the processor. For example, keys and data material may be passed to the encryption engine via register writes. Alternatively, a private channel for transfer of keys and data material may be included to reduce exposure from potentially untrusted processor activity.
[0041] The HMAC engine 210 may utilize, for example, the Secure Hash Algorithm (SHA) SHA-256 as a hash algorithm. SHA-256 is a member of the SHA-2 family of hash algorithms that results in a digest (or hash output) length of 256b, regardless of the data size of the input being hashed. After declaring the start of a hash request, data is sent to the HMAC engine 210, which zeroes out any internal state to an initial condition (e.g., 32b at a time). Once the data has been sent by a constituent client, the client may indicate completion of the hash request (with an optional partial word final write). In accordance with an exemplary portability interface scheme, the HMAC engine 210 generates the hash result and makes it available for register reads by the requesting client. The data transfer may be made available to the processor, or may be private to reduce exposure to potentially untrusted processor activity.
[0042] HMAC is a message authentication protocol layered on top of a hash function (e.g., SHA-256), where HMAC mixes in a secret key for encryption purposes. HMAC is a specific application of adding a secret key in a prescribed manner (e.g., twice) around the hash of the message (via SHA-256). To provide this functionality, a 256b key may be programmed into the circuitry before the message hashing begins. The timing of authentication completion may vary and may result in longer latency than when using native SHA-256. Again, the hash information or secret key may be made available to the processor for convenience or processing efficiency, or may be kept private in some manner to improve security.
[0043] The alert handler 208 is responsible for handling and responding to alerts (e.g., errors), including alerts provided by other components 114 (e.g., check modules). Alerts can be considered security-critical interrupts that need to be handled in a timely manner to respond to a sensed security threat. Unlike "standard" interrupts, alerts do not have to be handled solely by software running on the processor 106. An alert can trigger a first stage request to be treated as a "normal" interrupt by the software. However, if the software is unable to respond to and properly repair an interrupt triggered by an alert, the alert handler 208 triggers a second stage response. The second stage response can include taking security measures, examples of which include terminating processing, erasing or otherwise removing data, removing power from a circuit portion, or resetting an IC chip or part thereof. This ensures that the underlying problem (the sensed security threat) is addressed even if the processor 106 is busy, unresponsive, and / or under attack.
[0044] Thus, an alert can be implemented as an elevated interrupt-type signal or an alert indication that the alert handler 208 receives from other circuit components 114 and that indicates a potential security threat. In operation, the alert handler 208 can collect alerts from other circuit components 114 of the security circuit 112 and convert them into interrupts that the processor 106 can address. However, if the processor 106 does not clear the interrupt, the alert handler 208 provides a hardware response to address the potential security threat.
[0045] In general, the alert handler 208 may store some indication of an alert received from any of the components 114. In some cases, this indication may appear after the IC 104 or one or more circuit components 114 perform a reset. As a specific example, the alert handler 208 may write the time, location, or type of alert to a register to allow the operation of the IC 104 or circuit component 114 to continue normal operation. The resulting data may be discarded once the operation is completed or a predefined period of time has elapsed, so that privileged information is not exposed or the operation of the IC 104 is not altered by the underlying event that caused the alert. In this way, a malicious agent may not be able to determine when an alert is provided to the alert handler 208.
[0046] The security circuitry 112 may also include a RNG 222. In general, randomness can contribute to security functions by providing execution variations that prevent an attacker from predicting the right time to launch an attack. For example, random numbers can provide secret material used for authentication and encryption purposes. Incorporating the RNG 222 into algorithmic calculations can obscure sensitive data values. In general, the RNG 222 provides better performance as its number generation becomes more and more truly random, and can also be hardened against attacks. Some cryptographic processes may not require strong randomness guarantees. The RNG 222 may be implemented as a "true" RNG (TRNG) or may involve a design with analog parts to exploit some non-deterministic physical event or process. Exemplary TRNG designs rely on metastability, electronic noise, timing changes, thermal noise, quantum fluctuations, etc. The TRNG filters the resulting variable(s) and sends them to an entropy pool that the device can sample at a given time for the current randomization function. In some cases, the interface to the entropy pool may include a read request for available random bits. The TRNG interface indicates the number of bits available, and the requesting circuitry or software may read from this pool within the range of available bits. An interrupt or alert may be triggered if an attempt is made to read an entropy bit that is not available.
[0047] Two other components 114 include a timer 224 and a flash controller (not shown). The latter is described in the next paragraph. The timer 224 can support accurate performance by the processor 106, for example. The timer 224 is formed of multiple bits (e.g., 64 bits) and operates as a free-running timer with a frequency guaranteed within a certain percentage. The timer 224 can enable the circuit component 114 to determine an appropriate time interval for performing an operation. For example, the FSM 118 or other components of the encrypted IP block 116 can send or receive signals according to the timer 224. Another timer (not explicitly shown) can function as a watchdog timer to assist the processor 106 if the processor becomes unresponsive. The unresponsiveness can be due to unresponsive development code, security attacks, etc.
[0048] Although not shown, the flash controller may control the flash memory 206, which is available for storing code and data. The primary read path for this data may be located in a standard memory address space. However, because flash is not written to in a standard fashion, writes to that address space may be ignored. Instead, software interacts with the flash controller to write to the flash memory 206. Flash functions may include three primary commands: read, erase, and program. Read commands may be standardized and may use the address space of the chip memory. Erase commands are performed at the page level, with the page size being parameterizable by the flash controller. Upon receiving an erase request, the flash controller erases the contents of the target page, forcing the data to a "1" state (e.g., 0xFFFFFFFF per word). Software may then program individual words to any value. Because flash bits do not return to a "1" state unless another erase is performed, future contents are effectively changed by ANDing the current contents with the written value. Erase and program commands are relatively slow. Typical erase times are measured in milliseconds, while program times are in the microsecond range. Security is also a concern since sensitive data may be stored in the flash memory 206. Therefore, some degree of memory protection may be provided by the flash controller.
[0049] The security circuit 112 is shown in FIG. 2 with a particular set of circuit components 114. However, a given security circuit 112 may have more, fewer, or different circuit components. Also, the circuit components may be interconnected or operated differently from the exemplary manner described herein. Furthermore, some circuit components 114 may be omitted while other circuit components 114 are implemented in multiple instances. For example, the alert handler 208 may be duplicated or distributed, or there may be multiple cryptographic IP blocks 116 in any one security circuit 112. Furthermore, in cases where the security circuit 112 is an IC chip that forms only one core out of dozens of cores, the GPIO interface 216 may be omitted from the components 114 of the security circuit 112.
[0050] In aspects, any of the circuit components 114 may include a cryptographic processor 230 that performs the cryptographic operations of the component. In some implementations, the cryptographic processor 230 is separate from the processor 106 of FIG. 1. In other implementations, the cryptographic processor 230 is implemented within the processor 106. Additionally, any of the circuit components 114 may include a cryptographic processor 230 (or processing core) for performing the particular functions of that component.
[0051] 3 illustrates an exemplary encryption IP block 116 that can implement secure multi-rail control for sparsely encoded signals in accordance with one or more aspects. The encryption IP block 116 can provide symmetric encryption and decryption using one or more protocols and various key sizes, such as 128b, 192b or 256b. The components can select to encrypt or decrypt data arriving in 16 byte amounts, for example, that is encrypted or decrypted using different block cipher operating modes.
[0052] The encryption IP block 116 may be attached to the IC interconnect 122 as a peripheral module. Communication with the processor may be performed through a set of control and status registers 302 that may input or output data, keys, and / or status and control information. In aspects, the registers 302 may provide an interface through which a key manager may provide or receive keys without exposing the keys to a processor or other host attached to the IC interconnect 122. The encryption IP block 116 has multiple components communicatively coupled to each other by one or more rails 120. In aspects, the rails 120 carrying critical signals are represented by bold arrows in FIG. 3 and may include multiple rails implementing secure multi-rail control for sparsely encoded signals.
[0053] As shown, the cryptographic IP block 116 includes a main control FSM 304, which controls access to software, registers, and other portions of the IC, and a cryptographic core 306 (e.g., an embodiment of cryptographic processor 230), which includes a core control FSM 308, which controls operation of circuit components within the cryptographic core 306. The main control FSM 304 and the core control FSM 308 may communicate handshake signals over rails 120. In aspects, the main control FSM 304 may send signals to the core control FSM 308 to start and end operation of the cryptographic core 306. For example, when data is entered into the register 302, the main control FSM 304 may send a handshake signal along rails 120 to the core control FSM 308 to indicate that data has been entered into the register 302 for undergoing a cryptographic operation. In response, core control FSM 308 may send a handshake signal along rails 120 to main control FSM 304 to acknowledge or receive a request for cryptographic core 306 to perform an operation on the data input into register 302. Main control FSM 304 may then send a signal to register 302 to release the data to cryptographic core 306 and / or write lock register 302 so that it cannot input or output data.
[0054] To initiate an operation, core control FSM 308 may send a multiplexer (MUX) selector signal to MUX 310 to multiplex input data (e.g., using MUX 310-1) or an input key (e.g., using MUX 310-2). In aspects, the data and / or key may be masked based on randomness provided to MUX 310. Once the data and / or key are multiplexed, core control FSM 308 may send a write enable signal along rails 120 to status register 312 and / or key register 314 to allow the data / key to be written to the respective register.
[0055] The data and key may then be passed to datapath logic 316 where various operations may be performed on the values. For example, the data may undergo non-linear operations within datapath logic 316. Datapath logic 316 may include any number of S-boxes or Galois Field (GF) multipliers. Datapath logic 316 may perform multiple encryption / decryption on the data using the key. To protect against various attacks, datapath logic 316 may perform multiple parallel operations using multiple S-boxes. Core control FSM 308 may communicate various signals to circuit components within datapath logic 316 to enable or disable operations within datapath logic 316 or to acknowledge that an operation has been performed.
[0056] Once the data has passed through the datapath logic 316, it may be multiplexed again using a set of MUXs 318. The core control FSM 308 may send a MUX selector signal using rails 120 to control the operation of the MUXs 318. In doing so, after a first series of operations, the data may be multiplexed in MUX 318-1 to create the encrypted / decrypted data, and the key may be multiplexed in MUX 318-2 to create the round key used for encryption / decryption during the next series of operations. The data and round key may then be returned by MUX 310 to the status register 312 and key register 314 for another iteration of encryption / decryption.
[0057] Once the data has finished operating (e.g., the data has completed all encryption / decryption sequences), the core control FSM 308 may send a handshake signal over rails 120 to the primary control FSM 304 to indicate that the data is ready to be output to the register 302. In response, the primary control FSM 304 may send a handshake signal over rails 120 to acknowledge or receive a request to release the data to the register. The primary control FSM 304 may then send a signal (e.g., a write enable signal or a read enable signal) along rails 120 to the register 302 to allow the data to be written to the register and then released along interconnect 122. As a result, the data may be output from the encryption core 306 and written to the register 302.
[0058] As previously mentioned, various signals may be sent by the main control FSM 304 or core control FSM 308 to manage the operation of the cryptographic IP block 116, making handshaking and other control signals in the cryptographic and other IP blocks common targets for fault injection attacks. Successfully injecting a single fault into such a signal may cause the IC to prematurely terminate (e.g., terminate) a current operation, prematurely begin a new operation, or release intermediate data, thereby enabling a series of other attacks. In some instances, these attacks target the cryptographic IP blocks because they may hold privileged information or be used to implement security features such as secure boot processing or a root of trust.
[0059] In some cases, critical signals may be targeted by fault injection attacks (e.g., handshake and other control signals) that utilize a targeted laser to flip transistors in the IP block. For example, changing bits of a critical signal may redirect logic flow to terminate execution of a chip operation, release intermediate data associated with the chip operation, or repeat operations out of order. Information released as a result of a fault injection attack or logic redirection may provide an attacker with privileged information about the on-chip logic that can be used to circumvent security measures or design other attacks against the IC. With respect to encryption IP blocks, fault injection attacks may enable the release of encryption private keys to non-privileged individuals or systems. Thus, in some examples, the described secure multi-rail control logic may be implemented only within the encryption IP block. In other implementations, the described sparse encoding may be utilized at any point where sensitive operations are occurring on a hardware device, such as any operations involving access control or privilege level operations (e.g., power, register access, system reset). For example, the described secure multi-rail control logic may be utilized between any two components of an IC that perform such sensitive operations.
[0060] In general, critical signals may have bit representations that do not require a set minimum Hamming distance, or circuit synthesis tools may reduce the effectiveness of larger bit representations that adhere to the minimum Hamming distance. Thus, a fault injection attack may be able to change these signals to an unintended but otherwise valid state, thereby directly affecting the operation of the IC. For example, a handshake may be altered, out of order, or initiate an interaction between the encryption IP block 116 and the encryption core 306 if it is untimely (e.g., an untimely output or termination from the encryption core 306). As another example, manipulating the operation of a MUX device (e.g., MUX 310 or MUX 318) may directly affect the data output to the register 302 (e.g., affecting the type of data that is latched).
[0061] To protect against fault injection attacks, sparse encoding may be used for any critical signals that directly affect the operation of the circuit. For example, instead of communicating critical signals on a single wire (or rail) using a single bit representation, n wires (n greater than 1) may be used for these critical signals. In this way, each of the rails 120 used to connect circuit components may communicate a single bit of the sparsely encoded signal. Furthermore, critical signals may require the use of a minimum Hamming distance such that flipping a single bit (or a number of bits less than the Hamming distance) does not provide a valid solution. Specifically, an attacker may need to flip as many bits simultaneously (or in a short time) as the minimum Hamming distance to change a critical signal into a different valid signal. As a non-limiting example, the minimum Hamming distance may be between 2 and 5 bits, however, it should be understood that the minimum Hamming distance may be scaled to improve security enhancement. For a two-state signal (e.g., one bit, HIGH or LOW), the number of rails 120 (e.g., between any two components) may be the same as the minimum Hamming distance. Therefore, the number of rails 120 can be a scalable parameter for enhanced security.
[0062] In the non-security enhancing case, a bit representation of a valid signal may be 1'b1 and another valid signal may be 1'b0. Thus, flipping a single bit may change one valid signal to a different valid signal. In contrast, for the described sparse encoding example, a bit representation of one valid signal may be 3'b101 and a second valid signal may be 3'b010. In this case, the Hamming distance is 3 since 3 bits need to be changed to get a different valid signal. A larger Hamming distance may increase the security parameters of the encrypted IP block 116. This is because an attacker would need to flip multiple bits of a signal simultaneously to maliciously change the signal to a different valid signal. This may be particularly difficult. A fault injection attack using lasers may require separate lasers to each trigger different transistors at the same time (or within a short time) to flip multiple bits of a signal. Generally speaking, an attacker needs to insert as many faults as the Hamming distance to successfully trigger the attack (e.g., generate a successful signal).
[0063] In some implementations, a valid signal should be selected such that its bit representation does not include all bits with the same binary value. For example, in a three-bit scenario, the bit representation of a valid signal should not be 3'b111 or 3'b000. In general, an attacker may be more successful in flipping all bits in a signal to a single value than in flipping multiple single bits of a word to different values. For example, an attacker may be able to clear a word by converting all bits in the word to a binary value of 0 or a binary value of 1. Therefore, a valid signal may be defined to have both zero and non-zero bits to improve the security of an IC against fault injection attacks.
[0064] To communicate a valid signal as a mixture of HIGH and LOW bits, a negation rail may be implemented that includes negation logic. For example, if a first component desires to transmit a HIGH value (e.g., 1'b1), the HIGH 3-bit signal may be selected to be (3'b100). In this example, the rail transmitting the most significant bit is the positive rail since it transmits the same value communicated by the first component. In contrast, the rail transmitting the second most significant bit and the rail transmitting the third most significant bit are negative rails since they transmit values different from the value transmitted by the first component. In this example, a HIGH value (e.g., 1'b1) is transmitted as 3'b100 and a LOW value (e.g., 1'b0) is transmitted as 3'b011. As shown, once the multi-bit representation of the HIGH or LOW value is determined, the positive and negative rails may be determined. Once determined, the multi-bit signal may be automatically generated by placing an inverter on the negative rail or at the output of the transmitting component (e.g., FSM). As a result, a sparsely encoded signal may be represented as a combination of both HIGH and LOW signals.
[0065] In an embodiment, the inverter may be merged (e.g., consolidated) into a given rail. For example, by appropriately configuring (e.g., constraints, instructions) the synthesis tool, the synthesis tool may be caused to merge the inverter into the negated rail.
[0066] Given that increasing the Hamming distance of valid signal values may require a larger word size (e.g., more bits to describe the value), elements of the encryption IP block 116 may be designed to accommodate the greater number of bits used for each signal. For example, MUX devices (e.g., MUX 310 and MUX 318) that operate on the values may be required to support the greater number of bits at once. Additionally, other logic (e.g., datapath logic 316) or registers (e.g., register 302, status register 312, or key register 314) within the encryption IP block 116 may need to be modified to accommodate the greater Hamming distance. In this manner, components may operate natively in the higher bit format.
[0067] In aspects, the multi-bit size of the MUX control signal poses further challenges when utilizing sparse encoding to improve security. For example, the MUX control signal may require more bits to reach the same minimum Hamming distance compared to a control signal that can be represented by a single bit. In these cases, the number of rails 120 may be equal to or greater than the minimum Hamming distance.
[0068] As an example of a multi-bit multi-value MUX control signal, to achieve a minimum Hamming distance n for a ternary sparse signal, a suitable encoding may be determined along with the minimum number of bits required. For example, if n is 3, the following 5-bit values may be used: 5'b01110, 5'b11000, 5'b00001. For such a signal, all FSM rails may drive the full multi-bit signal (e.g., core control FSMs 308-1 through 308-n drive the full 5-bit signal). In addition, a check unit (e.g., check unit 320) may combine the n copies into a single signal, e.g., via a bitwise OR. As a result, if the first core control FSM drives 5'b01110 and the second core control FSM drives 5'b11000, the combined signal will be 5'b11110, which is invalid. In an additional embodiment, the check unit may compare all signals to each other before the bitwise OR combination.
[0069] While some components may simply be modified to accommodate larger bit signals, for FSMs responsible for driving control signals (e.g., main control FSM 304 and core control FSM 308), this modification may be more complicated if, for example, the FSM may include a circuit synthesis device that collapses the circuit logic to operate more efficiently. In these cases, circuit synthesis may reduce the effectiveness of redundant multi-bit representations used in the sparse encoding when performing operations through optimizations. This is further explained in FIG. 4. As such, the circuit synthesis tool may eliminate the sparse encoding, thus making the circuit more susceptible to fault injection attacks and potentially increasing the attack surface. To combat this, some techniques may create custom multi-bit gates that do not tolerate synthesis. Other techniques may disable the circuit synthesis tool to ensure that security measures are not unintentionally removed. However, each of these techniques may be suboptimal, as custom multi-bit gates may be costly, non-scalable, and space-disadvantaged. Furthermore, disabling the circuit synthesis tool may disable other circuit synthesis operations that allow for higher performance of the IC without impacting the security measures.
[0070] In contrast to these techniques, multi-rail control for sparsely encoded signals may utilize multiple copies of each FSM to operate natively in multi-bit format and eliminate circuit synthesis via multiple rails 120 and FSMs (e.g., main control FSM 304 or core control FSM 308). For example, each FSM (e.g., main control FSM 304 or core control FSM 308) may be implemented as multiple FSMs (e.g., main control FSMs 304-1 to 304-n or core control FSMs 308-1 to 304-n). Each of the multiple FSMs may be communicatively coupled to a single one of the rails 120 and may communicate and receive a single bit of the sparsely encoded signal. In this way, the entire signal is not dependent on a single logic operation with circuit synthesis. In addition, when multiple FSMs are used, custom multi-bit gates may not be required since identical copies of compatible FSMs may be placed in the circuit. In this way, these techniques may provide a scalable solution that may scale the number of FSMs and rails 120 based on security, space, and cost constraints. Moreover, these techniques can be scaled as attacks advance: for example, if an attacker can now flip three bits simultaneously, he could increase the number of FSMs, rails 120, and minimum Hamming distance to, say, four.
[0071] In addition to using sparse encoding to increase the minimum Hamming distance between valid signals, a check module 320 may be placed in an area where handshake or control signals are communicated. For example, a check module 320 may be placed on rail 120 between the main control FSM 304 and the register 302, a check module 320 may be placed on rail 120 between the main control FSM 304 and the core control FSM 308, or a check module 320 may be placed between the core control FSM 308 and a cryptographic core register (e.g., status register 312 or key register 314) or a MUX device (e.g., MUX 310 or MUX 318).
[0072] The check module 320 may monitor signals transmitted over the rails 120 and check to ensure that the communicated signals represent a valid encoding (e.g., no faults have been injected). For example, the check module of the check module 320 may store a set of predefined valid states that the signals passing over the rails 120 may represent. For signals that transmit either a HIGH or a LOW value, the check module may store a multi-bit representation of the HIGH value and a multi-bit representation of the LOW state. For signals that cannot be represented by only two states (e.g., MUX selector signals), the corresponding check module 320 may include a set of valid states having more than two valid states. In storing the predefined set of valid states, the check module 320 may compare the signals passing along the rails 120 to the predefined set of valid states to determine whether the signals match any valid state in the set of valid states. If the signals match any valid state, the IC or encryption IP block 116 may continue normal operation. If it is determined that the signal passed from the first component to the second component does not represent a valid encoding (e.g., it is not within the list of valid states because a fault injection attack has occurred), an error may be flagged and passed up the chain for error handling.
[0073] In aspects, the error handling may vary based on the particular implementation. For example, when an error is triggered by the check module (e.g., if a fault injection attack is suspected to have occurred), the control FSM (e.g., the main control FSM block and the core control FSM block of FIG. 3) may be locked or reset to limit the ability to change the processing running in the encryption IP block 116. In other embodiments, a shutdown or reset of the entire IC may be triggered when a sufficient number of errors are reached. In general, a device reset may be required to restore the state of a locked FSM or a locked chip. In another example, processing may continue to the end of the current processing, but the output is discarded. In this manner, an attacker may not know when an error is triggered by an attack, and the mechanisms of the security system may remain hidden. In general, an error detected by the check module 320 may be handled in multiple ways, as described with respect to the alert handler of FIG. 2. In general, however, the check module 320 may keep a record of the determined error by saving an indication of the error in a register of the IC to ensure that the encryption IP block 116 knows that an error has occurred.
[0074] By implementing the check module 320 along the rail 120, the encryption IP block 116 may determine if a fault injection attack has occurred and then the behavior of the circuit may have been altered by the attack. Furthermore, to successfully trigger a fault injection, an attacker may need to inject multiple faults into the signal, either simultaneously or within a short time period. For example, a valid solution passed initially may have a bit representation of 3'b100. During the first attack, the attacker may invert the first bit of the signal to have a bit representation of 3'b000. Then, in a second attack, the attacker may invert the second bit of the signal to obtain 3'b010. Finally, in a third attack, the attacker may invert the third bit of the signal to generate a valid solution 3'b011. Without the check module 320 in a location near the handshake or control signal, the encryption IP block 116 may function based on the valid signal 3'b011 even if it has been altered by a fault injection attack.
[0075] However, if the checking module 320 is implemented, the encryption IP block 116 may be able to identify the signal resulting from the first or second attack as an invalid signal and trigger an alert. Thus, the checking module may be able to determine a fault injection attack even when the final result is a valid signal. In some embodiments, the checking module may detect an attack even when bits are flipped in rapid succession. In this way, an attacker attempting to flip multiple bits to arrive at a different valid signal would need to flip each bit simultaneously, which may require multiple lasers and precise timing. Additionally, the checking module may increase the overall difficulty of successfully executing a fault injection attack, since the modified signal is checked immediately after the attack. In this way, a potential attacker may need to correctly flip multiple bits in one attempt.
[0076] In some implementations, the check modules 320 between the main control FSM 304 and the core control FSM 306 may be removed from the encryption IP block 116 because faults injected into these signals propagate to signals that control the data path (e.g., write enable signals or MUX selector signals) and are flagged by the check modules 320 located on those rails 120. This may save circuit space and cost without negatively impacting the security of the IC.
[0077] FIG. 4 illustrates an example logic 400 of an AND operation on sparsely encoded signals susceptible to fault injection attacks by circuit synthesis. In aspects, an AND operator may be used to determine whether a signal is a HIGH or LOW value (e.g., by using the AND operator on signals having a HIGH value as an input). In this example, the control logic shown utilizes a three-bit control signal. The logic 400 is used to perform an AND operation on signals A 402 and B 404. In aspects, signals A 402 or B 404 may generally represent any of the handshake or control signals described herein. Signal A 402 is a three-bit signal having a most significant bit A0 402-0, a second most significant bit A1 402-1, and a third most significant bit A2 402-2. Similarly, signal B is a three-bit signal having a most significant bit B0 404-0, a second most significant bit B1 404-1, and a third most significant bit B2 404-2. In aspects, signal A 402 and signal B 404 may be sparsely encoded signals implemented using a minimum Hamming distance requirement.
[0078] Logic 400 represents synthesized logic optimized for redundancy, similar to logic resulting from optimization with modern circuit synthesis tools. Each signal is initially carried by a rail in a 3-bit format indicated by an arrow (thick arrows indicate rails carrying multi-bit signals). Each bit is then separated from the signal, and the most significant bit and the second most significant bit of each signal are input to a NAND operation. Specifically, bit A0 402-0 and bit A1 402-1 are input to a NAND operator, and bit B0 404-0 and bit B1 404-1 are input to a NAND operator. The output is then input to a NOR operator and an OR operator, respectively, using the third most significant bit of each signal. As shown, the output of the NAND operation of bit A0 402-0, bit A1 402-1, and bit A2 402-2 is input to a NOR operator. Similarly, the output of a NAND operation of bit B0 404-0, bit B1 404-1, and bit B2 404-2 is input to an OR operator. The outputs of the NOR and OR operators are passed to the ORNOT and ANDNOT operators with their respective outputs being input to each operator. As a result, the output of the ANDNOT gate is bit C1 406-1, which is the second most significant bit of the output signal C 406, and the output of the ORNOT gate is bit C2 406-2, which is the third most significant bit of the output signal C 406. As shown, the 3-bit representation has been optimized into a 2-bit format, and one bit of the representation, in this case the most significant bit C0 (not shown), has been omitted from the output.
[0079] Throughout logic 400, there are multiple locations (indicated by dashed arrows in FIG. 4) where a single fault could be injected to flip the output of both bits in C 406 (e.g., C1 406-1, C2 406-2). Thus, the use of sparse encoding did not improve security against fault injection attacks as a result of circuit synthesis, but actually expanded the attack surface because now faults can be injected into multiple rails to successfully change the signal to different valid states. In this way, rather than brute-forcing multiple faults to succeed in an injection attack, the synthesis tool may simply move the points where a fault injection attack can succeed by using sparse encoding, potentially increasing the attack surface. Also, circuit synthesis has optimized away check modules located on the rails to eliminate redundant bits in the circuit. Thus, sparsely encoded signals must be carefully controlled to protect against possible fault injection attacks.
[0080] It should be noted that although a 3-bit implementation is shown, implementations using any number of bits may encounter similar problems caused by circuit synthesis. Thus, multi-rail control may be useful for implementing sparsely encoded signals regardless of the particular number of bits used.
[0081] FIG. 5 illustrates an example logic 500 for an AND operation using secure multi-rail control for sparsely encoded signals, in accordance with one or more aspects. In this example, the illustrated control logic utilizes a three-bit control signal. Similar to logic 400 of FIG. 4, logic 500 is used to perform an AND operation on signals A 402 and B 404. Signals A 402 or B 404 may generally represent any of the handshake or control signals described herein. Signal A 402 is a three-bit signal having a most significant bit A0 402-0, a second most significant bit A1 402-1, and a third most significant bit A2 402-2. Similarly, signal B is a three-bit signal having a most significant bit B0 404-0, a second most significant bit B1 404-1, and a third most significant bit B2 404-2. In this example, the rails carrying bit A2 402-2 and bit B2 404-2 are negated rails, and these bits are the negated values of the bits sent by the first component (e.g., if the first component sends a signal indicating a HIGH value, these bits carry a LOW value, and vice versa). The negated bits may be generated automatically using an inverter. In aspects, signal A 402 and signal B 404 may be sparsely encoded signals implemented using a minimum Hamming distance requirement. Unlike logic 400 of FIG. 4, each bit of the signal may be driven individually through its respective rail (e.g., by a single FSM of the multiple FSMs described in FIG. 3) to protect against circuit composition through different rails.
[0082] In the exemplary multi-rail control logic, every bit of the sparsely encoded signal is processed and communicated separately. In this way, the multi-rail control logic can operate natively in a multi-bit environment and reduce the effectiveness of fault injection attacks. In logic 500, the signal is first transmitted over a multi-bit rail (indicated by the thick arrow) as a 3-bit signal, and then each bit is separated and sent over an individual rail. Each set of corresponding bits of the signal (e.g., the most significant bit of A402, and the most significant bit of B404, the least significant bit of A402, and the least significant bit of B404, etc.) is input to an AND operator. For example, bit A0 402-0 and bit B0 404-0 are input to an AND operator, bit A1 402-1 and bit B1 404-1 are input to an AND operator, and bit A2 402-2 and bit B2 404-2 are input to a negated rail AND operator. In aspects, the negated rail AND operator is implemented on the negated rail. In the negated rail AND operator, the inputs are inverted, an AND operation is performed, and then the output of the AND operator is negated. Although described with respect to the AND operator, it should be noted that the negation operator may be performed with any other operation (e.g., OR, XOR, NOR, NAND) as well. For example, the inputs may be negated, an operation may be performed, and the output may be negated. The negated rail AND operator is further described in connection with FIG. 6. The outputs of the two AND operators and the negated rail AND operator are bit C0 406-0, bit C1 406-1, and bit C2 406-2, respectively. These bits are then collected into the output signal C 406.
[0083] The output signal C406 is then provided as an input to a checker 506 (e.g., an embodiment of the check module 320 of FIG. 3). The checker 506 may compare the output signal C406 to a set of predefined valid states stored in the IC. If the checker 506 determines that the signal C does not correspond to one of the predefined valid states stored in the IC, it may flag an error 502. The error may be passed up the chain for error handling, for example, as described with respect to the alert handler of FIG. 2. The output signal C504 may be provided to a second component of the IC to which the signal was sent. When the signal C406 is determined to correspond to one of the valid states, the output signal C504 may correspond to the output signal C406. When the output signal C406 does not correspond to a valid state, the output signal C504 may not be output, may be a different value that does not correspond to the output signal 406, or may be a value that corresponds to the output signal 406. In this manner, errors may be handled in multiple ways based on the embodiment.
[0084] Unlike logic 400 of FIG. 4, output signal C 406 of FIG. 5 has the same number of bits as input signal A 402 and input signal B 404. Thus, because output signal C 406 corresponds to a predetermined state format, checker 506 may be effective in determining whether the signal has been altered by a fault injection attack. Additionally, logic 500 does not have a location where each bit of output signal C 406 can be inverted by injecting a single fault. In this way, logic 500 may not be susceptible to single fault injection, like logic 400 of FIG. 4.
[0085] 6A and 6B illustrate example logic for implementing safe multi-rail control for sparsely encoded signals, according to one or more aspects. For example, FIG. 6A illustrates how control logic for multiple rails is automatically determined for positive and negative rails. The logic determines whether each rail is a positive or negative rail by determining whether a HIGH value is represented as a 1'b1 signal. If so, the rail is determined to be a positive rail, and appropriate positive rail logic (e.g., the AND operator described in FIG. 5) may be used to control operations on signals carried by the rail. If not, negative rail logic (e.g., the negative rail AND operator described in FIG. 5) may be used to control operations on signals carried by the rail.
[0086] FIG. 6B illustrates an example of a positive rail AND operator and a negative rail AND operator. For example, a positive rail operator may perform an AND operation on bits input to the operator to provide an output from the operation. In contrast, a negative rail AND operator may invert the inputs to the AND operation, perform an AND operation on the inverted inputs, and invert the resulting output. In this manner, the negative rail logic may use an inverter in the negative rail to generate true negated logic. Although described with respect to an AND operator, the negative rail logic may apply to any other operation. For example, a negative OR operator may negate the inputs to the operation and perform an OR operation to negate the output. This logic may similarly be applied to a NAND operation, an XOR operation, a NOR operation, or any other logical operation.
[0087] In an embodiment, a synthesis tool can be configured (e.g., constrained) to merge inverters into a given rail. For example, by placing a synthesis constraint on the and_n hierarchy (as shown in FIG. 6B), the synthesis tool can merge the inverter into the negated rail. In this way, the synthesis tool may not perform logic optimizations across the hierarchy boundary of, for example, the and_n module. As a result, maintenance of a single copy of the code (e.g., and_p) is achievable (e.g., and_n can be implemented as a wrapper that includes several inverters). Such an embodiment is useful when a system includes one or more FSMs with tens, hundreds, or thousands of logic gates. Also, in some cases, the logic in a rail may not be limited to simple logic gates (e.g., AND, NAND, OR, NOR). Such hierarchical constraints are described by way of example only, and any of a variety of additional configurations including hierarchical constraints may be implemented to remove redundant rails and / or security measures.
[0088] FIG. 7 illustrates an example method 700 of transmitting a sparsely encoded signal using multiple rails. Method 700 is illustrated as a set of blocks that specify operations that may be performed, but are not necessarily limited to the order or combinations shown for performing the operations by the respective blocks. Furthermore, any one or more of the operations may be repeated, combined, rearranged, or linked to provide a variety of additional and / or alternative methods. These techniques are not limited to performance by one entity, or multiple entities operating on one device. In some aspects, the operations or functions of method 700 are implemented or managed by the encryption IP block 116, the FSM (e.g., the main control FSM 304 and / or the core control FSM 308), and / or any circuit components 114. For clarity, the method is described with reference to elements of FIGS. 1 through 6.
[0089] At 702, a first signal indicative of a first bit of the sparsely encoded signal is transmitted by a first component to a second component via a first rail communicatively coupling the first component to the second component. In aspects, the first rail is a single rail of the plurality of rails 120. The first rail may be a positive rail or a negative rail. The first signal may be transmitted by a respective FSM of the plurality of FSMs (e.g., the main control FSM 304 or the core control FSM 308). The first component or the second component may be the main control FSM 304, the core control FSM 308, a register (e.g., the register 302, the status register 312, or the key register 314), a MUX device (e.g., the MUX 310 or the MUX 318), or any other circuit component 114 in the IC 104. In aspects, the sparsely encoded signal may indicate a handshake signal, a write enable signal, a register setting, or a MUX selector signal. Furthermore, the first signal may be a HIGH signal or a LOW signal.
[0090] At 704, a second signal indicative of a second bit of the sparsely encoded signal is transmitted by the first component to the second component via a first rail communicatively coupling the first and second components. The second rail may be a rail different from the first rail of the plurality of rails 120. In aspects, the second signal may be a HIGH signal or a LOW signal. In some implementations, at least one of the first rail or the second rail is a negative rail such that the sparsely encoded signal may be represented as a combination of HIGH and LOW values. In aspects, the first and second components each include multiple FSMs, with each FSM of the first component communicatively coupled to a respective FSM of the second component by a respective rail.
[0091] At 706, the first and second signals are received at a second component. In aspects, the first and second rails are coupled to a check module 320, and an output of the check module 320 is coupled to the second component. The check module 320 may store a number of predefined valid states in its own memory or in any other memory of the IC. The check module 320 may compare the sparsely encoded signal to the predefined valid states stored in the memory. If the sparsely encoded signal does not match the predefined valid states of the sparsely encoded signal, the check module 320 may output an error 502 to the alert handler 208. If the sparsely encoded signal matches the predefined valid states, the check module 320 may output the sparsely encoded signal to the second component.
[0092] At 708, IC 104 performs a circuit operation based on the first and second signals. For example, IC 104 may perform a write to memory, a MUX operation, or any other circuit operation. In aspects, the circuit operation may be used to encrypt or decrypt data. As such, the first and second components may be implemented within encryption IP block 116 of IC 104. Alternatively, either or both of the first and second components may be implemented outside encryption IP block 116 but within IC 104.
[0093] FIG. 8 illustrates various components of an exemplary electronic device 800 that can implement secure multi-rail control for sparsely encoded signals in accordance with one or more aspects. The electronic device 800 may be implemented as any one or more of fixed, mobile, standalone or embedded devices. It may also be implemented in any form of consumer, computer, portable, user, server, communication, telephone, navigation, gaming, audio, camera, messaging, media playback, and / or other types of electronic devices 800, such as a smartphone, shown in FIG. 1 as device 102. One or more of the illustrated components may be realized as separate components or as integrated components on at least one IC of the electronic device 800. In general, the various components of the electronic device 800 are coupled via an interconnect 122 and / or one or more fabrics that support communication between the components in accordance with one or more aspects of secure multi-rail control for sparsely encoded signals.
[0094] The electronic device 800 may include one or more communications transceivers 126 that enable wired and / or wireless communication of device data 110 (such as received data, transmitted data, or other information identified herein). Examples of communications transceivers 126 include Near Field Communication (NFC) transceivers, wireless personal area network (PAN) (WPAN) radios conforming to the various IEEE 802.15 (Bluetooth®) standards, wireless local area network (LAN) (WLAN) radios conforming to any of the various IEEE 802.11 (WiFi®) standards, wireless wide area network (WAN) (WWAN) radios for cellular phones (e.g., those conforming to the 3rd Generation Partnership Project (3GPP®)), wireless metropolitan area network (MAN) (WMAN) radios conforming to the various IEEE 802.16 (WiMAX®) standards, infrared (IR) transceivers conforming to the Infrared Data Association (IrDA) protocol, wired local area network (LAN) (WLAN) Ethernet transceivers, and the like.
[0095] The electronic device 800 may also include one or more data input ports 128, through which any type of data, media content, and / or other input may be received. Examples include user selectable inputs, messages, applications, music, television content, recorded video content, and any other type of audio, video, and / or image data received from any content and / or data source, including sensors such as microphones or cameras. The data input ports 128 may include USB ports, coaxial cable ports, fiber optic ports for fiber optic interconnects or cables, and other serial or parallel connectors (including internal connectors) for flash memory, DVDs, CDs, and the like. These data input ports 128 may be used to couple the electronic device to components, peripherals, or accessories such as keyboards, microphones, cameras, or other sensors.
[0096] The electronic device 800 of this example includes at least one processor 106 (e.g., any one or more of an application processor, a microprocessor, a digital signal processor (DSP), a controller, etc.), which may include a combined processor and memory system (e.g., implemented as part of an SoC) that processes (e.g., executes) computer-executable instructions to control the operation of the device. The processor 106 may be implemented as an application processor, an embedded controller, a microcontroller, a security processor, an artificial intelligence (AI) accelerator, etc. In general, a processor or processing system may be implemented at least in part in hardware, which may include components of integrated circuits or on-chip systems, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), complex programmable logic devices (CPLDs), and other implementations in silicon and / or other materials.
[0097] Alternatively or additionally, electronic device 800 may be implemented with any one or combination of electronic circuitry, which may include software, hardware, firmware, or fixed logic circuitry implemented in association with processing and control circuitry, generally designated 802 (as electronic circuitry 802). This electronic circuitry 802 may implement executable or hardware-based modules (not shown in FIG. 8), such as by processing / computer-executable instructions stored on a computer-readable medium, logic circuitry, and / or hardware (e.g., FPGAs, etc.).
[0098] In aspects, the electronic device 800 includes an interconnect 122, which may include any one or more of a system bus, interconnect, crossbar, data transfer system, or other switch fabric that couples various components within the device to enable various aspects of signaling and / or communication in a sparse encoding. The system bus or interconnect may include any one or combination of a variety of bus structures, examples of which include a memory bus or memory controller, a peripheral bus, a parity block, an error correction code (ECC) block, a universal serial bus, and / or a processor bus or local bus utilizing any of a variety of bus architectures.
[0099] The electronic device 800 also includes one or more memory devices 804 that enable data storage, examples of which include random access memory (RAM), non-volatile memory (e.g., read only memory (ROM), flash memory, erasable programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM)), and disk storage devices. Thus, the memory device(s) 804 may be distributed throughout various logical storage levels of the system and among various physical components. The memory device(s) 804 provide a data storage mechanism for storing device data 110, other types of code and / or data, as well as various device applications 806 (e.g., software applications or programs). For example, an operating system 808 may be maintained as software instructions in the memory device 804 and executed by the processor 106.
[0100] In some implementations, the electronic device 800 also includes an audio and / or video processing system 810, which processes audio data and / or passes audio and video data to an audio system 812 and / or a display system 814 (e.g., a video buffer, or a smartphone or camera screen). The audio system 812 and / or the display system 814 may include any device that processes, displays, and / or otherwise renders audio, video, display, and / or image data. The display data and audio signals may be communicated to the audio and / or display components via other similar communication links, such as an RF (radio frequency) link, an S-video link, an HDMI (high definition multimedia interface), a composite video link, a component video link, a DVI (digital video interface), an analog audio connection, a video bus, or a media data port 816. In some implementations, the audio system 812 and / or the display system 814 are external or separate components of the electronic device 800. Alternatively, for example, the display system 814 can be an integrated component of the example electronic device 800, such as part of an integrated touch interface.
[0101] The electronic device 800 of Figure 8 is an exemplary implementation of the apparatus 102 of Figure 1, i.e., an exemplary implementation of a device capable of implementing secure multi-rail control for sparsely encoded signals as described with reference to Figures 1 to 7. Thus, the electronic device 800 can include a security circuit 112, which can be a separate IC chip or can be included as part of another IC chip or device, such as the processor 106, the electronic circuit 802 or the memory device 804. Thus, one or more of the illustrated components can be integrated on the same IC chip, such as an SoC, or at least on a single printed circuit board (PCB).
[0102] The security circuit 112 may include an encryption IP block 116 that performs cryptographically secure operations using encryption or decryption. The encryption IP block 116 may include various circuit components configured to communicate sparsely encoded signals according to one or more aspects of secure multi-rail control for sparsely encoded signals. For example, the encryption IP block 116 may include multiple rails, each transmitting a single bit of the sparsely encoded signal. Each of the multiple rails may be communicatively coupled to one of multiple FSMs that communicate a respective bit of the sparsely encoded signal via a respective rail. Thus, the principles of secure multi-rail control for sparsely encoded signals described herein may be implemented by or in conjunction with the electronic device 800 of FIG. 8.
[0103] Unless otherwise indicated by context, the use of the word "or" herein may be considered as an "inclusive or" or as the use of a term permitting the inclusion or application of one or more items linked by the word "or" (e.g., the phrase "A or B" may be interpreted as permitting only "A", permitting only "B", or permitting both "A" and "B"). Also, as used herein, a phrase referring to "at least one" of a list of items refers to any combination of those items, including single members. For example, "at least one of a, b, or c" includes a, b, c, ab, ac, bc, and abc, as well as multiple combinations of the same elements (e.g., aa, aaa, aab, aac, abb, acc, bb, bbb, bbc, cc, and ccc, or other permutations of a, b, and c). Additionally, items depicted in the accompanying figures and terms discussed herein may refer to one or more items or terms, and thus may refer to the singular or plural forms of items and terms interchangeably herein. Although implementations of the secure cryptographic processor have been described in language specific to particular functions and / or methods, the subject matter of the appended claims is not necessarily limited to the specific functions or methods described. Rather, the specific functions and methods are disclosed as exemplary implementations of secure cryptographic processing.
[0104] An example of safe multi-rail control for sparsely encoded signals is given below.
[0105] Example 1: An integrated circuit comprising: a first component comprising a first plurality of finite state machines (FSMs); a second component; and a plurality of separate data lines configured to communicate a sparsely encoded signal having a predetermined minimum Hamming distance between the first component and the second component, each FSM of the first plurality of FSMs communicatively coupled to the second component by a respective data line of the plurality of separate data lines, each of the plurality of separate data lines configured to communicate a respective single bit of the sparsely encoded signal.
[0106] Example 2: The integrated circuit of any of the preceding examples, wherein the second component comprises a second plurality of FSMs. Optionally, each FSM of the first plurality of FSMs may be communicatively coupled to a respective FSM of the second plurality of FSMs by a respective data line of the plurality of separate data lines.
[0107] Example 3: The integrated circuit of any of the preceding examples, wherein the sparsely encoded signal indicates a handshake between the first component and the second component.
[0108] Example 4: The integrated circuit of any of the preceding examples, wherein the plurality of separate data lines includes at least one positive data line and at least one negative data line configured to carry a second signal different from a first signal carried by the at least one positive data line.
[0109] Example 5: An integrated circuit as described in any of the preceding examples, wherein one or more FSMs of the first plurality of FSMs are communicatively coupled to the second component by a respective negated data line of the at least one negated data line, and / or the respective negated data line or the one or more FSMs of the first plurality of FSMs includes an inversion circuit configured to invert an output of a single bit of the sparsely encoded signal, the single bit being communicated using the respective negated data line.
[0110] Example 6: An integrated circuit as described in any of the previous examples, wherein the at least one positive data line comprises logic that causes the first signal to perform a positive circuit operation and / or the at least one negative data line comprises logic that causes the second signal to perform a negative circuit operation.
[0111] Example 7: The integrated circuit of any of the preceding examples, wherein the plurality of separate data lines comprises a number of data lines equal to the minimum Hamming distance of the sparsely encoded signal.
[0112] Example 8: An integrated circuit as described in any of the preceding examples, further comprising an encryption intellectual property block including the first component, the second component, and the plurality of separate data lines.
[0113] Example 9: An integrated circuit as described in any of the preceding examples, further comprising a check module coupled to the plurality of separate data lines, the check module configured to determine validity of the sparsely encoded signal communicated using the plurality of separate data lines.
[0114] Example 10: An integrated circuit as described in any of the preceding examples, wherein the checking module is configured to: store a predefined set of valid states; compare the sparsely encoded signal communicated using the plurality of separate data lines with the predefined set of valid states; and determine, based on the comparison between the predefined set of valid states and the sparsely encoded signal, whether the sparsely encoded signal matches any one valid state of the predefined set of valid states.
[0115] Example 11: An integrated circuit as described in any of the preceding examples, wherein the checking module is configured to report an error in response to the determination that the sparsely encoded signal does not match any one valid state of the set of predefined valid states.
[0116] Example 12: The integrated circuit of any of the preceding examples, wherein the predefined set of valid states includes a high state and a low state.
[0117] Example 13: An integrated circuit as recited in any of the preceding examples, wherein the second component comprises a multiplexer device. Optionally, the sparsely encoded signal may indicate a multiplexer selector control signal between the first component and the second component.
[0118] Example 14: An integrated circuit as recited in any of the preceding examples, wherein the second component comprises a storage register, or a configuration and setting register. Optionally, the sparsely encoded signal indicates at least one of a register setting configured by the first component, or a writeable control signal between the first component and the second component.
[0119] Example 15: A method of communicating a sparsely encoded signal performed by an integrated circuit as described in any of the preceding examples. For example, the method may include transmitting, by a first component of the integrated circuit, a first signal indicative of a first bit of the sparsely encoded signal over a first rail communicatively coupling the first component to a second component of the integrated circuit, transmitting, by the first component of the integrated circuit, a second signal indicative of a second bit of the sparsely encoded signal over a second rail communicatively coupling the first component to the second component, receiving, by the second component of the integrated circuit, the first signal and the second signal, and performing a circuit operation based on the first signal and the second signal. The circuit operation may be performed by the integrated circuit.
[0120] Although aspects of the described apparatus, systems and methods for implementing safe multi-rail control for sparsely encoded signals have been described in feature- and / or method-specific language, the subject matter of the appended claims is not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as exemplary implementations of safe multi-rail control for sparsely encoded signals, and other equivalent features and methods are intended to be within the scope of the appended claims. Moreover, various aspects of safe multi-rail control for sparsely encoded signals have been described, with it being understood that each described aspect can be implemented independently or in conjunction with one or more other described aspects.
Claims
1. 1. An integrated circuit comprising: a first component comprising a first plurality of finite state machines (FSMs); A second component; and and a plurality of separate data lines configured to communicate a sparsely encoded signal having a predetermined minimum Hamming distance between the first component and the second component, each FSM of the first plurality of FSMs being communicatively coupled to the second component by a respective data line of the plurality of separate data lines, each of the plurality of separate data lines configured to communicate a single bit of the sparsely encoded signal.
2. the second component comprises a second plurality of FSMs; 2. The integrated circuit of claim 1, wherein each FSM of the first plurality of FSMs is communicatively coupled to a respective FSM of the second plurality of FSMs by a respective data line of the plurality of separate data lines.
3. The integrated circuit of claim 2 , wherein the sparsely encoded signal indicates a handshake between the first component and the second component.
4. The plurality of separate data lines include: At least one positive data line; and at least one negative data line configured to carry a second signal different from the first signal carried by the at least one positive data line.
5. one or more FSMs of the first plurality of FSMs are communicatively coupled to the second component by a respective negative data line of the at least one negative data line; 5. The integrated circuit of claim 4, wherein the respective negated data line or the one or more FSMs of the first plurality of FSMs includes an inversion circuit configured to invert an output of a single bit of the sparsely encoded signal, the single bit being communicated using the respective negated data line.
6. the at least one assert data line includes logic for causing the first signal to perform an assert circuit operation; 5. The integrated circuit of claim 4, wherein the at least one negate data line comprises logic that causes the second signal to perform a negate circuit operation.
7. The integrated circuit of claim 1 , wherein the plurality of separate data lines comprises a number of data lines equal to the minimum Hamming distance of the sparsely encoded signal.
8. 10. The integrated circuit of claim 1 further comprising an encryption intellectual property block including said first component, said second component and said plurality of separate data lines.
9. 10. The integrated circuit of claim 1, further comprising a check module coupled to the plurality of separate data lines, the check module configured to determine validity of the sparsely encoded signal communicated using the plurality of separate data lines.
10. The check module includes: Store a set of predefined valid states, comparing the predefined set of valid states with the sparsely encoded signal communicated using the plurality of separate data lines; 10. The integrated circuit of claim 9, configured to determine, based on the comparison between the set of predefined valid states and the sparsely encoded signal, whether the sparsely encoded signal matches any one valid state of the set of predefined valid states.
11. 11. The integrated circuit of claim 10, wherein the checking module is configured to report an error in response to determining that the sparsely encoded signal does not match any one valid state of the set of predefined valid states.
12. The integrated circuit of claim 10 , wherein the predefined set of valid states includes a high state and a low state.
13. the second component comprises a multiplexer device; 2. The integrated circuit of claim 1, wherein the sparsely encoded signal indicates a multiplexer selector control signal between the first component and the second component.
14. the second component comprises a storage register or a configuration and setting register; The sparsely encoded signal is A register setting configured by the first component; or 13. The integrated circuit of claim 1, further comprising: a writeable control signal between said first component and said second component.
15. A method of communicating sparsely encoded signals performed by an integrated circuit according to any one of claims 1 to 14.
Citation Information
Patent Citations
Address selection system
JP1979102839A
Secure data transmission between two modules
US20070055868A1
Shift register protected against physical attacks
US20200021427A1
Detection of laser fault injection attacks on cryptographic devices
US20210328761A1
