Abstract: An apparatus includes a temperature sensor, a voltage regulator, and a field programmable gate array (FPGA). The temperature sensor and the voltage regulator are adapted, respectively, to provide a temperature signal, and to provide at least one output voltage. The FPGA includes at least one circuit adapted to receive the at least one output voltage of the voltage regulator, and a set of monitor circuits adapted to provide indications of process and temperature for the at least one circuit. The FPGA further includes a controller adapted to derive a body-bias signal and a voltage-level signal from the temperature signal, from the indications of process and temperature for the at least one circuit, and from the at least one output voltage of the voltage regulator. The controller is further adapted to provide the body-bias signal to at least one transistor in the at least one circuit, and to provide the voltage-level signal to the voltage regulator.

Abstract: A kill switch is provided that, when triggered, may cause the programmable logic device (PLD) to become at least partially reset, disabled, or both. The kill switch may be implemented as a fuse or a volatile battery-backed memory bit. When, for example, a security threat is detected, the switch may be blown, and a reconfiguration of the device initiated in order to zero or clear some or all of the memory and programmable logic of the PLD.

Abstract: A programmable chip design tool is provided to enumerate and specify the security and/or redundancy constraints of a programmable chip design. A design is implemented with a high-level security or redundancy scheme, and the programmable chip design tool applies the scheme while simultaneously optimizing for desired metrics (logic density, routability, timing, power, etc.). An underlying assignment scheme as well as user interface components used to enter this assignment scheme are provided.

Abstract: Techniques for reconfiguring an integrated circuit (IC) are provided. The techniques may improve error detection in the partially reconfigurable IC. A cyclic redundancy check (CRC) value for a first configuration data is received by the IC and a second configuration data is generated based on the first configuration data and a prior configuration data stored in the IC. The first configuration data may be a partial reconfiguration data that is used to reconfigure at least a portion of the IC. A third configuration data is then generated based on the first and second configuration data and the prior configuration data. A second CRC value is calculated based on the third configuration data. The second CRC value, together with the first CRC value and a prior CRC value stored in the IC, is used to calculate an updated CRC value. The updated CRC value is stored in the IC.

Abstract: In an exemplary embodiment, an apparatus includes a first set of circuit elements and a second set of circuit elements. The first set of circuit elements is used in a first configuration of the apparatus, and the second set of circuit elements is used in a second configuration of the apparatus. The first configuration of the apparatus is switched to the second configuration of the apparatus in order to improve reliability of the apparatus.

Abstract: A programmable logic device includes logic blocks such as a logic array blocks (LAB) that can be configured as a random access memory (RAM) or as a lookup table (LUT). A mode flag is provided to indicate the mode of operation of configuration logic such as a configuration RAM (CRAM) used during partial reconfiguration of a logic block. An enable read flag is provided to indicate if values stored in the configuration logic are to be read out or a known state is to be read out during a data verification process. Thus, exclusion and inclusion of portions of a region of configuration logic from data verification and correction processes allow a region of configuration logic to store both a design state and a user defined state. Moreover, the region of configuration logic may be dynamically reconfigured from one state to another without causing verification errors.

Abstract: Circuits and methods to generate a True Random Number Generator (TRNG) with tamper-detection are presented. In one embodiment, the circuit includes two identical TRNG circuits and logic circuitry that combines and correlates the outputs of the two TRNG circuits. The two identical TRNG circuits are located in close proximity to each other inside an Integrated Circuit (IC). The logic circuitry analyzes the outputs of the two TRNG circuits and the historical values of the relation between the outputs of the two TRNG circuits to determine if the outputs are correlated. If the outputs are not correlated, the logic circuitry outputs a true random number sequence based on the combination of the two TRNG circuits. As a result, circuit tampering, such as changes in temperature or voltage supplies, is detected in the IC.

Abstract: Systems and methods are disclosed for preventing tampering of a programmable integrated circuit device. Generally, programmable devices, such as FPGAs, have two stages of operation; a configuration stage and a user mode stage. To prevent tampering and/or reverse engineering of a programmable device, various anti-tampering techniques may be employed during either stage of operation to disable the device and/or erase sensitive information stored on the device once tampering is suspected. One type of tampering involves bombarding the device with a number of false configuration attempts in order to decipher encrypted data. By utilizing a dirty bit and a sticky error counter, the device can keep track of the number of failed configuration attempts that have occurred and initiate anti-tampering operations when tampering is suspected while the device is still in the configuration stage of operation.

Abstract: Methods, circuits, and systems for preventing data remanence in memory systems are provided. Original data is stored in a first memory, which may be a static random access memory (SRAM). Data is additionally stored in a second memory. Data in the first memory is periodically inverted, preventing data remanence in the first memory. The data in the second memory is periodically inverted concurrently with the data in the first memory. The data in the second memory is used to keep track of the inversion state of the data in the first memory. The original data in the first memory can be reconstructed performing a logical exclusive-OR operation between the data in the first memory and the data in the second memory.

Abstract: Systems and methods are disclosed for securing a programmable integrated circuit device against an over-voltage attack. Generally, programmable devices, such as FPGAs, contain volatile memory registers that may store sensitive information. To prevent tampering and/or reverse engineering of such a programmable device, an over-voltage detection circuit may be employed to disable the device and/or erase the sensitive information stored on the device when an over-voltage attack is suspected. In particular, once the over-voltage detection circuit detects that the voltage applied to the programmable device exceeds a trigger voltage, it may cause logic circuitry to erase the sensitive information stored on the device. Desirably, the over-voltage detection circuit includes components arranged in such a way as to render current consumption negligible when the voltage applied to the programmable device, e.g., by a battery, remains below the trigger voltage.

Abstract: A kill switch is provided that, when triggered, may cause the programmable logic device (PLD) to become at least partially reset, disabled, or both. The kill switch may be implemented as a fuse or a volatile battery-backed memory bit. When, for example, a security threat is detected, the switch may be blown, and a reconfiguration of the device initiated in order to zero or clear some or all of the memory and programmable logic of the PLD.

Abstract: Systems and methods are provided for destroying or erasing circuitry elements, data, or both, such as transistors, volatile keys, or fuse blocks, located in an integrated circuit device. An initiation signal may be provided to induce latch-up in a circuitry element in response to a user command, a tampering event, or both. As a result of the latch-up effect, the circuitry element, data, or both may be destroyed or erased.

Abstract: Techniques of the present invention impede power consumption measurements of an encryption engine on a logic device by running the encryption engine with an independent clock. This clock produces a signal that is decoupled from and asynchronous to clock signals feeding other circuits on the device. The clock feeding the encryption engine is not accessible externally to the device. Circuits may be employed to intentionally slow down or add jitter to one or more of the clock signals.

Abstract: Memory elements are provided that exhibit immunity to soft error upset events when subjected to high-energy atomic particle strikes. The memory elements may each have ten transistors including two address transistors and four transistor pairs that are interconnected to form a bistable element. Clear lines such as true and complement clear lines may be routed to positive power supply terminals and ground power supply terminals associated with certain transistor pairs. During clear operations, some or all of the transistor pairs can be selectively depowered using the clear lines. This facilitates clear operations in which logic zero values are driven through the address transistors and reduces cross-bar current surges.

Abstract: Memory elements are provided that exhibit immunity to soft error upset events when subjected to high-energy atomic particle strikes. The memory elements may each have ten transistors including two address transistors and four transistor pairs that are interconnected to form a bistable element. Clear lines such as true and complement clear lines may be routed to positive power supply terminals and ground power supply terminals associated with certain transistor pairs. During clear operations, some or all of the transistor pairs can be selectively depowered using the clear lines. This facilitates clear operations in which logic zero values are driven through the address transistors and reduces cross-bar current surges.

Abstract: A metastability-hardened storage circuit includes at least one inverting circuit. The inverting circuit has a logical input. The logical input of the inverting circuit is split into a pair of physical inputs.

Abstract: A metastability-hardened storage circuit includes at least one inverting circuit. The inverting circuit has a logical input. The logical input of the inverting circuit is split into a pair of physical inputs.

Abstract: A metastability-hardened storage circuit includes at least one inverting circuit. The inverting circuit has a logical input. The logical input of the inverting circuit is split into a pair of physical inputs.

Abstract: Memory elements are provided that exhibit immunity to soft error upset events when subjected to high-energy atomic particle strikes. The memory elements may each have ten transistors including two address transistors and four transistor pairs that are interconnected to form a bistable element. Clear lines such as true and complement clear lines may be routed to positive power supply terminals and ground power supply terminals associated with certain transistor pairs. During clear operations, some or all of the transistor pairs can be selectively depowered using the clear lines. This facilitates clear operations in which logic zero values are driven through the address transistors and reduces cross-bar current surges.

Abstract: Partial reconfiguration techniques and reconfiguration circuitry are provided that allow portions of a memory cell array to be reconfigured with new reconfiguration data without disturbing other portions of the memory cell array. The memory cells may be loaded with configuration data on an integrated circuit such as a programmable logic device. Memory cell outputs may configure programmable logic. To avoid disturbing programmable logic operations for programmable logic that is unaffected by the reconfigured cells during reconfiguration, unaffected memory cells are not unnecessarily cleared. Only those memory cells that need to be cleared to conform to the new configuration data that is being loaded into the array need to be loaded with logic zero values during reconfiguration operations. After these clearing operations are complete, set operations may be performed to convert appropriate memory cells to logic one values to match the new configuration data.