Methods and apparatus for automatic fault detection
View Patent ↗Techniques and mechanisms are provided to monitor signals including critical signals at the endpoints, or leaves, of one or more signal trees in an integrated circuit device. Sensors or layers of sensors may be configured in fault detection circuitry to monitor signals and compare them to static or dynamically varying values. The fault detection circuits may include OR-gate daisy chains that output a fault detection signal to control circuitry if any signal at a particular leaf deviates from an expected signal. Fault detection circuits may also be configured to identify instances where two or more or N or more signals deviate from an expected signal. Mechanisms may also be provided to assure the reliability of fault detection circuitry itself.
1. An integrated circuit device comprising:
control logic operable to send a signal onto a signal tree included on the integrated circuit device, the signal tree comprising branches and leaves, wherein the branches and leaves include columns and rows of circuit lines configured to carry signals from a signal source to signal endpoints, wherein each leaf of the signal tree comprises a signal endpoint;
user logic operable to implement a user design on the integrated circuit device; and
fault detection circuitry on the integrated circuit device operable to monitor, using sensors, a plurality of signal values at each of the leaves on the signal tree such that the fault detection circuitry can detect faults at any point along the branches of the signal tree, wherein the fault detection circuitry is operable to generate a fault detect signal having a first value in response to determining that the first signal value of the plurality of signal values deviates from an expected signal value, and wherein the control logic implements a fault handling procedure upon receiving the fault detect signal having the first value.
2. The integrated circuit device of claim 1 , wherein the fault handling procedure comprises resetting the device.
3. The integrated circuit device of claim 1 , wherein the fault handling procedure comprises deleting cryptographic keys.
4. The integrated circuit device of claim 1 , wherein the signal is a test enable signal.
5. The integrated circuit device of claim 1 , wherein the signal is an address and data configuration signal.
6. The integrated circuit device of claim 1 , wherein the expected value is maintained in a fuse or configuration random access memory (CRAM).
7. The integrated circuit device of claim 1 , wherein the expected value is a variable and is provided by the control logic.
8. The integrated circuit device of claim 1 , wherein the expected value is a variable and is provided by user logic.
9. The integrated circuit device of claim 1 , wherein fault detection circuitry comprises a daisy chain of OR gates.
10. The integrated circuit device of claim 9 , wherein fault detection circuitry comprises layers of a plurality of sensors.
11. A method comprising:
monitoring a plurality of signals at a plurality of leaves corresponding to a plurality of signal trees on an integrated circuit device, the plurality of signal trees comprising branches and leaves, wherein the branches and leaves include columns and rows of circuit lines configured to carry signals from a signal source to signal endpoints, wherein each leaf of a signal tree in the plurality of signal trees comprises a signal endpoint, such that fault detection circuitry included on the integrated circuit device can detect faults at any point along the branches of a signal tree in the plurality of signal trees using sensors that monitor each of the leaves in the plurality of signal trees;
determining at the fault detection circuitry included on the integrated circuit device whether a first signal deviates from an expected value;
generating a fault detect signal with a first value in response to value of said first signal deviating from value of said expected value and a second value otherwise;
determining whether the fault detect signal should be masked; and
implementing a fault handling procedure using control logic associated with the device.
12. The method of claim 11 , wherein the fault handling procedure comprises resetting the device.
13. The method of claim 11 , wherein the fault handling procedure comprises deleting cryptographic keys.
14. The method of claim 11 , wherein the plurality of signals includes a test enable signal.
15. The method of claim 11 , wherein the plurality of signals includes an address and data configuration signal.
16. The method of claim 11 , wherein the expected value is maintained in a fuse or configuration random access memory (CRAM).
17. The method of claim 11 , wherein the expected value is a variable provided by control logic.
18. The method of claim 11 , wherein the expected value is a variable provided by user logic.
19. The method of claim 11 , wherein fault detection circuitry comprises a daisy chain of OR gates.
20. A system, comprising:
a processor;
memory coupled to the processor;
an integrated circuit device including control logic operable to send a signal onto a signal tree included on the integrated circuit device, the signal tree comprising branches and leaves, wherein the branches and leaves include columns and rows of circuit lines configured to carry signals from a signal source to signal endpoints, wherein each leaf of the signal tree comprises a signal endpoint, user logic operable to implement a user design on the integrated circuit device, and fault detection circuitry on the integrated circuit device operable to monitor, using sensors, a plurality of signal values at each of the leaves on the signal tree such that the fault detection circuitry can detect faults at any point along the branches of the signal tree, wherein the fault detection circuitry is operable to generate a fault detect signal having a first value in response to determining that the first signal value of the plurality of signal values deviates from an expected signal value, and wherein the control logic implements a fault handling procedure upon receiving the fault detect signal having the first value.