Multichip fault management

A network of interconnected processing circuit chips, a processing circuit chip in the network comprising:

• an internal signal ring;
• an external signal ring;
• and
• a plurality of circuits coupled to the internal and external signal rings, the plurality of circuits configured to:
• capture, via the external signal ring, external fault data originating from a first of the processing circuit chips in the network and comprising information about a fault detected at the first processing circuit chip,
• pass the captured external fault data via the external signal ring to a first communication path outside of the processing circuit chip for conveyance of the external fault data to one of the processing circuit chips in the network,
• transfer, from the internal signal ring to the external signal ring, internal fault data originating from an internal circuit of the processing circuit chip and comprising information about another fault detected at the internal circuit,
• pass the internal fault data via the external signal ring to a second communication path outside of the processing circuit chip for conveyance of the internal fault data to a second of the processing circuit chips,
• receive, from the second processing circuit chip, a query for determining an origin of the other fault;
• and
• transfer the query via the external signal ring to a control circuit of the processing circuit chip, the control circuit configured to:
• determine a validity of the query, and
• transfer the query from the control circuit to the internal signal ring for conveyance of the query via the internal signal ring to at least one data processing circuit of the processing circuit chip, in response to the determined validity indicating the query is intended for the processing circuit chip.

A device of a network of interconnected devices, the device comprising:

• an internal signal ring;
• an external signal ring;
• and
• a plurality of circuits coupled to the internal signal ring and the external signal ring, the plurality of circuits configured to:
• capture, via the external signal ring, external fault data originating from another device in the network and comprising information about a fault detected at the other device,
• pass the captured external fault data via the external signal ring to a first communication path outside of the device for conveyance of the external fault data to one of the devices in the network,
• transfer, from the internal signal ring to the external signal ring, internal fault data originating from an internal circuit of the device and comprising information about another fault detected at the internal circuit,
• pass the internal fault data via the external signal ring to a second communication path outside of the device for conveyance of the internal fault data to a host device in the network,
• receive, from the host device, a query for determining an origin of the fault, and
• transfer the query via the external signal ring to a control circuit of the device, the control circuit configured to:
• determine a validity of the query,
• transfer the query via the external signal ring to an inter-chip circuit of the plurality of circuits based on routing data, in response to the determined validity indicating the query is not intended for the device, and
• pass the query from the circuit to a third communication path outside of the device for conveyance of the query to the other device.