Multi-processor device with external interface failover

1 . A multi-processor integrated circuit (IC) chip, comprising:

interface circuitry to receive requests from at least one host;

a primary processor coupled to the interface circuitry during a normal mode of operation, the primary processor to process the requests received over the interface circuitry during the normal mode of operation;

a secondary processor selectively coupled to the interface circuitry;

failure detection circuitry coupled to the primary processor to detect a failure event associated with the primary processor and to initiate a failure mode of operation upon detection of the failure event;

wherein during the failure mode of operation, the interface circuitry decouples from the primary processor and directs requests received from the at least one host directly to the secondary processor;

wherein during the failure mode of operation, the secondary processor receives updated firmware from the at least one host; and

wherein the secondary processor performs recovery operations on the primary processor during the failure mode of operation comprising re-booting the primary processor using the received updated firmware.

2 . The multi-processor IC chip of claim 1 , wherein the secondary processor comprises:

a secure processor to perform cryptographic operations on the requests in an absence of the failure event associated with the primary processor.

3 . The multi-processor IC chip of claim 2 , wherein the secure processor comprises:

a hardware root-of-trust processor utilizing Public Key Infrastructure (PKI).

4 . The multi-processor IC chip of claim 1 , wherein the interface circuitry comprises:

a compute express link (CXL) interface managed by the primary processor; and

a sideband interface independent of the CXL interface.

5 . The multi-processor IC chip of claim 1 , wherein the failure detection circuitry is incorporated into the secondary processor.

6 . The multi-processor IC chip of claim 1 , wherein the failure detection circuitry comprises:

a timer circuit to detect timeout information indicative of a failure event.

7 . The multi-processor IC chip of claim 1 , wherein the failure detection circuitry comprises:

a heartbeat status register to store a value that is periodically updated by the primary processor in the normal mode of operation.

8 . The multi-processor IC chip of claim 1 , embodied as a CXL buffer chip.

9 . A multi-processor device, comprising:

external interface circuitry to receive requests from an external device;

a primary processor coupled to the external interface circuitry to process the requests in an absence of a failure event associated with the primary processor;

a secondary processor selectively coupled to the external interface circuitry, the secondary processor to selectively receive the requests from the external interface circuitry based on detection of the failure event associated with the primary processor;

wherein upon detection of the failure event, the external interface circuitry is configured to direct updated firmware received from the external device directly to the secondary processor; and

wherein the secondary processor performs recovery operations on the primary processor comprising re-booting the primary processor using the received updated firmware.

10 . The multi-processor device of claim 9 , further comprising:

failure detection circuitry coupled to the primary processor to detect the failure event.

11 . The multi-processor device of claim 10 , wherein the failure detection circuitry comprises:

a timer circuit to detect timeout information indicative of a failure event.

12 . The multi-processor device of claim 10 , wherein the failure detection circuitry comprises:

a heartbeat status register to store a value that is periodically updated by the primary processor in the absence of the failure event associated with the primary processor.

13 . The multi-processor device of claim 9 , wherein the secondary processor comprises:

a secure processor to perform cryptographic operations on the requests in the absence of the failure event associated with the primary processor.

14 . The multi-processor device of claim 13 , wherein the secure processor comprises:

a hardware root-of-trust processor utilizing Public Key Infrastructure (PKI).

15 . The multi-processor device of claim 9 , wherein the external interface circuitry comprises:

a compute express link (CXL) interface managed by the primary processor; and

a sideband interface independent of the CXL interface.

16 . The multi-processor device of claim 9 , wherein the multi-processor device is embodied as a single integrated circuit (IC) chip.

17 . The multi-processor device of claim 16 , wherein the single IC chip comprises:

a CXL buffer chip.

18 . A method of operation in a multi-processor device, the multi-processor device including an external interface, a primary processor, and a secondary processor, the method comprising:

operating the multi-processor device in a normal mode of operation such that the primary processor processes requests received from an external device by the external interface, and the secondary processor processes operations on behalf of the primary processor;

detecting a failure event associated with the primary processor;

operating the multi-processor device in a failure mode of operation upon detecting the failure event such that the external interface is configured to direct updated firmware received from the external device directly to the secondary processor; and

performing, by the secondary processor, recovery operations on the primary processor comprising re-booting the primary processor using the received updated firmware.

19 . The method of claim 18 , further comprising:

during the failure mode of operation, performing, by the secondary processor, a cryptographic signature verification before receiving the updated firmware from the external device.

20 . The method of claim 18 , wherein:

the performing recovery operations is carried out in response to commands from an external device.