Dynamic optical channel sparing in an industry standard input/output subsystem
Mechanisms, in a data processing system comprising an input/output subsystem implementing an industry standard optical bus, for handling a failure of an optical channel in an optical bus are provided. The mechanisms detect, by failure detection logic of the input/output (I/O) subsystem, failure of an optical channel of the optical bus. The mechanisms send, by a controller of the I/O subsystem, a control signal to channel routing logic of the I/O subsystem to control a routing of data signals between active bus lanes of the data processing system and optical channels of the optical bus in response to detecting the failure of the optical channel. The mechanisms control, by the channel routing logic, routing of data signals between the active bus lanes and the optical channels so as to remove the failed optical channel from further use and use a spare optical channel instead of the failed optical channel.
1. A method, in a data processing system comprising an input/output subsystem implementing an optical bus, for handling a failure of an optical channel in the optical bus, the method comprising:
detecting, by failure detection logic of the input/output (I/O) subsystem of the data processing system, the failure of the optical channel of the optical bus;
sending, by the failure detection logic to a controller of the I/O subsystem, a signal indicative of the failure;
sending, by the controller to register update logic, a control signal, wherein the control signal is in response to the signal indicative of the failure;
loading, by the register update logic in response to the control signal from the controller, a mux selection code adjusted for the failure detected by the failure detection logic;
sending, by the register update logic to the I/O subsystem, a control signal indicative of the mux selection code to update a fault register;
sending, by the controller of the I/O subsystem, a control signal to channel routing logic of the I/O subsystem to control a routing of data signals between active bus lanes of the data processing system and the optical channels of the optical bus in response to detecting the failure of the optical channel; and
controlling, by the channel routing logic of the I/O subsystem, routing of the data signals between the active bus lanes and the optical channels so as to remove the failed optical channel from further use and use a spare optical channel instead of the failed optical channel.
2. The method of claim 1 , wherein the optical bus is an optical bus that implements Peripheral Component Interconnect (PCI) or Peripheral Component Interconnect Express (PCIe).
3. The method of claim 1 , wherein the channel routing logic comprises a plurality of multiplexers, and wherein sending the control signal to the channel routing logic comprises sending a multiplexer select control signal to at least a subset of multiplexers in the plurality of multiplexers to change a switching of the subset of multiplexers.
4. The method of claim 1 , wherein detecting the failure of the optical channel comprises monitoring at least one characteristic of the optical channel and comparing the at least one characteristic to a failure detection criteria that indicates a predetermined criteria indicative of a failure of the optical channel.
5. The method of claim 4 , wherein the at least one characteristic is at least one of a power of a laser associated with the optical channel, a bias alert signal output associated with the optical channel, or a register state of a fault register associated with the optical channel.
6. The method of claim 1 , wherein detecting the failure of the optical channel comprises comparing a laser power of a laser associated with the optical channel to a predetermined threshold power value and detecting the failure in response to the laser power of the laser falling below the predetermined threshold power value.
7. The method of claim 1 , wherein a bandwidth of the active bus lanes remains constant after detection of the failure of the optical channel due to the controlling of the routing of the data signals between the active bus lanes and the optical channels.
8. The method of claim 1 , wherein the channel routing logic is integrated into one of a serializer/deserializer (SerDes) circuit of the I/O subsystem, an optical transceiver module of the I/O subsystem, a separate printed circuit board (PCB) from the controller and the failure detection logic, in firmware of the I/O subsystem, or in a field programmable gate array (FPGA) of the I/O subsystem.
9. The method of claim 1 , wherein controlling routing of the data signals between the active bus lanes and the optical channels comprises switching an operation of at least one multiplexer of the channel routing logic from routing the data signals through a first signal path through the at least one multiplexer to routing the data signals through a second signal path through the at least one multiplexer.
10. The method of claim 9 , wherein the channel routing logic comprises a plurality of multiplexers and the switching of the operation of at least one multiplexer comprises switching the operation of a sub-set of multiplexers within the plurality of multiplexers.