Edit

Storage Class Driver's RetryRequest Routine

  • Article

The underlying storage port driver is responsible for retrying requests if device errors occur that involve transmission of data on the bus, including bus-parity errors, selection time-outs, and target/controller-busy errors. If retry attempts fail, the storage port driver completes the request with an appropriate error and logs the I/O error, as well.

A storage class driver should never attempt to retry a request that the port driver has already failed due to any of the preceding errors.

A storage class driver is responsible for retrying requests that fail due to device-specific errors, target/controller errors other than target/controller-busy, bus resets, or request time-outs. In general, a RetryRequest routine can resubmit such an IRP to the next-lower driver and direct that the SRB be placed at the head of the port driver's LU-specific queue.

In particular, a RetryRequest routine should do the following:

  1. Ensure that a partial transfer request has the correct values set for the starting address and length.

  2. Zero the SrbStatus and ScsiStatus members of the SRB.

  3. Set up the SrbFlags member, as necessary for the device.

  4. Set up the I/O stack location for the port driver in the IRP as already described in Storage Class Driver's Dispatch Routines through Storage Class Driver's SplitTransferRequest Routine.

  5. Call IoSetCompletionRoutine for the IRP, because the driver's IoCompletion routine must free the SRB before the IRP returns. The IoCompletion routine might also need to retry the request more than once, or to call the driver's InterpretRequestSense or ReleaseQueue routine.

  6. Pass the request on to the next-lower driver with IoCallDriver.