On Thu, Jul 14, 2022 at 05:30:51PM +0200, Daniel Wagner wrote:
>On Thu, Jul 14, 2022 at 01:49:23PM +0530, Kanchan Joshi wrote:
>> If path is not available, retry is not done immediately rather we wait for
>> path to be available (as underlying controller may still be
>> resetting/connecting). List helped as command gets added into
>> it (and submitter/io_uring gets the control back), and retry is done
>> exact point in time.
>> But yes, it won't harm if we do couple of retries even if path is known
>> not to be available (somewhat like iopoll). As this situation is
>> not common. And with that scheme, we don't have to link io_uring_cmd.
>
>Stupid question does it only fail over immediately when the path is not
>available or any failure? If it fails over for everything it's possible
>the target gets the same request twice. FWIW, we are just debugging this
>scenario right now.

failover is only for path-related failure, and not otherwise.
you might want to take a look at nvme_decide_disposition routine where
it makes that decision.