[PATCH 1/2] nvmet: fix oops in nvmet_execute_passthru_cmd()

[Chaitanya Kulkarni]

On 8/6/20 12:41 PM, Sagi Grimberg wrote:
>> Christoph, Sagi how guys would like to have this fix?
>>
>> I personally like to avoid introducing new labels unless they are
>> extremely necessary.
> I think that having an explicit and reverse-ordered target error labels
> are always better.
> 

Okay will send v2 shortly, although it dose create an inconsistency 
between a call to nvmet_req_complete() and blk_put_request()  in 
nvmet_execute_passthru_cmd() and nvmet_passthru_req_done() which I
didn't understand why we should allow it.