On 2/28/22 12:05 PM, Lee Duncan wrote: >> - if (shost->hostt->scan_finished) { >> - if (scsi_queue_work(shost, &session->scan_work)) >> - atomic_inc(&ihost->nr_scans); >> - } >> ISCSI_DBG_TRANS_SESSION(session, "Completed unblocking session\n"); >> } .... > > I have no issue with this, but it seems kind of unrelated to speeding up session unblocking It's related because in the next patch we stop creating the host's transport wq which is used above in the chunk that was removed in __iscsi_unblock_session.