On Mon, Sep 13, 2021 at 10:17:48AM +0200, Jan H??ppner wrote: > I think, just like with some of the other changes, there is some > cleanup required before returning. I'll prepare a patch and > come back to you. If you are touching the dasd probe path anyway, can you look insto switching to use blk_mq_alloc_disk as well? Right now dasd allocates the request_queue and gendisk separately in different stages of initialization, but unlike SCSI which needs to probe using just the request_queue I can't find a good reason for that.