On Fri, Apr 28, 2017 at 10:54:27AM -0600, Jens Axboe wrote: > OK, I think we're getting there now. I dropped the RQF_RESERVED > and the tag helper, since mtip32xx can just call > blk_rq_is_passthrough() and know it's an internal request. > > We can do another cleanup series of mtip32xx making the > internal requests use the normal blk end_io infrastructure, > and use the provider helpers for mapping data. I think > that should be done separately as a cleanup, keeping this series > as a straight forward conversion. > > I've tested this on the following card: > > [ 56.394162] mtip32xx 0000:81:00.0: Model: P320h-MTFDGAR350SAH > > and it works fine for me. Tested this patchset against the latest linus tree, works fine on my system with mq-deadline scheduler enabled. Tested-by: Ming Lei <ming.lei@xxxxxxxxxx> Thanks, Ming
