dm mpath: remove unnecessary NVMe branching in favor of scsi_dh checks
authorMike Snitzer <snitzer@redhat.com>
Mon, 5 Mar 2018 19:10:11 +0000 (14:10 -0500)
committerMike Snitzer <snitzer@redhat.com>
Wed, 7 Mar 2018 01:23:58 +0000 (20:23 -0500)
commit8d47e65948ddea4398892946d9e50778a316b397
treecb699eae23c76bd9378e2e75de250e545b9ee2ad
parent99243b922c9ddb4976b8db2eeffb0aed6e06c6f9
dm mpath: remove unnecessary NVMe branching in favor of scsi_dh checks

This eliminates the "queue_mode" configuration's "nvme" mode.  There
wasn't anything NVMe-specific about that mode.  It was named "nvme"
because it was a short name for the mode.  But the entire point of the
mode was to optimize the multipath target for underlying devices that
are _not_ SCSI-based.  Devices that aren't SCSI have no need for the
various SCSI device handler (scsi_dh) specific code in DM multipath.

But rather than narrowly define this scsi_dh vs not branching in terms
of "nvme": invert the logic so that we're just checking whether a
multipath device is layered on SCSI devices with scsi_dh attached.

This allows any future storage technology to avoid scsi_dh specific code
in the multipath target too.

Fixes: 848b8aefd4 ("dm mpath: optimize NVMe bio-based support")
Suggested-by: Mikulas Patocka <mpatocka@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
drivers/md/dm-mpath.c