[PATCH] nvme: disable namespaces with unsupported metadata

[Keith Busch]

The only fabrics target that supports metadata is RDMA and usable only
if it is 8B per block, and formatted for protection information. If an
rdma target were to export a namespace with a different format (ex:
4k+64B), the driver will not be able to submit a valid command for that
format.

Suppress exporting these namespaces as read/writeable to the block
stack.

Cc: Max Gurtovoy <mgurtovoy at nvidia.com>
Cc: Sagi Grimberg <sagi at grimberg.me>
Signed-off-by: Keith Busch <kbusch at kernel.org>
---
 drivers/nvme/host/core.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index 4b5de8f5435a..3afbf73639c3 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1740,7 +1740,16 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)
 		 */
 		if (WARN_ON_ONCE(!(id->flbas & NVME_NS_FLBAS_META_EXT)))
 			return -EINVAL;
-		if (ctrl->max_integrity_segments)
+
+		/*
+		 * Currently rdma is the only fabrics nvme host that supports
+		 * metadata integrity segments, but will not use the metadata
+		 * payload if it doesn't satisfy 'nvme_ns_has_pi()'. Since the
+		 * initiator is not able to form requests that satisify the
+		 * format, do not support the namespace if the metadata does
+		 * not satisfy that requirement.
+		 */
+		if (ctrl->max_integrity_segments && nvme_ns_has_pi(ns))
 			ns->features |=
 				(NVME_NS_METADATA_SUPPORTED | NVME_NS_EXT_LBAS);
 	} else {
-- 
2.25.4