[PATCHv3] nvme: disable namespace access for unsupported metadata

Keith Busch kbusch at kernel.org
Tue Nov 30 08:14:54 PST 2021


The only fabrics target that supports metadata handling through the
separate integrity buffer is RDMA. It is currently usable only if the
size 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 valid read/write commands
for that namespace.

Suppress setting the metadata feature in the namespace so that the
gendisk capacity will be set to 0. This will prevent read/write access
through the block stack, but will continue to allow ioctl passthrough
commands.

Cc: Max Gurtovoy <mgurtovoy at nvidia.com>
Cc: Sagi Grimberg <sagi at grimberg.me>
Signed-off-by: Keith Busch <kbusch at kernel.org>
---
v2->v3:

  Retain support for the EXT_LBAS features for ioctl passthrough use.
  The namespace format is usable through that interface since user space
  provides a virtually contiguous buffer with the metadata already spliced
  in.

 drivers/nvme/host/core.c | 17 ++++++++++++++---
 1 file changed, 14 insertions(+), 3 deletions(-)

diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index f82c098b1a61..8ea3f3d3a279 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -1749,9 +1749,20 @@ 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)
-			ns->features |=
-				(NVME_NS_METADATA_SUPPORTED | NVME_NS_EXT_LBAS);
+
+		ns->features |= NVME_NS_EXT_LBAS;
+
+		/*
+		 * The current fabrics transport drivers support namespace
+		 * metadata formats only if nvme_ns_has_pi() returns true.
+		 * Suppress support for all other formats so the namespace will
+		 * have a 0 capacity and not be usable through the block stack.
+		 *
+		 * Note, this check will need to be modified if any drivers
+		 * gain the ability to use other metadata formats.
+		 */
+		if (ctrl->max_integrity_segments && nvme_ns_has_pi(ns))
+			ns->features |= NVME_NS_METADATA_SUPPORTED;
 	} else {
 		/*
 		 * For PCIe controllers, we can't easily remap the separate
-- 
2.25.4




More information about the Linux-nvme mailing list