nvmet: fix nvme status code when namespace is disabled
authorSagi Grimberg <sagi@grimberg.me>
Sun, 28 Apr 2024 09:25:40 +0000 (12:25 +0300)
committerKeith Busch <kbusch@kernel.org>
Wed, 1 May 2024 09:58:43 +0000 (02:58 -0700)
commit505363957fad35f7aed9a2b0d8dad73451a80fb5
tree9001d58a8cfb109cd948308d5d0d4c0c7dd539df
parent6825bdde44340c5a9121f6d6fa25cc885bd9e821
nvmet: fix nvme status code when namespace is disabled

If the user disabled a nvmet namespace, it is removed from the subsystem
namespaces list. When nvmet processes a command directed to an nsid that
was disabled, it cannot differentiate between a nsid that is disabled
vs. a non-existent namespace, and resorts to return NVME_SC_INVALID_NS
with the dnr bit set.

This translates to a non-retryable status for the host, which translates
to a user error. We should expect disabled namespaces to not cause an
I/O error in a multipath environment.

Address this by searching a configfs item for the namespace nvmet failed
to find, and if we found one, conclude that the namespace is disabled
(perhaps temporarily). Return NVME_SC_INTERNAL_PATH_ERROR in this case
and keep DNR bit cleared.

Reported-by: Jirong Feng <jirong.feng@easystack.cn>
Tested-by: Jirong Feng <jirong.feng@easystack.cn>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
drivers/nvme/target/configfs.c
drivers/nvme/target/core.c
drivers/nvme/target/nvmet.h