linux.git/drivers/nvme, branch v5.16-rc4

nvmet: use IOCB_NOWAIT only if the filesystem supports it

2021-11-25T14:02:40+00:00

Submit I/O requests with the IOCB_NOWAIT flag set only if
the underlying filesystem supports it.

Fixes: 50a909db36f2 ("nvmet: use IOCB_NOWAIT for file-ns buffered I/O")
Signed-off-by: Maurizio Lombardi 
Reviewed-by: Chaitanya Kulkarni 
Signed-off-by: Christoph Hellwig

nvme: fix write zeroes pi

2021-11-23T16:22:41+00:00

Write Zeroes sets PRACT when block integrity is enabled (as it should),
but neglects to also set the reftag which is expected by reads. This
causes protection errors on reads.

Fix this by setting the reftag for type 1 and 2 (for type 3, reads will
not check the reftag).

Signed-off-by: Klaus Jensen 
Reviewed-by: Martin K. Petersen 
Signed-off-by: Christoph Hellwig

nvme-fabrics: ignore invalid fast_io_fail_tmo values

2021-11-23T16:22:41+00:00

Valid fast_io_fail_tmo values are integers >= 0 or -1 (disabled).
Prevent userspace from setting arbitrary negative values.

Signed-off-by: Maurizio Lombardi 
Reviewed-by: Sagi Grimberg 
Reviewed-by: Chaitanya Kulkarni 
Signed-off-by: Christoph Hellwig

nvme-pci: add NO APST quirk for Kioxia device

2021-11-23T16:22:41+00:00

This particular Kioxia device times out and aborts I/O during any load,
but it's more easily observable with discards (fstrim).

The device gets to a state that is also not possible to use
"nvme set-feature" to disable APST.
Booting with nvme_core.default_ps_max_latency=0 solves the issue.

We had a dozen or so of these devices behaving this same way in
customer environments.

Signed-off-by: Enzo Matsumiya 
Signed-off-by: Christoph Hellwig

nvme-tcp: fix memory leak when freeing a queue

2021-11-23T16:22:41+00:00

Release the page frag cache when tearing down the io queues

Signed-off-by: Maurizio Lombardi 
Reviewed-by: Sagi Grimberg 
Reviewed-by: John Meneghini 
Signed-off-by: Christoph Hellwig

nvme-tcp: validate R2T PDU in nvme_tcp_handle_r2t()

2021-11-23T16:22:40+00:00

If maxh2cdata < r2t_length then driver will form multiple
H2CData PDUs, validate R2T PDU in nvme_tcp_handle_r2t() to
reuse nvme_tcp_setup_h2c_data_pdu().

Also set req->state to NVME_TCP_SEND_H2C_PDU in
nvme_tcp_setup_h2c_data_pdu().

Signed-off-by: Varun Prakash 
Reviewed-by: Sagi Grimberg 
Signed-off-by: Christoph Hellwig

nvmet-tcp: fix incomplete data digest send

2021-11-23T16:22:40+00:00

Current nvmet_try_send_ddgst() code does not check whether
all data digest bytes are transmitted, fix this by returning
-EAGAIN if all data digest bytes are not transmitted.

Fixes: 872d26a391da ("nvmet-tcp: add NVMe over TCP target driver")
Signed-off-by: Varun Prakash 
Reviewed-by: Sagi Grimberg 
Signed-off-by: Christoph Hellwig

nvmet-tcp: fix memory leak when performing a controller reset

2021-11-23T16:19:25+00:00

If a reset controller is executed while the initiator
is performing some I/O the driver may leak the memory allocated
for the commands' iovec.

Make sure that nvmet_tcp_uninit_data_in_cmds() releases
all the memory.

Signed-off-by: Maurizio Lombardi 
Reviewed-by: Keith Busch 
Reviewed-by: Sagi Grimberg 
Reviewed-by: John Meneghini 
Signed-off-by: Christoph Hellwig

nvmet-tcp: add an helper to free the cmd buffers

2021-11-23T16:19:25+00:00

Makes the code easier to read and to debug.

Sets the freed pointers to NULL, it will be useful
when destroying the queues to understand if the commands'
buffers have been released already or not.

Signed-off-by: Maurizio Lombardi 
Reviewed-by: Keith Busch 
Reviewed-by: Sagi Grimberg 
Reviewed-by: John Meneghini 
Signed-off-by: Christoph Hellwig

nvmet-tcp: fix a race condition between release_queue and io_work

2021-11-23T16:19:25+00:00

If the initiator executes a reset controller operation while
performing I/O, the target kernel will crash because of a race condition
between release_queue and io_work;
nvmet_tcp_uninit_data_in_cmds() may be executed while io_work
is running, calling flush_work() was not sufficient to
prevent this because io_work could requeue itself.

Fix this bug by using cancel_work_sync() to prevent io_work
from requeuing itself and set rcv_state to NVMET_TCP_RECV_ERR to
make sure we don't receive any more data from the socket.

Signed-off-by: Maurizio Lombardi 
Reviewed-by: Keith Busch 
Reviewed-by: Sagi Grimberg 
Reviewed-by: John Meneghini 
Signed-off-by: Christoph Hellwig