Split nvme_rdma_alloc_tagset into one helper for the admin tag_set and
one for the I/O tag set.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Split nvme_dev_add into a helper to actually allocate the tag set, and
one that just update the number of queues. Add a local variable for
the tag_set to clean up the code a bit.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Split nvme_alloc_admin_tags into a helper to actually allocate the
tag set, and one that just restarts the admin queue. Add a local
variable for the tag_set to clean up the code a bit.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
To allow for slightly better debugging, print the command name when
aborting an command.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If prp_list is NULL, nvme_unmap_sg will be performed, and the assignment
to first_dma is meaningless, so remove it.
Signed-off-by: Liu Song <liusong@linux.alibaba.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit 89b3d6e605 ("nvme: simplify the compat ioctl handling") removed
the initialization of compat_ioctl from the nvme block_device_operations
structures.
Presumably the expectation was that 32-bit ioctls would be directed
through the regular handler but this is not the case: failing to assign
.compat_ioctl actually means that the compat case is disabled entirely,
and any attempt to submit nvme ioctls from 32-bit userspace fails
outright with -ENOTTY.
For example:
% smartctl -x /dev/nvme0n1
[...]
Read NVMe Identify Controller failed: NVME_IOCTL_ADMIN_CMD: Inappropriate ioctl for device
The blkdev_compat_ptr_ioctl helper can be used to direct compat calls
through the main ioctl handler and makes things work again.
Fixes: 89b3d6e605 ("nvme: simplify the compat ioctl handling")
Signed-off-by: Nick Bowler <nbowler@draconx.ca>
Reviewed-by: Guixin Liu <kanie@linux.alibaba.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The entire content of constants.c if guarded by an ifdef, so switch to
just building the file conditionally instead.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Use command_id instead of req->tag in trace_nvme_complete_rq(),
because of commit e7006de6c2 ("nvme: code command_id with a genctr
for use authentication after release"), cmd->common.command_id is set to
((genctl & 0xf)< 12 | req->tag), no longer req->tag, which makes cid in
trace_nvme_complete_rq and trace_nvme_setup_cmd are not the same.
Fixes: e7006de6c2 ("nvme: code command_id with a genctr for use authentication after release")
Signed-off-by: Bean Huo <beanhuo@micron.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pass anagrpid as second argument. This is prep patch that allows reusing
this function for supporting unknown command sets.
Signed-off-by: Joel Granados <j.granados@samsung.com>
Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, command data is only sent in-capsule on the for admin or I/O
commands on queues that indicate support for it. Send fabrics command
data in-capsule for I/O queues as well to avoid needing a separate
H2CData PDU for the connect command.
This is optimization. Without this change, we send the connect command
capsule and data in separate PDUs (CapsuleCmd and H2CData), and must wait
for the controller to respond with an R2T PDU before sending the H2CData.
With the change, we send a single CapsuleCmd PDU that includes the data.
This reduces the number of bytes (and likely packets) sent across the network,
and simplifies the send state machine handling in the driver.
Signed-off-by: Caleb Sander <csander@purestorage.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In case many controllers start error recovery at the same time (i.e.,
when port is down and up), they may never succeed to reconnect again.
This is because the target can't handle all the connect requests at
three seconds (the arbitrary value set today). Even if some of the
connections are established, when a single queue fails to connect,
all the controller's queues are destroyed as well. So, on the
following reconnection attempts the number of connect requests may
remain the same. To fix this, remove the timeout and wait for RDMA-CM
event to abort/complete the connect request. RDMA-CM sends unreachable
event when a timeout of ~90 seconds is expired. This approach is used
at other RDMA-CM users like SRP and iSER at blocking mode. The commit
also renames NVME_RDMA_CONNECT_TIMEOUT_MS to NVME_RDMA_CM_TIMEOUT_MS.
Signed-off-by: Israel Rukshin <israelr@nvidia.com>
Reviewed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Implement NVMe-oF In-Band authentication according to NVMe TPAR 8006.
This patch adds two new fabric options 'dhchap_secret' to specify the
pre-shared key (in ASCII respresentation according to NVMe 2.0 section
8.13.5.8 'Secret representation') and 'dhchap_ctrl_secret' to specify
the pre-shared controller key for bi-directional authentication of both
the host and the controller.
Re-authentication can be triggered by writing the PSK into the new
controller sysfs attribute 'dhchap_secret' or 'dhchap_ctrl_secret'.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
[axboe: fold in clang build fix]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The 'connect' command might fail with NVME_SC_AUTH_REQUIRED, so we
should be decoding this error, too.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Himanshu Madhani <himanshu.madhani@oracle.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Only caller of the __nvme_submit_sync_cmd() with qid value not equal to
NVME_QID_ANY is nvmf_connect_io_queues(), where qid value is alway set
to > 0.
[1] __nvme_submit_sync_cmd() callers with qid parameter from :-
Caller | qid parameter
------------------------------------------------------
* nvme_fc_connect_io_queues() |
nvmf_connect_io_queue() | qid > 0
* nvme_rdma_start_io_queues() |
nvme_rdma_start_queue() |
nvmf_connect_io_queues() | qid > 0
* nvme_tcp_start_io_queues() |
nvme_tcp_start_queue() |
nvmf_connect_io_queues() | qid > 0
* nvme_loop_connect_io_queues() |
nvmf_connect_io_queues() | qid > 0
When qid value of the function parameter __nvme_submit_sync_cmd() is > 0
from above callers, we use blk_mq_alloc_request_hctx(), where we pass
last parameter as 0 if qid functional parameter value is set to 0 with
conditional operators, see 1002 :-
991 int __nvme_submit_sync_cmd(struct request_queue *q, struct nvme_command *cmd,
992 union nvme_result *result, void *buffer, unsigned bufflen,
993 int qid, int at_head, blk_mq_req_flags_t flags)
994 {
995 struct request *req;
996 int ret;
997
998 if (qid == NVME_QID_ANY)
999 req = blk_mq_alloc_request(q, nvme_req_op(cmd), flags);
1000 else
1001 req = blk_mq_alloc_request_hctx(q, nvme_req_op(cmd), flags,
1002 qid ? qid - 1 : 0);
1003
But qid function parameter value of the __nvme_submit_sync_cmd() will
never be 0 from above caller list see [1], and all the other callers of
__nvme_submit_sync_cmd() use NVME_QID_ANY as qid value :-
1. nvme_submit_sync_cmd()
2. nvme_features()
3. nvme_sec_submit()
4. nvmf_reg_read32()
5. nvmf_reg_read64()
6. nvmf_ref_write32()
7. nvmf_connect_admin_queue()
Remove the conditional operator to pass the qid as 0 in the call to
blk_mq_alloc_requst_hctx().
Signed-off-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The function __nvme_submit_sync_cmd() has following list of callers
that sets the timeout value to 0 :-
Callers | Timeout value
------------------------------------------------
nvme_submit_sync_cmd() | 0
nvme_features() | 0
nvme_sec_submit() | 0
nvmf_reg_read32() | 0
nvmf_reg_read64() | 0
nvmf_reg_write32() | 0
nvmf_connect_admin_queue() | 0
nvmf_connect_io_queue() | 0
Remove the timeout function parameter from __nvme_submit_sync_cmd() and
adjust the rest of code accordingly.
Signed-off-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In the NVM Express Revision 1.4 spec, Figure 145 describes possible
values for an AER with event type "Error" (value 000b). For a
Persistent Internal Error (value 03h), the host should perform a
controller reset.
Add support for this error using code that already exists for
doing a controller reset. As part of this support, introduce
two utility functions for parsing the AER type and subtype.
This new support was tested in a lab environment where we can
generate the persistent internal error on demand, and observe
both the Linux side and NVMe controller side to see that the
controller reset has been done.
Signed-off-by: Michael Kelley <mikelley@microsoft.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Pull block updates from Jens Axboe:
- Improve the type checking of request flags (Bart)
- Ensure queue mapping for a single queues always picks the right queue
(Bart)
- Sanitize the io priority handling (Jan)
- rq-qos race fix (Jinke)
- Reserved tags handling improvements (John)
- Separate memory alignment from file/disk offset aligment for O_DIRECT
(Keith)
- Add new ublk driver, userspace block driver using io_uring for
communication with the userspace backend (Ming)
- Use try_cmpxchg() to cleanup the code in various spots (Uros)
- Finally remove bdevname() (Christoph)
- Clean up the zoned device handling (Christoph)
- Clean up independent access range support (Christoph)
- Clean up and improve block sysfs handling (Christoph)
- Clean up and improve teardown of block devices.
This turns the usual two step process into something that is simpler
to implement and handle in block drivers (Christoph)
- Clean up chunk size handling (Christoph)
- Misc cleanups and fixes (Bart, Bo, Dan, GuoYong, Jason, Keith, Liu,
Ming, Sebastian, Yang, Ying)
* tag 'for-5.20/block-2022-07-29' of git://git.kernel.dk/linux-block: (178 commits)
ublk_drv: fix double shift bug
ublk_drv: make sure that correct flags(features) returned to userspace
ublk_drv: fix error handling of ublk_add_dev
ublk_drv: fix lockdep warning
block: remove __blk_get_queue
block: call blk_mq_exit_queue from disk_release for never added disks
blk-mq: fix error handling in __blk_mq_alloc_disk
ublk: defer disk allocation
ublk: rewrite ublk_ctrl_get_queue_affinity to not rely on hctx->cpumask
ublk: fold __ublk_create_dev into ublk_ctrl_add_dev
ublk: cleanup ublk_ctrl_uring_cmd
ublk: simplify ublk_ch_open and ublk_ch_release
ublk: remove the empty open and release block device operations
ublk: remove UBLK_IO_F_PREFLUSH
ublk: add a MAINTAINERS entry
block: don't allow the same type rq_qos add more than once
mmc: fix disk/queue leak in case of adding disk failure
ublk_drv: fix an IS_ERR() vs NULL check
ublk: remove UBLK_IO_F_INTEGRITY
ublk_drv: remove unneeded semicolon
...
The dma_map operations now support P2PDMA pages directly. So remove
the calls to pci_p2pdma_[un]map_sg_attrs() and replace them with calls
to dma_map_sgtable().
dma_map_sgtable() returns more complete error codes than dma_map_sg()
and allows differentiating EREMOTEIO errors in case an unsupported
P2PDMA transfer is requested. When this happens, return BLK_STS_TARGET
so the request isn't retried.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Reviewed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Introduce a supports_pci_p2pdma() operation in nvme_ctrl_ops to
replace the fixed NVME_F_PCI_P2PDMA flag such that the dma_map_ops
flags can be checked for PCI P2PDMA support.
Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The issue exists when multipath is enabled and the namespace is
shared, but all the other controller checks at nvme_is_unique_nsid()
are false. The reason for this issue is that nvme_is_unique_nsid()
returns false when is called from nvme_mpath_alloc_disk() due to an
uninitialized value of head->shared. The patch fixes it by setting
head->shared before nvme_mpath_alloc_disk() is called.
Fixes: 5974ea7ce0 ("nvme: allow duplicate NSIDs for private namespaces")
Signed-off-by: Israel Rukshin <israelr@nvidia.com>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
A reset on a live device experiencing a link error still needs to have
the queue freeze state started for the subsequent reinitialization. Skip
only the register read if the device is not present instead of bypassing
the freeze checks.
Fixes: b98235d3a4 ("nvme-pci: harden drive presence detect in nvme_dev_disable()")
Reported-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Tested-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
This will allow the trace event to know the full size of the data
intended to be copied and silence read overflow checks.
Reported-by: John Garry <john.garry@huawei.com>
Suggested-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
We encountered a problem that the disconnect command hangs.
After analyzing the log and stack, we found that the triggering
process is as follows:
CPU0 CPU1
nvme_rdma_error_recovery_work
nvme_rdma_teardown_io_queues
nvme_do_delete_ctrl nvme_stop_queues
nvme_remove_namespaces
--clear ctrl->namespaces
nvme_start_queues
--no ns in ctrl->namespaces
nvme_ns_remove return(because ctrl is deleting)
blk_freeze_queue
blk_mq_freeze_queue_wait
--wait for ns to unquiesce to clean infligt IO, hang forever
This problem was not found in older kernels because we will flush
err work in nvme_stop_ctrl before nvme_remove_namespaces.It does not
seem to be modified for functional reasons, the patch can be revert
to solve the problem.
Revert commit 794a4cb3d2 ("nvme: remove the .stop_ctrl callout")
Signed-off-by: Ruozhu Li <liruozhu@huawei.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
queue stoppage and inflight requests cancellation is fully fenced from
io_work and thus failing a request from this context. Hence we don't
need to try to guess from the socket retcode if this failure is because
the queue is about to be torn down or not.
We are perfectly safe to just fail it, the request will not be cancelled
later on.
This solves possible very long shutdown delays when the users issues a
'nvme disconnect-all'
Reported-by: Daniel Wagner <dwagner@suse.de>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Set the queue dying flag and call blk_mq_exit_queue from del_gendisk for
all disks that do not have separately allocated queues, and thus remove
the need to call blk_cleanup_queue for them.
Rename blk_cleanup_disk to blk_mq_destroy_queue to make it clear that
this function is intended only for separately allocated blk-mq queues.
This saves an extra queue freeze for devices without a separately
allocated queue.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Link: https://lore.kernel.org/r/20220619060552.1850436-6-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This device shares the PCI ID with the Samsung 970 Evo Plus that
does not need or want the quirks. Move the the quirk entry to the
core table based on the model number instead.
Fixes: bc360b0b16 ("nvme-pci: add quirks for Samsung X5 SSDs")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Pankaj Raghav <p.raghav@samsung.com>
The Micron MTFDKBA2T0TFH device reports the same subsysem NQN for
all devices. Add a quick to ignore it.
Signed-off-by: Leo Savernik <l.savernik@aon.at>
Reviewed-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Like commit 5611ec2b98 ("nvme-pci: prevent SK hynix PC400 from using
Write Zeroes command"), UMIS and Samsung has the same issue:
[ 6305.633887] blk_update_request: operation not supported error,
dev nvme0n1, sector 340812032 op 0x9:(WRITE_ZEROES) flags 0x0
phys_seg 0 prio class 0
So also disable Write Zeroes command on UMIS and Samsung.
Signed-off-by: rasheed.hsueh <rasheed.hsueh@lcfc.corp-partner.google.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
When ZHITAI TiPro7000 SSDs entered deepest power state(ps4)
it has the same APST sleep problem as Kingston A2000.
by chance the system crashes and displays the same dmesg info:
https://bugzilla.kernel.org/show_bug.cgi?id=195039#c65
As the Archlinux wiki suggest (enlat + exlat) < 25000 is fine
and my testing shows no system crashes ever since.
Therefore disabling the deepest power state will fix the APST sleep issue.
https://wiki.archlinux.org/title/Solid_state_drive/NVMe
This is the APST data from 'nvme id-ctrl /dev/nvme1'
NVME Identify Controller:
vid : 0x1e49
ssvid : 0x1e49
sn : [...]
mn : ZHITAI TiPro7000 1TB
fr : ZTA32F3Y
[...]
ps 0 : mp:3.50W operational enlat:5 exlat:5 rrt:0 rrl:0
rwt:0 rwl:0 idle_power:- active_power:-
ps 1 : mp:3.30W operational enlat:50 exlat:100 rrt:1 rrl:1
rwt:1 rwl:1 idle_power:- active_power:-
ps 2 : mp:2.80W operational enlat:50 exlat:200 rrt:2 rrl:2
rwt:2 rwl:2 idle_power:- active_power:-
ps 3 : mp:0.1500W non-operational enlat:500 exlat:5000 rrt:3 rrl:3
rwt:3 rwl:3 idle_power:- active_power:-
ps 4 : mp:0.0200W non-operational enlat:2000 exlat:60000 rrt:4 rrl:4
rwt:4 rwl:4 idle_power:- active_power:-
Signed-off-by: Ning Wang <ningwang35@outlook.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
ADATA XPG GAMMIX S50 drives report bogus eui64 values that appear to
be the same across drives in one system. Quirk them out so they are
not marked as "non globally unique" duplicates.
Signed-off-by: Stefan Reiter <stefan@pimaker.at>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Many users have encountered IO timeouts with a CSTS value of 0xffffffff,
which indicates a failure to read the register. While there are various
potential causes for this observation, faulty NVMe APST has been the
culprit quite frequently. Add the recommended troubleshooting steps in
the error output when this condition occurs.
Signed-off-by: Keith Busch <kbusch@kernel.org>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
The recent global id check is finding poorly implemented devices in the
wild. Include relavant device information in the output to help quicken
an appropriate quirk patch.
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Christoph Hellwig <hch@lst.de>
