nvme_tcp_init_connection() attempts to receive an ICResp PDU but only
checks that the return value from recvmsg() is non-negative. If the
sender closes the TCP connection or sends fewer than 128 bytes, this
check will pass even though the full PDU wasn't received.
Ensure the full ICResp PDU is received by checking that recvmsg()
returns the expected 128 bytes.
Additionally set the MSG_WAITALL flag for recvmsg(), as a sender could
split the ICResp over multiple TCP frames. Without MSG_WAITALL,
recvmsg() could return prematurely with only part of the PDU.
Fixes: 3f2304f8c6 ("nvme-tcp: add NVMe over TCP host driver")
Signed-off-by: Caleb Sander Mateos <csander@purestorage.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
When compiling with W=1, a warning result for the function
nvme_tcp_set_queue_io_cpu():
host/tcp.c:1578: warning: Function parameter or struct member 'queue'
not described in 'nvme_tcp_set_queue_io_cpu'
host/tcp.c:1578: warning: expecting prototype for Track the number of
queues assigned to each cpu using a global per(). Prototype was for
nvme_tcp_set_queue_io_cpu() instead
Avoid this warning by using the regular comment format for the function
nvme_tcp_set_queue_io_cpu() instead of the kdoc comment format.
Fixes: 3219378987 ("nvme-tcp: Fix I/O queue cpu spreading for multiple controllers")
Signed-off-by: Damien Le Moal <dlemoal@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Previously, the NVMe/TCP host driver did not handle the C2HTermReq PDU,
instead printing "unsupported pdu type (3)" when received. This patch adds
support for processing the C2HTermReq PDU, allowing the driver
to print the Fatal Error Status field.
Example of output:
nvme nvme4: Received C2HTermReq (FES = Invalid PDU Header Field)
Signed-off-by: Maurizio Lombardi <mlombard@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull block updates from Jens Axboe:
- NVMe pull requests via Keith:
- Target support for PCI-Endpoint transport (Damien)
- TCP IO queue spreading fixes (Sagi, Chaitanya)
- Target handling for "limited retry" flags (Guixen)
- Poll type fix (Yongsoo)
- Xarray storage error handling (Keisuke)
- Host memory buffer free size fix on error (Francis)
- MD pull requests via Song:
- Reintroduce md-linear (Yu Kuai)
- md-bitmap refactor and fix (Yu Kuai)
- Replace kmap_atomic with kmap_local_page (David Reaver)
- Quite a few queue freeze and debugfs deadlock fixes
Ming introduced lockdep support for this in the 6.13 kernel, and it
has (unsurprisingly) uncovered quite a few issues
- Use const attributes for IO schedulers
- Remove bio ioprio wrappers
- Fixes for stacked device atomic write support
- Refactor queue affinity helpers, in preparation for better supporting
isolated CPUs
- Cleanups of loop O_DIRECT handling
- Cleanup of BLK_MQ_F_* flags
- Add rotational support for null_blk
- Various fixes and cleanups
* tag 'for-6.14/block-20250118' of git://git.kernel.dk/linux: (106 commits)
block: Don't trim an atomic write
block: Add common atomic writes enable flag
md/md-linear: Fix a NULL vs IS_ERR() bug in linear_add()
block: limit disk max sectors to (LLONG_MAX >> 9)
block: Change blk_stack_atomic_writes_limits() unit_min check
block: Ensure start sector is aligned for stacking atomic writes
blk-mq: Move more error handling into blk_mq_submit_bio()
block: Reorder the request allocation code in blk_mq_submit_bio()
nvme: fix bogus kzalloc() return check in nvme_init_effects_log()
md/md-bitmap: move bitmap_{start, end}write to md upper layer
md/raid5: implement pers->bitmap_sector()
md: add a new callback pers->bitmap_sector()
md/md-bitmap: remove the last parameter for bimtap_ops->endwrite()
md/md-bitmap: factor behind write counters out from bitmap_{start/end}write()
md: Replace deprecated kmap_atomic() with kmap_local_page()
md: reintroduce md-linear
partitions: ldm: remove the initial kernel-doc notation
blk-cgroup: rwstat: fix kernel-doc warnings in header file
blk-cgroup: fix kernel-doc warnings in header file
nbd: fix partial sending
...
Since day-1 we are assigning the queue io_cpu very naively. We always
base the queue id (controller scope) and assign it its matching cpu
from the online mask. This works fine when the number of queues match
the number of cpu cores.
The problem starts when we have less queues than cpu cores. First, we
should take into account the mq_map and select a cpu within the cpus
that are assigned to this queue by the mq_map in order to minimize cross
numa cpu bouncing.
Second, even worse is that we don't take into account multiple
controllers may have assigned queues to a given cpu. As a result we may
simply compund more and more queues on the same set of cpus, which is
suboptimal.
We fix this by introducing global per-cpu counters that tracks the
number of queues assigned to each cpu, and we select the least used cpu
based on the mq_map and the per-cpu counters, and assign it as the queue
io_cpu.
The behavior for a single controller is slightly optimized by selecting
better cpu candidates by consulting with the mq_map, and multiple
controllers are spreading queues among cpu cores much better, resulting
in lower average cpu load, and less likelihood to hit hotspots.
Note that the accounting is not 100% perfect, but we don't need to be,
we're simply putting our best effort to select the best candidate cpu
core that we find at any given point.
Another byproduct is that every controller reset/reconnect may change
the queues io_cpu mapping, based on the current LRU accounting scheme.
Here is the baseline queue io_cpu assignment for 4 controllers, 2 queues
per controller, and 4 cpus on the host:
nvme1: queue 0: using cpu 0
nvme1: queue 1: using cpu 1
nvme2: queue 0: using cpu 0
nvme2: queue 1: using cpu 1
nvme3: queue 0: using cpu 0
nvme3: queue 1: using cpu 1
nvme4: queue 0: using cpu 0
nvme4: queue 1: using cpu 1
And this is the fixed io_cpu assignment:
nvme1: queue 0: using cpu 0
nvme1: queue 1: using cpu 2
nvme2: queue 0: using cpu 1
nvme2: queue 1: using cpu 3
nvme3: queue 0: using cpu 0
nvme3: queue 1: using cpu 2
nvme4: queue 0: using cpu 1
nvme4: queue 1: using cpu 3
Fixes: 3f2304f8c6 ("nvme-tcp: add NVMe over TCP host driver")
Suggested-by: Hannes Reinecke <hare@kernel.org>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
[fixed kbuild reported errors]
Signed-off-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Now when destroying the IO queue we call nvme_tcp_stop_io_queues()
twice, nvme_tcp_destroy_io_queues() has an unnecessary call. Here we
try to remove nvme_tcp_destroy_io_queues() and merge it into
nvme_tcp_teardown_io_queues(), simplify the code and align with
nvme-rdma, make it easy to maintaince.
Signed-off-by: Chunguang.xu <chunguang.xu@shopee.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
As nvme_tcp_teardown_io_queues() is the only one caller of
nvme_tcp_destroy_admin_queue(), so we can merge it into
nvme_tcp_teardown_io_queues() to simplify the code.
Signed-off-by: Chunguang.xu <chunguang.xu@shopee.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
As we quiesce admin_q in nvme_tcp_teardown_admin_queue(), so we should no
need to quiesce it in nvme_tcp_reaardown_io_queues(), make things simple.
Signed-off-by: Chunguang.xu <chunguang.xu@shopee.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Now while we create new ctrl failed, we have not free the
tagset occupied by admin_q, here try to fix it.
Fixes: fd1418de10 ("nvme-tcp: avoid open-coding nvme_tcp_teardown_admin_queue()")
Signed-off-by: Chunguang.xu <chunguang.xu@shopee.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
key_lookup() will always return a key, even if that key is revoked
or invalidated. So check for invalid keys before continuing.
Signed-off-by: Hannes Reinecke <hare@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
There is a difference between TLS configured (ie the user has
provisioned/requested a key) and TLS enabled (ie the connection
is encrypted with TLS). This becomes important for secure concatenation,
where the initial authentication is run on an unencrypted connection
(ie with TLS configured, but not enabled), and then the queue is reset to
run over TLS (ie TLS configured _and_ enabled).
So to differentiate between those two states store the generated
key in opts->tls_key (as we're using the same TLS key for all queues),
the key serial of the resulting TLS handshake in ctrl->tls_pskid
(to signal that TLS on the admin queue is enabled), and a simple
flag for the queues to indicated that TLS has been enabled.
Signed-off-by: Hannes Reinecke <hare@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Currently nvme_tcp_try_send_data() use sendpage_ok() in order to disable
MSG_SPLICE_PAGES, it check the first page of the iterator, the iterator
may represent contiguous pages.
MSG_SPLICE_PAGES enables skb_splice_from_iter() which checks all the
pages it sends with sendpage_ok().
When nvme_tcp_try_send_data() sends an iterator that the first page is
sendable, but one of the other pages isn't skb_splice_from_iter() warns
and aborts the data transfer.
Using the new helper sendpages_ok() in order to disable MSG_SPLICE_PAGES
solves the issue.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Ofir Gal <ofir.gal@volumez.com>
Link: https://lore.kernel.org/r/20240718084515.3833733-3-ofir.gal@volumez.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Scheduling reset_work after a nvme subsystem reset is expected to fail
on pcie, but this also prevents potential handling the platform's pcie
services may provide that might successfully recovering the link without
re-enumeration. Such examples include AER, DPC, and power's EEH.
Provide a pci specific operation that safely initiates a subsystem
reset, and instead of scheduling reset work, read back the status
register to trigger a pcie read error.
Since this only affects pci, the other fabrics drivers subscribe to a
generic nvmf subsystem reset that is exactly the same as before. The
loop fabric doesn't use it because nvmet doesn't support setting that
property anyway.
And since we're using the magic NSSR value in two places now, provide a
symbolic define for it.
Reported-by: Nilay Shroff <nilay@linux.ibm.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Combining both creates an ambiguous cleanup scenario for the caller if
an error is returned: does the device reference need to be dropped or
did the error occur before the device was initialized? If an error
occurs after the device is added, then the existing cleanup routines
will leak memory.
Furthermore, the nvme core is taking it upon itself to free the device's
kobj name under certain conditions rather than go through the core
device API. We shouldn't be peaking into these implementation details.
Split the device initialization from the addition to make it easier to
know the error handling actions, fix the existing memory leaks, and stop
the device layering violations.
Link: https://lore.kernel.org/linux-nvme/c4050a37-ecc9-462c-9772-65e25166f439@grimberg.me/
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Drivers must call nvme_uninit_ctrl after a successful nvme_init_ctrl.
Split the allocation side out to make the error handling boundary easier
to navigate. The nvme tcp driver's error handling had different returns
in the error goto label's, which harm readability.
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Makes clear max reconnects translated by ctrl loss tmo and reconnect delay.
Signed-off-by: Tokunori Ikegami <ikegami.t@gmail.com>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Returning a nvme status from nvme_tcp_setup_ctrl() indicates that the
association was established and we have received a status from the
controller; consequently we should honour the DNR bit. If not any future
reconnect attempts will just return the same error, so we can
short-circuit the reconnect attempts and fail the connection directly.
Signed-off-by: Hannes Reinecke <hare@suse.de>
[dwagner: - extended nvme_should_reconnect]
Signed-off-by: Daniel Wagner <dwagner@suse.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
TLS requires a strict pdu pacing via MSG_EOR to signal the end
of a record and subsequent encryption. If we do not set MSG_EOR
at the end of a sequence the record won't be closed, encryption
doesn't start, and we end up with a send stall as the message
will never be passed on to the TCP layer.
So do not check for the queue status when TLS is enabled but
rather make the MSG_MORE setting dependent on the current
request only.
Signed-off-by: Hannes Reinecke <hare@kernel.org>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull NVMe fixes from Keith:
"nvme updates for Linux 6.9
- Make an informative message less ominous (Keith)
- Enhanced trace decoding (Guixin)
- TCP updates (Hannes, Li)
- Fabrics connect deadlock fix (Chunguang)
- Platform API migration update (Uwe)
- A new device quirk (Jiawei)"
* tag 'nvme-6.9-2024-03-21' of git://git.infradead.org/nvme:
nvmet-rdma: remove NVMET_RDMA_REQ_INVALIDATE_RKEY flag
nvme: remove redundant BUILD_BUG_ON check
nvme/tcp: Add wq_unbound modparam for nvme_tcp_wq
nvme-tcp: Export the nvme_tcp_wq to sysfs
drivers/nvme: Add quirks for device 126f:2262
nvme: parse format command's lbafu when tracing
nvme: add tracing of reservation commands
nvme: parse zns command's zsa and zrasf to string
nvme: use nvme_disk_is_ns_head helper
nvme: fix reconnection fail due to reserved tag allocation
nvmet: add tracing of zns commands
nvmet: add tracing of authentication commands
nvme-apple: Convert to platform remove callback returning void
nvmet-tcp: do not continue for invalid icreq
nvme: change shutdown timeout setting message
The default nvme_tcp_wq will use all CPUs to process tasks. Sometimes it is
necessary to set CPU affinity to improve performance.
A new module parameter wq_unbound is added here. If set to true, users can
configure cpu affinity through
/sys/devices/virtual/workqueue/nvme_tcp_wq/cpumask.
Signed-off-by: Li Feng <fengli@smartx.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Make the workqueue userspace visible for easy viewing and configuration.
Signed-off-by: Li Feng <fengli@smartx.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Log hostnqn when connecting to nvme target.
As hostnqn could be changed, logging this information
in syslog at appropriate time may help in troubleshooting.
Signed-off-by: Nitin U. Yewale <nyewale@redhat.com>
Reviewed-by: John Meneghini <jmeneghi@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
nvme_opcode_str() currently supports admin, IO, and fabrics commands.
However, fabrics commands aren't allowed for the pci transport.
Currently the pci caller passes 0 as the fctype,
which means any fabrics command would be displayed as "Property Set".
Move fabrics command support into a function nvme_fabrics_opcode_str()
and remove the fctype argument to nvme_opcode_str().
This way, a fabrics command will display as "Unknown" for pci.
Convert the rdma and tcp transports to use nvme_fabrics_opcode_str().
Signed-off-by: Caleb Sander <csander@purestorage.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull block fixes from Jens Axboe:
- NVMe pull request via Keith:
- tcp, fc, and rdma target fixes (Maurizio, Daniel, Hannes,
Christoph)
- discard fixes and improvements (Christoph)
- timeout debug improvements (Keith, Max)
- various cleanups (Daniel, Max, Giuxen)
- trace event string fixes (Arnd)
- shadow doorbell setup on reset fix (William)
- a write zeroes quirk for SK Hynix (Jim)
- MD pull request via Song:
- Sparse warning since v6.0 (Bart)
- /proc/mdstat regression since v6.7 (Yu Kuai)
- Use symbolic error value (Christian)
- IO Priority documentation update (Christian)
- Fix for accessing queue limits without having entered the queue
(Christoph, me)
- Fix for loop dio support (Christoph)
- Move null_blk off deprecated ida interface (Christophe)
- Ensure nbd initializes full msghdr (Eric)
- Fix for a regression with the folio conversion, which is now easier
to hit because of an unrelated change (Matthew)
- Remove redundant check in virtio-blk (Li)
- Fix for a potential hang in sbitmap (Ming)
- Fix for partial zone appending (Damien)
- Misc changes and fixes (Bart, me, Kemeng, Dmitry)
* tag 'for-6.8/block-2024-01-18' of git://git.kernel.dk/linux: (45 commits)
Documentation: block: ioprio: Update schedulers
loop: fix the the direct I/O support check when used on top of block devices
blk-mq: Remove the hctx 'run' debugfs attribute
nbd: always initialize struct msghdr completely
block: Fix iterating over an empty bio with bio_for_each_folio_all
block: bio-integrity: fix kcalloc() arguments order
virtio_blk: remove duplicate check if queue is broken in virtblk_done
sbitmap: remove stale comment in sbq_calc_wake_batch
block: Correct a documentation comment in blk-cgroup.c
null_blk: Remove usage of the deprecated ida_simple_xx() API
block: ensure we hold a queue reference when using queue limits
blk-mq: rename blk_mq_can_use_cached_rq
block: print symbolic error name instead of error code
blk-mq: fix IO hang from sbitmap wakeup race
nvmet-rdma: avoid circular locking dependency on install_queue()
nvmet-tcp: avoid circular locking dependency on install_queue()
nvme-pci: set doorbell config before unquiescing
block: fix partial zone append completion handling in req_bio_endio()
block/iocost: silence warning on 'last_period' potentially being unused
md/raid1: Use blk_opf_t for read and write operations
...
Print the command_id along side blk-mq's tag to help match commands with
protocol wire traces and logs.
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
There is no requirement to call nvme_tcp_free_queue() for queue
deallocation if the pskid is null or the queue allocation fails, as
the NVME_TCP_Q_ALLOCATED flag would not be set in such scenarios.
Signed-off-by: Guixin Liu <kanie@linux.alibaba.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
A different CPU may be setting the ctrl->state value, so ensure proper
barriers to prevent optimizing to a stale state. Normally it isn't a
problem to observe the wrong state as it is merely advisory to take a
quicker path during initialization and error recovery, but seeing an old
state can report unexpected ENETRESET errors when a reset request was in
fact successful.
Reported-by: Minh Hoang <mh2022@meta.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Hannes Reinecke <hare@suse.de>
When CONFIG_NVME_KEYRING is enabled as a loadable module, but the TCP
host code is built-in, it fails to link:
arm-linux-gnueabi-ld: drivers/nvme/host/tcp.o: in function `nvme_tcp_setup_ctrl':
tcp.c:(.text+0x1940): undefined reference to `nvme_tls_psk_default'
The problem is that the compile-time conditionals are inconsistent here,
using a mix of #ifdef CONFIG_NVME_TCP_TLS, IS_ENABLED(CONFIG_NVME_TCP_TLS)
and IS_ENABLED(CONFIG_NVME_KEYRING) checks, with CONFIG_NVME_KEYRING
controlling whether the implementation is actually built.
Change it to use IS_ENABLED(CONFIG_NVME_KEYRING) checks consistently,
which should help readability and make it less error-prone. Combining
it with the check for the ctrl->opts->tls flag lets the compiler drop
all the TLS code in configurations without this feature, which also
helps runtime behavior in addition to avoiding the link failure.
To make it possible for the compiler to build the dead code, both
the tls_handshake_timeout variable and the TLS specific members
of nvme_tcp_queue need to be moved out of the #ifdef block as well,
but at least the former of these gets optimized out again.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20231122224719.4042108-4-arnd@kernel.org
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Stopping keep-alive not only stops the keep-alive workqueue,
but also needs to be synchronized with I/O termination as we
must not send a keep-alive command after all I/O had been
terminated.
So to avoid any regressions move the call to stop_keep_alive()
back to its original position and ensure that keep-alive is
correctly stopped failing to setup the admin queue.
Fixes: 4733b65d82 ("nvme: start keep-alive after admin queue setup")
Suggested-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
nvme_tcp_setup_ctrl() has an open-coded version of
nvme_tcp_teardown_admin_queue().
Signed-off-by: Hannes Reinecke <hare@suse.de>
Tested-by: Mark O'Donovan <shiftee@posteo.net>
Signed-off-by: Keith Busch <kbusch@kernel.org>
All error handling path end to the error handling path, except this one.
Go to the error handling branch as well here, otherwise 'icreq' and
'icresp' will leak.
Fixes: 2837966ab2 ("nvme-tcp: control message handling for recvmsg()")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Parse the fabrics options 'keyring' and 'tls_key' and store the
referenced keys in the options structure.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
When icreq/icresp fails we should be printing out a warning to
inform the user that the connection could not be established;
without it there won't be anything in the kernel message log,
just an error code returned to nvme-cli.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
kTLS is sending TLS ALERT messages as control messages for recvmsg().
As we can't do anything sensible with it just abort the connection
and let the userspace agent to a re-negotiation.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Add a fabrics option 'tls' and start the TLS handshake upcall
with the default PSK. When TLS is started the PSK key serial
number is displayed in the sysfs attribute 'tls_key'
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
When using the TLS upcall we need to allocate a socket file such
that the userspace daemon is able to use the socket.
Signed-off-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull block fixes from Jens Axboe:
- NVMe pull request via Keith:
- Fixes for request_queue state (Ming)
- Another uuid quirk (August)
- RCU poll fix for NVMe (Ming)
- Fix for an IO stall with polled IO (me)
- Fix for blk-iocost stats enable/disable accounting (Chengming)
- Regression fix for large pages for zram (Christoph)
* tag 'block-6.5-2023-08-11' of git://git.kernel.dk/linux:
nvme: core: don't hold rcu read lock in nvme_ns_chr_uring_cmd_iopoll
blk-iocost: fix queue stats accounting
block: don't make REQ_POLLED imply REQ_NOWAIT
block: get rid of unused plug->nowait flag
zram: take device and not only bvec offset into account
nvme-pci: add NVME_QUIRK_BOGUS_NID for Samsung PM9B1 256G and 512G
nvme-rdma: fix potential unbalanced freeze & unfreeze
nvme-tcp: fix potential unbalanced freeze & unfreeze
nvme: fix possible hang when removing a controller during error recovery
Move start_freeze into nvme_tcp_configure_io_queues(), and there is
at least two benefits:
1) fix unbalanced freeze and unfreeze, since re-connection work may
fail or be broken by removal
2) IO during error recovery can be failfast quickly because nvme fabrics
unquiesces queues after teardown.
One side-effect is that !mpath request may timeout during connecting
because of queue topo change, but that looks not one big deal:
1) same problem exists with current code base
2) compared with !mpath, mpath use case is dominant
Fixes: 2875b0aeca ("nvme-tcp: fix controller reset hang during traffic")
Cc: stable@vger.kernel.org
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Tested-by: Yi Zhang <yi.zhang@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull networking changes from Jakub Kicinski:
"WiFi 7 and sendpage changes are the biggest pieces of work for this
release. The latter will definitely require fixes but I think that we
got it to a reasonable point.
Core:
- Rework the sendpage & splice implementations
Instead of feeding data into sockets page by page extend sendmsg
handlers to support taking a reference on the data, controlled by a
new flag called MSG_SPLICE_PAGES
Rework the handling of unexpected-end-of-file to invoke an
additional callback instead of trying to predict what the right
combination of MORE/NOTLAST flags is
Remove the MSG_SENDPAGE_NOTLAST flag completely
- Implement SCM_PIDFD, a new type of CMSG type analogous to
SCM_CREDENTIALS, but it contains pidfd instead of plain pid
- Enable socket busy polling with CONFIG_RT
- Improve reliability and efficiency of reporting for ref_tracker
- Auto-generate a user space C library for various Netlink families
Protocols:
- Allow TCP to shrink the advertised window when necessary, prevent
sk_rcvbuf auto-tuning from growing the window all the way up to
tcp_rmem[2]
- Use per-VMA locking for "page-flipping" TCP receive zerocopy
- Prepare TCP for device-to-device data transfers, by making sure
that payloads are always attached to skbs as page frags
- Make the backoff time for the first N TCP SYN retransmissions
linear. Exponential backoff is unnecessarily conservative
- Create a new MPTCP getsockopt to retrieve all info
(MPTCP_FULL_INFO)
- Avoid waking up applications using TLS sockets until we have a full
record
- Allow using kernel memory for protocol ioctl callbacks, paving the
way to issuing ioctls over io_uring
- Add nolocalbypass option to VxLAN, forcing packets to be fully
encapsulated even if they are destined for a local IP address
- Make TCPv4 use consistent hash in TIME_WAIT and SYN_RECV. Ensure
in-kernel ECMP implementation (e.g. Open vSwitch) select the same
link for all packets. Support L4 symmetric hashing in Open vSwitch
- PPPoE: make number of hash bits configurable
- Allow DNS to be overwritten by DHCPACK in the in-kernel DHCP client
(ipconfig)
- Add layer 2 miss indication and filtering, allowing higher layers
(e.g. ACL filters) to make forwarding decisions based on whether
packet matched forwarding state in lower devices (bridge)
- Support matching on Connectivity Fault Management (CFM) packets
- Hide the "link becomes ready" IPv6 messages by demoting their
printk level to debug
- HSR: don't enable promiscuous mode if device offloads the proto
- Support active scanning in IEEE 802.15.4
- Continue work on Multi-Link Operation for WiFi 7
BPF:
- Add precision propagation for subprogs and callbacks. This allows
maintaining verification efficiency when subprograms are used, or
in fact passing the verifier at all for complex programs,
especially those using open-coded iterators
- Improve BPF's {g,s}setsockopt() length handling. Previously BPF
assumed the length is always equal to the amount of written data.
But some protos allow passing a NULL buffer to discover what the
output buffer *should* be, without writing anything
- Accept dynptr memory as memory arguments passed to helpers
- Add routing table ID to bpf_fib_lookup BPF helper
- Support O_PATH FDs in BPF_OBJ_PIN and BPF_OBJ_GET commands
- Drop bpf_capable() check in BPF_MAP_FREEZE command (used to mark
maps as read-only)
- Show target_{obj,btf}_id in tracing link fdinfo
- Addition of several new kfuncs (most of the names are
self-explanatory):
- Add a set of new dynptr kfuncs: bpf_dynptr_adjust(),
bpf_dynptr_is_null(), bpf_dynptr_is_rdonly(), bpf_dynptr_size()
and bpf_dynptr_clone().
- bpf_task_under_cgroup()
- bpf_sock_destroy() - force closing sockets
- bpf_cpumask_first_and(), rework bpf_cpumask_any*() kfuncs
Netfilter:
- Relax set/map validation checks in nf_tables. Allow checking
presence of an entry in a map without using the value
- Increase ip_vs_conn_tab_bits range for 64BIT builds
- Allow updating size of a set
- Improve NAT tuple selection when connection is closing
Driver API:
- Integrate netdev with LED subsystem, to allow configuring HW
"offloaded" blinking of LEDs based on link state and activity
(i.e. packets coming in and out)
- Support configuring rate selection pins of SFP modules
- Factor Clause 73 auto-negotiation code out of the drivers, provide
common helper routines
- Add more fool-proof helpers for managing lifetime of MDIO devices
associated with the PCS layer
- Allow drivers to report advanced statistics related to Time Aware
scheduler offload (taprio)
- Allow opting out of VF statistics in link dump, to allow more VFs
to fit into the message
- Split devlink instance and devlink port operations
New hardware / drivers:
- Ethernet:
- Synopsys EMAC4 IP support (stmmac)
- Marvell 88E6361 8 port (5x1GE + 3x2.5GE) switches
- Marvell 88E6250 7 port switches
- Microchip LAN8650/1 Rev.B0 PHYs
- MediaTek MT7981/MT7988 built-in 1GE PHY driver
- WiFi:
- Realtek RTL8192FU, 2.4 GHz, b/g/n mode, 2T2R, 300 Mbps
- Realtek RTL8723DS (SDIO variant)
- Realtek RTL8851BE
- CAN:
- Fintek F81604
Drivers:
- Ethernet NICs:
- Intel (100G, ice):
- support dynamic interrupt allocation
- use meta data match instead of VF MAC addr on slow-path
- nVidia/Mellanox:
- extend link aggregation to handle 4, rather than just 2 ports
- spawn sub-functions without any features by default
- OcteonTX2:
- support HTB (Tx scheduling/QoS) offload
- make RSS hash generation configurable
- support selecting Rx queue using TC filters
- Wangxun (ngbe/txgbe):
- add basic Tx/Rx packet offloads
- add phylink support (SFP/PCS control)
- Freescale/NXP (enetc):
- report TAPRIO packet statistics
- Solarflare/AMD:
- support matching on IP ToS and UDP source port of outer
header
- VxLAN and GENEVE tunnel encapsulation over IPv4 or IPv6
- add devlink dev info support for EF10
- Virtual NICs:
- Microsoft vNIC:
- size the Rx indirection table based on requested
configuration
- support VLAN tagging
- Amazon vNIC:
- try to reuse Rx buffers if not fully consumed, useful for ARM
servers running with 16kB pages
- Google vNIC:
- support TCP segmentation of >64kB frames
- Ethernet embedded switches:
- Marvell (mv88e6xxx):
- enable USXGMII (88E6191X)
- Microchip:
- lan966x: add support for Egress Stage 0 ACL engine
- lan966x: support mapping packet priority to internal switch
priority (based on PCP or DSCP)
- Ethernet PHYs:
- Broadcom PHYs:
- support for Wake-on-LAN for BCM54210E/B50212E
- report LPI counter
- Microsemi PHYs: support RGMII delay configuration (VSC85xx)
- Micrel PHYs: receive timestamp in the frame (LAN8841)
- Realtek PHYs: support optional external PHY clock
- Altera TSE PCS: merge the driver into Lynx PCS which it is a
variant of
- CAN: Kvaser PCIEcan:
- support packet timestamping
- WiFi:
- Intel (iwlwifi):
- major update for new firmware and Multi-Link Operation (MLO)
- configuration rework to drop test devices and split the
different families
- support for segmented PNVM images and power tables
- new vendor entries for PPAG (platform antenna gain) feature
- Qualcomm 802.11ax (ath11k):
- Multiple Basic Service Set Identifier (MBSSID) and Enhanced
MBSSID Advertisement (EMA) support in AP mode
- support factory test mode
- RealTek (rtw89):
- add RSSI based antenna diversity
- support U-NII-4 channels on 5 GHz band
- RealTek (rtl8xxxu):
- AP mode support for 8188f
- support USB RX aggregation for the newer chips"
* tag 'net-next-6.5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1602 commits)
net: scm: introduce and use scm_recv_unix helper
af_unix: Skip SCM_PIDFD if scm->pid is NULL.
net: lan743x: Simplify comparison
netlink: Add __sock_i_ino() for __netlink_diag_dump().
net: dsa: avoid suspicious RCU usage for synced VLAN-aware MAC addresses
Revert "af_unix: Call scm_recv() only after scm_set_cred()."
phylink: ReST-ify the phylink_pcs_neg_mode() kdoc
libceph: Partially revert changes to support MSG_SPLICE_PAGES
net: phy: mscc: fix packet loss due to RGMII delays
net: mana: use vmalloc_array and vcalloc
net: enetc: use vmalloc_array and vcalloc
ionic: use vmalloc_array and vcalloc
pds_core: use vmalloc_array and vcalloc
gve: use vmalloc_array and vcalloc
octeon_ep: use vmalloc_array and vcalloc
net: usb: qmi_wwan: add u-blox 0x1312 composition
perf trace: fix MSG_SPLICE_PAGES build error
ipvlan: Fix return value of ipvlan_queue_xmit()
netfilter: nf_tables: fix underflow in chain reference counter
netfilter: nf_tables: unbind non-anonymous set if rule construction fails
...
tcp and rdma transports have lots of duplicate code setting up the
different queue mappings. Add common helpers.
Cc: Chaitanya Kulkarni <kch@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Pull block updates from Jens Axboe:
- drbd patches, bringing us closer to unifying the out-of-tree version
and the in tree one (Andreas, Christoph)
- support for auto-quiesce for the s390 dasd driver (Stefan)
- MD pull request via Song:
- md/bitmap: Optimal last page size (Jon Derrick)
- Various raid10 fixes (Yu Kuai, Li Nan)
- md: add error_handlers for raid0 and linear (Mariusz Tkaczyk)
- NVMe pull request via Christoph:
- Drop redundant pci_enable_pcie_error_reporting (Bjorn Helgaas)
- Validate nvmet module parameters (Chaitanya Kulkarni)
- Fence TCP socket on receive error (Chris Leech)
- Fix async event trace event (Keith Busch)
- Minor cleanups (Chaitanya Kulkarni, zhenwei pi)
- Fix and cleanup nvmet Identify handling (Damien Le Moal,
Christoph Hellwig)
- Fix double blk_mq_complete_request race in the timeout handler
(Lei Yin)
- Fix irq locking in nvme-fcloop (Ming Lei)
- Remove queue mapping helper for rdma devices (Sagi Grimberg)
- use structured request attribute checks for nbd (Jakub)
- fix blk-crypto race conditions between keyslot management (Eric)
- add sed-opal support for reading read locking range attributes
(Ondrej)
- make fault injection configurable for null_blk (Akinobu)
- clean up the request insertion API (Christoph)
- clean up the queue running API (Christoph)
- blkg config helper cleanups (Tejun)
- lazy init support for blk-iolatency (Tejun)
- various fixes and tweaks to ublk (Ming)
- remove hybrid polling. It hasn't really been useful since we got
async polled IO support, and these days we don't support sync polled
IO at all (Keith)
- misc fixes, cleanups, improvements (Zhong, Ondrej, Colin, Chengming,
Chaitanya, me)
* tag 'for-6.4/block-2023-04-21' of git://git.kernel.dk/linux: (118 commits)
nbd: fix incomplete validation of ioctl arg
ublk: don't return 0 in case of any failure
sed-opal: geometry feature reporting command
null_blk: Always check queue mode setting from configfs
block: ublk: switch to ioctl command encoding
blk-mq: fix the blk_mq_add_to_requeue_list call in blk_kick_flush
block, bfq: Fix division by zero error on zero wsum
fault-inject: fix build error when FAULT_INJECTION_CONFIGFS=y and CONFIGFS_FS=m
block: store bdev->bd_disk->fops->submit_bio state in bdev
block: re-arrange the struct block_device fields for better layout
md/raid5: remove unused working_disks variable
md/raid10: don't call bio_start_io_acct twice for bio which experienced read error
md/raid10: fix memleak of md thread
md/raid10: fix memleak for 'conf->bio_split'
md/raid10: fix leak of 'r10bio->remaining' for recovery
md/raid10: don't BUG_ON() in raise_barrier()
md: fix soft lockup in status_resync
md: add error_handlers for raid0 and linear
md: Use optimal I/O size for last bitmap page
md: Fix types in sb writer
...
Ensure that no further socket reads occur after a receive processing
error, either from io_work being re-scheduled or nvme_tcp_poll.
Failing to do so can result in unrecognised PDU payloads or TCP stream
garbage being processed as a C2H data PDU, and potentially start copying
the payload to an invalid destination after looking up a request using a
bogus command id.
Signed-off-by: Chris Leech <cleech@redhat.com>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: John Meneghini <jmeneghi@redhat.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
When we allocate a nvme-tcp queue, we set the data_ready callback before
we actually need to use it. This creates the potential that if a stray
controller sends us data on the socket before we connect, we can trigger
the io_work and start consuming the socket.
In this case reported: we failed to allocate one of the io queues, and
as we start releasing the queues that we already allocated, we get
a UAF [1] from the io_work which is running before it should really.
Fix this by setting the socket ops callbacks only before we start the
queue, so that we can't accidentally schedule the io_work in the
initialization phase before the queue started. While we are at it,
rename nvme_tcp_restore_sock_calls to pair with nvme_tcp_setup_sock_ops.
[1]:
[16802.107284] nvme nvme4: starting error recovery
[16802.109166] nvme nvme4: Reconnecting in 10 seconds...
[16812.173535] nvme nvme4: failed to connect socket: -111
[16812.173745] nvme nvme4: Failed reconnect attempt 1
[16812.173747] nvme nvme4: Reconnecting in 10 seconds...
[16822.413555] nvme nvme4: failed to connect socket: -111
[16822.413762] nvme nvme4: Failed reconnect attempt 2
[16822.413765] nvme nvme4: Reconnecting in 10 seconds...
[16832.661274] nvme nvme4: creating 32 I/O queues.
[16833.919887] BUG: kernel NULL pointer dereference, address: 0000000000000088
[16833.920068] nvme nvme4: Failed reconnect attempt 3
[16833.920094] #PF: supervisor write access in kernel mode
[16833.920261] nvme nvme4: Reconnecting in 10 seconds...
[16833.920368] #PF: error_code(0x0002) - not-present page
[16833.921086] Workqueue: nvme_tcp_wq nvme_tcp_io_work [nvme_tcp]
[16833.921191] RIP: 0010:_raw_spin_lock_bh+0x17/0x30
...
[16833.923138] Call Trace:
[16833.923271] <TASK>
[16833.923402] lock_sock_nested+0x1e/0x50
[16833.923545] nvme_tcp_try_recv+0x40/0xa0 [nvme_tcp]
[16833.923685] nvme_tcp_io_work+0x68/0xa0 [nvme_tcp]
[16833.923824] process_one_work+0x1e8/0x390
[16833.923969] worker_thread+0x53/0x3d0
[16833.924104] ? process_one_work+0x390/0x390
[16833.924240] kthread+0x124/0x150
[16833.924376] ? set_kthread_struct+0x50/0x50
[16833.924518] ret_from_fork+0x1f/0x30
[16833.924655] </TASK>
Reported-by: Yanjun Zhang <zhangyanjun@cestc.cn>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Tested-by: Yanjun Zhang <zhangyanjun@cestc.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
