At the moment you have to use sendmsg for vectorized send.
While this works it's suboptimal as it also means you need to
allocate a struct msghdr that needs to be kept alive until a
submission happens. We can remove this limitation by just
allowing to use send directly.
Signed-off-by: Norman Maurer <norman_maurer@apple.com>
Link: https://lore.kernel.org/r/20250729065952.26646-1-norman_maurer@apple.com
[axboe: remove -EINVAL return for SENDMSG and SEND_VECTORIZED]
[axboe: allow send_zc to set SEND_VECTORIZED too]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
An errant ';' slipped into that definition, which will cause some
compilers to complain when it's used in an application:
timestamp.c:257:45: error: empty expression statement has no effect; remove unnecessary ';' to silence this warning [-Werror,-Wextra-semi-stmt]
257 | hwts = cqe->flags & IORING_CQE_F_TSTAMP_HW;
| ^
Fixes: 9e4ed359b8 ("io_uring/netcmd: add tx timestamping cmd support")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add a new socket command which returns tx time stamps to the user. It
provide an alternative to the existing error queue recvmsg interface.
The command works in a polled multishot mode, which means io_uring will
poll the socket and keep posting timestamps until the request is
cancelled or fails in any other way (e.g. with no space in the CQ). It
reuses the net infra and grabs timestamps from the socket's error queue.
The command requires IORING_SETUP_CQE32. All non-final CQEs (marked with
IORING_CQE_F_MORE) have cqe->res set to the tskey, and the upper 16 bits
of cqe->flags keep tstype (i.e. offset by IORING_CQE_BUFFER_SHIFT). The
timevalue is store in the upper part of the extended CQE. The final
completion won't have IORING_CQE_F_MORE and will have cqe->res storing
0/error.
Suggested-by: Vadim Fedorenko <vadim.fedorenko@linux.dev>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/92ee66e6b33b8de062a977843d825f58f21ecd37.1750065793.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
To test and profile the overhead of io_uring task_work and the various
types of it, add IORING_NOP_TW which tells nop to signal completions
through task_work rather than complete them inline.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=k8SF
-----END PGP SIGNATURE-----
Merge tag 'for-6.16/io_uring-20250523' of git://git.kernel.dk/linux
Pull io_uring updates from Jens Axboe:
- Avoid indirect function calls in io-wq for executing and freeing
work.
The design of io-wq is such that it can be a generic mechanism, but
as it's just used by io_uring now, may as well avoid these indirect
calls
- Clean up registered buffers for networking
- Add support for IORING_OP_PIPE. Pretty straight forward, allows
creating pipes with io_uring, particularly useful for having these be
instantiated as direct descriptors
- Clean up the coalescing support fore registered buffers
- Add support for multiple interface queues for zero-copy rx
networking. As this feature was merged for 6.15 it supported just a
single ifq per ring
- Clean up the eventfd support
- Add dma-buf support to zero-copy rx
- Clean up and improving the request draining support
- Clean up provided buffer support, most notably with an eye toward
making the legacy support less intrusive
- Minor fdinfo cleanups, dropping support for dumping what credentials
are registered
- Improve support for overflow CQE handling, getting rid of GFP_ATOMIC
for allocating overflow entries where possible
- Improve detection of cases where io-wq doesn't need to spawn a new
worker unnecessarily
- Various little cleanups
* tag 'for-6.16/io_uring-20250523' of git://git.kernel.dk/linux: (59 commits)
io_uring/cmd: warn on reg buf imports by ineligible cmds
io_uring/io-wq: only create a new worker if it can make progress
io_uring/io-wq: ignore non-busy worker going to sleep
io_uring/io-wq: move hash helpers to the top
trace/io_uring: fix io_uring_local_work_run ctx documentation
io_uring: finish IOU_OK -> IOU_COMPLETE transition
io_uring: add new helpers for posting overflows
io_uring: pass in struct io_big_cqe to io_alloc_ocqe()
io_uring: make io_alloc_ocqe() take a struct io_cqe pointer
io_uring: split alloc and add of overflow
io_uring: open code io_req_cqe_overflow()
io_uring/fdinfo: get rid of dumping credentials
io_uring/fdinfo: only compile if CONFIG_PROC_FS is set
io_uring/kbuf: unify legacy buf provision and removal
io_uring/kbuf: refactor __io_remove_buffers
io_uring/kbuf: don't compute size twice on prep
io_uring/kbuf: drop extra vars in io_register_pbuf_ring
io_uring/kbuf: use mem_is_zero()
io_uring/kbuf: account ring io_buffer_list memory
io_uring: drain based on allocates reqs
...
Allow userspace to pass a per-I/O write stream in the SQE:
__u8 write_stream;
The __u8 type matches the size the filesystems and block layer support.
Application can query the supported values from the block devices
max_write_streams sysfs attribute. Unsupported values are ignored by
file operations that do not support write streams or rejected with an
error by those that support them.
Reviewed-by: Hannes Reinecke <hare@suse.de>
Reviewed-by: Nitesh Shetty <nj.shetty@samsung.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Keith Busch <kbusch@kernel.org>
Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
Link: https://lore.kernel.org/r/20250506121732.8211-7-joshi.k@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This works just like pipe2(2), except it also supports fixed file
descriptors. Used in a similar fashion as for other fd instantiating
opcodes (like accept, socket, open, etc), where sqe->file_slot is set
appropriately if two direct descriptors are desired rather than a set
of normal file descriptors.
sqe->addr must be set to a pointer to an array of 2 integers, which
is where the fixed/normal file descriptors are copied to.
sqe->pipe_flags contains flags, same as what is allowed for pipe2(2).
Future expansion of per-op private flags can go in sqe->ioprio,
like we do for other opcodes that take both a "syscall" flag set and
an io_uring opcode specific flag set.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
IORING_OP_RECV_ZC requests take a zcrx object id via sqe::zcrx_ifq_idx,
which binds it to the corresponding if / queue. However, we don't return
that id back to the user. It's fine as currently there can be only one
zcrx and the user assumes that its id should be 0, but as we'll need
multiple zcrx objects in the future let's explicitly pass it back on
registration.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/8714667d370651962f7d1a169032e5f02682a73e.1744722517.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=vOm1
-----END PGP SIGNATURE-----
Merge tag 'for-6.15/io_uring-reg-vec-20250327' of git://git.kernel.dk/linux
Pull more io_uring updates from Jens Axboe:
"Final separate updates for io_uring.
This started out as a series of cleanups improvements and improvements
for registered buffers, but as the last series of the io_uring changes
for 6.15, it also collected a few fixes for the other branches on top:
- Add support for vectored fixed/registered buffers.
Previously only single segments have been supported for commands,
now vectored variants are supported as well. This series includes
networking and file read/write support.
- Small series unifying return codes across multi and single shot.
- Small series cleaning up registerd buffer importing.
- Adding support for vectored registered buffers for uring_cmd.
- Fix for io-wq handling of command reissue.
- Various little fixes and tweaks"
* tag 'for-6.15/io_uring-reg-vec-20250327' of git://git.kernel.dk/linux: (25 commits)
io_uring/net: fix io_req_post_cqe abuse by send bundle
io_uring/net: use REQ_F_IMPORT_BUFFER for send_zc
io_uring: move min_events sanitisation
io_uring: rename "min" arg in io_iopoll_check()
io_uring: open code __io_post_aux_cqe()
io_uring: defer iowq cqe overflow via task_work
io_uring: fix retry handling off iowq
io_uring/net: only import send_zc buffer once
io_uring/cmd: introduce io_uring_cmd_import_fixed_vec
io_uring/cmd: add iovec cache for commands
io_uring/cmd: don't expose entire cmd async data
io_uring: rename the data cmd cache
io_uring: rely on io_prep_reg_vec for iovec placement
io_uring: introduce io_prep_reg_iovec()
io_uring: unify STOP_MULTISHOT with IOU_OK
io_uring: return -EAGAIN to continue multishot
io_uring: cap cached iovec/bvec size
io_uring/net: implement vectored reg bufs for zctx
io_uring/net: convert to struct iou_vec
io_uring/net: pull vec alloc out of msghdr import
...
-----BEGIN PGP SIGNATURE-----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=AZkA
-----END PGP SIGNATURE-----
Merge tag 'for-6.15/io_uring-epoll-wait-20250325' of git://git.kernel.dk/linux
Pull io_uring epoll support from Jens Axboe:
"This adds support for reading epoll events via io_uring.
While this may seem counter-intuitive (and/or productive), the
reasoning here is that quite a few existing epoll event loops can
easily do a partial conversion to a completion based model, but are
still stuck with one (or few) event types that remain readiness based.
For that case, they then need to add the io_uring fd to the epoll
context, and continue to rely on epoll_wait(2) for waiting on events.
This misses out on the finer grained waiting that io_uring can do, to
reduce context switches and wait for multiple events in one batch
reliably.
With adding support for reaping epoll events via io_uring, the whole
legacy readiness based event types can still be reaped via epoll, with
the overall waiting in the loop be driven by io_uring"
* tag 'for-6.15/io_uring-epoll-wait-20250325' of git://git.kernel.dk/linux:
io_uring/epoll: add support for IORING_OP_EPOLL_WAIT
io_uring/epoll: remove CONFIG_EPOLL guards
-----BEGIN PGP SIGNATURE-----
iQJEBAABCAAuFiEEwPw5LcreJtl1+l5K99NY+ylx4KYFAmfjTP8QHGF4Ym9lQGtl
cm5lbC5kawAKCRD301j7KXHgpm6oEACnpGL52FAKTVj14GDqFo6Pq0Jmnh07x8qj
mpHFPwxfWAzRiuNyji2iS9ecS2cnlkixNyMWZipXRi4KJAUjJH6YDd7IofUI3Glf
6v7b6srFSvsWJIJ8LdkJHLHAJuzYnJvFZ8apwgQczEDqgHq7BAunM1sVQ+mydjYk
EXT4kN6DSBOPzwr9GAay52f8nXhbqdHfT+YTGHPHg+QToojL6gD7vvW57w/QqD/x
91hJef1z01cSIsDZOxA0EUeD+9bBAHpoamr/e3IOOCVYCN6hy0dGa9g0QGbbpVyE
AeU4FGZLV9J8OOfvHVraDt5Wn3IXxYaMu22dSn1S6tVinwjXhJR2LAA+t4fGHAkt
i38LjOsIbopSQn/cNhzwC8UZcHLqnVsdDolHlHzsVFVfcpck2/4JFpUeP8QhWgrk
f9tY12QUf/oEaWm0/sUCHZNFxpIGeFA5FFXf0Z92clnzBuiuWoesBNvxqY/2DeZn
IDNXiv+Trxr6kFEjTpzPeuxbWrn4PJ7afQSAFcEmOCguk5riM+zJZNIKg0TxUHSS
tt6sfxmTP1DhgDKad5kT3MLyzOcx47Kbjf4dj6KmRnD+3DGwwN2F7X7R1GJylPSp
RLOzJ+Ouuy9UmBN6JMsT4BmR9+FJTVirADU926d/ZqCTtRV8Tnq/6HHmKmmr4CR0
THJ0PJqQjg==
=MOve
-----END PGP SIGNATURE-----
Merge tag 'for-6.15/io_uring-rx-zc-20250325' of git://git.kernel.dk/linux
Pull io_uring zero-copy receive support from Jens Axboe:
"This adds support for zero-copy receive with io_uring, enabling fast
bulk receive of data directly into application memory, rather than
needing to copy the data out of kernel memory.
While this version only supports host memory as that was the initial
target, other memory types are planned as well, with notably GPU
memory coming next.
This work depends on some networking components which were queued up
on the networking side, but have now landed in your tree.
This is the work of Pavel Begunkov and David Wei. From the v14 posting:
'We configure a page pool that a driver uses to fill a hw rx queue
to hand out user pages instead of kernel pages. Any data that ends
up hitting this hw rx queue will thus be dma'd into userspace
memory directly, without needing to be bounced through kernel
memory. 'Reading' data out of a socket instead becomes a
_notification_ mechanism, where the kernel tells userspace where
the data is. The overall approach is similar to the devmem TCP
proposal
This relies on hw header/data split, flow steering and RSS to
ensure packet headers remain in kernel memory and only desired
flows hit a hw rx queue configured for zero copy. Configuring this
is outside of the scope of this patchset.
We share netdev core infra with devmem TCP. The main difference is
that io_uring is used for the uAPI and the lifetime of all objects
are bound to an io_uring instance. Data is 'read' using a new
io_uring request type. When done, data is returned via a new shared
refill queue. A zero copy page pool refills a hw rx queue from this
refill queue directly. Of course, the lifetime of these data
buffers are managed by io_uring rather than the networking stack,
with different refcounting rules.
This patchset is the first step adding basic zero copy support. We
will extend this iteratively with new features e.g. dynamically
allocated zero copy areas, THP support, dmabuf support, improved
copy fallback, general optimisations and more'
In a local setup, I was able to saturate a 200G link with a single CPU
core, and at netdev conf 0x19 earlier this month, Jamal reported
188Gbit of bandwidth using a single core (no HT, including soft-irq).
Safe to say the efficiency is there, as bigger links would be needed
to find the per-core limit, and it's considerably more efficient and
faster than the existing devmem solution"
* tag 'for-6.15/io_uring-rx-zc-20250325' of git://git.kernel.dk/linux:
io_uring/zcrx: add selftest case for recvzc with read limit
io_uring/zcrx: add a read limit to recvzc requests
io_uring: add missing IORING_MAP_OFF_ZCRX_REGION in io_uring_mmap
io_uring: Rename KConfig to Kconfig
io_uring/zcrx: fix leaks on failed registration
io_uring/zcrx: recheck ifq on shutdown
io_uring/zcrx: add selftest
net: add documentation for io_uring zcrx
io_uring/zcrx: add copy fallback
io_uring/zcrx: throttle receive requests
io_uring/zcrx: set pp memory provider for an rx queue
io_uring/zcrx: add io_recvzc request
io_uring/zcrx: dma-map area for the device
io_uring/zcrx: implement zerocopy receive pp memory provider
io_uring/zcrx: grab a net device
io_uring/zcrx: add io_zcrx_area
io_uring/zcrx: add interface queue and refill queue
By default, io_uring marks a waiting task as being in iowait, if it's
sleeping waiting on events and there are pending requests. This isn't
necessarily always useful, and may be confusing on non-storage setups
where iowait isn't expected. It can also cause extra power usage, by
preventing the CPU from entering lower sleep states.
This adds a new enter flag, IORING_ENTER_NO_IOWAIT. If set, then
io_uring will not account the sleeping task as being in iowait. If the
kernel supports this feature, then it will be marked by having the
IORING_FEAT_NO_IOWAIT feature flag set.
As the kernel currently does not support separating the iowait
accounting and CPU frequency boosting, the IORING_ENTER_NO_IOWAIT
controls both of these at the same time. In the future, if those do end
up being split, then it'd be possible to control them separately.
However, it seems more likely that the kernel will decouple iowait and
CPU frequency boosting anyway.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
For existing epoll event loops that can't fully convert to io_uring,
the used approach is usually to add the io_uring fd to the epoll
instance and use epoll_wait() to wait on both "legacy" and io_uring
events. While this work, it isn't optimal as:
1) epoll_wait() is pretty limited in what it can do. It does not support
partial reaping of events, or waiting on a batch of events.
2) When an io_uring ring is added to an epoll instance, it activates the
io_uring "I'm being polled" logic which slows things down.
Rather than use this approach, with EPOLL_WAIT support added to io_uring,
event loops can use the normal io_uring wait logic for everything, as
long as an epoll wait request has been armed with io_uring.
Note that IORING_OP_EPOLL_WAIT does NOT take a timeout value, as this
is an async request. Waiting on io_uring events in general has various
timeout parameters, and those are the ones that should be used when
waiting on any kind of request. If events are immediately available for
reaping, then This opcode will return those immediately. If none are
available, then it will post an async completion when they become
available.
cqe->res will contain either an error code (< 0 value) for a malformed
request, invalid epoll instance, etc. It will return a positive result
indicating how many events were reaped.
IORING_OP_EPOLL_WAIT requests may be canceled using the normal io_uring
cancelation infrastructure.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This is obviously not that important, but when changes are synced back
from the kernel to liburing, the codespell CI ends up erroring because
of this misspelling. Let's just correct it and avoid this biting us
again on an import.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add io_uring opcode OP_RECV_ZC for doing zero copy reads out of a
socket. Only the connection should be land on the specific rx queue set
up for zero copy, and the socket must be handled by the io_uring
instance that the rx queue was registered for zero copy with. That's
because neither net_iovs / buffers from our queue can be read by outside
applications, nor zero copy is possible if traffic for the zero copy
connection goes to another queue. This coordination is outside of the
scope of this patch series. Also, any traffic directed to the zero copy
enabled queue is immediately visible to the application, which is why
CAP_NET_ADMIN is required at the registration step.
Of course, no data is actually read out of the socket, it has already
been copied by the netdev into userspace memory via DMA. OP_RECV_ZC
reads skbs out of the socket and checks that its frags are indeed
net_iovs that belong to io_uring. A cqe is queued for each one of these
frags.
Recall that each cqe is a big cqe, with the top half being an
io_uring_zcrx_cqe. The cqe res field contains the len or error. The
lower IORING_ZCRX_AREA_SHIFT bits of the struct io_uring_zcrx_cqe::off
field contain the offset relative to the start of the zero copy area.
The upper part of the off field is trivially zero, and will be used
to carry the area id.
For now, there is no limit as to how much work each OP_RECV_ZC request
does. It will attempt to drain a socket of all available data. This
request always operates in multishot mode.
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: David Wei <dw@davidwei.uk>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20250215000947.789731-7-dw@davidwei.uk
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add io_zcrx_area that represents a region of userspace memory that is
used for zero copy. During ifq registration, userspace passes in the
uaddr and len of userspace memory, which is then pinned by the kernel.
Each net_iov is mapped to one of these pages.
The freelist is a spinlock protected list that keeps track of all the
net_iovs/pages that aren't used.
For now, there is only one area per ifq and area registration happens
implicitly as part of ifq registration. There is no API for
adding/removing areas yet. The struct for area registration is there for
future extensibility once we support multiple areas and TCP devmem.
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: David Wei <dw@davidwei.uk>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20250215000947.789731-3-dw@davidwei.uk
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add a new object called an interface queue (ifq) that represents a net
rx queue that has been configured for zero copy. Each ifq is registered
using a new registration opcode IORING_REGISTER_ZCRX_IFQ.
The refill queue is allocated by the kernel and mapped by userspace
using a new offset IORING_OFF_RQ_RING, in a similar fashion to the main
SQ/CQ. It is used by userspace to return buffers that it is done with,
which will then be re-used by the netdev again.
The main CQ ring is used to notify userspace of received data by using
the upper 16 bytes of a big CQE as a new struct io_uring_zcrx_cqe. Each
entry contains the offset + len to the data.
For now, each io_uring instance only has a single ifq.
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: David Wei <dw@davidwei.uk>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20250215000947.789731-2-dw@davidwei.uk
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After commit 9a213d3b80c0, we can pass additional attributes along with
read/write. However, userspace doesn't know that. Add a new feature flag
IORING_FEAT_RW_ATTR, to notify the userspace that the kernel has this
ability.
Signed-off-by: Anuj Gupta <anuj20.g@samsung.com>
Reviewed-by: Li Zetao <lizetao1@huawei.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Tested-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20241205062109.1788-1-anuj20.g@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add the ability to pass additional attributes along with read/write.
Application can prepare attibute specific information and pass its
address using the SQE field:
__u64 attr_ptr;
Along with setting a mask indicating attributes being passed:
__u64 attr_type_mask;
Overall 64 attributes are allowed and currently one attribute
'IORING_RW_ATTR_FLAG_PI' is supported.
With PI attribute, userspace can pass following information:
- flags: integrity check flags IO_INTEGRITY_CHK_{GUARD/APPTAG/REFTAG}
- len: length of PI/metadata buffer
- addr: address of metadata buffer
- seed: seed value for reftag remapping
- app_tag: application defined 16b value
Process this information to prepare uio_meta_descriptor and pass it down
using kiocb->private.
PI attribute is supported only for direct IO.
Signed-off-by: Anuj Gupta <anuj20.g@samsung.com>
Signed-off-by: Kanchan Joshi <joshi.k@samsung.com>
Link: https://lore.kernel.org/r/20241128112240.8867-7-anuj20.g@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now we've got a more generic region registration API, place
IORING_ENTER_EXT_ARG_REG and re-enable it.
First, the user has to register a region with the
IORING_MEM_REGION_REG_WAIT_ARG flag set. It can only be done for a
ring in a disabled state, aka IORING_SETUP_R_DISABLED, to avoid races
with already running waiters. With that we should have stable constant
values for ctx->cq_wait_{size,arg} in io_get_ext_arg_reg() and hence no
READ_ONCE required.
The other API difference is that we're now passing byte offsets instead
of indexes. The user _must_ align all offsets / pointers to the native
word size, failing to do so might but not necessarily has to lead to a
failure usually returned as -EFAULT. liburing will be hiding this
details from users.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/81822c1b4ffbe8ad391b4f9ad1564def0d26d990.1731689588.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Regions will serve multiple purposes. First, with it we can decouple
ring/etc. object creation from registration / mapping of the memory they
will be placed in. We already have hacks that allow to put both SQ and
CQ into the same huge page, in the future we should be able to:
region = create_region(io_ring);
create_pbuf_ring(io_uring, region, offset=0);
create_pbuf_ring(io_uring, region, offset=N);
The second use case is efficiently passing parameters. The following
patch enables back on top of regions IORING_ENTER_EXT_ARG_REG, which
optimises wait arguments. It'll also be useful for request arguments
replacing iovecs, msghdr, etc. pointers. Eventually it would also be
handy for BPF as well if it comes to fruition.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0798cf3a14fad19cfc96fc9feca5f3e11481691d.1731689588.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We've got a good number of mappings we share with the userspace, that
includes the main rings, provided buffer rings, upcoming rings for
zerocopy rx and more. All of them duplicate user argument parsing and
some internal details as well (page pinnning, huge page optimisations,
mmap'ing, etc.)
Introduce a notion of regions. For userspace for now it's just a new
structure called struct io_uring_region_desc which is supposed to
parameterise all such mapping / queue creations. A region either
represents a user provided chunk of memory, in which case the user_addr
field should point to it, or a request for the kernel to allocate the
memory, in which case the user would need to mmap it after using the
offset returned in the mmap_offset field. With a uniform userspace API
we can avoid additional boiler plate code and apply future optimisation
to all of them at once.
Internally, there is a new structure struct io_mapped_region holding all
relevant runtime information and some helpers to work with it. This
patch limits it to user provided regions.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0e6fe25818dfbaebd1bd90b870a6cac503fe1a24.1731689588.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A new hybrid poll is implemented on the io_uring layer. Once an IO is
issued, it will not poll immediately, but rather block first and re-run
before IO complete, then poll to reap IO. While this poll method could
be a suboptimal solution when running on a single thread, it offers
performance lower than regular polling but higher than IRQ, and CPU
utilization is also lower than polling.
To use hybrid polling, the ring must be setup with both the
IORING_SETUP_IOPOLL and IORING_SETUP_HYBRID)IOPOLL flags set. Hybrid
polling has the same restrictions as IOPOLL, in that commands must
explicitly support it.
Signed-off-by: hexue <xue01.he@samsung.com>
Link: https://lore.kernel.org/r/20241101091957.564220-2-xue01.he@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently cloning a buffer table will fail if the destination already has
a table. But it should be possible to use it to replace existing elements.
Add a IORING_REGISTER_DST_REPLACE cloning flag, which if set, will allow
the destination to already having a buffer table. If that is the case,
then entries designated by offset + nr buffers will be replaced if they
already exist.
Note that it's allowed to use IORING_REGISTER_DST_REPLACE and not have
an existing table, in which case it'll work just like not having the
flag set and an empty table - it'll just assign the newly created table
for that case.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Right now buffer cloning is an all-or-nothing kind of thing - either the
whole table is cloned from a source to a destination ring, or nothing at
all.
However, it's not always desired to clone the whole thing. Allow for
the application to specify a source and destination offset, and a
number of buffers to clone. If the destination offset is non-zero, then
allocate sparse nodes upfront.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Useful for testing performance/efficiency impact of registered files
and buffers, vs (particularly) non-registered files.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Generally applications have 1 or a few waits of waiting, yet they pass
in a struct io_uring_getevents_arg every time. This needs to get copied
and, in turn, the timeout value needs to get copied.
Rather than do this for every invocation, allow the application to
register a fixed set of wait regions that can simply be indexed when
asking the kernel to wait on events.
At ring setup time, the application can register a number of these wait
regions and initialize region/index 0 upfront:
struct io_uring_reg_wait *reg;
reg = io_uring_setup_reg_wait(ring, nr_regions, &ret);
/* set timeout and mark as set, sigmask/sigmask_sz as needed */
reg->ts.tv_sec = 0;
reg->ts.tv_nsec = 100000;
reg->flags = IORING_REG_WAIT_TS;
where nr_regions >= 1 && nr_regions <= PAGE_SIZE / sizeof(*reg). The
above initializes index 0, but 63 other regions can be initialized,
if needed. Now, instead of doing:
struct __kernel_timespec timeout = { .tv_nsec = 100000, };
io_uring_submit_and_wait_timeout(ring, &cqe, nr, &t, NULL);
to wait for events for each submit_and_wait, or just wait, operation, it
can just reference the above region at offset 0 and do:
io_uring_submit_and_wait_reg(ring, &cqe, nr, 0);
to achieve the same goal of waiting 100usec without needing to copy
both struct io_uring_getevents_arg (24b) and struct __kernel_timeout
(16b) for each invocation. Struct io_uring_reg_wait looks as follows:
struct io_uring_reg_wait {
struct __kernel_timespec ts;
__u32 min_wait_usec;
__u32 flags;
__u64 sigmask;
__u32 sigmask_sz;
__u32 pad[3];
__u64 pad2[2];
};
embedding the timeout itself in the region, rather than passing it as
a pointer as well. Note that the signal mask is still passed as a
pointer, both for compatability reasons, but also because there doesn't
seem to be a lot of high frequency waits scenarios that involve setting
and resetting the signal mask for each wait.
The application is free to modify any region before a wait call, or it
can use keep multiple regions with different settings to avoid needing to
modify the same one for wait calls. Up to a page size of regions is mapped
by default, allowing PAGE_SIZE / 64 available regions for use.
The registered region must fit within a page. On a 4kb page size system,
that allows for 64 wait regions if a full page is used, as the size of
struct io_uring_reg_wait is 64b. The region registered must be aligned
to io_uring_reg_wait in size. It's valid to register less than 64
entries.
In network performance testing with zero-copy, this reduced the time
spent waiting on the TX side from 3.12% to 0.3% and the RX side from 4.4%
to 0.3%.
Wait regions are fixed for the lifetime of the ring - once registered,
they are persistent until the ring is torn down. The regions support
minimum wait timeout as well as the regular waits.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Once a ring has been created, the size of the CQ and SQ rings are fixed.
Usually this isn't a problem on the SQ ring side, as it merely controls
the available number of requests that can be submitted in a single
system call, and there's rarely a need to change that.
For the CQ ring, it's a different story. For most efficient use of
io_uring, it's important that the CQ ring never overflows. This means
that applications must size it for the worst case scenario, which can
be wasteful.
Add IORING_REGISTER_RESIZE_RINGS, which allows an application to resize
the existing rings. It takes a struct io_uring_params argument, the same
one which is used to setup the ring initially, and resizes rings
according to the sizes given.
Certain properties are always inherited from the original ring setup,
like SQE128/CQE32 and other setup options. The implementation only
allows flag associated with how the CQ ring is sized and clamped.
Existing unconsumed SQE and CQE entries are copied as part of the
process. If either the SQ or CQ resized destination ring cannot hold the
entries already present in the source rings, then the operation is failed
with -EOVERFLOW. Any register op holds ->uring_lock, which prevents new
submissions, and the internal mapping holds the completion lock as well
across moving CQ ring state.
To prevent races between mmap and ring resizing, add a mutex that's
solely used to serialize ring resize and mmap. mmap_sem can't be used
here, as as fork'ed process may be doing mmaps on the ring as well.
The ctx->resize_lock is held across mmap operations, and the resize
will grab it before swapping out the already mapped new data.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Normally MSG_RING requires both a source and a destination ring. But
some users don't always have a ring avilable to send a message from, yet
they still need to notify a target ring.
Add support for using io_uring_register(2) without having a source ring,
using a file descriptor of -1 for that. Internally those are called
blind registration opcodes. Implement IORING_REGISTER_SEND_MSG_RING as a
blind opcode, which simply takes an sqe that the application can put on
the stack and use the normal liburing helpers to initialize it. Then the
app can call:
io_uring_register(-1, IORING_REGISTER_SEND_MSG_RING, &sqe, 1);
and get the same behavior in terms of the target, where a CQE is posted
with the details given in the sqe.
For now this takes a single sqe pointer argument, and hence arg must
be set to that, and nr_args must be 1. Could easily be extended to take
an array of sqes, but for now let's keep it simple.
Link: https://lore.kernel.org/r/20240924115932.116167-3-axboe@kernel.dk
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A recent commit added support for copying registered buffers from one
ring to another. But that term is a bit confusing, as no copying of
buffer data is done here. What is being done is simply cloning the
buffer registrations from one ring to another.
Rename it while we still can, so that it's more descriptive. No
functional changes in this patch.
Fixes: 7cc2a6eadc ("io_uring: add IORING_REGISTER_COPY_BUFFERS method")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Buffers can get registered with io_uring, which allows to skip the
repeated pin_pages, unpin/unref pages for each O_DIRECT operation. This
reduces the overhead of O_DIRECT IO.
However, registrering buffers can take some time. Normally this isn't an
issue as it's done at initialization time (and hence less critical), but
for cases where rings can be created and destroyed as part of an IO
thread pool, registering the same buffers for multiple rings become a
more time sensitive proposition. As an example, let's say an application
has an IO memory pool of 500G. Initial registration takes:
Got 500 huge pages (each 1024MB)
Registered 500 pages in 409 msec
or about 0.4 seconds. If we go higher to 900 1GB huge pages being
registered:
Registered 900 pages in 738 msec
which is, as expected, a fully linear scaling.
Rather than have each ring pin/map/register the same buffer pool,
provide an io_uring_register(2) opcode to simply duplicate the buffers
that are registered with another ring. Adding the same 900GB of
registered buffers to the target ring can then be accomplished in:
Copied 900 pages in 17 usec
While timing differs a bit, this provides around a 25,000-40,000x
speedup for this use case.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
By default, any recv/read operation that uses provided buffers will
consume at least 1 buffer fully (and maybe more, in case of bundles).
This adds support for incremental consumption, meaning that an
application may add large buffers, and each read/recv will just consume
the part of the buffer that it needs.
For example, let's say an application registers 1MB buffers in a
provided buffer ring, for streaming receives. If it gets a short recv,
then the full 1MB buffer will be consumed and passed back to the
application. With incremental consumption, only the part that was
actually used is consumed, and the buffer remains the current one.
This means that both the application and the kernel needs to keep track
of what the current receive point is. Each recv will still pass back a
buffer ID and the size consumed, the only difference is that before the
next receive would always be the next buffer in the ring. Now the same
buffer ID may return multiple receives, each at an offset into that
buffer from where the previous receive left off. Example:
Application registers a provided buffer ring, and adds two 32K buffers
to the ring.
Buffer1 address: 0x1000000 (buffer ID 0)
Buffer2 address: 0x2000000 (buffer ID 1)
A recv completion is received with the following values:
cqe->res 0x1000 (4k bytes received)
cqe->flags 0x11 (CQE_F_BUFFER|CQE_F_BUF_MORE set, buffer ID 0)
and the application now knows that 4096b of data is available at
0x1000000, the start of that buffer, and that more data from this buffer
will be coming. Now the next receive comes in:
cqe->res 0x2010 (8k bytes received)
cqe->flags 0x11 (CQE_F_BUFFER|CQE_F_BUF_MORE set, buffer ID 0)
which tells the application that 8k is available where the last
completion left off, at 0x1001000. Next completion is:
cqe->res 0x5000 (20k bytes received)
cqe->flags 0x1 (CQE_F_BUFFER set, buffer ID 0)
and the application now knows that 20k of data is available at
0x1003000, which is where the previous receive ended. CQE_F_BUF_MORE
isn't set, as no more data is available in this buffer ID. The next
completion is then:
cqe->res 0x1000 (4k bytes received)
cqe->flags 0x10001 (CQE_F_BUFFER|CQE_F_BUF_MORE set, buffer ID 1)
which tells the application that buffer ID 1 is now the current one,
hence there's 4k of valid data at 0x2000000. 0x2001000 will be the next
receive point for this buffer ID.
When a buffer will be reused by future CQE completions,
IORING_CQE_BUF_MORE will be set in cqe->flags. This tells the application
that the kernel isn't done with the buffer yet, and that it should expect
more completions for this buffer ID. Will only be set by provided buffer
rings setup with IOU_PBUF_RING INC, as that's the only type of buffer
that will see multiple consecutive completions for the same buffer ID.
For any other provided buffer type, any completion that passes back
a buffer to the application is final.
Once a buffer has been fully consumed, the buffer ring head is
incremented and the next receive will indicate the next buffer ID in the
CQE cflags.
On the send side, the application can manage how much data is sent from
an existing buffer by setting sqe->len to the desired send length.
An application can request incremental consumption by setting
IOU_PBUF_RING_INC in the provided buffer ring registration. Outside of
that, any provided buffer ring setup and buffer additions is done like
before, no changes there. The only change is in how an application may
see multiple completions for the same buffer ID, hence needing to know
where the next receive will happen.
Note that like existing provided buffer rings, this should not be used
with IOSQE_ASYNC, as both really require the ring to remain locked over
the duration of the buffer selection and the operation completion. It
will consume a buffer otherwise regardless of the size of the IO done.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Expose min_wait_usec in io_uring_getevents_arg, replacing the pad member
that is currently in there. The value is in usecs, which is explained in
the name as well.
Note that if min_wait_usec and a normal timeout is used in conjunction,
the normal timeout is still relative to the base time. For example, if
min_wait_usec is set to 100 and the normal timeout is 1000, the max
total time waited is still 1000. This also means that if the normal
timeout is shorter than min_wait_usec, then only the min_wait_usec will
take effect.
See previous commit for an explanation of how this works.
IORING_FEAT_MIN_TIMEOUT is added as a feature flag for this, as
applications doing submit_and_wait_timeout() style operations will
generally not see the -EINVAL from the wait side as they return the
number of IOs submitted. Only if no IOs are submitted will the -EINVAL
bubble back up to the application.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add a new registration opcode IORING_REGISTER_CLOCK, which allows the
user to select which clock id it wants to use with CQ waiting timeouts.
It only allows a subset of all posix clocks and currently supports
CLOCK_MONOTONIC and CLOCK_BOOTTIME.
Suggested-by: Lewis Baker <lewissbaker@gmail.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/98f2bc8a3c36cdf8f0e6a275245e81e903459703.1723039801.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In addition to current relative timeouts for the waiting loop, where the
timespec argument specifies the maximum time it can wait for, add
support for the absolute mode, with the value carrying a CLOCK_MONOTONIC
absolute time until which we should return control back to the user.
Suggested-by: Lewis Baker <lewissbaker@gmail.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/4d5b74d67ada882590b2e42aa3aa7117bbf6b55f.1723039801.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
IORING_OP_LISTEN provides the semantic of listen(2) via io_uring. While
this is an essentially synchronous system call, the main point is to
enable a network path to execute fully with io_uring registered and
descriptorless files.
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
Link: https://lore.kernel.org/r/20240614163047.31581-4-krisman@suse.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
IORING_OP_BIND provides the semantic of bind(2) via io_uring. While
this is an essentially synchronous system call, the main point is to
enable a network path to execute fully with io_uring registered and
descriptorless files.
Signed-off-by: Gabriel Krisman Bertazi <krisman@suse.de>
Link: https://lore.kernel.org/r/20240614163047.31581-3-krisman@suse.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Support to inject result for NOP so that we can inject failure from
userspace. It is very helpful for covering failure handling code in
io_uring core change.
With nop flags, it becomes possible to add more test features on NOP in
future.
Suggested-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Ming Lei <ming.lei@redhat.com>
Link: https://lore.kernel.org/r/20240510035031.78874-3-ming.lei@redhat.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Similarly to how polling first is supported for receive, it makes sense
to provide the same for accept. An accept operation does a lot of
expensive setup, like allocating an fd, a socket/inode, etc. If no
connection request is already pending, this is wasted and will just be
cleaned up and freed, only to retry via the usual poll trigger.
Add IORING_ACCEPT_POLL_FIRST, which tells accept to only initiate the
accept request if poll says we have something to accept.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This allows the caller to perform a non-blocking attempt, similarly to
how recvmsg has MSG_DONTWAIT. If set, and we get -EAGAIN on a connection
attempt, propagate the result to userspace rather than arm poll and
wait for a retry.
Suggested-by: Norman Maurer <norman_maurer@apple.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If IORING_OP_RECV is used with provided buffers, the caller may also set
IORING_RECVSEND_BUNDLE to turn it into a multi-buffer recv. This grabs
buffers available and receives into them, posting a single completion for
all of it.
This can be used with multishot receive as well, or without it.
Now that both send and receive support bundles, add a feature flag for
it as well. If IORING_FEAT_RECVSEND_BUNDLE is set after registering the
ring, then the kernel supports bundles for recv and send.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
