We currently spin in iopoll() when requests to be iopolled are for
same file(device), while one device may have multiple hardware queues.
given an example:
hw_queue_0 | hw_queue_1
req(30us) req(10us)
If we first spin on iopolling for the hw_queue_0. the avg latency would
be (30us + 30us) / 2 = 30us. While if we do round robin, the avg
latency would be (30us + 10us) / 2 = 20us since we reap the request in
hw_queue_1 in time. So it's better to do spinning only when requests
are in same hardware queue.
Signed-off-by: Hao Xu <haoxu@linux.alibaba.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Most of requests are allocated from an internal cache, so it's waste of
time fully initialising them every time. Instead, let's pre-init some of
the fields we can during initial allocation (e.g. kmalloc(), see
io_alloc_req()) and keep them valid on request recycling. There are four
of them in this patch:
->ctx is always stays the same
->link is NULL on free, it's an invariant
->result is not even needed to init, just a precaution
->async_data we now clean in io_dismantle_req() as it's likely to
never be allocated.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/892ba0e71309bba9fe9e0142472330bbf9d8f05d.1624739600.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since cancellation got moved before exit_signals(), there is no one left
who can call io_run_task_work() with PF_EXIING set, so remove the check.
Note that __io_req_task_submit() still needs a similar check.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/f7f305ececb1e6044ea649fb983ca754805bb884.1624739600.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
gcc 11 goes a weird path and duplicates most of io_arm_poll_handler()
for READ and WRITE cases. Help it and move all pollin vs pollout
specific bits under a single if-else, so there is no temptation for this
kind of unfolding.
before vs after:
text data bss dec hex filename
85362 12650 8 98020 17ee4 ./fs/io_uring.o
85186 12650 8 97844 17e34 ./fs/io_uring.o
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1deea0037293a922a0358e2958384b2e42437885.1624739600.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
It is quite frequent that when an operation fails and returns EAGAIN,
the data becomes available between that failure and the call to
vfs_poll() done by io_arm_poll_handler().
Detecting the situation and reissuing the operation is much faster
than going ahead and push the operation to the io-wq.
Performance improvement testing has been performed with:
Single thread, 1 TCP connection receiving a 5 Mbps stream, no sqpoll.
4 measurements have been taken:
1. The time it takes to process a read request when data is already available
2. The time it takes to process by calling twice io_issue_sqe() after vfs_poll() indicated that data was available
3. The time it takes to execute io_queue_async_work()
4. The time it takes to complete a read request asynchronously
2.25% of all the read operations did use the new path.
ready data (baseline)
avg 3657.94182918628
min 580
max 20098
stddev 1213.15975908162
reissue completion
average 7882.67567567568
min 2316
max 28811
stddev 1982.79172973284
insert io-wq time
average 8983.82276995305
min 3324
max 87816
stddev 2551.60056552038
async time completion
average 24670.4758861127
min 10758
max 102612
stddev 3483.92416873804
Conclusion:
On average reissuing the sqe with the patch code is 1.1uSec faster and
in the worse case scenario 59uSec faster than placing the request on
io-wq
On average completion time by reissuing the sqe with the patch code is
16.79uSec faster and in the worse case scenario 73.8uSec faster than
async completion.
Signed-off-by: Olivier Langlois <olivier@trillion01.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/9e8441419bb1b8f3c3fcc607b2713efecdef2136.1624364038.git.olivier@trillion01.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We can't support IOPOLL with non-pollable request types, and we should
check for unused/reserved fields like we do for other request types.
Fixes: 14a1143b68 ("io_uring: add support for IORING_OP_UNLINKAT")
Cc: stable@vger.kernel.org
Reported-by: Dmitry Kadashev <dkadashev@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We can't support IOPOLL with non-pollable request types, and we should
check for unused/reserved fields like we do for other request types.
Fixes: 80a261fd00 ("io_uring: add support for IORING_OP_RENAMEAT")
Cc: stable@vger.kernel.org
Reported-by: Dmitry Kadashev <dkadashev@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The magic number used to cap the number of entries extracted from an
io_uring instance SQ before moving to the other instances is an
interesting parameter to experiment with.
A define has been created to make it easy to change its value from a
single location.
Signed-off-by: Olivier Langlois <olivier@trillion01.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/b401640063e77ad3e9f921e09c9b3ac10a8bb923.1624473200.git.olivier@trillion01.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If task_state is cleared, io_req_task_work_add() will go the slow path
adding a task_work, setting the task_state, waking up the task and so
on. Not to mention it's expensive. tctx_task_work() first clears the
state and then executes all the work items queued, so if any of them
resubmits or adds new task_work items, it would unnecessarily go through
the slow path of io_req_task_work_add().
Let's clear the ->task_state at the end. We still have to check
->task_list for emptiness afterward to synchronise with
io_req_task_work_add(), do that, and set the state back if we're going
to retry, because clearing not-ours task_state on the next iteration
would be buggy.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1ef72cdac7022adf0cd7ce4bfe3bb5c82a62eb93.1623949695.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Entering tctx_task_work() with empty task_list is a strange scenario,
that can happen only on rare occasion during task exit, so let's not
check for task_list emptiness in advance and do it do-while style. The
code still correct for the empty case, just would do extra work about
which we don't care.
Do extra step and do the check before cond_resched(), so we don't
resched if have nothing to execute.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/c4173e288e69793d03c7d7ce826f9d28afba718a.1623949695.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't need a full copy of tctx->task_list in tctx_task_work(), but
only a first one, so just assign node directly.
Taking into account that task_works are run in a context of a task,
it's very unlikely to first see non-empty tctx->task_list and then
splice it empty, can only happen with task_work cancellations that is
not-normal slow path anyway. Hence, get rid of the check in the end,
it's there not for validity but "performance" purposes.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d076c83fedb8253baf43acb23b8fafd7c5da1714.1623949695.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
tctx_task_work() tries to fetch a next batch of requests, but before it
would flush completions from the previous batch that may be sub-optimal.
E.g. io_req_task_queue() executes a head of the link where all the
linked may be enqueued through the same io_req_task_queue(). And there
are more cases for that.
Do the flushing at the end, so it can cache completions of several waves
of a single tctx_task_work(), and do the flush at the very end.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/3cac83934e4fbce520ff8025c3524398b3ae0270.1623949695.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently, if req->creds is not NULL, then there are creds assigned.
Track the invariant with a new flag in req->flags. No need to clear the
field at init, and also cleanup can be efficiently moved into
io_clean_op().
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/5f8baeb8d3b909487f555542350e2eac97005556.1623949695.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io-wq defaults to per-node masks for IO workers. This works fine by
default, but isn't particularly handy for workloads that prefer more
specific affinities, for either performance or isolation reasons.
This adds IORING_REGISTER_IOWQ_AFF that allows the user to pass in a CPU
mask that is then applied to IO thread workers, and an
IORING_UNREGISTER_IOWQ_AFF that simply resets the masks back to the
default of per-node.
Note that no care is given to existing IO threads, they will need to go
through a reschedule before the affinity is correct if they are already
running or sleeping.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The req pointer uniquely identify a specific request.
Having it in traces can provide valuable insights that is not possible
to have if the calling process is reusing the same user_data value.
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Olivier Langlois <olivier@trillion01.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In most cases io_commit_cqring() is just an smp_store_release(), and
it's hot enough, especially for IRQ rw, to want it to save on a function
call. Mark it inline and extract a non-inlined slow path doing drain
and timeout flushing. The inlined part is pretty slim to not cause
binary bloating.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7350f8b6b92caa50a48a80be39909f0d83eddd93.1623772051.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Place all drain_next logic into io_drain_req(), so it's never executed
if there was no drained requests before. The only thing we need is to
set ->drain_active if we see a request with IOSQE_IO_DRAIN, do that in
io_init_req() where flags are definitely in registers.
Also, all drain-related code is encapsulated in io_drain_req(), makes it
cleaner.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/68bf4f7395ddaafbf1a26bd97b57d57d45a9f900.1623772051.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
->drain_used is one way, which is not optimal if users use DRAIN but
very rarely. However, we can just clear it in io_drain_req() when all
drained before requests are gone. Also rename the flag to reflect the
change and be more clear about it.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7f37a240857546a94df6348507edddacab150460.1623772051.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
fs/io_uring.c: In function 'io_alloc_page_table':
include/linux/minmax.h:20:28: warning: comparison of distinct pointer
types lacks a cast
Cast everything to size_t using min_t.
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Fixes: 9123c8ffce ("io_uring: add helpers for 2 level table alloc")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/50f420a956bca070a43810d4a805293ed54f39d8.1623759527.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The sqe_ptr argument has been gone since 709b302fad (io_uring:
simplify io_get_sqring, 2020-04-08), made the return value of the
function. Update the comment accordingly.
Signed-off-by: Fam Zheng <fam.zheng@bytedance.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20210604164256.12242-1-fam.zheng@bytedance.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Replace drain checks with one-way flag set upon seeing the first
IOSQE_IO_DRAIN request. There are several places where it cuts cycles
well:
1) It's much faster than the fast check with two
conditions in io_drain_req() including pretty complex
list_empty_careful().
2) We can mark io_queue_sqe() inline now, that's a huge win.
3) It replaces timeout and drain checks in io_commit_cqring() with a
single flags test. Also great not touching ->defer_list there without a
reason so limiting cache bouncing.
It adds a small amount of overhead to drain path, but it's negligible.
The main nuisance is that once it meets any DRAIN request in io_uring
instance lifetime it will _always_ go through a slower path, so
drain-less and offset-mode timeout less applications are preferable.
The overhead in that case would be not big, but it's worth to bear in
mind.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/98d2fff8c4da5144bb0d08499f591d4768128ea3.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
->uring_lock is prevalently used for submission, even though it protects
many other things like iopoll, registeration, selected bufs, and more.
And it's placed together with ->cq_wait poked on completion and CQ
waiting sides. Move them apart, ->uring_lock goes to the submission
data, and cq_wait to completion related chunk. The last one requires
some reshuffling so everything needed by io_cqring_ev_posted*() is in
one cacheline.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/dea5e845caee4c98aa0922b46d713154d81f7bd8.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We use several wait_queue_head's for different purposes, but namings are
confusing. First rename ctx->cq_wait into ctx->poll_wait, because this
one is used for polling an io_uring instance. Then rename ctx->wait into
ctx->cq_wait, which is responsible for CQE waiting.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/47b97a097780c86c67b20b6ccc4e077523dce682.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There are no users of ->sq_check_overflow, only ->cq_check_overflow is
used. Combine it and move out of completion related part of struct
io_ring_ctx.
A not so obvious benefit of it is fitting all completion side fields
into a single cacheline. It was taking 2 lines before with 56B padding,
and io_cqring_ev_posted*() were still touching both of them.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/25927394964df31d113e3c729416af573afff5f5.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
submit_state.link is used only to assemble a link and not used for
actual submission, so clear it before io_queue_sqe() in io_submit_sqe(),
awhile it's hot and in caches and queueing doesn't spoil it. May also
potentially help compiler with spilling or to do other optimisations.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1579939426f3ad6b55af3005b1389bbbed7d780d.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_commit_cqring() might be very hot and we definitely don't want to
touch ->timeout_list there, because 1) it's shared with the submission
side so might lead to cache bouncing and 2) may need to load an extra
cache line, especially for IRQ completions.
We're interested in it at the completion side only when there are
offset-mode timeouts, which are not so popular. Replace
list_empty(->timeout_list) hot path check with a new one-way flag, which
is set when we prepare the first offset-mode timeout.
note: the flag sits in the same line as briefly used after ->rings
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e4892ec68b71a69f92ffbea4a1499be3ec0d463b.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Kill ->cached_sq_dropped and wire DRAIN sequence number correction via
->cq_extra, which is there exactly for that purpose. User visible
dropped counter will be populated by incrementing it instead of keeping
a copy, similarly as it was done not so long ago with cq_overflow.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/088aceb2707a534d531e2770267c4498e0507cc1.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since moving locked_free_* out of struct io_submit_state
ctx->submit_state is accessed on submission side only, so move it into
the submission section. Same goes for rsrc table pointers/nodes/etc.,
they must be taken and checked during submission because sync'ed by
uring_lock, so move them there as well.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/8a5899a50afc6ccca63249e716f580b246f3dec6.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
ctx->flags are heavily used by both, completion and submission sides, so
move it out from the ctx fields related to submissions. Instead, place
it together with ctx->refs, because it's already cacheline-aligned and
so pads lots of space, and both almost never change. Also, in most
occasions they are accessed together as refs are taken at submission
time and put back during completion.
Do same with ctx->rings, where the pointer itself is never modified
apart from ring init/free.
Note: in percpu mode, struct percpu_ref doesn't modify the struct itself
but takes indirection with ref->percpu_count_ptr.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/4c48c173e63d35591383ba2b87e8b8e8dfdbd23d.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
sq_array and sq_sqes are always used together, however they are in
different cachelines, where the borderline is right before
cq_overflow_list is rather rarely touched. Move the fields together so
it loads only one cacheline.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/3ef2411a94874da06492506a8897eff679244f49.1623709150.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Static analysis is warning that the sizeof being used is should be
of *data->tags[i] and not data->tags[i]. Although these are the same
size on 64 bit systems it is not a portable assumption to assume
this is true for all cases. Fix this by using a temporary pointer
tag_slot to make the code a clearer.
Addresses-Coverity: ("Sizeof not portable")
Fixes: d878c81610 ("io_uring: hide rsrc tag copy into generic helpers")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20210615130011.57387-1-colin.king@canonical.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
tctx in submission part is always synchronised because is executed from
the task's context, so we can batch allocate tctx/task references and
store them across syscall boundaries. It avoids enough of operations,
including an atomic for getting task ref and a percpu_counter_add()
function call, which still fallback to spinlock for large batching
cases (around >=32). Should be good for SQPOLL submitting in small
portions and coming at some moment bpf submissions.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/14b327b973410a3eec1f702ecf650e100513aca9.1623634181.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There is a complaint against sys_io_uring_enter() blocking if it submits
stdin reads. The problem is in __io_file_supports_async(), which
sees that it's a cdev and allows it to be processed inline.
Punt char devices using generic rules of io_file_supports_async(),
including checking for presence of *_iter() versions of rw callbacks.
Apparently, it will affect most of cdevs with some exceptions like
null and zero devices.
Cc: stable@vger.kernel.org
Reported-by: Birk Hirdman <lonjil@gmail.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d60270856b8a4560a639ef5f76e55eb563633599.1623236455.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Relax buffer registration restictions, which filters out file backed
memory, and allow shmem/memfd as they have normal anonymous pages
underneath.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
struct io_submit_state contains struct io_comp_state and so
locked_free_*, that renders cachelines around ->locked_free* being
invalidated on most non-inline completions, that may terrorise caches if
submissions and completions are done by different tasks.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/290cb5412b76892e8631978ee8ab9db0c6290dd5.1621201931.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There are two copies of cq_overflow, shared with userspace and internal
cached one. It was needed for DRAIN accounting, but now we have yet
another knob to tune the accounting, i.e. cq_extra, and we can throw
away the internal counter and just increment the one in the shared ring.
If user modifies it as so never gets the right overflow value ever
again, it's its problem, even though before we would have restored it
back by next overflow.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/8427965f5175dd051febc63804909861109ce859.1621201931.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
ring has two types of resource-related fields: used for request
submission, and field needed for update/registration. Reshuffle them
into these two groups for better locality and readability. The second
group is not in the hot path, so it's natural to place them somewhere in
the end. Also update an outdated comment.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/05b34795bb4440f4ec4510f08abd5a31830f8ca0.1621201931.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The main difference is in req_set_fail_links() renamed into
req_set_fail(), which now sets REQ_F_FAIL_LINK/REQ_F_FAIL flag
unconditional on whether it has been a link or not. It only matters in
io_disarm_next(), which already handles it well, and all calls to it
have a fast path checking REQ_F_LINK/HARDLINK.
It looks cleaner, and sheds binary size
text data bss dec hex filename
84235 12390 8 96633 17979 ./fs/io_uring.o
84151 12414 8 96573 1793d ./fs/io_uring.o
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e2224154dd6e53b665ac835d29436b177872fa10.1621201931.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Going through submission in __io_sq_thread() and still having a full SQ
is rather unexpected, so remove a check for SQ fullness and just wake up
whoever wait on sqo_sq_wait. Also skip if it doesn't do submission in
the first place, likely may to happen for SQPOLL sharing and/or IOPOLL.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e2e91751e87b1a39f8d63ef884aaff578123f61e.1621201931.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add IORING_FEAT_RSRC_TAGS indicating that io_uring supports a bunch of
new IORING_REGISTER operations, in particular
IORING_REGISTER_[FILES[,UPDATE]2,BUFFERS[2,UPDATE]] that support rsrc
tagging, and also indicating implemented dynamic fixed buffer updates.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/9b995d4045b6c6b4ab7510ca124fd25ac2203af7.1623339162.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There are ABI moments about recently added rsrc registration/update and
tagging that might become a nuisance in the future. First,
IORING_REGISTER_RSRC[_UPD] hide different types of resources under it,
so breaks fine control over them by restrictions. It works for now, but
once those are wanted under restrictions it would require a rework.
It was also inconvenient trying to fit a new resource not supporting
all the features (e.g. dynamic update) into the interface, so better
to return to IORING_REGISTER_* top level dispatching.
Second, register/update were considered to accept a type of resource,
however that's not a good idea because there might be several ways of
registration of a single resource type, e.g. we may want to add
non-contig buffers or anything more exquisite as dma mapped memory.
So, remove IORING_RSRC_[FILE,BUFFER] out of the ABI, and place them
internally for now to limit changes.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/9b554897a7c17ad6e3becc48dfed2f7af9f423d5.1623339162.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
As Andres reports "... io_sqe_buffer_register() doesn't initialize imu.
io_buffer_account_pin() does imu->acct_pages++, before calling
io_account_mem(ctx, imu->acct_pages).", leading to evevntual -ENOMEM.
Initialise the field.
Reported-by: Andres Freund <andres@anarazel.de>
Fixes: 41edf1a5ec ("io_uring: keep table of pointers to ubufs")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/438a6f46739ae5e05d9c75a0c8fa235320ff367c.1622285901.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Commit ba5ef6dc8a ("io_uring: fortify tctx/io_wq cleanup") introduced
setting tctx->io_wq to NULL a bit earlier. This has caused KCSAN to
detect a data race between accesses to tctx->io_wq:
write to 0xffff88811d8df330 of 8 bytes by task 3709 on cpu 1:
io_uring_clean_tctx fs/io_uring.c:9042 [inline]
__io_uring_cancel fs/io_uring.c:9136
io_uring_files_cancel include/linux/io_uring.h:16 [inline]
do_exit kernel/exit.c:781
do_group_exit kernel/exit.c:923
get_signal kernel/signal.c:2835
arch_do_signal_or_restart arch/x86/kernel/signal.c:789
handle_signal_work kernel/entry/common.c:147 [inline]
exit_to_user_mode_loop kernel/entry/common.c:171 [inline]
...
read to 0xffff88811d8df330 of 8 bytes by task 6412 on cpu 0:
io_uring_try_cancel_iowq fs/io_uring.c:8911 [inline]
io_uring_try_cancel_requests fs/io_uring.c:8933
io_ring_exit_work fs/io_uring.c:8736
process_one_work kernel/workqueue.c:2276
...
With the config used, KCSAN only reports data races with value changes:
this implies that in the case here we also know that tctx->io_wq was
non-NULL. Therefore, depending on interleaving, we may end up with:
[CPU 0] | [CPU 1]
io_uring_try_cancel_iowq() | io_uring_clean_tctx()
if (!tctx->io_wq) // false | ...
... | tctx->io_wq = NULL
io_wq_cancel_cb(tctx->io_wq, ...) | ...
-> NULL-deref |
Note: It is likely that thus far we've gotten lucky and the compiler
optimizes the double-read into a single read into a register -- but this
is never guaranteed, and can easily change with a different config!
Fix the data race by restoring the previous behaviour, where both
setting io_wq to NULL and put of the wq are _serialized_ after
concurrent io_uring_try_cancel_iowq() via acquisition of the uring_lock
and removal of the node in io_uring_del_task_file().
Fixes: ba5ef6dc8a ("io_uring: fortify tctx/io_wq cleanup")
Suggested-by: Pavel Begunkov <asml.silence@gmail.com>
Reported-by: syzbot+bf2b3d0435b9b728946c@syzkaller.appspotmail.com
Signed-off-by: Marco Elver <elver@google.com>
Cc: Jens Axboe <axboe@kernel.dk>
Link: https://lore.kernel.org/r/20210527092547.2656514-1-elver@google.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There is an old problem with io-wq cancellation where requests should be
killed and are in io-wq but are not discoverable, e.g. in @next_hashed
or @linked vars of io_worker_handle_work(). It adds some unreliability
to individual request canellation, but also may potentially get
__io_uring_cancel() stuck. For instance:
1) An __io_uring_cancel()'s cancellation round have not found any
request but there are some as desribed.
2) __io_uring_cancel() goes to sleep
3) Then workers wake up and try to execute those hidden requests
that happen to be unbound.
As we already cancel all requests of io-wq there, set IO_WQ_BIT_EXIT
in advance, so preventing 3) from executing unbound requests. The
workers will initially break looping because of getting a signal as they
are threads of the dying/exec()'ing user task.
Cc: stable@vger.kernel.org
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/abfcf8c54cb9e8f7bfbad7e9a0cc5433cc70bdc2.1621781238.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't want anyone poking into tctx->io_wq awhile it's being destroyed
by io_wq_put_and_exit(), and even though it shouldn't even happen, if
buggy would be preferable to get a NULL-deref instead of subtle delayed
failure or UAF.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/827b021de17926fd807610b3e53a5a5fa8530856.1621513214.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Since recent changes instead of storing a large array of struct
io_mapped_ubuf, we store pointers to them, that is 4 times slimmer and
we should not to so worry about restricting max number of registererd
buffer slots, increase the limit 4 times.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d3dee1da37f46da416aa96a16bf9e5094e10584d.1620990371.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
WARNING: CPU: 0 PID: 10242 at lib/refcount.c:28 refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
RIP: 0010:refcount_warn_saturate+0x15b/0x1a0 lib/refcount.c:28
Call Trace:
__refcount_sub_and_test include/linux/refcount.h:283 [inline]
__refcount_dec_and_test include/linux/refcount.h:315 [inline]
refcount_dec_and_test include/linux/refcount.h:333 [inline]
io_put_req fs/io_uring.c:2140 [inline]
io_queue_linked_timeout fs/io_uring.c:6300 [inline]
__io_queue_sqe+0xbef/0xec0 fs/io_uring.c:6354
io_submit_sqe fs/io_uring.c:6534 [inline]
io_submit_sqes+0x2bbd/0x7c50 fs/io_uring.c:6660
__do_sys_io_uring_enter fs/io_uring.c:9240 [inline]
__se_sys_io_uring_enter+0x256/0x1d60 fs/io_uring.c:9182
io_link_timeout_fn() should put only one reference of the linked timeout
request, however in case of racing with the master request's completion
first io_req_complete() puts one and then io_put_req_deferred() is
called.
Cc: stable@vger.kernel.org # 5.12+
Fixes: 9ae1f8dd37 ("io_uring: fix inconsistent lock state")
Reported-by: syzbot+a2910119328ce8e7996f@syzkaller.appspotmail.com
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/ff51018ff29de5ffa76f09273ef48cb24c720368.1620417627.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Read and write operations are capped to MAX_RW_COUNT. Some read ops rely on
that limit, and that is not guaranteed by the IORING_OP_PROVIDE_BUFFERS.
Truncate those lengths when doing io_add_buffers, so buffer addresses still
use the uncapped length.
Also, take the chance and change struct io_buffer len member to __u32, so
it matches struct io_provide_buffer len member.
This fixes CVE-2021-3491, also reported as ZDI-CAN-13546.
Fixes: ddf0322db7 ("io_uring: add IORING_OP_PROVIDE_BUFFERS")
Reported-by: Billy Jheng Bing-Jhong (@st424204)
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@canonical.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Currently the -EINVAL error return path is leaking memory allocated
to data. Fix this by not returning immediately but instead setting
the error return variable to -EINVAL and breaking out of the loop.
Kudos to Pavel Begunkov for suggesting a correct fix.
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/20210429104602.62676-1-colin.king@canonical.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Allow empty reg buffer slots any request using which should fail. This
allows users to not register all buffers in advance, but do it lazily
and/or on demand via updates. That is achieved by setting iov_base and
iov_len to zero for registration and/or buffer updates. Empty buffer
can't have a non-zero tag.
Implementation details: to not add extra overhead to io_import_fixed(),
create a dummy buffer crafted to fail any request using it, and set it
to all empty buffer slots.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7e95e4d700082baaf010c648c72ac764c9cc8826.1619611868.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
CQE flags take one byte that we store in req->flags together with other
REQ_F_* internal flags. CQE flags are copied directly into req and then
verified that requires some handling on failures, e.g. to make sure that
that copy doesn't set some of the internal flags.
Move all internal flags to take bits after the first byte, so we don't
need extra handling and make it safer overall.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/b8b5b02d1ab9d786fcc7db4a3fe86db6b70b8987.1619536280.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
-----BEGIN PGP SIGNATURE-----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=ZJ8L
-----END PGP SIGNATURE-----
Merge tag 'for-5.13/io_uring-2021-04-27' of git://git.kernel.dk/linux-block
Pull io_uring updates from Jens Axboe:
- Support for multi-shot mode for POLL requests
- More efficient reference counting. This is shamelessly stolen from
the mm side. Even though referencing is mostly single/dual user, the
128 count was retained to keep the code the same. Maybe this
should/could be made generic at some point.
- Removal of the need to have a manager thread for each ring. The
manager threads only job was checking and creating new io-threads as
needed, instead we handle this from the queue path.
- Allow SQPOLL without CAP_SYS_ADMIN or CAP_SYS_NICE. Since 5.12, this
thread is "just" a regular application thread, so no need to restrict
use of it anymore.
- Cleanup of how internal async poll data lifetime is managed.
- Fix for syzbot reported crash on SQPOLL cancelation.
- Make buffer registration more like file registrations, which includes
flexibility in avoiding full set unregistration and re-registration.
- Fix for io-wq affinity setting.
- Be a bit more defensive in task->pf_io_worker setup.
- Various SQPOLL fixes.
- Cleanup of SQPOLL creds handling.
- Improvements to in-flight request tracking.
- File registration cleanups.
- Tons of cleanups and little fixes
* tag 'for-5.13/io_uring-2021-04-27' of git://git.kernel.dk/linux-block: (156 commits)
io_uring: maintain drain logic for multishot poll requests
io_uring: Check current->io_uring in io_uring_cancel_sqpoll
io_uring: fix NULL reg-buffer
io_uring: simplify SQPOLL cancellations
io_uring: fix work_exit sqpoll cancellations
io_uring: Fix uninitialized variable up.resv
io_uring: fix invalid error check after malloc
io_uring: io_sq_thread() no longer needs to reset current->pf_io_worker
kernel: always initialize task->pf_io_worker to NULL
io_uring: update sq_thread_idle after ctx deleted
io_uring: add full-fledged dynamic buffers support
io_uring: implement fixed buffers registration similar to fixed files
io_uring: prepare fixed rw for dynanic buffers
io_uring: keep table of pointers to ubufs
io_uring: add generic rsrc update with tags
io_uring: add IORING_REGISTER_RSRC
io_uring: enumerate dynamic resources
io_uring: add generic path for rsrc update
io_uring: preparation for rsrc tagging
io_uring: decouple CQE filling from requests
...
Now that we have multishot poll requests, one SQE can emit multiple
CQEs. given below example:
sqe0(multishot poll)-->sqe1-->sqe2(drain req)
sqe2 is designed to issue after sqe0 and sqe1 completed, but since sqe0
is a multishot poll request, sqe2 may be issued after sqe0's event
triggered twice before sqe1 completed. This isn't what users leverage
drain requests for.
Here the solution is to wait for multishot poll requests fully
completed.
To achieve this, we should reconsider the req_need_defer equation, the
original one is:
all_sqes(excluding dropped ones) == all_cqes(including dropped ones)
This means we issue a drain request when all the previous submitted
SQEs have generated their CQEs.
Now we should consider multishot requests, we deduct all the multishot
CQEs except the cancellation one, In this way a multishot poll request
behave like a normal request, so:
all_sqes == all_cqes - multishot_cqes(except cancellations)
Here we introduce cq_extra for it.
Signed-off-by: Hao Xu <haoxu@linux.alibaba.com>
Link: https://lore.kernel.org/r/1618298439-136286-1-git-send-email-haoxu@linux.alibaba.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
syzkaller identified KASAN: null-ptr-deref Write in
io_uring_cancel_sqpoll.
io_uring_cancel_sqpoll is called by io_sq_thread before calling
io_uring_alloc_task_context. This leads to current->io_uring being NULL.
io_uring_cancel_sqpoll should not have to deal with threads where
current->io_uring is NULL.
In order to cast a wider safety net, perform input sanitisation directly
in io_uring_cancel_sqpoll and return for NULL value of current->io_uring.
This is safe since if current->io_uring isn't set, then there's no way
for the task to have submitted any requests.
Reported-by: syzbot+be51ca5a4d97f017cd50@syzkaller.appspotmail.com
Cc: stable@vger.kernel.org
Signed-off-by: Palash Oswal <hello@oswalpalash.com>
Link: https://lore.kernel.org/r/20210427125148.21816-1-hello@oswalpalash.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_import_fixed() doesn't expect a registered buffer slot to be NULL and
would fail stumbling on it. We don't allow it, but if during
__io_sqe_buffers_update() rsrc removal succeeds but following register
fails, we'll get such a situation.
Do it atomically and don't remove buffers until we sure that a new one
can be set.
Fixes: 634d00df5e ("io_uring: add full-fledged dynamic buffers support")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/830020f9c387acddd51962a3123b5566571b8c6d.1619446608.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
All sqpoll rings (even sharing sqpoll task) are currently dead bound
to the task that created them, iow when owner task dies it kills all
its SQPOLL rings and their inflight requests via task_work infra. It's
neither the nicist way nor the most convenient as adds extra
locking/waiting and dependencies.
Leave it alone and rely on SIGKILL being delivered on its thread group
exit, so there are only two cases left:
1) thread group is dying, so sqpoll task gets a signal and exit itself
cancelling all requests.
2) an sqpoll ring is dying. Because refs_kill() is called the sqpoll not
going to submit any new request, and that's what we need. And
io_ring_exit_work() will do all the cancellation itself before
actually killing ctx, so sqpoll doesn't need to worry about it.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/3cd7f166b9c326a2c932b70e71a655b03257b366.1619389911.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
After closing an SQPOLL ring, io_ring_exit_work() kicks in and starts
doing cancellations via io_uring_try_cancel_requests(). It will go
through io_uring_try_cancel_iowq(), which uses ctx->tctx_list, but as
SQPOLL task don't have a ctx note, its io-wq won't be reachable and so
is left not cancelled.
It will eventually cancelled when one of the tasks dies, but if a thread
group survives for long and changes rings, it will spawn lots of
unreclaimed resources and live locked works.
Cancel SQPOLL task's io-wq separately in io_ring_exit_work().
Cc: stable@vger.kernel.org
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/a71a7fe345135d684025bb529d5cb1d8d6b46e10.1619389911.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The variable up.resv is not initialized and is being checking for a
non-zero value in the call to _io_register_rsrc_update. Fix this by
explicitly setting the variable to 0.
Addresses-Coverity: ("Uninitialized scalar variable)"
Fixes: c3bdad0271 ("io_uring: add generic rsrc update with tags")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20210426094735.8320-1-colin.king@canonical.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
With dynamic buffer updates, registered buffers in the table may change
at any moment. First of all we want to prevent future races between
updating and importing (i.e. io_import_fixed()), where the latter one
may happen without uring_lock held, e.g. from io-wq.
Save the first loaded io_mapped_ubuf buffer and reuse.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/21a2302d07766ae956640b6f753292c45200fe8f.1619356238.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of keeping a table of ubufs convert them into pointers to ubuf,
so we can atomically read one pointer and be sure that the content of
ubuf won't change.
Because it was already dynamically allocating imu->bvec, throw both
imu and bvec into a single structure so they can be allocated together.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/b96efa4c5febadeccf41d0e849ac099f4c83b0d3.1619356238.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add a new io_uring_register() opcode for rsrc registeration. Instead of
accepting a pointer to resources, fds or iovecs, it @arg is now pointing
to a struct io_uring_rsrc_register, and the second argument tells how
large that struct is to make it easily extendible by adding new fields.
All that is done mainly to be able to pass in a pointer with tags. Pass
it in and enable CQE posting for file resources. Doesn't support setting
tags on update yet.
A design choice made here is to not post CQEs on rsrc de-registration,
but only when we updated-removed it by rsrc dynamic update.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/c498aaec32a4bb277b2406b9069662c02cdda98c.1619356238.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We need a way to notify userspace when a lazily removed resource
actually died out. This will be done by associating a tag, which is u64
exactly like req->user_data, with each rsrc (e.g. buffer of file). A CQE
will be posted once a resource is actually put down.
Tag 0 is a special value set by default, for whcih it don't generate an
CQE, so providing the old behaviour.
Don't expose it to the userspace yet, but prepare internally, allocate
buffers, add all posting hooks, etc.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/2e6beec5eabe7216bb61fb93cdf5aaf65812a9b0.1619356238.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
REQ_F_INFLIGHT deaccounting doesn't do any spinlocking or resource
freeing anymore, so it's safe to move it into the normal cleanup flow,
i.e. into io_clean_op(), so making it cleaner.
Also move io_req_needs_clean() to be first in io_dismantle_req() so it
doesn't reload req->flags.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/90653a3a5de4107e3a00536fa4c2ea5f2c38a4ac.1618916549.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
[ 736.982891] INFO: task iou-sqp-4294:4295 blocked for more than 122 seconds.
[ 736.982897] Call Trace:
[ 736.982901] schedule+0x68/0xe0
[ 736.982903] io_uring_cancel_sqpoll+0xdb/0x110
[ 736.982908] io_sqpoll_cancel_cb+0x24/0x30
[ 736.982911] io_run_task_work_head+0x28/0x50
[ 736.982913] io_sq_thread+0x4e3/0x720
We call io_uring_cancel_sqpoll() one by one for each ctx either in
sq_thread() itself or via task works, and it's intended to cancel all
requests of a specified context. However the function uses per-task
counters to track the number of inflight requests, so it counts more
requests than available via currect io_uring ctx and goes to sleep for
them to appear (e.g. from IRQ), that will never happen.
Cancel a bit more than before, i.e. all ctxs that share sqpoll
and continue to use shared counters. Don't forget that we should not
remove ctx from the list before running that task_work sqpoll-cancel,
otherwise the function wouldn't be able to find the context and will
hang.
Reported-by: Joakim Hassila <joj@mac.com>
Reported-by: Jens Axboe <axboe@kernel.dk>
Fixes: 37d1e2e364 ("io_uring: move SQPOLL thread io-wq forked worker")
Cc: stable@vger.kernel.org
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/1bded7e6c6b32e0bae25fce36be2868e46b116a0.1618752958.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Colin reported before possible overflow and sign extension problems in
io_provide_buffers_prep(). As Linus pointed out previous attempt did nothing
useful, see d81269fecb ("io_uring: fix provide_buffers sign extension").
Do that with help of check_<op>_overflow helpers. And fix struct
io_provide_buf::len type, as it doesn't make much sense to keep it
signed.
Reported-by: Colin Ian King <colin.king@canonical.com>
Fixes: efe68c1ca8 ("io_uring: validate the full range of provided buffers for access")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/46538827e70fce5f6cdb50897cff4cacc490f380.1618488258.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't fail submission attempts if there are CQEs in the overflow
backlog, but give away the decision making to the userspace. It
might be very inconvenient to the userspace, especially if
submission and completion are done by different threads.
We can remove it because of recent changes, where requests
are now not locked by the backlog, backlog entries are allocated
separately, so they take less space and cgroup accounted.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
A hand-edit while applying this patch on top of a new base resulted in
a reverted check for re-issue, resulting in spurious -EAGAIN errors.
Fixes: 8c130827f4 ("io_uring: don't alter iopoll reissue fail ret code")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We manage these separately right now, just tie it to the request lifetime
and make it be part of the usual REQ_F_NEED_CLEANUP logic.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We have this in two spots right now, which is a bit fragile. In
preparation for moving REQ_F_POLLED cleanup into the same spot, move
the check into a separate helper so we only have it once.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The re-add handling isn't correct for the multi wait case, so let's
just disable it for now explicitly until we can get that sorted out. This
just turns it into a one-shot request. Since we pass back whether or not
a poll request terminates in multishot mode on completion, this should
not break properly behaving applications that check for IORING_CQE_F_MORE
on completion.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
[ 245.463317] INFO: task iou-sqp-1374:1377 blocked for more than 122 seconds.
[ 245.463334] task:iou-sqp-1374 state:D flags:0x00004000
[ 245.463345] Call Trace:
[ 245.463352] __schedule+0x36b/0x950
[ 245.463376] schedule+0x68/0xe0
[ 245.463385] __io_uring_cancel+0xfb/0x1a0
[ 245.463407] do_exit+0xc0/0xb40
[ 245.463423] io_sq_thread+0x49b/0x710
[ 245.463445] ret_from_fork+0x22/0x30
It happens when sqpoll forgot to run park_task_work and goes to exit,
then exiting user may remove ctx from sqd_list, and so corresponding
io_sq_thread() -> io_uring_cancel_sqpoll() won't be executed. Hopefully
it just stucks in do_exit() in this case.
Fixes: dbe1bdbb39 ("io_uring: handle signals for IO threads like a normal thread")
Reported-by: Joakim Hassila <joj@mac.com>
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Having poll update function as a part of IORING_OP_POLL_ADD is not
great, we have to do hack around struct layouts and add some overhead in
the way of more popular POLL_ADD. Even more serious drawback is that
POLL_ADD requires file and always grabs it, and so poll update, which
doesn't need it.
Incorporate poll update into IORING_OP_POLL_REMOVE instead of
IORING_OP_POLL_ADD. It also more consistent with timeout remove/update.
Fixes: b69de288e9 ("io_uring: allow events and user_data update of running poll requests")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Isolate poll mask SQE parsing and preparations into a new function,
which will be reused shortly.
Fixes: b69de288e9 ("io_uring: allow events and user_data update of running poll requests")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't allow REQ_OP_POLL_REMOVE to kill apoll requests, users should not
know about it. Also, remove weird -EACCESS in io_poll_update(), it
shouldn't know anything about apoll, and have to work even if happened
to have a poll and an async poll'ed request with same user_data.
Fixes: b69de288e9 ("io_uring: allow events and user_data update of running poll requests")
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't reinit io_ring_exit_work()'s exit work/completions on each
iteration, that's wasteful. Also add list_rotate_left(), so if we failed
to complete the task job, we don't try it again and again but defer it
until others are processed.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The only way to get out of io_iopoll_getevents() and continue iterating
is to have empty iopoll_list, otherwise the main loop would just exit.
So, instead of the unlock on 8th time heuristic, do that based on
iopoll_list.
Also, as no one can add new requests to iopoll_list while
io_iopoll_check() hold uring_lock, it's useless to spin with the list
empty, return in that case.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/5b8ebe84f5fff7ffa1f708952dfef7fc78b668e2.1618278933.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
As CQE overflows are now untied from requests and so don't hold any
ref, we don't need to handle exiting/exec'ing cases there anymore.
Moreover, it's much nicer in regards to userspace to save overflowed
CQEs whenever possible, so remove failing on in_idle.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d873b7dab75c7f3039ead9628a745bea01f2cfd2.1618278933.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't save return values of hrtimer_try_to_cancel() in a variable, but
use right away. It's in general safer to not have an intermediate
variable, which may be reused and passed out wrongly, but it be
contracted out. Also clean io_timeout_extract().
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d2566ef7ce632e6882dc13e022a26249b3fd30b5.1618278933.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If io_sqe_files_unregister() faults on io_rsrc_ref_quiesce(), it will
fail to do unregister leaving files referenced. And that may well happen
because of a strayed signal or just because it does allocations inside.
In io_ring_ctx_free() do an unsafe version of unregister, as it's
guaranteed to not have requests by that point and so quiesce is useless.
Cc: stable@vger.kernel.org
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e696e9eade571b51997d0dc1d01f144c6d685c05.1618278933.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Revert of revert of "io_uring: wait potential ->release() on resurrect",
which adds a helper for resurrect not racing completion reinit, as was
removed because of a strange bug with no clear root or link to the
patch.
Was improved, instead of rcu_synchronize(), just wait_for_completion()
because we're at 0 refs and it will happen very shortly. Specifically
use non-interruptible version to ignore all pending signals that may
have ended prior interruptible wait.
This reverts commit cb5e1b8130.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7a080c20f686d026efade810b116b72f88abaff9.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
req_set_fail_links() condition checking is bulky. Even though it's
always in a slow path, it's inlined and generates lots of extra code,
simplify it be moving HARDLINK checking into helpers killing linked
requests.
text data bss dec hex filename
before: 79318 12330 8 91656 16608 ./fs/io_uring.o
after: 79126 12330 8 91464 16548 ./fs/io_uring.o
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/96a9387db658a9d5a44ecbfd57c2a62cb888c9b6.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We don't need to store file tables in rsrc nodes, for now it's easier to
handle tables not generically, so move file tables into the context. A
nice side effect is having one less pointer dereference for request with
fixed file initialisation.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/de9fc4cd3545f24c26c03be4556f58ba3d18b9c3.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
In preparation for more changes do a little cleanup of
io_sqe_buffers_register(). Move all args/invariant checking into it from
io_buffers_map_alloc(), because it's confusing. And add a bit more
cleaning for the loop.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/93292cb9708c8455e5070cc855861d94e11ca042.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Randy reports the following error on CONFIG_BLOCK not being set:
../fs/io_uring.c: In function ‘kiocb_done’:
../fs/io_uring.c:2766:7: error: implicit declaration of function ‘io_resubmit_prep’; did you mean ‘io_put_req’? [-Werror=implicit-function-declaration]
if (io_resubmit_prep(req)) {
Provide a dummy stub for io_resubmit_prep() like we do for
io_rw_should_reissue(), which also helps remove an ifdef sequence from
io_complete_rw_iopoll() as well.
Fixes: 8c130827f4 ("io_uring: don't alter iopoll reissue fail ret code")
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
There are three cases where we much care about performance of
io_cqring_fill_event() -- flushing inline completions, iopoll and
io_req_complete_post(). Inline a hot part of fill_event() into them.
All others are not as important and we don't want to bloat binary for
them, so add a noinline version of the function for all other use
use cases.
nops test(batch=32): 16.932 vs 17.822 KIOPS
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/a11d59424bf4417aca33f5ec21008bb3b0ebd11e.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Take advantage of delayed/inline completion flushing and pass right
issue flags for completion of open, open2, fadvise and poll remove
opcodes. All others either already use it or always punted and never
executed inline.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0badc7512e82f7350b73bb09abbebbecbdd5dab8.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of keeping requests in a inflight_list, just track them with a
per tctx atomic counter. Apart from it being much easier and more
consistent with task cancel, it frees ->inflight_entry from being shared
between iopoll and cancel-track, so less headache for us.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/3c2ee0863cd7eeefa605f3eaff4c1c461a6f1157.1618101759.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
hash_del() works well with non-hashed nodes, there's no need to check
if it is hashed first.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Remove error parameter from io_poll_complete(), 0 is always passed,
and do a bit of cleaning on top.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_poll_complete() always fills an event (even an overflowed one), so we
always should do io_cqring_ev_posted() afterwards. And that's what is
currently happening, because second EPOLLONESHOT check is always true,
it can't return !done for oneshots.
Remove those branching, it's much easier to read.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We currently allow racy updates to multishot requests, but we can end up
double adding the poll request if both completion and update does it.
Ensure that we skip re-add on the update side if someone else is
completing it.
Fixes: b69de288e9 ("io_uring: allow events and user_data update of running poll requests")
Reported-by: Joakim Hassila <joj@mac.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Add struct io_fixed_file representing a single registered file, first to
hide ugly struct file **, which may be misleading, and secondly to
retype it to unsigned long as conversions to it and back to file * for
handling and masking FFS_* flags are getting nasty.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/78669731a605a7614c577c3de552631cfaf0869a.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Introduce a heler io_free_file_tables() doing all the cleaning, there
are several places where it's hand coded. Also move all allocations into
io_sqe_alloc_file_tables() and rename it, so all of it is in one place.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/502a84ebf41ff119b095e59661e678eacb752bf8.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Don't put linked timeout req in io_async_find_and_cancel() but do it in
io_link_timeout_fn(), so we have only one point for that and won't have
to do it differently as it's now (put vs put_deferred). Btw, improve a
bit io_async_find_and_cancel()'s locking.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d75b70957f245275ab7cba83e0ac9c1b86aae78a.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Replace a hand-coded overflow check with a specialised function. Even
though compilers are smart enough to generate identical binary (i.e.
check carry bit), but it's more foolproof and conveys the intention
better.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/e437dcdc929bacbb6f11a4824ecbbf17225cb82a.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
rsrc_data refs should always be valid for potential submitters,
io_rsrc_ref_quiesce() restores it before unlocking, so
percpu_ref_is_dying() check in io_sqe_files_unregister() does nothing
and misleading. Concurrent quiesce is prevented with
struct io_rsrc_data::quiesce.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/bf97055e1748ee3a382e66daf384a469eb90b931.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If we're going to ever support multiple types of resources we need
shared rsrc nodes to not bloat requests, that is implemented in this
patch. It also gives a nicer API and saves one pointer dereference
in io_req_set_rsrc_node().
We may say that all requests bound to a resource belong to one and only
one rsrc node, and considering that nodes are removed and recycled
strictly in-order, this separates requests into generations, where
generation are changed on each node switch (i.e. io_rsrc_node_switch()).
The API is simple, io_rsrc_node_switch() switches to a new generation if
needed, and also optionally kills a passed in io_rsrc_data. Each call to
io_rsrc_node_switch() have to be preceded with
io_rsrc_node_switch_start(). The start function is idempotent and should
not necessarily be followed by switch.
One difference is that once a node was set it will always retain a valid
rsrc node, even on unregister. It may be a nuisance at the moment, but
makes much sense for multiple types of resources. Another thing changed
is that nodes are bound to/associated with a io_rsrc_data later just
before killing (i.e. switching).
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/7e9c693b4b9a2f47aa784b616ce29843021bb65a.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io_rsrc_node_get() and io_rsrc_node_set() are always used together,
merge them into one so most users don't even see io_rsrc_node and don't
need to care about it.
It helped to catch io_sqe_files_register() inferring rsrc data argument
for get and set differently, not a problem but a good sign.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/0827b080b2e61b3dec795380f7e1a1995595d41f.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Keep it consistent with update and use io_rsrc_node_prealloc() +
io_rsrc_node_get() in io_sqe_files_register() as well, that will be used
in future patches, not as error prone and allows to deduplicate
rsrc_node init.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/cf87321e6be5e38f4dc7fe5079d2aa6945b1ace0.1617287883.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Joakim reports that in some conditions he sees a multishot poll request
being canceled, and that it coincides with getting -EALREADY on
modification. As part of the poll update procedure, there's a small window
where the request is marked as canceled, and if this coincides with the
event actually triggering, then we can get a spurious -ECANCELED and
termination of the multishot request.
Don't mark the poll request as being canceled for update. We also don't
care if we race on removal unless it's a one-shot request, we can safely
updated for either case.
Fixes: b69de288e9 ("io_uring: allow events and user_data update of running poll requests")
Reported-by: Joakim Hassila <joj@mac.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Now that we have any worker being attached to the original task as
threads, accounting of CPU time is directly attributed to the original
task as well. This means that we no longer have to restrict SQPOLL to
needing elevated privileges, as it's really no different from just having
the task spawn a busy looping thread in userspace.
Reported-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
io-wq relies on a manager thread to create/fork new workers, as needed.
But there's really no strong need for it anymore. We have the following
cases that fork a new worker:
1) Work queue. This is done from the task itself always, and it's trivial
to create a worker off that path, if needed.
2) All workers have gone to sleep, and we have more work. This is called
off the sched out path. For this case, use a task_work items to queue
a fork-worker operation.
3) Hashed work completion. Don't think we need to do anything off this
case. If need be, it could just use approach 2 as well.
Part of this change is incrementing the running worker count before the
fork, to avoid cases where we observe we need a worker and then queue
creation of one. Then new work comes in, we fork a new one. That last
queue operation should have waited for the previous worker to come up,
it's quite possible we don't even need it. Hence move the worker running
from before we fork it off to more efficiently handle that case.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
This adds two new POLL_ADD flags, IORING_POLL_UPDATE_EVENTS and
IORING_POLL_UPDATE_USER_DATA. As with the other POLL_ADD flag, these are
masked into sqe->len. If set, the POLL_ADD will have the following
behavior:
- sqe->addr must contain the the user_data of the poll request that
needs to be modified. This field is otherwise invalid for a POLL_ADD
command.
- If IORING_POLL_UPDATE_EVENTS is set, sqe->poll_events must contain the
new mask for the existing poll request. There are no checks for whether
these are identical or not, if a matching poll request is found, then it
is re-armed with the new mask.
- If IORING_POLL_UPDATE_USER_DATA is set, sqe->off must contain the new
user_data for the existing poll request.
A POLL_ADD with any of these flags set may complete with any of the
following results:
1) 0, which means that we successfully found the existing poll request
specified, and performed the re-arm procedure. Any error from that
re-arm will be exposed as a completion event for that original poll
request, not for the update request.
2) -ENOENT, if no existing poll request was found with the given
user_data.
3) -EALREADY, if the existing poll request was already in the process of
being removed/canceled/completing.
4) -EACCES, if an attempt was made to modify an internal poll request
(eg not one originally issued ass IORING_OP_POLL_ADD).
The usual -EINVAL cases apply as well, if any invalid fields are set
in the sqe for this command type.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We'll need this helper for another purpose, for now just abstract it
out and have io_poll_cancel() use it for lookups.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
If we hit overflow and fail to allocate an overflow entry for the
completion, terminate the multishot poll mode.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
The default io_uring poll mode is one-shot, where once the event triggers,
the poll command is completed and won't trigger any further events. If
we're doing repeated polling on the same file or socket, then it can be
more efficient to do multishot, where we keep triggering whenever the
event becomes true.
This deviates from the usual norm of having one CQE per SQE submitted. Add
a CQE flag, IORING_CQE_F_MORE, which tells the application to expect
further completion events from the submitted SQE. Right now the only user
of this is POLL_ADD in multishot mode.
Since sqe->poll_events is using the space that we normally use for adding
flags to commands, use sqe->len for the flag space for POLL_ADD. Multishot
mode is selected by setting IORING_POLL_ADD_MULTI in sqe->len. An
application should expect more CQEs for the specificed SQE if the CQE is
flagged with IORING_CQE_F_MORE. In multishot mode, only cancelation or an
error will terminate the poll request, in which case the flag will be
cleared.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
We should be including the completion flags for better introspection on
exactly what completion event was logged.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of using a request itself for overflowed CQE stashing, allocate a
separate entry. The disadvantage is that the allocation may fail and it
will be accounted as lost (see rings->cq_overflow), so we lose reliability
in case of memory pressure if the application is driving the CQ ring into
overflow. However, it opens a way for for multiple CQEs per an SQE and
even generating SQE-less CQEs.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
[axboe: use GFP_ATOMIC | __GFP_ACCOUNT]
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Instead of masking these in as part of regular POLL_ADD prep, do it in
io_init_poll_iocb(), and include NVAL as that's generally unmaskable,
and RDHUP alongside the HUP that is already set.
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Expect read/write to succeed and create a hot path for this case, in
particular hide all error handling with resubmission under a single
check with the desired result.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Move iov_iter_revert() resetting iterator in case of -EIOCBQUEUED into
io_resubmit_prep(), so we don't do heavy revert in hot path, also saves
a couple of checks.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Hao Xu