Extract a helper for io_work_get_acct() and io_wqe_get_acct() to avoid
duplication.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

struct io_uring_task::sqpoll is not used anymore, kill it
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

io_match_task() matches all requests with PF_EXITING task, even though
those may be valid requests. It was necessary for SQPOLL cancellation,
but now it kills all requests before exiting via
io_uring_cancel_sqpoll(), so it's not needed.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

REQ_F_LINK_TIMEOUT is a hint that to look for linked timeouts to cancel,
we're leaving it even when it's already fired. Hence don't care to clear
it in io_kill_linked_timeout(), it's safe and is called only once.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Inline io_task_work_add() into io_req_task_work_add(). They both work
with a request, so keeping them separate doesn't make things much more
clear, but merging allows optimise it. Apart from small wins like not
reading req->ctx or not calculating @notify in the hot path, i.e. with
tctx->task_state set, it avoids doing wake_up_process() for every single
add, but only after actually done task_work_add().
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

io_put_file() doesn't do a good job at generating a good code. Inline
it, so we can check REQ_F_FIXED_FILE first, prioritising FIXED_FILE case
over requests without files, and saving a memory load in that case.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Reshuffle io_dismantle_req() checks to put most of slow path stuff under
a single if.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Inline io_clean_op(), leaving __io_clean_op() but renaming it. This will
be used in following patches.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Both io_req_complete_failed() and __io_req_task_cancel() do the same
thing: set failure flag, put both req refs and emit an CQE. The former
one is a bit more advance as it puts req back into a req cache, so make
it to take over __io_req_task_cancel() and remove the last one.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Add a new helper io_flush_cached_locked_reqs() that splices
locked_free_list to free_list, and does it right doing all sync and
invariant reinit.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We don't care about ret value in io_free_req_deferred(), make the code a
bit more concise.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

One big omission is that io_put_req() haven't been marked inline, and at
least gcc 9 doesn't inline it, not to mention that it's really hot and
extra function call is intolerable, especially when it doesn't put a
final ref.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

There are two problems:
1) we always allocate refnodes in advance and free them if those
haven't been used. It's expensive, takes two allocations, where one of
them is percpu. And it may be pretty common not actually using them.

2) Current API with allocating a refnode and setting some of the fields
is error prone, we don't ever want to have a file node runninng fixed
buffer callback...

Solve both with pre-init/get API. Pre-init just leaves the node for
later if not used, and for get (i.e. io_rsrc_refnode_get()), you need to
explicitly pass all arguments setting callbacks/etc., so it's more
resilient.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Emphasize that return value of io_flush_cached_reqs() depends on number
of requests in the cache. It looks nicer and might help tools from
false-negative analyses.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Move the case of successfully issued request by doing that check first.
It's not much of a difference, just generates slightly better code for
me.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Inline __io_queue_linked_timeout(), we don't need it
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Don't do a function call (io_dismantle_req()) in the middle and place it
to near other function calls, otherwise may lead to excessive register
spilling.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

First of all, w need to set tctx->sqpoll only when we add a new entry
into ->xa, so move it from the hot path. Also extract a hot path for
io_uring_add_task_file() as an inline helper.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Add unlikely annotations, because my compiler pretty much mispredicts
every first check, and apart jumping around in the fast path, it also
generates extra instructions, like in advance setting ret value.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

__tctx_task_work() guarantees that ctx won't be killed while running
task_works, so we can remove now unnecessary ctx pinning for internally
armed polling.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We can set canceled == true and complete out-of-line, ensure that we catch
that and correctly return -ECANCELED if the poll operation got canceled.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The correct function is io_iopoll_complete(), which deals with completions
of IOPOLL requests, not io_poll_complete().
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We have to dig quite deep to check for particularly whether or not a
file supports a fast-path nonblock attempt. For fixed files, we can do
this lookup once and cache the state instead.

This adds two new bits to track whether we support async read/write
attempt, and lines up the REQ_F_ISREG bit with those two. The file slot
re-uses the last 3 (or 2, for 32-bit) of the file pointer to cache that
state, and then we mask it in when we go and use a fixed file.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We don't allow them at registration time, so limit the check for needing
inflight tracking in io_file_get() to the non-fixed path.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Use a more comprehensible() max instead of hand coding it with ifs in
io_sqd_update_thread_idle().
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

io_uring manipulates references twice for each request, and hence is very
sensitive to performance of the reference count. This commit borrows a
trick from:

commit f958d7b5
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Thu Apr 11 10:06:20 2019 -0700

    mm: make page ref count overflow check tighter and more explicit

and switches to atomic_t for references, while still retaining overflow
and underflow checks.

This is good for a 2-3% increase in peak IOPS on a single core. Before:

IOPS=2970879, IOS/call=31/31, inflight=128 (128)
IOPS=2952597, IOS/call=31/31, inflight=128 (128)
IOPS=2943904, IOS/call=31/31, inflight=128 (128)
IOPS=2930006, IOS/call=31/31, inflight=96 (96)

and after:

IOPS=3054354f, IOS/call=31/31, inflight=128 (128)
IOPS=3059038, IOS/call=31/31, inflight=128 (128)
IOPS=3060320, IOS/call=31/31, inflight=128 (128)
IOPS=3068256, IOS/call=31/31, inflight=96 (96)
Signed-off-by: NJens Axboe <axboe@kernel.dk>

No functional changes in this patch, just in preparation for handling the
references a bit more efficiently.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

If not for async_data NULL check, io_resubmit_prep() is already an rw
specific version of io_req_prep_async(), but slower because 1) it always
goes through io_import_iovec() even if following io_setup_async_rw() the
result 2) instead of initialising iovec/iter in-place it does it
on-stack and then copies with io_setup_async_rw().
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Merge two function and do renaming in favour of the second one, it
relays the meaning better.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

needs_async_data controls allocation of async_data, and used in two
cases. 1) when async setup requires it (by io_req_prep_async() or
handler themselves), and 2) when op always needs additional space to
operate, like timeouts do.

Opcode preps already don't bother about the second case and do
allocation unconditionally, restrict needs_async_data to the first case
only and rename it into needs_async_setup.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
[axboe: update for IOPOLL fix]
Signed-off-by: NJens Axboe <axboe@kernel.dk>

All opcode handlers pretty well know whether they need async data or
not, and can skip testing for needs_async_data. The exception is rw
the generic path, but those test the flag by hand anyway. So, check the
flag and make io_alloc_async_data() allocating unconditionally.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

IORING_OP_[SEND,RECV] don't need async setup neither will get into
io_req_prep_async(). Remove them from io_req_prep_async() and remove
needs_async_data checks from the related setup functions.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

__io_cqring_fill_event() takes cflags as long to squeeze it into u32 in
an CQE, awhile all users pass int or unsigned. Replace it with unsigned
int and store it as u32 in struct io_completion to match CQE.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Always complete request holding the mutex instead of doing that strange
dancing with conditional ordering.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Add a simple helper doing CQE posting, marking request for link-failure,
and putting both submission and completion references.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

io_fixed_file_slot() and io_file_from_index() behave pretty similarly,
DRY and call one from another.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Use io_req_task_queue_fail() on the fail path of io_req_task_queue().
It's unlikely to happen, so don't care about additional overhead, but
allows to keep all the req->result invariant in a single function.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Don't bother to take a ctx->refs for io_req_task_cancel() because it
take uring_lock before putting a request, and the context is promised to
stay alive until unlock happens.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Pull btrfs fix from David Sterba:
 "One more patch that we'd like to get to 5.12 before release.

  It's changing where and how the superblock is stored in the zoned
  mode. It is an on-disk format change but so far there are no
  implications for users as the proper mkfs support hasn't been merged
  and is waiting for the kernel side to settle.

  Until now, the superblocks were derived from the zone index, but zone
  size can differ per device. This is changed to be based on fixed
  offset values, to make it independent of the device zone size.

  The work on that got a bit delayed, we discussed the exact locations
  to support potential device sizes and usecases. (Partially delayed
  also due to my vacation.) Having that in the same release where the
  zoned mode is declared usable is highly desired, there are userspace
  projects that need to be updated to recognize the feature. Pushing
  that to the next release would make things harder to test"

* tag 'for-5.12-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: zoned: move superblock logging zone location