We currently merge async work items if we see a strict sequential hit.
This helps avoid unnecessary workqueue switches when we don't need
them. We can extend this merging to cover cases where it's not a strict
sequential hit, but the IO still fits within the same page. If an
application is doing multiple requests within the same page, we don't
want separate workers waiting on the same page to complete IO. It's much
faster to let the first worker bring in the page, then operate on that
page from the same worker to complete the next request(s).
Reviewed-by: NJeff Moyer <jmoyer@redhat.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

All the popular filesystems need to grab the inode lock for buffered
writes. With io_uring punting buffered writes to async context, we
observe a lot of contention with all workers hamming this mutex.

For buffered writes, we generally don't need a lot of parallelism on
the submission side, as the flushing will take care of that for us.
Hence we don't need a deep queue on the write side, as long as we
can safely punt from the original submission context.

Add a workqueue with a limit of 2 that we can use for buffered writes.
This greatly improves the performance and efficiency of higher queue
depth buffered async writes with io_uring.
Reported-by: NAndres Freund <andres@anarazel.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Add a helper for queueing a request for async execution, in preparation
for optimizing it.

No functional change in this patch.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

For some applications that end up using a submit-and-wait type of
approach for certain batches of IO, we can make that a bit more
efficient by allowing the application to block for the last IO
submission. This prevents an async when we don't need it, as the
application will be blocking for the completion event(s) anyway.

Typical use cases are using the liburing
io_uring_submit_and_wait() API, or just using io_uring_enter()
doing both submissions and completions. As a specific example,
RocksDB doing MultiGet() is sped up quite a bit with this
change.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

To support the link with drain, we need to do two parts.

There is an sqes:

    0     1     2     3     4     5     6
 +-----+-----+-----+-----+-----+-----+-----+
 |  N  |  L  |  L  | L+D |  N  |  N  |  N  |
 +-----+-----+-----+-----+-----+-----+-----+

First, we need to ensure that the io before the link is completed,
there is a easy way is set drain flag to the link list's head, so
all subsequent io will be inserted into the defer_list.

	+-----+
    (0) |  N  |
	+-----+
           |          (2)         (3)         (4)
	+-----+     +-----+     +-----+     +-----+
    (1) | L+D | --> |  L  | --> | L+D | --> |  N  |
	+-----+     +-----+     +-----+     +-----+
           |
	+-----+
    (5) |  N  |
	+-----+
           |
	+-----+
    (6) |  N  |
	+-----+

Second, ensure that the following IO will not be completed first,
an easy way is to create a mirror of drain io and insert it into
defer_list, in this way, as long as drain io is not processed, the
following io in the defer_list will not be actively process.

	+-----+
    (0) |  N  |
	+-----+
           |          (2)         (3)         (4)
	+-----+     +-----+     +-----+     +-----+
    (1) | L+D | --> |  L  | --> | L+D | --> |  N  |
	+-----+     +-----+     +-----+     +-----+
           |
	+-----+
   ('3) |  D  |   <== This is a shadow of (3)
	+-----+
           |
	+-----+
    (5) |  N  |
	+-----+
           |
	+-----+
    (6) |  N  |
	+-----+
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Sqo_thread will get sqring in batches, which will cause
ctx->cached_sq_head to be added in batches. if one of these
sqes is set with the DRAIN flag, then he will never get a
chance to process, and finally sqo_thread will not exit.

Fixes: de0617e4 ("io_uring: add support for marking commands as draining")
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

After commit 75b28aff we can get by with just a single mmap to
map both the sq and cq ring. However, userspace doesn't know that.

Add a features variable to io_uring_params, and notify userspace
that the kernel has this ability. This can then be used in liburing
(or in applications directly) to avoid the second mmap.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Both the sq and the cq rings have sizes just over a power of two, and
the sq ring is significantly smaller. By bundling them in a single
alllocation, we get the sq ring for free.

This also means that IORING_OFF_SQ_RING and IORING_OFF_CQ_RING now mean
the same thing. If we indicate this to userspace, we can save a mmap
call.
Signed-off-by: NHristo Venev <hristo@venev.name>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

For pages that were retained via get_user_pages*(), release those pages
via the new put_user_page*() routines, instead of via put_page() or
release_pages().

This is part a tree-wide conversion, as described in commit fc1d8e7c
("mm: introduce put_user_page*(), placeholder versions").

Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: linux-fsdevel@vger.kernel.org
Cc: linux-block@vger.kernel.org
Signed-off-by: NJohn Hubbard <jhubbard@nvidia.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The outer poll loop checks for whether we need to reschedule, and
returns to userspace if we do. However, it's possible to get stuck
in the inner loop as well, if the CPU we are running on needs to
reschedule to finish the IO work.

Add the need_resched() check in the inner loop as well. This fixes
a potential hang if the kernel is configured with
CONFIG_PREEMPT_VOLUNTARY=y.
Reported-by: NSagi Grimberg <sagi@grimberg.me>
Reviewed-by: NSagi Grimberg <sagi@grimberg.me>
Tested-by: NSagi Grimberg <sagi@grimberg.me>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We need to check if we have CQEs pending before starting a poll loop,
as those could be the events we will be spinning for (and hence we'll
find none). This can happen if a CQE triggers an error, or if it is
found by eg an IRQ before we get a chance to find it through polling.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

If a request issue ends up being punted to async context to avoid
blocking, we can get into a situation where the original application
enters the poll loop for that very request before it has been issued.
This should not be an issue, except that the polling will hold the
io_uring uring_ctx mutex for the duration of the poll. When the async
worker has actually issued the request, it needs to acquire this mutex
to add the request to the poll issued list. Since the application
polling is already holding this mutex, the workqueue sleeps on the
mutex forever, and the application thus never gets a chance to poll for
the very request it was interested in.

Fix this by ensuring that the polling drops the uring_ctx occasionally
if it's not making any progress.
Reported-by: NJeffrey M. Birnbaum <jmbnyc@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This patch may fix two issues:

First, when IOSQE_IO_DRAIN set, the next IOs need to be inserted into
defer list to delay execution, but link io will be actively scheduled to
run by calling io_queue_sqe.

Second, when multiple LINK_IOs are inserted together with defer_list,
the LINK_IO is no longer keep order.

   |-------------|
   |   LINK_IO   |      ----> insert to defer_list  -----------
   |-------------|                                            |
   |   LINK_IO   |      ----> insert to defer_list  ----------|
   |-------------|                                            |
   |   LINK_IO   |      ----> insert to defer_list  ----------|
   |-------------|                                            |
   |   NORMAL_IO |      ----> insert to defer_list  ----------|
   |-------------|                                            |
                                                              |
                              queue_work at same time   <-----|

Fixes: 9e645e11 ("io_uring: add support for sqe links")
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Commit bd11b3a3 ("io_uring: don't use iov_iter_advance() for fixed
buffers") introduced an optimization to avoid using the slow
iov_iter_advance by manually populating the iov_iter iterator in some
cases.

However, the computation of the iterator count field was erroneous: The
first bvec was always accounted for an extent of page size even if the
bvec length was smaller.

In consequence, some I/O operations on fixed buffers were unable to
operate on the full extent of the buffer, consistently skipping some
bytes at the end of it.

Fixes: bd11b3a3 ("io_uring: don't use iov_iter_advance() for fixed buffers")
Cc: stable@vger.kernel.org
Signed-off-by: NAleix Roca Nonell <aleix.rocanonell@bsc.es>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

[root@localhost ~]# ./liburing/test/link

QEMU Standard PC report that:

[   29.379892] CPU: 0 PID: 84 Comm: kworker/u2:2 Not tainted 5.3.0-rc2-00051-g4010b622-dirty #86
[   29.379902] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-1 04/01/2014
[   29.379913] Workqueue: io_ring-wq io_sq_wq_submit_work
[   29.379929] Call Trace:
[   29.379953]  dump_stack+0xa9/0x10e
[   29.379970]  ? io_sq_wq_submit_work+0xbf4/0xe90
[   29.379986]  print_address_description.cold.6+0x9/0x317
[   29.379999]  ? io_sq_wq_submit_work+0xbf4/0xe90
[   29.380010]  ? io_sq_wq_submit_work+0xbf4/0xe90
[   29.380026]  __kasan_report.cold.7+0x1a/0x34
[   29.380044]  ? io_sq_wq_submit_work+0xbf4/0xe90
[   29.380061]  kasan_report+0xe/0x12
[   29.380076]  io_sq_wq_submit_work+0xbf4/0xe90
[   29.380104]  ? io_sq_thread+0xaf0/0xaf0
[   29.380152]  process_one_work+0xb59/0x19e0
[   29.380184]  ? pwq_dec_nr_in_flight+0x2c0/0x2c0
[   29.380221]  worker_thread+0x8c/0xf40
[   29.380248]  ? __kthread_parkme+0xab/0x110
[   29.380265]  ? process_one_work+0x19e0/0x19e0
[   29.380278]  kthread+0x30b/0x3d0
[   29.380292]  ? kthread_create_on_node+0xe0/0xe0
[   29.380311]  ret_from_fork+0x3a/0x50

[   29.380635] Allocated by task 209:
[   29.381255]  save_stack+0x19/0x80
[   29.381268]  __kasan_kmalloc.constprop.6+0xc1/0xd0
[   29.381279]  kmem_cache_alloc+0xc0/0x240
[   29.381289]  io_submit_sqe+0x11bc/0x1c70
[   29.381300]  io_ring_submit+0x174/0x3c0
[   29.381311]  __x64_sys_io_uring_enter+0x601/0x780
[   29.381322]  do_syscall_64+0x9f/0x4d0
[   29.381336]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

[   29.381633] Freed by task 84:
[   29.382186]  save_stack+0x19/0x80
[   29.382198]  __kasan_slab_free+0x11d/0x160
[   29.382210]  kmem_cache_free+0x8c/0x2f0
[   29.382220]  io_put_req+0x22/0x30
[   29.382230]  io_sq_wq_submit_work+0x28b/0xe90
[   29.382241]  process_one_work+0xb59/0x19e0
[   29.382251]  worker_thread+0x8c/0xf40
[   29.382262]  kthread+0x30b/0x3d0
[   29.382272]  ret_from_fork+0x3a/0x50

[   29.382569] The buggy address belongs to the object at ffff888067172140
                which belongs to the cache io_kiocb of size 224
[   29.384692] The buggy address is located 120 bytes inside of
                224-byte region [ffff888067172140, ffff888067172220)
[   29.386723] The buggy address belongs to the page:
[   29.387575] page:ffffea00019c5c80 refcount:1 mapcount:0 mapping:ffff88806ace5180 index:0x0
[   29.387587] flags: 0x100000000000200(slab)
[   29.387603] raw: 0100000000000200 dead000000000100 dead000000000122 ffff88806ace5180
[   29.387617] raw: 0000000000000000 00000000800c000c 00000001ffffffff 0000000000000000
[   29.387624] page dumped because: kasan: bad access detected

[   29.387920] Memory state around the buggy address:
[   29.388771]  ffff888067172080: fb fb fb fb fb fb fb fb fb fb fb fb fc fc fc fc
[   29.390062]  ffff888067172100: fc fc fc fc fc fc fc fc fb fb fb fb fb fb fb fb
[   29.391325] >ffff888067172180: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[   29.392578]                                         ^
[   29.393480]  ffff888067172200: fb fb fb fb fc fc fc fc fc fc fc fc fc fc fc fc
[   29.394744]  ffff888067172280: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[   29.396003] ==================================================================
[   29.397260] Disabling lock debugging due to kernel taint

io_sq_wq_submit_work free and read req again.

Cc: Zhengyuan Liu <liuzhengyuan@kylinos.cn>
Cc: linux-block@vger.kernel.org
Cc: stable@vger.kernel.org
Fixes: f7b76ac9 ("io_uring: fix counter inc/dec mismatch in async_list")
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Daniel reports that when testing an http server that uses io_uring
to poll for incoming connections, sometimes it hard crashes. This is
due to an uninitialized list member for the io_uring request. Normally
this doesn't trigger and none of the test cases caught it.
Reported-by: NDaniel Kozak <kozzi11@gmail.com>
Tested-by: NDaniel Kozak <kozzi11@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We are using PAGE_SIZE as the unit to determine if the total len in
async_list has exceeded max_pages, it's not fair for smaller io sizes.
For example, if we are doing 1k-size io streams, we will never exceed
max_pages since len >>= PAGE_SHIFT always gets zero. So use original
bytes to make it more accurate.
Signed-off-by: NZhengyuan Liu <liuzhengyuan@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Hrvoje reports that when a large fixed buffer is registered and IO is
being done to the latter pages of said buffer, the IO submission time
is much worse:

reading to the start of the buffer: 11238 ns
reading to the end of the buffer:   1039879 ns

In fact, it's worse by two orders of magnitude. The reason for that is
how io_uring figures out how to setup the iov_iter. We point the iter
at the first bvec, and then use iov_iter_advance() to fast-forward to
the offset within that buffer we need.

However, that is abysmally slow, as it entails iterating the bvecs
that we setup as part of buffer registration. There's really no need
to use this generic helper, as we know it's a BVEC type iterator, and
we also know that each bvec is PAGE_SIZE in size, apart from possibly
the first and last. Hence we can just use a shift on the offset to
find the right index, and then adjust the iov_iter appropriately.
After this fix, the timings are:

reading to the start of the buffer: 10135 ns
reading to the end of the buffer:   1377 ns

Or about an 755x improvement for the tail page.
Reported-by: NHrvoje Zeba <zeba.hrvoje@gmail.com>
Tested-by: NHrvoje Zeba <zeba.hrvoje@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

There is a hang issue while using fio to do some basic test. The issue
can be easily reproduced using the below script:

        while true
        do
                fio  --ioengine=io_uring  -rw=write -bs=4k -numjobs=1 \
                     -size=1G -iodepth=64 -name=uring   --filename=/dev/zero
        done

After several minutes (or more), fio would block at
io_uring_enter->io_cqring_wait in order to waiting for previously
committed sqes to be completed and can't return to user anymore until
we send a SIGTERM to fio. After receiving SIGTERM, fio hangs at
io_ring_ctx_wait_and_kill with a backtrace like this:

        [54133.243816] Call Trace:
        [54133.243842]  __schedule+0x3a0/0x790
        [54133.243868]  schedule+0x38/0xa0
        [54133.243880]  schedule_timeout+0x218/0x3b0
        [54133.243891]  ? sched_clock+0x9/0x10
        [54133.243903]  ? wait_for_completion+0xa3/0x130
        [54133.243916]  ? _raw_spin_unlock_irq+0x2c/0x40
        [54133.243930]  ? trace_hardirqs_on+0x3f/0xe0
        [54133.243951]  wait_for_completion+0xab/0x130
        [54133.243962]  ? wake_up_q+0x70/0x70
        [54133.243984]  io_ring_ctx_wait_and_kill+0xa0/0x1d0
        [54133.243998]  io_uring_release+0x20/0x30
        [54133.244008]  __fput+0xcf/0x270
        [54133.244029]  ____fput+0xe/0x10
        [54133.244040]  task_work_run+0x7f/0xa0
        [54133.244056]  do_exit+0x305/0xc40
        [54133.244067]  ? get_signal+0x13b/0xbd0
        [54133.244088]  do_group_exit+0x50/0xd0
        [54133.244103]  get_signal+0x18d/0xbd0
        [54133.244112]  ? _raw_spin_unlock_irqrestore+0x36/0x60
        [54133.244142]  do_signal+0x34/0x720
        [54133.244171]  ? exit_to_usermode_loop+0x7e/0x130
        [54133.244190]  exit_to_usermode_loop+0xc0/0x130
        [54133.244209]  do_syscall_64+0x16b/0x1d0
        [54133.244221]  entry_SYSCALL_64_after_hwframe+0x49/0xbe

The reason is that we had added a req to ctx->pending_async at the very
end, but it didn't get a chance to be processed. How could this happen?

        fio#cpu0                                        wq#cpu1

        io_add_to_prev_work                    io_sq_wq_submit_work

          atomic_read() <<< 1

                                                  atomic_dec_return() << 1->0
                                                  list_empty();    <<< true;

          list_add_tail()
          atomic_read() << 0 or 1?

As atomic_ops.rst states, atomic_read does not guarantee that the
runtime modification by any other thread is visible yet, so we must take
care of that with a proper implicit or explicit memory barrier.

This issue was detected with the help of Jackie's <liuyun01@kylinos.cn>

Fixes: 31b51510 ("io_uring: allow workqueue item to handle multiple buffered requests")
Signed-off-by: NZhengyuan Liu <liuzhengyuan@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

task->saved_sigmask and ->restore_sigmask are only used in the ret-from-
syscall paths.  This means that set_user_sigmask() can save ->blocked in
->saved_sigmask and do set_restore_sigmask() to indicate that ->blocked
was modified.

This way the callers do not need 2 sigset_t's passed to set/restore and
restore_user_sigmask() renamed to restore_saved_sigmask_unless() turns
into the trivial helper which just calls restore_saved_sigmask().

Link: http://lkml.kernel.org/r/20190606113206.GA9464@redhat.comSigned-off-by: NOleg Nesterov <oleg@redhat.com>
Cc: Deepa Dinamani <deepa.kernel@gmail.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Eric Wong <e@80x24.org>
Cc: Jason Baron <jbaron@akamai.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Eric W. Biederman <ebiederm@xmission.com>
Cc: David Laight <David.Laight@aculab.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

We could queue a work for each req in defer and link list without
increasing async_list->cnt, so we shouldn't decrease it while exiting
from workqueue as well if we didn't process the req in async list.

Thanks to Jens Axboe <axboe@kernel.dk> for his guidance.

Fixes: 31b51510 ("io_uring: allow workqueue item to handle multiple buffered requests")
Signed-off-by: NZhengyuan Liu <liuzhengyuan@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

sq->cached_sq_head and cq->cached_cq_tail are both unsigned int. If
cached_sq_head overflows before cached_cq_tail, then we may miss a
barrier req. As cached_cq_tail always follows cached_sq_head, the NQ
should be enough.

Cc: stable@vger.kernel.org
Fixes: de0617e4 ("io_uring: add support for marking commands as draining")
Signed-off-by: NZhengyuan Liu <liuzhengyuan@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

INFO: task syz-executor.5:8634 blocked for more than 143 seconds.
       Not tainted 5.2.0-rc5+ #3
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
syz-executor.5  D25632  8634   8224 0x00004004
Call Trace:
  context_switch kernel/sched/core.c:2818 [inline]
  __schedule+0x658/0x9e0 kernel/sched/core.c:3445
  schedule+0x131/0x1d0 kernel/sched/core.c:3509
  schedule_timeout+0x9a/0x2b0 kernel/time/timer.c:1783
  do_wait_for_common+0x35e/0x5a0 kernel/sched/completion.c:83
  __wait_for_common kernel/sched/completion.c:104 [inline]
  wait_for_common kernel/sched/completion.c:115 [inline]
  wait_for_completion+0x47/0x60 kernel/sched/completion.c:136
  kthread_stop+0xb4/0x150 kernel/kthread.c:559
  io_sq_thread_stop fs/io_uring.c:2252 [inline]
  io_finish_async fs/io_uring.c:2259 [inline]
  io_ring_ctx_free fs/io_uring.c:2770 [inline]
  io_ring_ctx_wait_and_kill+0x268/0x880 fs/io_uring.c:2834
  io_uring_release+0x5d/0x70 fs/io_uring.c:2842
  __fput+0x2e4/0x740 fs/file_table.c:280
  ____fput+0x15/0x20 fs/file_table.c:313
  task_work_run+0x17e/0x1b0 kernel/task_work.c:113
  tracehook_notify_resume include/linux/tracehook.h:185 [inline]
  exit_to_usermode_loop arch/x86/entry/common.c:168 [inline]
  prepare_exit_to_usermode+0x402/0x4f0 arch/x86/entry/common.c:199
  syscall_return_slowpath+0x110/0x440 arch/x86/entry/common.c:279
  do_syscall_64+0x126/0x140 arch/x86/entry/common.c:304
  entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x412fb1
Code: 80 3b 7c 0f 84 c7 02 00 00 c7 85 d0 00 00 00 00 00 00 00 48 8b 05 cf
a6 24 00 49 8b 14 24 41 b9 cb 2a 44 00 48 89 ee 48 89 df <48> 85 c0 4c 0f
45 c8 45 31 c0 31 c9 e8 0e 5b 00 00 85 c0 41 89 c7
RSP: 002b:00007ffe7ee6a180 EFLAGS: 00000293 ORIG_RAX: 0000000000000003
RAX: 0000000000000000 RBX: 0000000000000004 RCX: 0000000000412fb1
RDX: 0000001b2d920000 RSI: 0000000000000000 RDI: 0000000000000003
RBP: 0000000000000001 R08: 00000000f3a3e1f8 R09: 00000000f3a3e1fc
R10: 00007ffe7ee6a260 R11: 0000000000000293 R12: 000000000075c9a0
R13: 000000000075c9a0 R14: 0000000000024c00 R15: 000000000075bf2c

=============================================

There is an wrong logic, when kthread_park running
in front of io_sq_thread.

CPU#0					CPU#1

io_sq_thread_stop:			int kthread(void *_create):

kthread_park()
					__kthread_parkme(self);	 <<< Wrong
kthread_stop()
    << wait for self->exited
    << clear_bit KTHREAD_SHOULD_PARK

					ret = threadfn(data);
					   |
					   |- io_sq_thread
					       |- kthread_should_park()	<< false
					       |- schedule() <<< nobody wake up

stuck CPU#0				stuck CPU#1

So, use a new variable sqo_thread_started to ensure that io_sq_thread
run first, then io_sq_thread_stop.

Reported-by: syzbot+94324416c485d422fe15@syzkaller.appspotmail.com
Suggested-by: NJens Axboe <axboe@kernel.dk>
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This is done through IORING_OP_RECVMSG. This opcode uses the same
sqe->msg_flags that IORING_OP_SENDMSG added, and we pass in the
msghdr struct in the sqe->addr field as well.

We use MSG_DONTWAIT to force an inline fast path if recvmsg() doesn't
block, and punt to async execution if it would have.
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This is done through IORING_OP_SENDMSG. There's a new sqe->msg_flags
for the flags argument, and the msghdr struct is passed in the
sqe->addr field.

We use MSG_DONTWAIT to force an inline fast path if sendmsg() doesn't
block, and punt to async execution if it would have.
Acked-by: NDavid S. Miller <davem@davemloft.net>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

If we pass pages through an iov_iter we always already have a reference
in the caller.  Thus remove the ITER_BVEC_FLAG_NO_REF and don't take
reference to pages by default for bvec backed iov_iters.
Reviewed-by: NMinwoo Im <minwoo.im.dev@gmail.com>
Reviewed-by: NJohannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: NChristoph Hellwig <hch@lst.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This is the minimal fix for stable, I'll send cleanups later.

Commit 854a6ed5 ("signal: Add restore_user_sigmask()") introduced
the visible change which breaks user-space: a signal temporary unblocked
by set_user_sigmask() can be delivered even if the caller returns
success or timeout.

Change restore_user_sigmask() to accept the additional "interrupted"
argument which should be used instead of signal_pending() check, and
update the callers.

Eric said:

: For clarity.  I don't think this is required by posix, or fundamentally to
: remove the races in select.  It is what linux has always done and we have
: applications who care so I agree this fix is needed.
:
: Further in any case where the semantic change that this patch rolls back
: (aka where allowing a signal to be delivered and the select like call to
: complete) would be advantage we can do as well if not better by using
: signalfd.
:
: Michael is there any chance we can get this guarantee of the linux
: implementation of pselect and friends clearly documented.  The guarantee
: that if the system call completes successfully we are guaranteed that no
: signal that is unblocked by using sigmask will be delivered?

Link: http://lkml.kernel.org/r/20190604134117.GA29963@redhat.com
Fixes: 854a6ed5 ("signal: Add restore_user_sigmask()")
Signed-off-by: NOleg Nesterov <oleg@redhat.com>
Reported-by: NEric Wong <e@80x24.org>
Tested-by: NEric Wong <e@80x24.org>
Acked-by: N"Eric W. Biederman" <ebiederm@xmission.com>
Acked-by: NArnd Bergmann <arnd@arndb.de>
Acked-by: NDeepa Dinamani <deepa.kernel@gmail.com>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Jason Baron <jbaron@akamai.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: David Laight <David.Laight@ACULAB.COM>
Cc: <stable@vger.kernel.org>	[5.0+]
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

With SQE links, we can create chains of dependent SQEs. One example
would be queueing an SQE that's a read from one file descriptor, with
the linked SQE being a write to another with the same set of buffers.

An SQE link will not stall the pipeline, it'll just ensure that
dependent SQEs aren't issued before the previous link has completed.

Any error at submission or completion time will break the chain of SQEs.
For completions, this also includes short reads or writes, as the next
SQE could depend on the previous one being fully completed.

Any SQE in a chain that gets canceled due to any of the above errors,
will get an CQE fill with -ECANCELED as the error value.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Stephen reports:

I hit the following General Protection Fault when testing io_uring via
the io_uring engine in fio. This was on a VM running 5.2-rc5 and the
latest version of fio. The issue occurs for both null_blk and fake NVMe
drives. I have not tested bare metal or real NVMe SSDs. The fio script
used is given below.

[io_uring]
time_based=1
runtime=60
filename=/dev/nvme2n1 (note /dev/nullb0 also fails)
ioengine=io_uring
bs=4k
rw=readwrite
direct=1
fixedbufs=1
sqthread_poll=1
sqthread_poll_cpu=0

general protection fault: 0000 [#1] SMP PTI
CPU: 0 PID: 872 Comm: io_uring-sq Not tainted 5.2.0-rc5-cpacket-io-uring #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014
RIP: 0010:fput_many+0x7/0x90
Code: 01 48 85 ff 74 17 55 48 89 e5 53 48 8b 1f e8 a0 f9 ff ff 48 85 db 48 89 df 75 f0 5b 5d f3 c3 0f 1f 40 00 0f 1f 44 00 00 89 f6 <f0> 48 29 77 38 74 01 c3 55 48 89 e5 53 48 89 fb 65 48 \

RSP: 0018:ffffadeb817ebc50 EFLAGS: 00010246
RAX: 0000000000000004 RBX: ffff8f46ad477480 RCX: 0000000000001805
RDX: 0000000000000000 RSI: 0000000000000001 RDI: f18b51b9a39552b5
RBP: ffffadeb817ebc58 R08: ffff8f46b7a318c0 R09: 000000000000015d
R10: ffffadeb817ebce8 R11: 0000000000000020 R12: ffff8f46ad4cd000
R13: 00000000fffffff7 R14: ffffadeb817ebe30 R15: 0000000000000004
FS:  0000000000000000(0000) GS:ffff8f46b7a00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 000055828f0bbbf0 CR3: 0000000232176004 CR4: 00000000003606f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 ? fput+0x13/0x20
 io_free_req+0x20/0x40
 io_put_req+0x1b/0x20
 io_submit_sqe+0x40a/0x680
 ? __switch_to_asm+0x34/0x70
 ? __switch_to_asm+0x40/0x70
 io_submit_sqes+0xb9/0x160
 ? io_submit_sqes+0xb9/0x160
 ? __switch_to_asm+0x40/0x70
 ? __switch_to_asm+0x34/0x70
 ? __schedule+0x3f2/0x6a0
 ? __switch_to_asm+0x34/0x70
 io_sq_thread+0x1af/0x470
 ? __switch_to_asm+0x34/0x70
 ? wait_woken+0x80/0x80
 ? __switch_to+0x85/0x410
 ? __switch_to_asm+0x40/0x70
 ? __switch_to_asm+0x34/0x70
 ? __schedule+0x3f2/0x6a0
 kthread+0x105/0x140
 ? io_submit_sqes+0x160/0x160
 ? kthread+0x105/0x140
 ? io_submit_sqes+0x160/0x160
 ? kthread_destroy_worker+0x50/0x50
 ret_from_fork+0x35/0x40

which occurs because using a kernel side submission thread isn't valid
without using fixed files (registered through io_uring_register()). This
causes io_uring to put the request after logging an error, but before
the file field is set in the request. If it happens to be non-zero, we
attempt to fput() garbage.

Fix this by ensuring that req->file is initialized when the request is
allocated.

Cc: stable@vger.kernel.org # 5.1+
Reported-by: NStephen Bates <sbates@raithlin.com>
Tested-by: NStephen Bates <sbates@raithlin.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Opening and closing an io_uring instance leaks a UNIX domain socket
inode.  This is because the ->file of the io_uring instance's internal
UNIX domain socket is set to point to the io_uring file, but then
sock_release() sees the non-NULL ->file and assumes the inode reference
is held by the file so doesn't call iput().  That's not the case here,
since the reference is still meant to be held by the socket; the actual
inode of the io_uring file is different.

Fix this leak by NULL-ing out ->file before releasing the socket.

Reported-by: syzbot+111cb28d9f583693aefa@syzkaller.appspotmail.com
Fixes: 2b188cc1 ("Add io_uring IO interface")
Cc: <stable@vger.kernel.org> # v5.1+
Signed-off-by: NEric Biggers <ebiggers@google.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We can encounter a short read when we're doing buffered reads and the
data is partially cached. Right now we just return the short read, but
that forces the application to read that CQE, then issue another SQE
to finish the read. That read will not be cached, and hence will result
in an async punt.

It's more efficient to do that async punt from within the kernel, as
that will the not need two round trips more to the kernel.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Currently these functions return < 0 on error, and 0 for success.
Change that so that we return < 0 on error, but number of bytes
for success.

Some callers already treat the return value that way, others need a
slight tweak.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

If io_copy_iov() fails, it will break the loop and report success,
albeit partially completed operation.
Signed-off-by: NPavel Begunkov <asml.silence@gmail.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The previous patch has ensured that io_cqring_events contain
smp_rmb memory barriers, Now we can use wait_event_interruptible
to keep the code simple.
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Whenever smp_rmb is required to use io_cqring_events,
keep smp_rmb inside the function io_cqring_events.
Signed-off-by: NJackie Liu <liuyun01@kylinos.cn>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

This fixes couple of races which lead to infinite wait of park completion
with the following backtraces:

  [20801.303319] Call Trace:
  [20801.303321]  ? __schedule+0x284/0x650
  [20801.303323]  schedule+0x33/0xc0
  [20801.303324]  schedule_timeout+0x1bc/0x210
  [20801.303326]  ? schedule+0x3d/0xc0
  [20801.303327]  ? schedule_timeout+0x1bc/0x210
  [20801.303329]  ? preempt_count_add+0x79/0xb0
  [20801.303330]  wait_for_completion+0xa5/0x120
  [20801.303331]  ? wake_up_q+0x70/0x70
  [20801.303333]  kthread_park+0x48/0x80
  [20801.303335]  io_finish_async+0x2c/0x70
  [20801.303336]  io_ring_ctx_wait_and_kill+0x95/0x180
  [20801.303338]  io_uring_release+0x1c/0x20
  [20801.303339]  __fput+0xad/0x210
  [20801.303341]  task_work_run+0x8f/0xb0
  [20801.303342]  exit_to_usermode_loop+0xa0/0xb0
  [20801.303343]  do_syscall_64+0xe0/0x100
  [20801.303349]  entry_SYSCALL_64_after_hwframe+0x44/0xa9

  [20801.303380] Call Trace:
  [20801.303383]  ? __schedule+0x284/0x650
  [20801.303384]  schedule+0x33/0xc0
  [20801.303386]  io_sq_thread+0x38a/0x410
  [20801.303388]  ? __switch_to_asm+0x40/0x70
  [20801.303390]  ? wait_woken+0x80/0x80
  [20801.303392]  ? _raw_spin_lock_irqsave+0x17/0x40
  [20801.303394]  ? io_submit_sqes+0x120/0x120
  [20801.303395]  kthread+0x112/0x130
  [20801.303396]  ? kthread_create_on_node+0x60/0x60
  [20801.303398]  ret_from_fork+0x35/0x40

 o kthread_park() waits for park completion, so io_sq_thread() loop
   should check kthread_should_park() along with khread_should_stop(),
   otherwise if kthread_park() is called before prepare_to_wait()
   the following schedule() never returns:

   CPU#0                    CPU#1

   io_sq_thread_stop():     io_sq_thread():

                               while(!kthread_should_stop() && !ctx->sqo_stop) {

      ctx->sqo_stop = 1;
      kthread_park()

	                            prepare_to_wait();
                                    if (kthread_should_stop() {
				    }
                                    schedule();   <<< nobody checks park flag,
				                  <<< so schedule and never return

 o if the flag ctx->sqo_stop is observed by the io_sq_thread() loop
   it is quite possible, that kthread_should_park() check and the
   following kthread_parkme() is never called, because kthread_park()
   has not been yet called, but few moments later is is called and
   waits there for park completion, which never happens, because
   kthread has already exited:

   CPU#0                    CPU#1

   io_sq_thread_stop():     io_sq_thread():

      ctx->sqo_stop = 1;
                               while(!kthread_should_stop() && !ctx->sqo_stop) {
                                   <<< observe sqo_stop and exit the loop
			       }

			       if (kthread_should_park())
			           kthread_parkme();  <<< never called, since was
					              <<< never parked

      kthread_park()           <<< waits forever for park completion

In the current patch we quit the loop by only kthread_should_park()
check (kthread_park() is synchronous, so kthread_should_stop() is
never observed), and we abandon ->sqo_stop flag, since it is racy.
At the end of the io_sq_thread() we unconditionally call parmke(),
since we've exited the loop by the park flag.
Signed-off-by: NRoman Penyaev <rpenyaev@suse.de>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: linux-block@vger.kernel.org
Signed-off-by: NJens Axboe <axboe@kernel.dk>

We always pass in 0 for the cqe flags argument, since the support for
"this read hit page cache" hint was dropped.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

The test case we have is rightfully failing with the current kernel:

io_uring_setup(1, 0x7ffe2cafebe0), flags: IORING_SETUP_SQPOLL|IORING_SETUP_SQ_AFF, resv: 0x00000000 0x00000000 0x00000000 0x00000000 0x00000000, sq_thread_cpu: 4
expected -1, got 3

This is in a vm, and CPU3 is the last valid one, hence asking for 4
should fail the setup with -EINVAL, not succeed. The problem is that
we're using array_index_nospec() with nr_cpu_ids as the index, hence we
wrap and end up using CPU0 instead of CPU4. This makes the setup
succeed where it should be failing.

We don't need to use array_index_nospec() as we're not indexing any
array with this. Instead just compare with nr_cpu_ids directly. This
is fine as we're checking with cpu_online() afterwards.
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Pach series "Add FOLL_LONGTERM to GUP fast and use it".

HFI1, qib, and mthca, use get_user_pages_fast() due to its performance
advantages.  These pages can be held for a significant time.  But
get_user_pages_fast() does not protect against mapping FS DAX pages.

Introduce FOLL_LONGTERM and use this flag in get_user_pages_fast() which
retains the performance while also adding the FS DAX checks.  XDP has also
shown interest in using this functionality.[1]

In addition we change get_user_pages() to use the new FOLL_LONGTERM flag
and remove the specialized get_user_pages_longterm call.

[1] https://lkml.org/lkml/2019/3/19/939

"longterm" is a relative thing and at this point is probably a misnomer.
This is really flagging a pin which is going to be given to hardware and
can't move.  I've thought of a couple of alternative names but I think we
have to settle on if we are going to use FL_LAYOUT or something else to
solve the "longterm" problem.  Then I think we can change the flag to a
better name.

Secondly, it depends on how often you are registering memory.  I have
spoken with some RDMA users who consider MR in the performance path...
For the overall application performance.  I don't have the numbers as the
tests for HFI1 were done a long time ago.  But there was a significant
advantage.  Some of which is probably due to the fact that you don't have
to hold mmap_sem.

Finally, architecturally I think it would be good for everyone to use
*_fast.  There are patches submitted to the RDMA list which would allow
the use of *_fast (they reworking the use of mmap_sem) and as soon as they
are accepted I'll submit a patch to convert the RDMA core as well.  Also
to this point others are looking to use *_fast.

As an aside, Jasons pointed out in my previous submission that *_fast and
*_unlocked look very much the same.  I agree and I think further cleanup
will be coming.  But I'm focused on getting the final solution for DAX at
the moment.

This patch (of 7):

This patch starts a series which aims to support FOLL_LONGTERM in
get_user_pages_fast().  Some callers who would like to do a longterm (user
controlled pin) of pages with the fast variant of GUP for performance
purposes.

Rather than have a separate get_user_pages_longterm() call, introduce
FOLL_LONGTERM and change the longterm callers to use it.

This patch does not change any functionality.  In the short term
"longterm" or user controlled pins are unsafe for Filesystems and FS DAX
in particular has been blocked.  However, callers of get_user_pages_fast()
were not "protected".

FOLL_LONGTERM can _only_ be supported with get_user_pages[_fast]() as it
requires vmas to determine if DAX is in use.

NOTE: In merging with the CMA changes we opt to change the
get_user_pages() call in check_and_migrate_cma_pages() to a call of
__get_user_pages_locked() on the newly migrated pages.  This makes the
code read better in that we are calling __get_user_pages_locked() on the
pages before and after a potential migration.

As a side affect some of the interfaces are cleaned up but this is not the
primary purpose of the series.

In review[1] it was asked:

<quote>
> This I don't get - if you do lock down long term mappings performance
> of the actual get_user_pages call shouldn't matter to start with.
>
> What do I miss?

A couple of points.

First "longterm" is a relative thing and at this point is probably a
misnomer.  This is really flagging a pin which is going to be given to
hardware and can't move.  I've thought of a couple of alternative names
but I think we have to settle on if we are going to use FL_LAYOUT or
something else to solve the "longterm" problem.  Then I think we can
change the flag to a better name.

Second, It depends on how often you are registering memory.  I have spoken
with some RDMA users who consider MR in the performance path...  For the
overall application performance.  I don't have the numbers as the tests
for HFI1 were done a long time ago.  But there was a significant
advantage.  Some of which is probably due to the fact that you don't have
to hold mmap_sem.

Finally, architecturally I think it would be good for everyone to use
*_fast.  There are patches submitted to the RDMA list which would allow
the use of *_fast (they reworking the use of mmap_sem) and as soon as they
are accepted I'll submit a patch to convert the RDMA core as well.  Also
to this point others are looking to use *_fast.

As an asside, Jasons pointed out in my previous submission that *_fast and
*_unlocked look very much the same.  I agree and I think further cleanup
will be coming.  But I'm focused on getting the final solution for DAX at
the moment.

</quote>

[1] https://lore.kernel.org/lkml/20190220180255.GA12020@iweiny-DESK2.sc.intel.com/T/#md6abad2569f3bf6c1f03686c8097ab6563e94965

[ira.weiny@intel.com: v3]
  Link: http://lkml.kernel.org/r/20190328084422.29911-2-ira.weiny@intel.com
Link: http://lkml.kernel.org/r/20190328084422.29911-2-ira.weiny@intel.com
Link: http://lkml.kernel.org/r/20190317183438.2057-2-ira.weiny@intel.comSigned-off-by: NIra Weiny <ira.weiny@intel.com>
Reviewed-by: NAndrew Morton <akpm@linux-foundation.org>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.ibm.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Paul Mackerras <paulus@samba.org>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Rich Felker <dalias@libc.org>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: James Hogan <jhogan@kernel.org>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NLinus Torvalds <torvalds@linux-foundation.org>

When punting to workers the SQE gets copied after the initial try.
There is a race condition between reading SQE data for the initial try
and copying it for punting it to the workers.

For example io_rw_done calls kiocb->ki_complete even if it was prepared
for IORING_OP_FSYNC (and would be NULL).

The easiest solution for now is to alway prepare again in the worker.

req->file is safe to prepare though as long as it is checked before use.
Signed-off-by: NStefan Bühler <source@stbuehler.de>
Signed-off-by: NJens Axboe <axboe@kernel.dk>