kernel/linux.git/include/linux/io_uring/cmd.h, branch v6.12.80

io_uring: Change res2 parameter type in io_uring_cmd_done

2024-12-14T19:03:26+00:00

commit a07d2d7930c75e6bf88683b376d09ab1f3fed2aa upstream. Change the type of the res2 parameter in io_uring_cmd_done from ssize_t to u64. This aligns the parameter type with io_req_set_cqe32_extra, which expects u64 arguments. The change eliminates potential issues on 32-bit architectures where ssize_t might be 32-bit. Only user of passing res2 is drivers/nvme/host/ioctl.c and it actually passes u64. Fixes: ee692a21e9bf ("fs,io_uring: add infrastructure for uring-cmd") Cc: stable@vger.kernel.org Reviewed-by: Kanchan Joshi Tested-by: Li Zetao Reviewed-by: Li Zetao Signed-off-by: Bernd Schubert Link: https://lore.kernel.org/r/20241203-io_uring_cmd_done-res2-as-u64-v2-1-5e59ae617151@ddn.com Signed-off-by: Jens Axboe Signed-off-by: Greg Kroah-Hartman

io_uring/cmd: give inline space in request to cmds

2024-09-11T16:44:10+00:00

Some io_uring commands can use some inline space in io_kiocb. We have 32 bytes in struct io_uring_cmd, expose it. Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/7ca779a61ee5e166e535d70df9c7f07b15d8a0ce.1726072086.git.asml.silence@gmail.com Signed-off-by: Jens Axboe

io_uring/cmd: expose iowq to cmds

2024-09-11T16:44:10+00:00

When an io_uring request needs blocking context we offload it to the io_uring's thread pool called io-wq. We can get there off ->uring_cmd by returning -EAGAIN, but there is no straightforward way of doing that from an asynchronous callback. Add a helper that would transfer a command to a blocking context. Note, we do an extra hop via task_work before io_queue_iowq(), that's a limitation of io_uring infra we have that can likely be lifted later if that would ever become a problem. Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/f735f807d7c8ba50c9452c69dfe5d3e9e535037b.1726072086.git.asml.silence@gmail.com Signed-off-by: Jens Axboe

nvme/io_uring: use helper for polled completions

2024-04-15T14:10:24+00:00

NVMe is making up issue_flags, which is a no-no in general, and to make matters worse, they are completely the wrong ones. For a pure polled request, which it does check for, we're already inside the ctx->uring_lock when the completions are run off io_do_iopoll(). Hence the correct flag would be '0' rather than IO_URING_F_UNLOCKED. Reviewed-by: Pavel Begunkov Signed-off-by: Jens Axboe

io_uring/cmd: document some uring_cmd related helpers

2024-04-15T14:10:24+00:00

Add comments warning users that they're only allowed to pass issue_flags that were given from io_uring. Signed-off-by: Pavel Begunkov Reviewed-by: Ming Lei Tested-by: Ming Lei Link: https://lore.kernel.org/r/82ff8a45f2c3eb5f3a04a33f0692e5e4a1320455.1710799188.git.asml.silence@gmail.com Signed-off-by: Jens Axboe

Merge tag 'for-6.8/io_uring-2024-01-08' of git://git.kernel.dk/linux

2024-01-11T22:19:23+00:00

Pull io_uring updates from Jens Axboe: "Mostly just come fixes and cleanups, but one feature as well. In detail: - Harden the check for handling IOPOLL based on return (Pavel) - Various minor optimizations (Pavel) - Drop remnants of SCM_RIGHTS fd passing support, now that it's no longer supported since 6.7 (me) - Fix for a case where bytes_done wasn't initialized properly on a failure condition for read/write requests (me) - Move the register related code to a separate file (me) - Add support for returning the provided ring buffer head (me) - Add support for adding a direct descriptor to the normal file table (me, Christian Brauner) - Fix for ensuring pending task_work for a ring with DEFER_TASKRUN is run even if we timeout waiting (me)" * tag 'for-6.8/io_uring-2024-01-08' of git://git.kernel.dk/linux: io_uring: ensure local task_work is run on wait timeout io_uring/kbuf: add method for returning provided buffer ring head io_uring/rw: ensure io->bytes_done is always initialized io_uring: drop any code related to SCM_RIGHTS io_uring/unix: drop usage of io_uring socket io_uring/register: move io_uring_register(2) related code to register.c io_uring/openclose: add support for IORING_OP_FIXED_FD_INSTALL io_uring/cmd: inline io_uring_cmd_get_task io_uring/cmd: inline io_uring_cmd_do_in_task_lazy io_uring: split out cmd api into a separate header io_uring: optimise ltimeout for inline execution io_uring: don't check iopoll if request completes

io_uring/cmd: inline io_uring_cmd_get_task

2023-12-12T14:42:52+00:00

With io_uring_types.h we see all required definitions to inline io_uring_cmd_get_task(). Signed-off-by: Pavel Begunkov Reviewed-by: Ming Lei Link: https://lore.kernel.org/r/aa8e317f09e651a5f3e72f8c0ad3902084c1f930.1701391955.git.asml.silence@gmail.com Signed-off-by: Jens Axboe

io_uring/cmd: inline io_uring_cmd_do_in_task_lazy

2023-12-12T14:42:52+00:00

Now as we can easily include io_uring_types.h, move IOU_F_TWQ_LAZY_WAKE and inline io_uring_cmd_do_in_task_lazy(). Signed-off-by: Pavel Begunkov Reviewed-by: Ming Lei Link: https://lore.kernel.org/r/2ec9fb31dd192d1c5cf26d0a2dec5657d88a8e48.1701391955.git.asml.silence@gmail.com Signed-off-by: Jens Axboe

io_uring: split out cmd api into a separate header

2023-12-12T14:42:52+00:00

linux/io_uring.h is slowly becoming a rubbish bin where we put anything exposed to other subsystems. For instance, the task exit hooks and io_uring cmd infra are completely orthogonal and don't need each other's definitions. Start cleaning it up by splitting out all command bits into a new header file. Signed-off-by: Pavel Begunkov Link: https://lore.kernel.org/r/7ec50bae6e21f371d3850796e716917fc141225a.1701391955.git.asml.silence@gmail.com Signed-off-by: Jens Axboe