]> Git Repo - linux.git/commit - io_uring/io_uring.h
io_uring: refactor io_fill_cqe_req_aux
authorPavel Begunkov <[email protected]>
Mon, 18 Mar 2024 22:00:31 +0000 (22:00 +0000)
committerJens Axboe <[email protected]>
Mon, 15 Apr 2024 14:10:24 +0000 (08:10 -0600)
commite5c12945be5016d681ff305ea7306fef5902219d
tree3688ce241b26c9a85466651773aa43684ce55ea8
parent8e5b3b89ecaf6d9295e561c225b35c574a5e0fe7
io_uring: refactor io_fill_cqe_req_aux

The restriction on multishot execution context disallowing io-wq is
driven by rules of io_fill_cqe_req_aux(), it should only be called in
the master task context, either from the syscall path or in task_work.
Since task_work now always takes the ctx lock implying
IO_URING_F_COMPLETE_DEFER, we can just assume that the function is
always called with its defer argument set to true.

Kill the argument. Also rename the function for more consistency as
"fill" in CQE related functions was usually meant for raw interfaces
only copying data into the CQ without any locking, waking the user
and other accounting "post" functions take care of.

Signed-off-by: Pavel Begunkov <[email protected]>
Tested-by: Ming Lei <[email protected]>
Link: https://lore.kernel.org/r/93423d106c33116c7d06bf277f651aa68b427328.1710799188.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <[email protected]>
io_uring/io_uring.c
io_uring/io_uring.h
io_uring/net.c
io_uring/poll.c
io_uring/rw.c
io_uring/timeout.c
This page took 0.045043 seconds and 4 git commands to generate.