/
githubmirror
/
liburing
Обзор
Документация
Войти
/
githubmirror
/
liburing
Код
Запросы
0
Пакеты
0
Релизы
0
Аналитика
Безопасность
master
man/io_uring_resize_rings.3
116 строк
3 KB
Jens Axboe
Update ring resizing for DEFER_TASKRUN
19 дек 2024, 19:45
19 дек 2024, 19:45
69e0bc2
Код
Авторство
О чём код?
.\" Copyright (C) 2022 Jens Axboe <axboe@kernel.dk> .\" .\" SPDX-License-Identifier: LGPL-2.0-or-later .\" .TH io_uring_resize_rings 3 "October 29, 2024" "liburing-2.9" "liburing Manual" .SH NAME io_uring_resize_rings \- resize the SQ and CQ rings .SH SYNOPSIS .nf .B #include <liburing.h> .PP .BI "int io_uring_resize_rings(struct io_uring *" ring ", .BI " struct io_uring_params *" p "); .PP .SH DESCRIPTION .PP The .BR io_uring_resize_rings (3) function performs resizes the SQ and/or CQ ring associated with .I ring by the parameters specified in .I p . The .I p argument must be filled in with the appropriate information for the resize operations, most notably the .IR sq_entries and .IR cq_entries fields must be filled out. The .IR flags field can also be set, see below for potential values that may be used with a resize operation. It's fairly uncommon to need to resize the SQ ring, but not that uncommon that the CQ ring would need resizing. For networked workloads, it can be hard to appropriately size the CQ ring upfront, as it's not always known what load a given ring will see. If overflow conditions are seen for the CQ ring, then resizing it larger may be a good idea. When a ring is resized, any pending SQ or CQ entries are copied along the way. It is not legal to resize a CQ ring that is in an overflow condition, and attempting to do so will fail. Currently doesn't support resizing rings setup with .B IORING_SETUP_NO_MMAP . This is purely a liburing limitation, the kernel does support it. Also note that ring resizing is currently only supported on rings setup with .B IORING_SETUP_DEFER_TASKRUN . Attempting to resize differently configured rings will result in an .BR -EINVAL error. Valid flags in .IR flags : .TP .B IORING_SETUP_CQSIZE If this isn't set, then the CQ ring size is set based on the specified SQ ring size. The default is twice as many CQ ring entries as there are SQ ring entries. If set, then .IR cq_entries will be used to size the CQ ring. .TP .B IORING_SETUP_CLAMP If set, then SQ and CQ ring entries are clamped to the maximum allowable size, if they exceed that. If not set, setting sizes too large will cause the operation to fail. .PP Other flags are inherited from the way the ring was setup, that includes flags like .B IORING_SETUP_NO_SQARRAY , .B IORING_SETUP_SQE128 , .B IORING_SETUP_CQE32 , and .B IORING_SETUP_NO_MMAP . Other fields in .IR p should be cleared to zero. Available since kernel 6.13. Also see .BR io_uring_setup (2) for a detailed description of the setup flags. .SH RETURN VALUE .PP Returns 0 on success, and < 0 on failure. Potential common failure cases: .TP .B -EEXIST Attempting to resize a ring setup with .B IORING_SETUP_SINGLE_ISSUER and the resizing task is different from the one that created/enabled the ring. .TP .B -EFAULT Copying of .I p was unsuccessful. .TP .B -EINVAL Invalid flags were specified for the operation .TP .B -EINVAL Attempt to resize a ring not setup with .BR IORING_SETUP_DEFER_TASKRUN . .TP .B -EOVERFLOW The values specified for SQ or CQ entries would cause an overflow. .SH SEE ALSO .BR io_uring_setup (2)