liburing
liburing copied to clipboard
axboe
1. Kernel polling has consumed one logical core. If the monitoring thread loops io_uring_peek_cqe, it will consume another logical core. 2. create extra threads equal to the num of io_uring...
This is just a behavior I have noticed and can't get past. Below 4kb ping/pong between two TCP processes, I get significantly better perf with io_uring - from 30% to...
I'm noticing that there's no authoritative "from the source" docs outlining what the best practices would be for one aiming to use liburing in a multi-threaded scenario. We should first...
I'm hoping to take advantage of io_uring to minimize syscall count, but I can't get anywhere close to non-io_uring performance. My application is a fast version of rm: every directory...
In my case, io_uring is used to replace UDP send() , all work fine without `SQPOLL`: ``` sqe = io_uring_get_sqe(ring); io_uring_prep_send(sqe, p->fd, data, size, msg_flags); ret = io_uring_submit_and_wait(ring, 1); if...
https://github.com/torvalds/linux/commit/7ef66d186eb95f987a97fb3329b65c840e2dc9bf removed `IORING_CQE_F_MSG` and I think there could be a reasonable justification for returning it. I am prototyping a thin event loop library over io_uring which abstract some implementation details...
Can io_uring_prep_read/write return -EAGAIN if timeout is not set?
Hello! This could be a dumb question but I wanted to know if this performance difference is expected. My program is written in Zig (the stdlib io_uring library is based...
Confused that man pages say any io_uring request that takes a file descriptor can use a direct descriptor but many io_uring requests explicitly disallow them. Example of a man page:...
Would it be possible to have `MSG_WAITALL` for splice, similar to the behavir of write?
