From: Linus Torvalds Date: Tue, 9 Dec 2025 00:07:28 +0000 (+0900) Subject: Merge tag 'io_uring-6.19-20251208' of git://git.kernel.org/pub/scm/linux/kernel/git... X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=cfd4039213e7b5a828c5b78e1b5235cac91af53d;p=thirdparty%2Flinux.git Merge tag 'io_uring-6.19-20251208' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux Pull io_uring updates from Jens Axboe: "Followup set of fixes for io_uring for this merge window. These are either later fixes, or cleanups that don't make sense to defer. This pull request contains: - Fix for a recent regression in io-wq worker creation - Tracing cleanup - Use READ_ONCE/WRITE_ONCE consistently for ring mapped kbufs. Mostly for documentation purposes, indicating that they are shared with userspace - Fix for POLL_ADD losing a completion, if the request is updated and now is triggerable - eg, if POLLIN is set with the updated, and the polled file is readable - In conjunction with the above fix, also unify how poll wait queue entries are deleted with the head update. We had 3 different spots doing both the list deletion and head write, with one of them nicely documented. Abstract that into a helper and use it consistently - Small series from Joanne fixing an issue with buffer cloning, and cleaning up the arg validation" * tag 'io_uring-6.19-20251208' of git://git.kernel.org/pub/scm/linux/kernel/git/axboe/linux: io_uring/poll: unify poll waitqueue entry and list removal io_uring/kbuf: use WRITE_ONCE() for userspace-shared buffer ring fields io_uring/kbuf: use READ_ONCE() for userspace-mapped memory io_uring/rsrc: fix lost entries after cloned range io_uring/rsrc: rename misleading src_node variable in io_clone_buffers() io_uring/rsrc: clean up buffer cloning arg validation io_uring/trace: rename io_uring_queue_async_work event "rw" field io_uring/io-wq: always retry worker create on ERESTART* io_uring/poll: correctly handle io_poll_add() return value on update --- cfd4039213e7b5a828c5b78e1b5235cac91af53d