Hi Linus,
Fixes for io_uring that should go into the 6.13-rc4 kernel release. This
pull request contains:
- Fix for a file ref leak for registered ring fds
- Turn the ->timeout_lock into a raw spinlock, as it nests under the
io-wq lock which is a raw spinlock as it's called from the scheduler
side
- Limit ring resizing to DEFER_TASKRUN for now. We will broaden this in
the future, but for now, ensure that it's only feasible on rings with
a single user
- Add sanity check for io-wq enqueuing
Please pull!
The following changes since commit 99d6af6e8a22b792e1845b186f943cd10bb4b7b0:
io_uring/rsrc: don't put/free empty buffers (2024-12-12 08:01:52 -0700)
are available in the Git repository at:
git://git.kernel.dk/linux.git tags/io_uring-6.13-20241220
for you to fetch changes up to dbd2ca9367eb19bc5e269b8c58b0b1514ada9156:
io_uring: check if iowq is killed before queuing (2024-12-19 13:31:53 -0700)
----------------------------------------------------------------
io_uring-6.13-20241220
----------------------------------------------------------------
Jann Horn (1):
io_uring: Fix registered ring file refcount leak
Jens Axboe (2):
io_uring: make ctx->timeout_lock a raw spinlock
io_uring/register: limit ring resizing to DEFER_TASKRUN
Pavel Begunkov (1):
io_uring: check if iowq is killed before queuing
include/linux/io_uring.h | 4 +---
include/linux/io_uring_types.h | 2 +-
io_uring/io_uring.c | 17 +++++++++++------
io_uring/register.c | 3 +++
io_uring/timeout.c | 40 ++++++++++++++++++++--------------------
5 files changed, 36 insertions(+), 30 deletions(-)
--
Jens Axboe
