Hi Linus,
Just fixes for the MSG_RING opcode. Nothing really major:
- Fix an overflow missing serialization around posting CQEs to the
target ring (me)
- Disable MSG_RING on a ring that isn't enabled yet. There's nothing
really wrong with allowing it, but 1) it's somewhat odd as nobody can
receive them yet, and 2) it means that using the right delivery
mechanism might change. As nobody should be sending CQEs to a ring
that isn't enabled yet, let's just disable it (Pavel)
- Tweak to when we decide to post remotely or not for MSG_RING (Pavel)
Please pull!
The following changes since commit 544d163d659d45a206d8929370d5a2984e546cb7:
io_uring: lock overflowing for IOPOLL (2023-01-13 07:32:46 -0700)
are available in the Git repository at:
git://git.kernel.dk/linux.git tags/io_uring-6.2-2023-01-20
for you to fetch changes up to 8579538c89e33ce78be2feb41e07489c8cbf8f31:
io_uring/msg_ring: fix remote queue to disabled ring (2023-01-20 09:49:34 -0700)
----------------------------------------------------------------
io_uring-6.2-2023-01-20
----------------------------------------------------------------
Jens Axboe (2):
io_uring/msg_ring: move double lock/unlock helpers higher up
io_uring/msg_ring: fix missing lock on overflow for IOPOLL
Pavel Begunkov (2):
io_uring/msg_ring: fix flagging remote execution
io_uring/msg_ring: fix remote queue to disabled ring
io_uring/io_uring.c | 4 +-
io_uring/msg_ring.c | 130 +++++++++++++++++++++++++++++++++-------------------
2 files changed, 84 insertions(+), 50 deletions(-)
--
Jens Axboe
