[PATCHSET v4 0/5] Allow non-atomic allocs for overflows

[Jens Axboe <axboe@kernel.dk>]

Hi,

This is heavily inspired by the series that Pavel posted here:

https://lore.kernel.org/io-uring/cover.1747209332.git.asml.silence@gmail.com/

Basically this patchset tries to accomplish two things:

1) Enable GFP_KERNEL alloc of the overflow entries, when possible.

2) Make the overflow side pollute the fast/common path as little as
   possible.

Also does some cleanups, like passing more appropriate and easily
readable arguments to the overflow handling, rather than need 3..5
arguments of various user_data/res/cflags/extra1/extra1 being passed
along.

 include/linux/io_uring_types.h |  2 +-
 io_uring/io_uring.c            | 99 ++++++++++++++++++++++------------
 2 files changed, 65 insertions(+), 36 deletions(-)

Since v3:
- Fix typos in commit message.
- Rename io_cqe_overflow_lockless() -> io_cqe_overflow(), and
  io_cqe_overflow() -> io_cqe_overflow_locked().
- Make io_init_cqe() a static inline function rather than a macro.

-- 
Jens Axboe