Hi Linus,
On top of the core and xattr branches, this pull request adds support
for socket(2) for io_uring. This is handy when using direct / registered
file descriptors with io_uring.
Outside of those two patches, a small series from Dylan on top that
improves the tracing by providing a text representation of the opcode
rather than needing to decode this by reading the header file every
time. Sits in this branch as it was the last opcode added (until it
wasn't...).
Please pull!
The following changes since commit 0200ce6a57c5de802f4e438485c14cc9d63d5f4b:
io_uring: fix trace for reduced sqe padding (2022-04-24 18:18:46 -0600)
are available in the Git repository at:
git://git.kernel.dk/linux-block.git tags/for-5.19/io_uring-socket-2022-05-22
for you to fetch changes up to 033b87d24f7257c45506bd043ad85ed24a9925e2:
io_uring: use the text representation of ops in trace (2022-04-28 17:06:03 -0600)
----------------------------------------------------------------
for-5.19/io_uring-socket-2022-05-22
----------------------------------------------------------------
Dylan Yudaken (4):
io_uring: add type to op enum
io_uring: add io_uring_get_opcode
io_uring: rename op -> opcode
io_uring: use the text representation of ops in trace
Jens Axboe (2):
net: add __sys_socket_file()
io_uring: add socket(2) support
fs/io_uring.c | 177 ++++++++++++++++++++++++++++++++++++++++
include/linux/io_uring.h | 5 ++
include/linux/socket.h | 1 +
include/trace/events/io_uring.h | 36 ++++----
include/uapi/linux/io_uring.h | 3 +-
net/socket.c | 52 +++++++++---
6 files changed, 248 insertions(+), 26 deletions(-)
--
Jens Axboe
