[PATCH] io_uring: enable EPOLLEXCLUSIVE for accept poll

[PATCH] io_uring: enable EPOLLEXCLUSIVE for accept poll

[Dylan Yudaken]

When polling sockets for accept, use EPOLLEXCLUSIVE. This is helpful
when multiple accept SQEs are submitted.

For O_NONBLOCK sockets multiple queued SQEs would previously have all
completed at once, but most with -EAGAIN as the result. Now only one
wakes up and completes.

For sockets without O_NONBLOCK there is no user facing change, but
internally the extra requests would previously be queued onto a worker
thread as they would wake up with no connection waiting, and be
punted. Now they do not wake up unnecessarily.

Co-developed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Dylan Yudaken <dylany@fb.com>
---
 fs/io_uring.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 63299896555d..fced35d3ee85 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -968,6 +968,7 @@ struct io_op_def {
 	/* set if opcode supports polled "wait" */
 	unsigned		pollin : 1;
 	unsigned		pollout : 1;
+	unsigned		poll_exclusive : 1;
 	/* op supports buffer selection */
 	unsigned		buffer_select : 1;
 	/* do prep async if is going to be punted */
@@ -1062,6 +1063,7 @@ static const struct io_op_def io_op_defs[] = {
 		.needs_file		= 1,
 		.unbound_nonreg_file	= 1,
 		.pollin			= 1,
+		.poll_exclusive		= 1,
 	},
 	[IORING_OP_ASYNC_CANCEL] = {
 		.audit_skip		= 1,
@@ -6280,7 +6282,8 @@ static int io_arm_poll_handler(struct io_kiocb *req, unsigned issue_flags)
 	} else {
 		mask |= POLLOUT | POLLWRNORM;
 	}
-
+	if (def->poll_exclusive)
+		mask |= EPOLLEXCLUSIVE;
 	if (!(issue_flags & IO_URING_F_UNLOCKED) &&
 	    !list_empty(&ctx->apoll_cache)) {
 		apoll = list_first_entry(&ctx->apoll_cache, struct async_poll,

base-commit: 6d4809f49bec8dc3b244debcb97f78239e52f5bb
-- 
2.30.2

Re: [PATCH] io_uring: enable EPOLLEXCLUSIVE for accept poll

[Jens Axboe]

On Fri, 25 Mar 2022 02:37:55 -0700, Dylan Yudaken wrote:
> When polling sockets for accept, use EPOLLEXCLUSIVE. This is helpful
> when multiple accept SQEs are submitted.
> 
> For O_NONBLOCK sockets multiple queued SQEs would previously have all
> completed at once, but most with -EAGAIN as the result. Now only one
> wakes up and completes.
> 
> [...]

Applied, thanks!

[1/1] io_uring: enable EPOLLEXCLUSIVE for accept poll
      commit: 52dd86406dfa322c8d42b3a4328858abdc6f1d85

Best regards,
-- 
Jens Axboe