public inbox for [email protected]
 help / color / mirror / Atom feed
* [PATCH v2] io_uring: add EPOLLEXCLUSIVE flag to aoid thundering herd type behavior
@ 2020-06-11  4:30 Jiufei Xue
  2020-06-11 15:07 ` Jens Axboe
  0 siblings, 1 reply; 2+ messages in thread
From: Jiufei Xue @ 2020-06-11  4:30 UTC (permalink / raw)
  To: io-uring; +Cc: axboe, joseph.qi

From: Jiufei Xue <[email protected]>

Applications can use this flag to avoid accept thundering herd.
And poll_events should be changed to 32 bits to cover EPOLLEXCLUSIVE.

Signed-off-by: Jiufei Xue <[email protected]>
---
 fs/io_uring.c                 | 13 +++++++++----
 include/uapi/linux/io_uring.h |  2 +-
 tools/io_uring/liburing.h     |  2 +-
 3 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 47790a2..03951ec 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -4289,7 +4289,11 @@ static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
 
 	pt->error = 0;
 	poll->head = head;
-	add_wait_queue(head, &poll->wait);
+
+	if (poll->events & EPOLLEXCLUSIVE)
+		add_wait_queue_exclusive(head, &poll->wait);
+	else
+		add_wait_queue(head, &poll->wait);
 }
 
 static void io_async_queue_proc(struct file *file, struct wait_queue_head *head,
@@ -4602,7 +4606,7 @@ static void io_poll_queue_proc(struct file *file, struct wait_queue_head *head,
 static int io_poll_add_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
 	struct io_poll_iocb *poll = &req->poll;
-	u16 events;
+	u32 events;
 
 	if (unlikely(req->ctx->flags & IORING_SETUP_IOPOLL))
 		return -EINVAL;
@@ -4612,7 +4616,8 @@ static int io_poll_add_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe
 		return -EBADF;
 
 	events = READ_ONCE(sqe->poll_events);
-	poll->events = demangle_poll(events) | EPOLLERR | EPOLLHUP;
+	poll->events = demangle_poll(events) | EPOLLERR | EPOLLHUP |
+		       (events & EPOLLEXCLUSIVE);
 
 	get_task_struct(current);
 	req->task = current;
@@ -8196,7 +8201,7 @@ static int __init io_uring_init(void)
 	BUILD_BUG_SQE_ELEM(28, /* compat */   int, rw_flags);
 	BUILD_BUG_SQE_ELEM(28, /* compat */ __u32, rw_flags);
 	BUILD_BUG_SQE_ELEM(28, __u32,  fsync_flags);
-	BUILD_BUG_SQE_ELEM(28, __u16,  poll_events);
+	BUILD_BUG_SQE_ELEM(28, __u32,  poll_events);
 	BUILD_BUG_SQE_ELEM(28, __u32,  sync_range_flags);
 	BUILD_BUG_SQE_ELEM(28, __u32,  msg_flags);
 	BUILD_BUG_SQE_ELEM(28, __u32,  timeout_flags);
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 92c2269..afc7edd 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -31,7 +31,7 @@ struct io_uring_sqe {
 	union {
 		__kernel_rwf_t	rw_flags;
 		__u32		fsync_flags;
-		__u16		poll_events;
+		__u32		poll_events;
 		__u32		sync_range_flags;
 		__u32		msg_flags;
 		__u32		timeout_flags;
diff --git a/tools/io_uring/liburing.h b/tools/io_uring/liburing.h
index 5f305c8..094b9ec 100644
--- a/tools/io_uring/liburing.h
+++ b/tools/io_uring/liburing.h
@@ -145,7 +145,7 @@ static inline void io_uring_prep_write_fixed(struct io_uring_sqe *sqe, int fd,
 }
 
 static inline void io_uring_prep_poll_add(struct io_uring_sqe *sqe, int fd,
-					  short poll_mask)
+					  unsigned poll_mask)
 {
 	memset(sqe, 0, sizeof(*sqe));
 	sqe->opcode = IORING_OP_POLL_ADD;
-- 
1.8.3.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* Re: [PATCH v2] io_uring: add EPOLLEXCLUSIVE flag to aoid thundering herd type behavior
  2020-06-11  4:30 [PATCH v2] io_uring: add EPOLLEXCLUSIVE flag to aoid thundering herd type behavior Jiufei Xue
@ 2020-06-11 15:07 ` Jens Axboe
  0 siblings, 0 replies; 2+ messages in thread
From: Jens Axboe @ 2020-06-11 15:07 UTC (permalink / raw)
  To: Jiufei Xue, io-uring; +Cc: joseph.qi

On 6/10/20 10:30 PM, Jiufei Xue wrote:
> From: Jiufei Xue <[email protected]>
> 
> Applications can use this flag to avoid accept thundering herd.
> And poll_events should be changed to 32 bits to cover EPOLLEXCLUSIVE.
> 
> Signed-off-by: Jiufei Xue <[email protected]>

Both of your patches seem to have a weird setup where the
From is alibaba.linux.com, and the SOB is linux.alibaba.com.
I'm assuming the latter is the correct one, as that's where
the email came from and the former doesn't have an MX record.

This is 5.9 material for me. With the above fixed up, please
turn this into a series where 1/2 is changing the poll type
to be 32-bits, and 2/2 is the functional change that is in
this patch.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2020-06-11 15:08 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-06-11  4:30 [PATCH v2] io_uring: add EPOLLEXCLUSIVE flag to aoid thundering herd type behavior Jiufei Xue
2020-06-11 15:07 ` Jens Axboe

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox