public inbox for [email protected]
 help / color / mirror / Atom feed
From: Jens Axboe <[email protected]>
To: Mark Papadakis <[email protected]>
Cc: [email protected]
Subject: Re: io_uring and spurious wake-ups from eventfd
Date: Wed, 8 Jan 2020 10:20:01 -0700	[thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>

On 1/8/20 9:50 AM, Jens Axboe wrote:
> On 1/8/20 9:46 AM, Mark Papadakis wrote:
>> Thus sounds perfect!
> 
> I'll try and cook this up, if you can test it.

Something like the below.

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 06fdda43163d..70478888bb16 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -206,6 +206,7 @@ struct io_ring_ctx {
 		bool			account_mem;
 		bool			cq_overflow_flushed;
 		bool			drain_next;
+		bool			eventfd_async;
 
 		/*
 		 * Ring buffer of indices into array of io_uring_sqe, which is
@@ -960,13 +961,20 @@ static struct io_uring_cqe *io_get_cqring(struct io_ring_ctx *ctx)
 	return &rings->cqes[tail & ctx->cq_mask];
 }
 
+static inline bool io_should_trigger_evfd(struct io_ring_ctx *ctx)
+{
+	if (!ctx->eventfd_async)
+		return true;
+	return io_wq_current_is_worker() || in_interrupt();
+}
+
 static void io_cqring_ev_posted(struct io_ring_ctx *ctx)
 {
 	if (waitqueue_active(&ctx->wait))
 		wake_up(&ctx->wait);
 	if (waitqueue_active(&ctx->sqo_wait))
 		wake_up(&ctx->sqo_wait);
-	if (ctx->cq_ev_fd)
+	if (ctx->cq_ev_fd && io_should_trigger_evfd(ctx))
 		eventfd_signal(ctx->cq_ev_fd, 1);
 }
 
@@ -6552,10 +6560,17 @@ static int __io_uring_register(struct io_ring_ctx *ctx, unsigned opcode,
 		ret = io_sqe_files_update(ctx, arg, nr_args);
 		break;
 	case IORING_REGISTER_EVENTFD:
+	case IORING_REGISTER_EVENTFD_ASYNC:
 		ret = -EINVAL;
 		if (nr_args != 1)
 			break;
 		ret = io_eventfd_register(ctx, arg);
+		if (ret)
+			break;
+		if (opcode == IORING_REGISTER_EVENTFD_ASYNC)
+			ctx->eventfd_async = true;
+		else
+			ctx->eventfd_async = false;
 		break;
 	case IORING_UNREGISTER_EVENTFD:
 		ret = -EINVAL;
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index be64c9757ff1..02e0b2d59b63 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -192,6 +192,7 @@ struct io_uring_params {
 #define IORING_REGISTER_EVENTFD		4
 #define IORING_UNREGISTER_EVENTFD	5
 #define IORING_REGISTER_FILES_UPDATE	6
+#define IORING_REGISTER_EVENTFD_ASYNC	7
 
 struct io_uring_files_update {
 	__u32 offset;

-- 
Jens Axboe


  reply	other threads:[~2020-01-08 17:20 UTC|newest]

Thread overview: 11+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-01-07 15:55 io_uring and spurious wake-ups from eventfd Mark Papadakis
2020-01-07 20:26 ` Jens Axboe
2020-01-07 20:34   ` Jens Axboe
2020-01-08  7:36     ` Mark Papadakis
2020-01-08 16:24       ` Jens Axboe
2020-01-08 16:46         ` Mark Papadakis
2020-01-08 16:50           ` Jens Axboe
2020-01-08 17:20             ` Jens Axboe [this message]
2020-01-08 18:08               ` Jens Axboe
2020-01-09  6:09         ` Daurnimator
2020-01-09 15:14           ` Jens Axboe

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    [email protected] \
    [email protected] \
    [email protected] \
    [email protected] \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox