public inbox for [email protected]
 help / color / mirror / Atom feed
From: Xiaoguang Wang <[email protected]>
To: [email protected]
Cc: [email protected], [email protected],
	Xiaoguang Wang <[email protected]>
Subject: [LIBURING v2] Check cq ring overflow status
Date: Thu,  9 Jul 2020 09:16:20 +0800	[thread overview]
Message-ID: <[email protected]> (raw)

If cq ring has been overflowed, need to enter kernel to flush cqes.

Signed-off-by: Xiaoguang Wang <[email protected]>
---
 src/include/liburing/io_uring.h |  1 +
 src/queue.c                     | 17 +++++++++++++----
 2 files changed, 14 insertions(+), 4 deletions(-)

diff --git a/src/include/liburing/io_uring.h b/src/include/liburing/io_uring.h
index 6a73522..8302717 100644
--- a/src/include/liburing/io_uring.h
+++ b/src/include/liburing/io_uring.h
@@ -202,6 +202,7 @@ struct io_sqring_offsets {
  * sq_ring->flags
  */
 #define IORING_SQ_NEED_WAKEUP	(1U << 0) /* needs io_uring_enter wakeup */
+#define IORING_SQ_CQ_OVERFLOW	(1U << 1) /* CQ ring is overflown */
 
 struct io_cqring_offsets {
 	__u32 head;
diff --git a/src/queue.c b/src/queue.c
index 88e0294..1f00251 100644
--- a/src/queue.c
+++ b/src/queue.c
@@ -32,6 +32,11 @@ static inline bool sq_ring_needs_enter(struct io_uring *ring,
 	return false;
 }
 
+static inline bool cq_ring_needs_flush(struct io_uring *ring)
+{
+	return IO_URING_READ_ONCE(*ring->sq.kflags) & IORING_SQ_CQ_OVERFLOW;
+}
+
 static int __io_uring_peek_cqe(struct io_uring *ring,
 			       struct io_uring_cqe **cqe_ptr)
 {
@@ -67,22 +72,26 @@ int __io_uring_get_cqe(struct io_uring *ring, struct io_uring_cqe **cqe_ptr,
 	int ret = 0, err;
 
 	do {
+		bool cq_overflow_flush = false;
 		unsigned flags = 0;
 
 		err = __io_uring_peek_cqe(ring, &cqe);
 		if (err)
 			break;
 		if (!cqe && !to_wait && !submit) {
-			err = -EAGAIN;
-			break;
+			if (!cq_ring_needs_flush(ring)) {
+				err = -EAGAIN;
+				break;
+			}
+			cq_overflow_flush = true;
 		}
 		if (wait_nr && cqe)
 			wait_nr--;
-		if (wait_nr)
+		if (wait_nr || cq_overflow_flush)
 			flags = IORING_ENTER_GETEVENTS;
 		if (submit)
 			sq_ring_needs_enter(ring, submit, &flags);
-		if (wait_nr || submit)
+		if (wait_nr || submit || cq_overflow_flush)
 			ret = __sys_io_uring_enter(ring->ring_fd, submit,
 						   wait_nr, flags, sigmask);
 		if (ret < 0) {
-- 
2.17.2


             reply	other threads:[~2020-07-09  1:16 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-09  1:16 Xiaoguang Wang [this message]
2020-07-09  1:30 ` [LIBURING v2] Check cq ring overflow status 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 \
    --in-reply-to=20200709011620.13666-1-xiaoguang.wang@linux.alibaba.com \
    [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