public inbox for [email protected]
 help / color / mirror / Atom feed
From: Pavel Begunkov <[email protected]>
To: [email protected]
Cc: Jens Axboe <[email protected]>, [email protected]
Subject: [PATCH for-next v2 5/6] io_uring: refactor poll arm error handling
Date: Thu, 23 Jun 2022 14:24:48 +0100	[thread overview]
Message-ID: <018cacdaef5fe95d7dc56b32e85d752cab7607f6.1655990418.git.asml.silence@gmail.com> (raw)
In-Reply-To: <[email protected]>

__io_arm_poll_handler() errors parsing is a horror, in case it failed it
returns 0 and the caller is expected to look at ipt.error, which already
led us to a number of problems before.

When it returns a valid mask, leave it as it's not, i.e. return 1 and
store the mask in ipt.result_mask. In case of a failure that can be
handled inline return an error code (negative value), and return 0 if
__io_arm_poll_handler() took ownership of the request and will complete
it.

Signed-off-by: Pavel Begunkov <[email protected]>
---
 io_uring/poll.c | 44 +++++++++++++++++++++-----------------------
 1 file changed, 21 insertions(+), 23 deletions(-)

diff --git a/io_uring/poll.c b/io_uring/poll.c
index 80113b036c88..3f3ae3b1505f 100644
--- a/io_uring/poll.c
+++ b/io_uring/poll.c
@@ -435,6 +435,12 @@ static void io_poll_queue_proc(struct file *file, struct wait_queue_head *head,
 			(struct io_poll **) &pt->req->async_data);
 }
 
+/*
+ * Returns 0 when it's handed over for polling. The caller owns the requests if
+ * it returns non-zero, but otherwise should not touch it. Negative values
+ * contain an error code. When the result is >0, the polling has completed
+ * inline and ipt.result_mask is set to the mask.
+ */
 static int __io_arm_poll_handler(struct io_kiocb *req,
 				 struct io_poll *poll,
 				 struct io_poll_table *ipt, __poll_t mask)
@@ -461,6 +467,17 @@ static int __io_arm_poll_handler(struct io_kiocb *req,
 	atomic_set(&req->poll_refs, 1);
 	mask = vfs_poll(req->file, &ipt->pt) & poll->events;
 
+	if (unlikely(ipt->error || !ipt->nr_entries)) {
+		io_poll_remove_entries(req);
+
+		if (mask && (poll->events & EPOLLET)) {
+			ipt->result_mask = mask;
+			return 1;
+		} else {
+			return ipt->error ?: -EINVAL;
+		}
+	}
+
 	if (mask &&
 	   ((poll->events & (EPOLLET|EPOLLONESHOT)) == (EPOLLET|EPOLLONESHOT))) {
 		io_poll_remove_entries(req);
@@ -469,25 +486,12 @@ static int __io_arm_poll_handler(struct io_kiocb *req,
 		return 1;
 	}
 
-	if (!mask && unlikely(ipt->error || !ipt->nr_entries)) {
-		io_poll_remove_entries(req);
-		if (!ipt->error)
-			ipt->error = -EINVAL;
-		return 0;
-	}
-
 	if (req->flags & REQ_F_HASH_LOCKED)
 		io_poll_req_insert_locked(req);
 	else
 		io_poll_req_insert(req);
 
 	if (mask && (poll->events & EPOLLET)) {
-		/* can't multishot if failed, just queue the event we've got */
-		if (unlikely(ipt->error || !ipt->nr_entries)) {
-			poll->events |= EPOLLONESHOT;
-			req->apoll_events |= EPOLLONESHOT;
-			ipt->error = 0;
-		}
 		__io_poll_execute(req, mask);
 		return 0;
 	}
@@ -582,9 +586,8 @@ int io_arm_poll_handler(struct io_kiocb *req, unsigned issue_flags)
 	io_kbuf_recycle(req, issue_flags);
 
 	ret = __io_arm_poll_handler(req, &apoll->poll, &ipt, mask);
-	if (ret || ipt.error)
-		return ret ? IO_APOLL_READY : IO_APOLL_ABORTED;
-
+	if (ret)
+		return ret > 0 ? IO_APOLL_READY : IO_APOLL_ABORTED;
 	trace_io_uring_poll_arm(req, mask, apoll->poll.events);
 	return IO_APOLL_OK;
 }
@@ -815,16 +818,11 @@ int io_poll_add(struct io_kiocb *req, unsigned int issue_flags)
 		req->flags &= ~REQ_F_HASH_LOCKED;
 
 	ret = __io_arm_poll_handler(req, poll, &ipt, poll->events);
-	if (ret) {
+	if (ret > 0) {
 		io_req_set_res(req, ipt.result_mask, 0);
 		return IOU_OK;
 	}
-	if (ipt.error) {
-		req_set_fail(req);
-		return ipt.error;
-	}
-
-	return IOU_ISSUE_SKIP_COMPLETE;
+	return ret ?: IOU_ISSUE_SKIP_COMPLETE;
 }
 
 int io_poll_remove(struct io_kiocb *req, unsigned int issue_flags)
-- 
2.36.1


  parent reply	other threads:[~2022-06-23 13:25 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-06-23 13:24 [PATCH for-next v2 0/6] poll cleanups and optimisations Pavel Begunkov
2022-06-23 13:24 ` [PATCH for-next v2 1/6] io_uring: clean poll ->private flagging Pavel Begunkov
2022-06-23 13:24 ` [PATCH for-next v2 2/6] io_uring: remove events caching atavisms Pavel Begunkov
2022-06-23 13:24 ` [PATCH for-next v2 3/6] io_uring: add a helper for apoll alloc Pavel Begunkov
2022-06-23 13:24 ` [PATCH for-next v2 4/6] io_uring: change arm poll return values Pavel Begunkov
2022-06-23 13:24 ` Pavel Begunkov [this message]
2022-06-23 13:24 ` [PATCH for-next v2 6/6] io_uring: optimise submission side poll_refs Pavel Begunkov
2022-06-23 13:47 ` [PATCH for-next v2 0/6] poll cleanups and optimisations Jens Axboe
2022-06-23 13:59 ` 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=018cacdaef5fe95d7dc56b32e85d752cab7607f6.1655990418.git.asml.silence@gmail.com \
    [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