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-6.0] io_uring/rw: fix short rw error handling
Date: Fri,  9 Sep 2022 12:11:49 +0100	[thread overview]
Message-ID: <89473c1a9205760c4fa6d158058da7b594a815f0.1662721607.git.asml.silence@gmail.com> (raw)

We have a couple of problems, first reports of unexpected link breakage
for reads when cqe->res indicates that the IO was done in full. The
reason here is partial IO with retries.

TL;DR; we compare the result in __io_complete_rw_common() against
req->cqe.res, but req->cqe.res doesn't store the full length but rather
the length left to be done. So, when we pass the full corrected result
via kiocb_done() -> __io_complete_rw_common(), it fails.

The second problem is that we don't try to correct res in
io_complete_rw(), which, for instance, might be a problem for O_DIRECT
but when a prefix of data was cached in the page cache. We also
definitely don't want to pass a corrected result into io_rw_done().

The fix here is to leave __io_complete_rw_common() alone, always pass
not corrected result into it and fix it up as the last step just before
actually finishing the I/O.

Cc: [email protected]
Signed-off-by: Pavel Begunkov <[email protected]>
---
 io_uring/rw.c | 30 ++++++++++++++++++------------
 1 file changed, 18 insertions(+), 12 deletions(-)

diff --git a/io_uring/rw.c b/io_uring/rw.c
index 1babd77da79c..1e18a44adcf5 100644
--- a/io_uring/rw.c
+++ b/io_uring/rw.c
@@ -206,6 +206,20 @@ static bool __io_complete_rw_common(struct io_kiocb *req, long res)
 	return false;
 }
 
+static inline unsigned io_fixup_rw_res(struct io_kiocb *req, unsigned res)
+{
+	struct io_async_rw *io = req->async_data;
+
+	/* add previously done IO, if any */
+	if (req_has_async_data(req) && io->bytes_done > 0) {
+		if (res < 0)
+			res = io->bytes_done;
+		else
+			res += io->bytes_done;
+	}
+	return res;
+}
+
 static void io_complete_rw(struct kiocb *kiocb, long res)
 {
 	struct io_rw *rw = container_of(kiocb, struct io_rw, kiocb);
@@ -213,7 +227,7 @@ static void io_complete_rw(struct kiocb *kiocb, long res)
 
 	if (__io_complete_rw_common(req, res))
 		return;
-	io_req_set_res(req, res, 0);
+	io_req_set_res(req, io_fixup_rw_res(req, res), 0);
 	req->io_task_work.func = io_req_task_complete;
 	io_req_task_work_add(req);
 }
@@ -240,22 +254,14 @@ static void io_complete_rw_iopoll(struct kiocb *kiocb, long res)
 static int kiocb_done(struct io_kiocb *req, ssize_t ret,
 		       unsigned int issue_flags)
 {
-	struct io_async_rw *io = req->async_data;
 	struct io_rw *rw = io_kiocb_to_cmd(req, struct io_rw);
-
-	/* add previously done IO, if any */
-	if (req_has_async_data(req) && io->bytes_done > 0) {
-		if (ret < 0)
-			ret = io->bytes_done;
-		else
-			ret += io->bytes_done;
-	}
+	unsigned final_ret = io_fixup_rw_res(req, ret);
 
 	if (req->flags & REQ_F_CUR_POS)
 		req->file->f_pos = rw->kiocb.ki_pos;
 	if (ret >= 0 && (rw->kiocb.ki_complete == io_complete_rw)) {
 		if (!__io_complete_rw_common(req, ret)) {
-			io_req_set_res(req, req->cqe.res,
+			io_req_set_res(req, final_ret,
 				       io_put_kbuf(req, issue_flags));
 			return IOU_OK;
 		}
@@ -268,7 +274,7 @@ static int kiocb_done(struct io_kiocb *req, ssize_t ret,
 		if (io_resubmit_prep(req))
 			io_req_task_queue_reissue(req);
 		else
-			io_req_task_queue_fail(req, ret);
+			io_req_task_queue_fail(req, final_ret);
 	}
 	return IOU_ISSUE_SKIP_COMPLETE;
 }
-- 
2.37.2


             reply	other threads:[~2022-09-09 11:14 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-09 11:11 Pavel Begunkov [this message]
2022-09-09 14:59 ` [PATCH for-6.0] io_uring/rw: fix short rw error handling 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=89473c1a9205760c4fa6d158058da7b594a815f0.1662721607.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