[PATCH 1/1] io_uring: don't wait when under-submitting

From: Pavel Begunkov <asml.silence@gmail.com>
To: Jens Axboe <axboe@kernel.dk>,
	io-uring@vger.kernel.org, linux-kernel@vger.kernel.org
Subject: [PATCH 1/1] io_uring: don't wait when under-submitting
Date: Fri, 13 Dec 2019 10:51:00 +0300	[thread overview]
Message-ID: <5caa38be87f069eb4cc921d58ee1a98ff5d53978.1576223348.git.asml.silence@gmail.com> (raw)

There is no reliable way to submit and wait in a single syscall, as
io_submit_sqes() may under-consume sqes (in case of an early error).
Then it will wait for not-yet-submitted requests, deadlocking the user
in most cases.

Don't wait/poll if can't submit all sqes, and return -EAGAIN

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---

I wonder, why it doesn't return 2 error codes for submission and waiting
separately? It's a bit puzzling figuring out what to return. I guess the
same with the userspace side.

 fs/io_uring.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 42de210be631..82152ea13fe2 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -4877,6 +4877,11 @@ SYSCALL_DEFINE6(io_uring_enter, unsigned int, fd, u32, to_submit,
 		submitted = io_submit_sqes(ctx, to_submit, f.file, fd,
 					   &cur_mm, false);
 		mutex_unlock(&ctx->uring_lock);
+
+		if (submitted != to_submit) {
+			submitted = -EAGAIN;
+			goto out;
+		}
 	}
 	if (flags & IORING_ENTER_GETEVENTS) {
 		unsigned nr_events = 0;
@@ -4890,6 +4895,7 @@ SYSCALL_DEFINE6(io_uring_enter, unsigned int, fd, u32, to_submit,
 		}
 	}
 
+out:
 	percpu_ref_put(&ctx->refs);
 out_fput:
 	fdput(f);
-- 
2.24.0

