From 71e32edd2210d0304e93ac110814b5a4b3a81dc0 Mon Sep 17 00:00:00 2001
From: Jens Axboe <axboe@kernel.dk>
Date: Sun, 12 Sep 2021 06:45:07 -0600
Subject: [PATCH] io_uring: ensure symmetry in handling iter types in loop_rw_iter()
Git-commit: 16c8d2df7ec0eed31b7d3b61cb13206a7fb930cc
Patch-mainline: v5.15-rc2
References: stable-5.14.7 bsc#1190664 CVE-2021-41073

commit 16c8d2df7ec0eed31b7d3b61cb13206a7fb930cc upstream.

When setting up the next segment, we check what type the iter is and
handle it accordingly. However, when incrementing and processed amount
we do not, and both iter advance and addr/len are adjusted, regardless
of type. Split the increment side just like we do on the setup side.

Fixes: 4017eb91a9e7 ("io_uring: make loop_rw_iter() use original user supplied pointers")
Cc: stable@vger.kernel.org
Reported-by: Valentina Palmiotti <vpalmiotti@gmail.com>
Reviewed-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Takashi Iwai <tiwai@suse.de>

---
 fs/io_uring.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index c5d4638f6d7f..e5698b3beb0e 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -3107,12 +3107,15 @@ static ssize_t loop_rw_iter(int rw, struct io_kiocb *req, struct iov_iter *iter)
 				ret = nr;
 			break;
 		}
+		if (!iov_iter_is_bvec(iter)) {
+			iov_iter_advance(iter, nr);
+		} else {
+			req->rw.len -= nr;
+			req->rw.addr += nr;
+		}
 		ret += nr;
 		if (nr != iovec.iov_len)
 			break;
-		req->rw.len -= nr;
-		req->rw.addr += nr;
-		iov_iter_advance(iter, nr);
 	}
 
 	return ret;
-- 
2.26.2