[PATCH 1/1] io_uring: don't modify req->poll for rw

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



__io_queue_proc() is used by both poll and apoll, so we should not
access req->poll directly but selecting right struct io_poll_iocb
depending on use case.

Reported-and-tested-by: syzbot+a84b8783366ecb1c65d0@xxxxxxxxxxxxxxxxxxxxxxxxx
Fixes: ea6a693d862d ("io_uring: disable multishot poll for double poll add cases")
Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx>
---
 fs/io_uring.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index e387ce687f4d..63f6b11d271b 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -5016,10 +5016,10 @@ static void __io_queue_proc(struct io_poll_iocb *poll, struct io_poll_table *pt,
 		 * Can't handle multishot for double wait for now, turn it
 		 * into one-shot mode.
 		 */
-		if (!(req->poll.events & EPOLLONESHOT))
-			req->poll.events |= EPOLLONESHOT;
+		if (!(poll_one->events & EPOLLONESHOT))
+			poll_one->events |= EPOLLONESHOT;
 		/* double add on the same waitqueue head, ignore */
-		if (poll->head == head)
+		if (poll_one->head == head)
 			return;
 		poll = kmalloc(sizeof(*poll), GFP_ATOMIC);
 		if (!poll) {
-- 
2.31.1




[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux