Skip to content

Commit 38fc96a

Browse files
isilenceaxboe
authored andcommitted
io_uring/rw: fix downgraded mshot read
The io-wq path can downgrade a multishot request to oneshot mode, however io_read_mshot() doesn't handle that and would still post multiple CQEs. That's not allowed, because io_req_post_cqe() requires stricter context requirements. The described can only happen with pollable files that don't support FMODE_NOWAIT, which is an odd combination, so if even allowed it should be fairly rare. Cc: [email protected] Reported-by: chase xd <[email protected]> Fixes: bee1d5b ("io_uring: disable io-wq execution of multishot NOWAIT requests") Signed-off-by: Pavel Begunkov <[email protected]> Link: https://lore.kernel.org/r/c5c8c4a50a882fd581257b81bf52eee260ac29fd.1735407848.git.asml.silence@gmail.com Signed-off-by: Jens Axboe <[email protected]>
1 parent e33ac68 commit 38fc96a

File tree

1 file changed

+2
-0
lines changed

1 file changed

+2
-0
lines changed

io_uring/rw.c

+2
Original file line numberDiff line numberDiff line change
@@ -983,6 +983,8 @@ int io_read_mshot(struct io_kiocb *req, unsigned int issue_flags)
983983
io_kbuf_recycle(req, issue_flags);
984984
if (ret < 0)
985985
req_set_fail(req);
986+
} else if (!(req->flags & REQ_F_APOLL_MULTISHOT)) {
987+
cflags = io_put_kbuf(req, ret, issue_flags);
986988
} else {
987989
/*
988990
* Any successful return value will keep the multishot read

0 commit comments

Comments
 (0)