From: Pavel Begunkov asml.silence@gmail.com
[ Upstream commit f0e20b8943509d81200cef5e30af2adfddba0f5c ]
There is a recipe to deadlock the kernel: submit a timeout sqe with a linked_timeout (e.g. test_single_link_timeout_ception() from liburing), and SIGKILL the process.
Then, io_kill_timeouts() takes @ctx->completion_lock, but the timeout isn't flagged with REQ_F_COMP_LOCKED, and will try to double grab it during io_put_free() to cancel the linked timeout. Probably, the same can happen with another io_kill_timeout() call site, that is io_commit_cqring().
Signed-off-by: Pavel Begunkov asml.silence@gmail.com Signed-off-by: Jens Axboe axboe@kernel.dk Signed-off-by: Sasha Levin sashal@kernel.org --- fs/io_uring.c | 1 + 1 file changed, 1 insertion(+)
diff --git a/fs/io_uring.c b/fs/io_uring.c index 2547c6395d5e4..44ae2641b4b06 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -688,6 +688,7 @@ static void io_kill_timeout(struct io_kiocb *req) if (ret != -1) { atomic_inc(&req->ctx->cq_timeouts); list_del_init(&req->list); + req->flags |= REQ_F_COMP_LOCKED; io_cqring_fill_event(req, 0); io_put_req(req); }