Commit 2ccc92f4 authored by Pavel Begunkov's avatar Pavel Begunkov Committed by Jens Axboe
Browse files

io_uring: add completion locking for iopoll 



There are pieces of code that may allow iopoll to race filling cqes,
temporarily add spinlocking around posting events.

Cc: stable@vger.kernel.org
Signed-off-by: default avatarPavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/84d86b5c117feda075471c5c9e65208e0dccf5d0.1669203009.git.asml.silence@gmail.com


Signed-off-by: default avatarJens Axboe <axboe@kernel.dk>
parent 6c16fe3c

io_uring/rw.c

+3 −2
Original line number Diff line number Diff line
@@ -1049,6 +1049,7 @@ int io_do_iopoll(struct io_ring_ctx *ctx, bool force_nonspin) 
	else if (!pos)

		return 0;



	spin_lock(&ctx->completion_lock);

	prev = start;

	wq_list_for_each_resume(pos, prev) {

		struct io_kiocb *req = container_of(pos, struct io_kiocb, comp_list);
@@ -1063,11 +1064,11 @@ int io_do_iopoll(struct io_ring_ctx *ctx, bool force_nonspin) 
		req->cqe.flags = io_put_kbuf(req, 0);

		__io_fill_cqe_req(req->ctx, req);

	}



	io_commit_cqring(ctx);

	spin_unlock(&ctx->completion_lock);

	if (unlikely(!nr_events))

		return 0;



	io_commit_cqring(ctx);

	io_cqring_ev_posted_iopoll(ctx);

	pos = start ? start->next : ctx->iopoll_list.first;

	wq_list_cut(&ctx->iopoll_list, prev, start);