io_uring: fix complete_post use ctx after free

If io_req_complete_post() put not a final ref, we can't rely on the
request's ctx ref, and so ctx may potentially be freed while
complete_post() is in io_cqring_ev_posted()/etc.

In that case get an additional ctx reference, and put it in the end, so
protecting following io_cqring_ev_posted(). And also prolong ctx
lifetime until spin_unlock happens, as we do with mutexes, so added
percpu_ref_get() doesn't race with ctx free.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>

fs/io_uring.c

@@ -1550,14 +1550,17 @@ static void io_req_complete_post(struct io_kiocb *req, long res,
 		io_put_task(req->task, 1);
 		list_add(&req->compl.list, &cs->locked_free_list);
 		cs->locked_free_nr++;
-	} else
+	} else {
+		if (!percpu_ref_tryget(&ctx->refs))
 			req = NULL;
+	}
 	io_commit_cqring(ctx);
 	spin_unlock_irqrestore(&ctx->completion_lock, flags);
-	io_cqring_ev_posted(ctx);
 
-	if (req)
+	if (req) {
+		io_cqring_ev_posted(ctx);
 		percpu_ref_put(&ctx->refs);
+	}
 }
 
 static void io_req_complete_state(struct io_kiocb *req, long res,
@@ -8373,11 +8376,13 @@ static void io_ring_ctx_free(struct io_ring_ctx *ctx)
 {
 	/*
 	 * Some may use context even when all refs and requests have been put,
-	 * and they are free to do so while still holding uring_lock, see
-	 * __io_req_task_submit(). Wait for them to finish.
+	 * and they are free to do so while still holding uring_lock or
+	 * completion_lock, see __io_req_task_submit(). Wait for them to finish.
 	 */
 	mutex_lock(&ctx->uring_lock);
 	mutex_unlock(&ctx->uring_lock);
+	spin_lock_irq(&ctx->completion_lock);
+	spin_unlock_irq(&ctx->completion_lock);
 
 	io_sq_thread_finish(ctx);
 	io_sqe_buffers_unregister(ctx);