io_uring: don't fail overflow on in_idle

As CQE overflows are now untied from requests and so don't hold any
ref, we don't need to handle exiting/exec'ing cases there anymore.
Moreover, it's much nicer in regards to userspace to save overflowed
CQEs whenever possible, so remove failing on in_idle.
Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/d873b7dab75c7f3039ead9628a745bea01f2cfd2.1618278933.git.asml.silence@gmail.comSigned-off-by: Jens Axboe <axboe@kernel.dk>

fs/io_uring.c

@@ -1513,13 +1513,18 @@ static bool io_cqring_event_overflow(struct io_kiocb *req, long res,
 				     unsigned int cflags)
 {
 	struct io_ring_ctx *ctx = req->ctx;
-
-	if (!atomic_read(&req->task->io_uring->in_idle)) {
 	struct io_overflow_cqe *ocqe;
 
 	ocqe = kmalloc(sizeof(*ocqe), GFP_ATOMIC | __GFP_ACCOUNT);
-		if (!ocqe)
-			goto overflow;
+	if (!ocqe) {
+		/*
+		 * If we're in ring overflow flush mode, or in task cancel mode,
+		 * or cannot allocate an overflow entry, then we need to drop it
+		 * on the floor.
+		 */
+		WRITE_ONCE(ctx->rings->cq_overflow, ++ctx->cached_cq_overflow);
+		return false;
+	}
 	if (list_empty(&ctx->cq_overflow_list)) {
 		set_bit(0, &ctx->sq_check_overflow);
 		set_bit(0, &ctx->cq_check_overflow);
@@ -1530,15 +1535,6 @@ static bool io_cqring_event_overflow(struct io_kiocb *req, long res,
 	ocqe->cqe.flags = cflags;
 	list_add_tail(&ocqe->list, &ctx->cq_overflow_list);
 	return true;
-	}
-overflow:
-	/*
-	 * If we're in ring overflow flush mode, or in task cancel mode,
-	 * or cannot allocate an overflow entry, then we need to drop it
-	 * on the floor.
-	 */
-	WRITE_ONCE(ctx->rings->cq_overflow, ++ctx->cached_cq_overflow);
-	return false;
 }
 
 static inline bool __io_cqring_fill_event(struct io_kiocb *req, long res,