io_uring: never overflow io_aux_cqe

Now all callers of io_aux_cqe() set allow_overflow to false, remove the
parameter and not allow overflowing auxilary multishot cqes.

When CQ is full the function callers and all multishot requests in
general are expected to complete the request. That prevents indefinite
in-background grows of the overflow list and let's the userspace to
handle the backlog at its own pace.

Resubmitting a request should also be faster than accounting a bunch of
overflows, so it should be better for perf when it happens, but a well
behaving userspace should be trying to avoid overflows in any case.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
Link: https://lore.kernel.org/r/bb20d14d708ea174721e58bb53786b0521e4dd6d.1691757663.git.asml.silence@gmail.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>

Author: isilence

io_uring/io_uring.c

@@ -939,15 +939,18 @@ bool io_post_aux_cqe(struct io_ring_ctx *ctx, u64 user_data, s32 res, u32 cflags
 	return __io_post_aux_cqe(ctx, user_data, res, cflags, true);
 }
 
-bool io_aux_cqe(const struct io_kiocb *req, bool defer, s32 res, u32 cflags,
-		bool allow_overflow)
+/*
+ * A helper for multishot requests posting additional CQEs.
+ * Should only be used from a task_work including IO_URING_F_MULTISHOT.
+ */
+bool io_fill_cqe_req_aux(struct io_kiocb *req, bool defer, s32 res, u32 cflags)
 {
 	struct io_ring_ctx *ctx = req->ctx;
 	u64 user_data = req->cqe.user_data;
 	struct io_uring_cqe *cqe;
 
 	if (!defer)
-		return __io_post_aux_cqe(ctx, user_data, res, cflags, allow_overflow);
+		return __io_post_aux_cqe(ctx, user_data, res, cflags, false);
 
 	lockdep_assert_held(&ctx->uring_lock);
 
@@ -962,7 +965,7 @@ bool io_aux_cqe(const struct io_kiocb *req, bool defer, s32 res, u32 cflags,
 	 * however it's main job is to prevent unbounded posted completions,
 	 * and in that it works just as well.
 	 */
-	if (!allow_overflow && test_bit(IO_CHECK_CQ_OVERFLOW_BIT, &ctx->check_cq))
+	if (test_bit(IO_CHECK_CQ_OVERFLOW_BIT, &ctx->check_cq))
 		return false;
 
 	cqe = &ctx->submit_state.cqes[ctx->submit_state.cqes_count++];

io_uring/io_uring.h

@@ -44,8 +44,7 @@ int io_run_task_work_sig(struct io_ring_ctx *ctx);
 void io_req_defer_failed(struct io_kiocb *req, s32 res);
 void io_req_complete_post(struct io_kiocb *req, unsigned issue_flags);
 bool io_post_aux_cqe(struct io_ring_ctx *ctx, u64 user_data, s32 res, u32 cflags);
-bool io_aux_cqe(const struct io_kiocb *req, bool defer, s32 res, u32 cflags,
-		bool allow_overflow);
+bool io_fill_cqe_req_aux(struct io_kiocb *req, bool defer, s32 res, u32 cflags);
 void __io_commit_cqring_flush(struct io_ring_ctx *ctx);
 
 struct page **io_pin_pages(unsigned long ubuf, unsigned long len, int *npages);

io_uring/net.c

@@ -641,8 +641,8 @@ static inline bool io_recv_finish(struct io_kiocb *req, int *ret,
 	}
 
 	if (!mshot_finished) {
-		if (io_aux_cqe(req, issue_flags & IO_URING_F_COMPLETE_DEFER,
-			       *ret, cflags | IORING_CQE_F_MORE, false)) {
+		if (io_fill_cqe_req_aux(req, issue_flags & IO_URING_F_COMPLETE_DEFER,
+					*ret, cflags | IORING_CQE_F_MORE)) {
 			io_recv_prep_retry(req);
 			/* Known not-empty or unknown state, retry */
 			if (cflags & IORING_CQE_F_SOCK_NONEMPTY ||
@@ -1366,8 +1366,8 @@ int io_accept(struct io_kiocb *req, unsigned int issue_flags)
 
 	if (ret < 0)
 		return ret;
-	if (io_aux_cqe(req, issue_flags & IO_URING_F_COMPLETE_DEFER, ret,
-		       IORING_CQE_F_MORE, false))
+	if (io_fill_cqe_req_aux(req, issue_flags & IO_URING_F_COMPLETE_DEFER,
+				ret, IORING_CQE_F_MORE))
 		goto retry;
 
 	return -ECANCELED;

io_uring/poll.c

@@ -300,8 +300,8 @@ static int io_poll_check_events(struct io_kiocb *req, struct io_tw_state *ts)
 			__poll_t mask = mangle_poll(req->cqe.res &
 						    req->apoll_events);
 
-			if (!io_aux_cqe(req, ts->locked, mask,
-					IORING_CQE_F_MORE, false)) {
+			if (!io_fill_cqe_req_aux(req, ts->locked, mask,
+						 IORING_CQE_F_MORE)) {
 				io_req_set_res(req, mask, 0);
 				return IOU_POLL_REMOVE_POLL_USE_RES;
 			}

io_uring/timeout.c

@@ -73,8 +73,8 @@ static void io_timeout_complete(struct io_kiocb *req, struct io_tw_state *ts)
 
 	if (!io_timeout_finish(timeout, data)) {
 		bool filled;
-		filled = io_aux_cqe(req, ts->locked, -ETIME, IORING_CQE_F_MORE,
-				    false);
+		filled = io_fill_cqe_req_aux(req, ts->locked, -ETIME,
+					     IORING_CQE_F_MORE);
 		if (filled) {
 			/* re-arm timer */
 			spin_lock_irq(&ctx->timeout_lock);