We already avoid flushing overflows in io_cqring_wait_schedule() but only return an error for the outer loop to handle it. Minimise it even further by moving all ->check_cq parsing there. Signed-off-by: Pavel Begunkov <asml.silence@xxxxxxxxx> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> --- io_uring/io_uring.c | 32 ++++++++++++++++++-------------- 1 file changed, 18 insertions(+), 14 deletions(-) diff --git a/io_uring/io_uring.c b/io_uring/io_uring.c index 3bb3e9889717..067e3577ac9b 100644 --- a/io_uring/io_uring.c +++ b/io_uring/io_uring.c @@ -2468,21 +2468,13 @@ static inline int io_cqring_wait_schedule(struct io_ring_ctx *ctx, ktime_t *timeout) { int ret; - unsigned long check_cq; + if (unlikely(READ_ONCE(ctx->check_cq))) + return 1; /* make sure we run task_work before checking for signals */ ret = io_run_task_work_sig(ctx); if (ret || io_should_wake(iowq)) return ret; - - check_cq = READ_ONCE(ctx->check_cq); - if (unlikely(check_cq)) { - /* let the caller flush overflows, retry */ - if (check_cq & BIT(IO_CHECK_CQ_OVERFLOW_BIT)) - return 1; - if (check_cq & BIT(IO_CHECK_CQ_DROPPED_BIT)) - return -EBADR; - } if (!schedule_hrtimeout(timeout, HRTIMER_MODE_ABS)) return -ETIME; @@ -2548,13 +2540,25 @@ static int io_cqring_wait(struct io_ring_ctx *ctx, int min_events, trace_io_uring_cqring_wait(ctx, min_events); do { - if (test_bit(IO_CHECK_CQ_OVERFLOW_BIT, &ctx->check_cq)) { - finish_wait(&ctx->cq_wait, &iowq.wq); - io_cqring_do_overflow_flush(ctx); - } + unsigned long check_cq; + prepare_to_wait_exclusive(&ctx->cq_wait, &iowq.wq, TASK_INTERRUPTIBLE); ret = io_cqring_wait_schedule(ctx, &iowq, &timeout); + + check_cq = READ_ONCE(ctx->check_cq); + if (unlikely(check_cq)) { + /* let the caller flush overflows, retry */ + if (check_cq & BIT(IO_CHECK_CQ_OVERFLOW_BIT)) { + finish_wait(&ctx->cq_wait, &iowq.wq); + io_cqring_do_overflow_flush(ctx); + } + if (check_cq & BIT(IO_CHECK_CQ_DROPPED_BIT)) { + ret = -EBADR; + break; + } + } + if (__io_cqring_events_user(ctx) >= min_events) break; cond_resched(); -- 2.38.1