commit | 902e4460bb45466df95228913ae799a17e86e470 | [log] [tgz] |
---|---|---|
author | Jens Axboe <axboe@kernel.dk> | Sun Nov 10 15:28:23 2019 -0700 |
committer | Jens Axboe <axboe@kernel.dk> | Sun Nov 10 15:29:47 2019 -0700 |
tree | 60a04435c1a2c4a34a96a82fe87aba7da23d46b1 | |
parent | 82efd43f751e4fda7515a8d6950e82c02a5c7280 [diff] |
Fix io_uring_get_sqe() for IORING_SETUP_SQPOLL If we use kernel side polling, we can't use the flush done by io_uring_submit() as the source of truth for whether or not we can grab more SQE entries. We need to use the kernel side for that, as the SQPOLL thread could not have caught up yet. Signed-off-by: Jens Axboe <axboe@kernel.dk>