On 1/7/25 1:52 PM, Christoph Hellwig wrote: > On Tue, Jan 07, 2025 at 12:28:29PM +0530, Nilay Shroff wrote: >>> - blk_mq_freeze_queue(ns->disk->queue); >>> lim = queue_limits_start_update(ns->disk->queue); >>> nvme_set_ctrl_limits(ns->ctrl, &lim); >>> + >>> + blk_mq_freeze_queue(ns->disk->queue); >>> ret = queue_limits_commit_update(ns->disk->queue, &lim); >>> set_disk_ro(ns->disk, nvme_ns_is_readonly(ns, info)); >>> blk_mq_unfreeze_queue(ns->disk->queue); >> >> I think we should freeze queue before nvme_set_ctrl_limits(). > > Why? > The nvme_set_ctrl_limits() sets certain queue limits which are also used during IO processing. For instance, ->max_segment_size is used while submitting bio. Also, if we look at the code before your patch, nvme_set_ctrl_limits() is called when the queue is freezed. Thanks, --Nilay
