[RFC v2 2/3] nvme: move ns id info to struct nvme_ns_head
Christoph Hellwig
hch at lst.de
Sun Dec 3 23:51:34 PST 2023
> + if (head) {
> pr_err_ratelimited("%s: %s(0x%x) @ LBA %llu, %llu blocks, %s (sct 0x%x / sc 0x%x) %s%s\n",
> - ns->disk ? ns->disk->disk_name : "?",
> + head->disk ? head->disk->disk_name : "?",
> nvme_get_opcode_str(nr->cmd->common.opcode),
> nr->cmd->common.opcode,
> - (unsigned long long)nvme_sect_to_lba(ns, blk_rq_pos(req)),
> - (unsigned long long)blk_rq_bytes(req) >> ns->lba_shift,
> + (unsigned long long)nvme_sect_to_lba(head, blk_rq_pos(req)),
Please avoid the overly long line here. Best done by removing the pointless
unsigned long long casts, as u64 is always an unsigned long long in the
kernel these days.
> - u64 slba = nvme_sect_to_lba(ns, bio->bi_iter.bi_sector);
> - u32 nlb = bio->bi_iter.bi_size >> ns->lba_shift;
> + u64 slba = nvme_sect_to_lba(head, bio->bi_iter.bi_sector);
Please avoid the overly long line here as well.
> +static void nvme_set_ref_tag(struct nvme_ns_head *head, struct nvme_command *cmnd,
.. and here. I'm going to stop now, please also fix up all other
places.
> void nvme_failover_req(struct request *req)
> {
> - struct nvme_ns *ns = req->q->queuedata;
> + struct nvme_ns_head *head = req->q->queuedata;
> + struct nvme_ctrl *ctrl = nvme_req(req)->ctrl;
> + struct nvme_ns *ns;
> u16 status = nvme_req(req)->status & 0x7ff;
> unsigned long flags;
> struct bio *bio;
>
> - nvme_mpath_clear_current_path(ns);
> + nvme_mpath_clear_current_path(head);
>
> /*
> * If we got back an ANA error, we know the controller is alive but not
> * ready to serve this namespace. Kick of a re-read of the ANA
> * information page, and just try any other available path for now.
> */
> - if (nvme_is_ana_error(status) && ns->ctrl->ana_log_buf) {
> + if (nvme_is_ana_error(status) && ctrl->ana_log_buf) {
> + ns = nvme_find_get_ns(ctrl, head->ns_id);
This looks unrelated.
> -bool nvme_mpath_clear_current_path(struct nvme_ns *ns)
> +bool nvme_mpath_clear_current_path(struct nvme_ns_head *head)
> {
> - struct nvme_ns_head *head = ns->head;
> bool changed = false;
> int node;
>
> @@ -181,7 +183,7 @@ bool nvme_mpath_clear_current_path(struct nvme_ns *ns)
> goto out;
>
> for_each_node(node) {
> - if (ns == rcu_access_pointer(head->current_path[node])) {
> + if (head == rcu_access_pointer(head->current_path[node])->head) {
and this can't work. We need to check the actual ns for the path
selection, as that's kindof the point.
More information about the Linux-nvme
mailing list