From: Christoph Hellwig <[email protected]>
To: Jens Axboe <[email protected]>
Cc: Christian Brauner <[email protected]>,
Keith Busch <[email protected]>, Sagi Grimberg <[email protected]>,
Kanchan Joshi <[email protected]>,
Hui Qi <[email protected]>,
Nitesh Shetty <[email protected]>, Jan Kara <[email protected]>,
Pavel Begunkov <[email protected]>,
[email protected], [email protected],
[email protected], [email protected],
[email protected]
Subject: [PATCH 14/15] nvme: enable FDP support
Date: Tue, 19 Nov 2024 13:16:28 +0100 [thread overview]
Message-ID: <[email protected]> (raw)
In-Reply-To: <[email protected]>
Wire up the block level write streams to the NVMe Flexible Data Placement
(FDP) feature as ratified in TP 4146a.
Based on code from Kanchan Joshi <[email protected]>,
Hui Qi <[email protected]>, Nitesh Shetty <[email protected]> and
Keith Busch <[email protected]>, but a lot of it has been rewritten to
fit the block layer write stream infrastructure.
Signed-off-by: Christoph Hellwig <[email protected]>
---
drivers/nvme/host/core.c | 129 +++++++++++++++++++++++++++++++++++++++
drivers/nvme/host/nvme.h | 4 ++
2 files changed, 133 insertions(+)
diff --git a/drivers/nvme/host/core.c b/drivers/nvme/host/core.c
index b61225201b47..543bbe7de063 100644
--- a/drivers/nvme/host/core.c
+++ b/drivers/nvme/host/core.c
@@ -673,6 +673,7 @@ static void nvme_free_ns_head(struct kref *ref)
ida_free(&head->subsys->ns_ida, head->instance);
cleanup_srcu_struct(&head->srcu);
nvme_put_subsystem(head->subsys);
+ kfree(head->plids);
kfree(head);
}
@@ -990,6 +991,15 @@ static inline blk_status_t nvme_setup_rw(struct nvme_ns *ns,
if (req->cmd_flags & REQ_RAHEAD)
dsmgmt |= NVME_RW_DSM_FREQ_PREFETCH;
+ if (op == nvme_cmd_write && ns->head->nr_plids) {
+ u16 write_stream = req->bio->bi_write_stream;
+
+ if (WARN_ON_ONCE(write_stream > ns->head->nr_plids))
+ return BLK_STS_INVAL;
+ dsmgmt |= ns->head->plids[write_stream - 1] << 16;
+ control |= NVME_RW_DTYPE_DPLCMT;
+ }
+
if (req->cmd_flags & REQ_ATOMIC && !nvme_valid_atomic_write(req))
return BLK_STS_INVAL;
@@ -2142,6 +2152,107 @@ static int nvme_update_ns_info_generic(struct nvme_ns *ns,
return ret;
}
+static int nvme_read_fdp_config(struct nvme_ns *ns, struct nvme_ns_info *info)
+{
+ struct nvme_fdp_config result;
+ struct nvme_fdp_config_log *log;
+ struct nvme_fdp_config_desc *configs;
+ size_t log_size;
+ int error;
+
+ error = nvme_get_features(ns->ctrl, NVME_FEAT_FDP, info->endgid, NULL,
+ 0, &result);
+ if (error)
+ return error;
+
+ if (!(result.flags & FDPCFG_FDPE)) {
+ dev_warn(ns->ctrl->device, "FDP not enable in current config\n");
+ return -EINVAL;
+ }
+
+ log_size = sizeof(*log) + (result.fdpcidx + 1) * sizeof(*configs);
+ log = kmalloc(log_size, GFP_KERNEL);
+ if (!log)
+ return -ENOMEM;
+
+ error = nvme_get_log_lsi(ns->ctrl, info->nsid, NVME_LOG_FDP_CONFIGS,
+ 0, 0, log, log_size, 0, info->endgid);
+ if (error) {
+ dev_warn(ns->ctrl->device,
+ "failed to read FDP config log: 0x%x\n", error);
+ goto out_free_log;
+ }
+
+ if (le32_to_cpu(log->size) < log_size) {
+ dev_warn(ns->ctrl->device, "FDP log too small: %d vs %zd\n",
+ le32_to_cpu(log->size), log_size);
+ error = -EINVAL;
+ goto out_free_log;
+ }
+
+ configs = (struct nvme_fdp_config_desc *)(log + 1);
+ if (le32_to_cpu(configs[result.fdpcidx].nrg) > 1) {
+ dev_warn(ns->ctrl->device, "FDP NRG > 1 not supported\n");
+ return -EINVAL;
+ }
+ ns->head->runs = le64_to_cpu(configs[result.fdpcidx].runs);
+
+out_free_log:
+ kfree(log);
+ return error;
+}
+
+static int nvme_fetch_fdp_plids(struct nvme_ns *ns, u32 nsid)
+{
+ struct nvme_ns_head *head = ns->head;
+ struct nvme_fdp_ruh_status *ruhs;
+ const unsigned int max_nr_plids = S8_MAX - 1;
+ size_t size = struct_size(ruhs, ruhsd, max_nr_plids);
+ struct nvme_command c = {
+ .imr.opcode = nvme_cmd_io_mgmt_recv,
+ .imr.nsid = cpu_to_le32(nsid),
+ .imr.mo = NVME_IO_MGMT_RECV_MO_RUHS,
+ .imr.numd = cpu_to_le32(nvme_bytes_to_numd(size)),
+ };
+ int ret, i;
+
+ ruhs = kzalloc(size, GFP_KERNEL);
+ if (!ruhs)
+ return -ENOMEM;
+
+ ret = nvme_submit_sync_cmd(ns->queue, &c, ruhs, size);
+ if (ret) {
+ dev_warn(ns->ctrl->device,
+ "failed to read FDP reclaim unit handles: 0x%x\n", ret);
+ goto out;
+ }
+
+ ns->head->nr_plids = le16_to_cpu(ruhs->nruhsd);
+ if (!ns->head->nr_plids)
+ goto out;
+
+ if (ns->head->nr_plids > max_nr_plids) {
+ dev_info(ns->ctrl->device,
+ "capping max write streams from %d to %d\n",
+ ns->head->nr_plids, max_nr_plids);
+ ns->head->nr_plids = max_nr_plids;
+ }
+
+ head->plids = kcalloc(ns->head->nr_plids, sizeof(head->plids),
+ GFP_KERNEL);
+ if (!head->plids) {
+ ret = -ENOMEM;
+ goto out;
+ }
+
+ for (i = 0; i < ns->head->nr_plids; i++)
+ head->plids[i] = le16_to_cpu(ruhs->ruhsd[i].pid);
+
+out:
+ kfree(ruhs);
+ return ret;
+}
+
static int nvme_update_ns_info_block(struct nvme_ns *ns,
struct nvme_ns_info *info)
{
@@ -2178,6 +2289,18 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
goto out;
}
+ if (!(ns->ctrl->ctratt & NVME_CTRL_ATTR_FDPS)) {
+ ns->head->nr_plids = 0;
+ kfree(ns->head->plids);
+ ns->head->plids = NULL;
+ } else if (!ns->head->plids) {
+ ret = nvme_read_fdp_config(ns, info);
+ if (!ret)
+ ret = nvme_fetch_fdp_plids(ns, info->nsid);
+ if (ret < 0)
+ goto out;
+ }
+
blk_mq_freeze_queue(ns->disk->queue);
ns->head->lba_shift = id->lbaf[lbaf].ds;
ns->head->nuse = le64_to_cpu(id->nuse);
@@ -2211,6 +2334,10 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,
if (!nvme_init_integrity(ns->head, &lim, info))
capacity = 0;
+ lim.max_write_streams = ns->head->nr_plids;
+ if (lim.max_write_streams)
+ lim.write_stream_granularity = ns->head->runs;
+
ret = queue_limits_commit_update(ns->disk->queue, &lim);
if (ret) {
blk_mq_unfreeze_queue(ns->disk->queue);
@@ -2313,6 +2440,8 @@ static int nvme_update_ns_info(struct nvme_ns *ns, struct nvme_ns_info *info)
ns->head->disk->flags |= GENHD_FL_HIDDEN;
else
nvme_init_integrity(ns->head, &lim, info);
+ lim.max_write_streams = ns_lim->max_write_streams;
+ lim.write_stream_granularity = ns_lim->write_stream_granularity;
ret = queue_limits_commit_update(ns->head->disk->queue, &lim);
set_capacity_and_notify(ns->head->disk, get_capacity(ns->disk));
diff --git a/drivers/nvme/host/nvme.h b/drivers/nvme/host/nvme.h
index 8cea8416b0d2..f10aa0cb6df5 100644
--- a/drivers/nvme/host/nvme.h
+++ b/drivers/nvme/host/nvme.h
@@ -493,6 +493,10 @@ struct nvme_ns_head {
struct gendisk *disk;
u16 endgid;
+ u16 nr_plids;
+ u16 *plids;
+ u64 runs;
+
#ifdef CONFIG_NVME_MULTIPATH
struct bio_list requeue_list;
spinlock_t requeue_lock;
--
2.45.2
next prev parent reply other threads:[~2024-11-19 12:17 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-11-19 12:16 support block layer write streams and FDP Christoph Hellwig
2024-11-19 12:16 ` [PATCH 01/15] fs: add write stream information to statx Christoph Hellwig
2024-11-19 12:16 ` [PATCH 02/15] fs: add a write stream field to the kiocb Christoph Hellwig
2024-11-19 12:16 ` [PATCH 03/15] io_uring: enable passing a per-io write stream Christoph Hellwig
2024-11-19 12:16 ` [PATCH 04/15] block: don't bother checking the data direction for merges Christoph Hellwig
2024-11-19 12:16 ` [PATCH 05/15] block: req->bio is always set in the merge code Christoph Hellwig
2024-11-19 12:16 ` [PATCH 06/15] block: add a bi_write_stream field Christoph Hellwig
2024-11-19 12:16 ` [PATCH 07/15] block: introduce max_write_streams queue limit Christoph Hellwig
2024-11-19 12:16 ` [PATCH 08/15] block: introduce a write_stream_granularity " Christoph Hellwig
2024-11-19 12:16 ` [PATCH 09/15] block: expose write streams for block device nodes Christoph Hellwig
2024-11-19 12:16 ` [PATCH 10/15] nvme: store the endurance group id in struct nvme_ns_head Christoph Hellwig
2024-11-19 12:16 ` [PATCH 11/15] nvme: pass a void pointer to nvme_get/set_features for the result Christoph Hellwig
2024-11-19 12:16 ` [PATCH 12/15] nvme: add a nvme_get_log_lsi helper Christoph Hellwig
2024-11-19 12:16 ` [PATCH 13/15] nvme.h: add FDP definitions Christoph Hellwig
2024-11-19 12:16 ` Christoph Hellwig [this message]
2024-11-19 18:17 ` [PATCH 14/15] nvme: enable FDP support Keith Busch
2024-11-19 18:24 ` Christoph Hellwig
2024-11-19 22:49 ` Keith Busch
2024-11-20 6:03 ` Christoph Hellwig
2024-11-19 12:16 ` [PATCH 15/15] RFC: block: allow write streams on partitions Christoph Hellwig
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
[email protected] \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox