From: Keith Busch <[email protected]>
To: Christoph Hellwig <[email protected]>
Cc: Jens Axboe <[email protected]>,
Christian Brauner <[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: Re: [PATCH 14/15] nvme: enable FDP support
Date: Tue, 19 Nov 2024 11:17:36 -0700 [thread overview]
Message-ID: <ZzzWQFyq0Sv7cuHb@kbusch-mbp> (raw)
In-Reply-To: <[email protected]>
On Tue, Nov 19, 2024 at 01:16:28PM +0100, Christoph Hellwig wrote:
> +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");
Why not support multiple reclaim groups?
> + return -EINVAL;
> + }
> + ns->head->runs = le64_to_cpu(configs[result.fdpcidx].runs);
The config descriptors are variable length, so you can't just index into
it. You have to read each index individually to get the next index's offset.
Something like:
struct nvme_fdp_config_desc *configs;
void *l;
int i;
...
l = log + 1;
for (i = 0; i < result.fdpcidx; i++) {
configs = l;
l += le16_to_cpu(configs->size);
}
ns->head->runs = le64_to_cpu(configs->runs);
next prev parent reply other threads:[~2024-11-19 18:17 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 ` [PATCH 14/15] nvme: enable FDP support Christoph Hellwig
2024-11-19 18:17 ` Keith Busch [this message]
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 \
--in-reply-to=ZzzWQFyq0Sv7cuHb@kbusch-mbp \
[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