public inbox for linux-nvme@lists.infradead.org
 help / color / mirror / Atom feed
From: Damien Le Moal <dlemoal@kernel.org>
To: Mike Christie <michael.christie@oracle.com>,
	chaitanyak@nvidia.com, kbusch@kernel.org, hch@lst.de,
	sagi@grimberg.me, joao.m.martins@oracle.com,
	linux-nvme@lists.infradead.org, kvm@vger.kernel.org,
	kwankhede@nvidia.com, alex.williamson@redhat.com,
	mlevitsk@redhat.com
Subject: Re: [PATCH RFC 03/11] nvmet: Add nvmet_fabrics_ops flag to indicate SGLs not supported
Date: Thu, 13 Mar 2025 18:02:29 +0900	[thread overview]
Message-ID: <970e0d79-f338-4803-92c4-255156a8257e@kernel.org> (raw)
In-Reply-To: <20250313052222.178524-4-michael.christie@oracle.com>

On 3/13/25 14:18, Mike Christie wrote:
> The nvmet_mdev_pci driver does not initially support SGLs. In some
> prelim testing I don't think there will be a perf gain (the virt related
> interface may be the major bottleneck so I may not notice) so I wasn't
> sure if they will be required/needed. This adds a nvmet_fabrics_ops flag
> so we can tell nvmet core to tell the host we do not supports SGLS.

That is a major spec violation as NVMe fabrics mandates SGL support.
So at the very least, this needs a big fat warning comment explaining what this
is about.

> 
> Signed-off-by: Mike Christie <michael.christie@oracle.com>
> ---
>  drivers/nvme/target/admin-cmd.c | 13 +++++++------
>  drivers/nvme/target/nvmet.h     |  1 +
>  2 files changed, 8 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/nvme/target/admin-cmd.c b/drivers/nvme/target/admin-cmd.c
> index acc138bbf8f2..486ed6f7b717 100644
> --- a/drivers/nvme/target/admin-cmd.c
> +++ b/drivers/nvme/target/admin-cmd.c
> @@ -755,12 +755,13 @@ static void nvmet_execute_identify_ctrl(struct nvmet_req *req)
>  	id->awun = 0;
>  	id->awupf = 0;
>  
> -	/* we always support SGLs */
> -	id->sgls = cpu_to_le32(NVME_CTRL_SGLS_BYTE_ALIGNED);
> -	if (ctrl->ops->flags & NVMF_KEYED_SGLS)
> -		id->sgls |= cpu_to_le32(NVME_CTRL_SGLS_KSDBDS);
> -	if (req->port->inline_data_size)
> -		id->sgls |= cpu_to_le32(NVME_CTRL_SGLS_SAOS);
> +	if (!(ctrl->ops->flags & NVMF_SGLS_NOT_SUPP)) {
> +		id->sgls = cpu_to_le32(NVME_CTRL_SGLS_BYTE_ALIGNED);
> +		if (ctrl->ops->flags & NVMF_KEYED_SGLS)
> +			id->sgls |= cpu_to_le32(NVME_CTRL_SGLS_KSDBDS);
> +		if (req->port->inline_data_size)
> +			id->sgls |= cpu_to_le32(NVME_CTRL_SGLS_SAOS);
> +	}
>  
>  	strscpy(id->subnqn, ctrl->subsys->subsysnqn, sizeof(id->subnqn));
>  
> diff --git a/drivers/nvme/target/nvmet.h b/drivers/nvme/target/nvmet.h
> index fcf4f460dc9a..ec3d10eb316a 100644
> --- a/drivers/nvme/target/nvmet.h
> +++ b/drivers/nvme/target/nvmet.h
> @@ -403,6 +403,7 @@ struct nvmet_fabrics_ops {
>  	unsigned int flags;
>  #define NVMF_KEYED_SGLS			(1 << 0)
>  #define NVMF_METADATA_SUPPORTED		(1 << 1)
> +#define NVMF_SGLS_NOT_SUPP		(1 << 2)
>  	void (*queue_response)(struct nvmet_req *req);
>  	int (*add_port)(struct nvmet_port *port);
>  	void (*remove_port)(struct nvmet_port *port);


-- 
Damien Le Moal
Western Digital Research


  parent reply	other threads:[~2025-03-13  9:04 UTC|newest]

Thread overview: 29+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-03-13  5:18 [PATCH RFC 00/11] nvmet: Add NVMe target mdev/vfio driver Mike Christie
2025-03-13  5:18 ` [PATCH RFC 01/11] nvmet: Remove duplicate uuid_copy Mike Christie
2025-03-13  6:36   ` Christoph Hellwig
2025-03-13  8:59   ` Damien Le Moal
2025-03-13 17:20   ` Keith Busch
2025-03-13  5:18 ` [PATCH RFC 02/11] nvmet: Export nvmet_add_async_event and add definitions Mike Christie
2025-03-13  6:36   ` Christoph Hellwig
2025-03-13 17:50     ` Mike Christie
2025-03-13  5:18 ` [PATCH RFC 03/11] nvmet: Add nvmet_fabrics_ops flag to indicate SGLs not supported Mike Christie
2025-03-13  6:37   ` Christoph Hellwig
2025-03-13  9:02   ` Damien Le Moal [this message]
2025-03-13  9:13     ` Christoph Hellwig
2025-03-13  9:16       ` Damien Le Moal
2025-03-13 17:19         ` Mike Christie
2025-03-13  5:18 ` [PATCH RFC 04/11] nvmet: Add function to get nvmet_fabrics_ops from trtype Mike Christie
2025-03-13  9:03   ` Damien Le Moal
2025-03-13  5:18 ` [PATCH RFC 05/11] nvmet: Add function to print trtype Mike Christie
2025-03-13  5:18 ` [PATCH RFC 06/11] nvmet: Allow nvmet_alloc_ctrl users to specify the cntlid Mike Christie
2025-03-13  5:18 ` [PATCH RFC 07/11] nvmet: Add static controller support to configfs Mike Christie
2025-03-13  5:18 ` [PATCH RFC 08/11] nvmet: Add shadow doorbell support Mike Christie
2025-03-13  5:18 ` [PATCH RFC 09/11] nvmet: Add helpers to find and get static controllers Mike Christie
2025-03-13  5:18 ` [PATCH RFC 10/11] nvmet: Add addr fam and trtype for mdev pci driver Mike Christie
2025-03-13  6:42   ` Christoph Hellwig
2025-03-13 17:56     ` Mike Christie
2025-03-13  5:18 ` [PATCH RFC 11/11] nvmet: Add nvmet-mdev-pci driver Mike Christie
2025-03-13  5:32 ` [PATCH RFC 00/11] nvmet: Add NVMe target mdev/vfio driver Damien Le Moal
2025-03-13  6:47 ` Christoph Hellwig
2025-03-13 17:17   ` Mike Christie
2025-03-14  8:31     ` Hannes Reinecke

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=970e0d79-f338-4803-92c4-255156a8257e@kernel.org \
    --to=dlemoal@kernel.org \
    --cc=alex.williamson@redhat.com \
    --cc=chaitanyak@nvidia.com \
    --cc=hch@lst.de \
    --cc=joao.m.martins@oracle.com \
    --cc=kbusch@kernel.org \
    --cc=kvm@vger.kernel.org \
    --cc=kwankhede@nvidia.com \
    --cc=linux-nvme@lists.infradead.org \
    --cc=michael.christie@oracle.com \
    --cc=mlevitsk@redhat.com \
    --cc=sagi@grimberg.me \
    /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