All of lore.kernel.org
 help / color / mirror / Atom feed
From: "Javier González" <javier@javigon.com>
To: Christoph Hellwig <hch@lst.de>
Cc: linux-nvme@lists.infradead.org, linux-block@vger.kernel.org,
	kbusch@kernel.org, sagi@grimberg.me, minwoo.im.dev@gmail.com
Subject: Re: [PATCH V6 1/2] nvme: enable char device per namespace
Date: Wed, 3 Mar 2021 11:02:12 +0100	[thread overview]
Message-ID: <20210303100212.e43jgjvuomgybmy2@mpHalley.localdomain> (raw)
In-Reply-To: <20210303091022.GA12784@lst.de>

On 03.03.2021 10:10, Christoph Hellwig wrote:
>On Mon, Mar 01, 2021 at 08:24:51PM +0100, javier@javigon.com wrote:
>> From: Javier González <javier.gonz@samsung.com>
>>
>> Create a char device per NVMe namespace. This char device is always
>> initialized, independently of whether the features implemented by the
>> device are supported by the kernel. User-space can therefore always
>> issue IOCTLs to the NVMe driver using the char device.
>>
>> The char device is presented as /dev/nvme-generic-XcYnZ. This naming
>> scheme follows the convention of the hidden device (nvmeXcYnZ). Support
>> for multipath will follow.
>
>So I'm a little worried about the "support for multipath will follow" as
>this has implications for the naming scheme, and our policy of how we
>allow access to a namespace.

I maintained the message to follow the original commit. In this
patchset multipath is implemented. I maintains the same naming scheme.

We can remove this comment as multipath is supported in the existing
series.

>
>Ignoring some of the deprecated historic mistakes I think the policy
>should be:
>
> - admin commands that often are controller specific should usually
>   go to a controller-specific device, the existing /dev/nvmeX
>   devices
> - I/O commands and admin command that do specific a nsid should go
>   through a per-namespace node that is multipath aware and not
>   controller specific

Sounds good.

The current implementation re-routes IOCTLs to the char device through
the existing bdev IOCTLs, so I believe we follow this policy already. We
basically default to current behavior.

And I assume that for legacy, namespace IOCTLs to the controller will
still be routed to the namespace (assuming a single namespace).

>Which also makes me wonder about patch 2 in the series that seems
>somewhat dangerous.   Can we clearly state the policy implemented?

Patch 2 is the one that exposes the existing logic for multipath. How do
you think we should do it instead?

WARNING: multiple messages have this Message-ID (diff)
From: "Javier González" <javier@javigon.com>
To: Christoph Hellwig <hch@lst.de>
Cc: linux-nvme@lists.infradead.org, linux-block@vger.kernel.org,
	kbusch@kernel.org, sagi@grimberg.me, minwoo.im.dev@gmail.com
Subject: Re: [PATCH V6 1/2] nvme: enable char device per namespace
Date: Wed, 3 Mar 2021 11:02:12 +0100	[thread overview]
Message-ID: <20210303100212.e43jgjvuomgybmy2@mpHalley.localdomain> (raw)
In-Reply-To: <20210303091022.GA12784@lst.de>

On 03.03.2021 10:10, Christoph Hellwig wrote:
>On Mon, Mar 01, 2021 at 08:24:51PM +0100, javier@javigon.com wrote:
>> From: Javier González <javier.gonz@samsung.com>
>>
>> Create a char device per NVMe namespace. This char device is always
>> initialized, independently of whether the features implemented by the
>> device are supported by the kernel. User-space can therefore always
>> issue IOCTLs to the NVMe driver using the char device.
>>
>> The char device is presented as /dev/nvme-generic-XcYnZ. This naming
>> scheme follows the convention of the hidden device (nvmeXcYnZ). Support
>> for multipath will follow.
>
>So I'm a little worried about the "support for multipath will follow" as
>this has implications for the naming scheme, and our policy of how we
>allow access to a namespace.

I maintained the message to follow the original commit. In this
patchset multipath is implemented. I maintains the same naming scheme.

We can remove this comment as multipath is supported in the existing
series.

>
>Ignoring some of the deprecated historic mistakes I think the policy
>should be:
>
> - admin commands that often are controller specific should usually
>   go to a controller-specific device, the existing /dev/nvmeX
>   devices
> - I/O commands and admin command that do specific a nsid should go
>   through a per-namespace node that is multipath aware and not
>   controller specific

Sounds good.

The current implementation re-routes IOCTLs to the char device through
the existing bdev IOCTLs, so I believe we follow this policy already. We
basically default to current behavior.

And I assume that for legacy, namespace IOCTLs to the controller will
still be routed to the namespace (assuming a single namespace).

>Which also makes me wonder about patch 2 in the series that seems
>somewhat dangerous.   Can we clearly state the policy implemented?

Patch 2 is the one that exposes the existing logic for multipath. How do
you think we should do it instead?

_______________________________________________
Linux-nvme mailing list
Linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme

  reply	other threads:[~2021-03-03 16:41 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-03-01 19:24 [PATCH V6 0/2] nvme: enable char device per namespace javier
2021-03-01 19:24 ` javier
2021-03-01 19:24 ` [PATCH V6 1/2] " javier
2021-03-01 19:24   ` javier
2021-03-03  9:10   ` Christoph Hellwig
2021-03-03  9:10     ` Christoph Hellwig
2021-03-03 10:02     ` Javier González [this message]
2021-03-03 10:02       ` Javier González
2021-03-09 11:31       ` Christoph Hellwig
2021-03-09 11:31         ` Christoph Hellwig
2021-03-09 12:42         ` Javier González
2021-03-09 12:42           ` Javier González
2021-03-09 15:05           ` Christoph Hellwig
2021-03-09 15:05             ` Christoph Hellwig
2021-03-09 21:18             ` Javier González
2021-03-09 21:18               ` Javier González
2021-03-24 12:29   ` Niklas Cassel
2021-03-24 12:29     ` Niklas Cassel
2021-03-25  2:09     ` Minwoo Im
2021-03-25  2:09       ` Minwoo Im
2021-03-25  8:26       ` hch
2021-03-25  8:26         ` hch
2021-03-25  8:39         ` Niklas Cassel
2021-03-25  8:39           ` Niklas Cassel
2021-03-25  9:34           ` Minwoo Im
2021-03-25  9:34             ` Minwoo Im
2021-03-25 12:25   ` Niklas Cassel
2021-03-25 12:25     ` Niklas Cassel
2021-03-25 12:33     ` Minwoo Im
2021-03-25 12:33       ` Minwoo Im
2021-03-25 15:14     ` Keith Busch
2021-03-25 15:14       ` Keith Busch
2021-03-01 19:24 ` [PATCH V6 2/2] nvme: allow open for nvme-generic char device javier
2021-03-01 19:24   ` javier

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=20210303100212.e43jgjvuomgybmy2@mpHalley.localdomain \
    --to=javier@javigon.com \
    --cc=hch@lst.de \
    --cc=kbusch@kernel.org \
    --cc=linux-block@vger.kernel.org \
    --cc=linux-nvme@lists.infradead.org \
    --cc=minwoo.im.dev@gmail.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.