From: Christoph Hellwig <hch@lst.de>
To: Max Gurtovoy <maxg@mellanox.com>
Cc: axboe@kernel.dk, jsmart2021@gmail.com, sagi@grimberg.me,
martin.petersen@oracle.com, shlomin@mellanox.com,
linux-rdma@vger.kernel.org, israelr@mellanox.com,
vladimirk@mellanox.com, linux-nvme@lists.infradead.org,
idanb@mellanox.com, jgg@mellanox.com, oren@mellanox.com,
kbusch@kernel.org, hch@lst.de
Subject: Re: [PATCH 01/17] nvme: introduce namespace features flag
Date: Tue, 21 Apr 2020 13:59:12 +0200 [thread overview]
Message-ID: <20200421115912.GB26432@lst.de> (raw)
In-Reply-To: <20200327171545.98970-3-maxg@mellanox.com>
On Fri, Mar 27, 2020 at 08:15:29PM +0300, Max Gurtovoy wrote:
> From: Israel Rukshin <israelr@mellanox.com>
>
> Centralize all the metadata checks to one place and make the code more
> readable. Introduce a new enum nvme_ns_features for that matter.
> The features flag description:
> - NVME_NS_EXT_LBAS - NVMe namespace supports extended LBA format.
> - NVME_NS_MD_HOST_SUPPORTED - NVMe namespace supports getting metadata
> from host's block layer.
> - NVME_NS_MD_CTRL_SUPPORTED - NVMe namespace supports metadata actions
> by the controller (generate/strip).
So whole I like the ->features flag, the defintion of these two
metadata related features really confuses me.
Here are my vague ideas to improve the situation:
> -static inline bool nvme_ns_has_pi(struct nvme_ns *ns)
> -{
> - return ns->pi_type && ns->ms == sizeof(struct t10_pi_tuple);
> -}
This function I think is generally useful, I'd rather keep iţ, document
it with a comment and remove the new NVME_NS_MD_CTRL_SUPPORTED
flag.
> - if (ns->ms && !ns->ext &&
> - (ns->ctrl->ops->flags & NVME_F_METADATA_SUPPORTED))
> + if (ns->features & NVME_NS_MD_HOST_SUPPORTED)
> nvme_init_integrity(disk, ns->ms, ns->pi_type);
> - if ((ns->ms && !nvme_ns_has_pi(ns) && !blk_get_integrity(disk)) ||
> - ns->lba_shift > PAGE_SHIFT)
> +
> + if ((ns->ms && !(ns->features & NVME_NS_MD_CTRL_SUPPORTED) &&
> + !(ns->features & NVME_NS_MD_HOST_SUPPORTED) &&
> + !blk_get_integrity(disk)) || ns->lba_shift > PAGE_SHIFT)
> capacity = 0;
I find this very confusing. Can we do something like:
/*
* The block layer can't support LBA sizes larger than the page size
* yet, so catch this early and don't allow block I/O.
*/
if (ns->lba_shift > PAGE_SHIFT)
capacity = 0;
/*
* Register a metadata profile for PI, or the plain non-integrity NVMe
* metadata masquerading as Typ 0 if supported, otherwise reject block
* I/O to namespaces with metadata except when the namespace supports
* PI, as it can strip/insert in that case.
*/
if (ns->ms) {
if (IS_ENABLED(CONFIG_BLK_DEV_INTEGRITY) &&
(ns->features & NVME_NS_MD_HOST_SUPPORTED))
nvme_init_integrity(disk, ns->ms, ns->pi_type);
else if (!nvme_ns_has_pi(ns))
capacity = 0;
}
> + if (ns->ms) {
> + if (id->flbas & NVME_NS_FLBAS_META_EXT)
> + ns->features |= NVME_NS_EXT_LBAS;
> +
> + /*
> + * For PCI, Extended logical block will be generated by the
> + * controller.
> + */
> + if (ns->ctrl->ops->flags & NVME_F_METADATA_SUPPORTED) {
> + if (!(ns->features & NVME_NS_EXT_LBAS))
> + ns->features |= NVME_NS_MD_HOST_SUPPORTED;
> + }
Maybe:
> + if (ns->ms) {
> + if (id->flbas & NVME_NS_FLBAS_META_EXT)
> + ns->features |= NVME_NS_EXT_LBAS;
> +
> + /*
> + * For PCI, Extended logical block will be generated by the
> + * controller.
> + */
> + if (ns->ctrl->ops->flags & NVME_F_METADATA_SUPPORTED) {
> + if (!(ns->features & NVME_NS_EXT_LBAS))
> + ns->features |= NVME_NS_MD_HOST_SUPPORTED;
> + }
This looks a little strange now, but I guess it will make more sense
with the fabrics addition. I'll take another look later in the series.
> +enum nvme_ns_features {
> + NVME_NS_EXT_LBAS = 1 << 0,
> + NVME_NS_MD_HOST_SUPPORTED = 1 << 1,
> + NVME_NS_MD_CTRL_SUPPORTED = 1 << 2,
> +};
Please document the meaning of each flag. I also suspect that just
moving ext to a flag first and than adding the NVME_NS_MD_HOST_SUPPORTED
bit might make more sense. I'd also rename NVME_NS_MD_HOST_SUPPORTED
to NVME_NS_METADATA_SUPPORTED.
_______________________________________________
linux-nvme mailing list
linux-nvme@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-nvme
next prev parent reply other threads:[~2020-04-21 11:59 UTC|newest]
Thread overview: 59+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-27 17:15 [PATCH 00/17 V5] nvme-rdma/nvmet-rdma: Add metadata/T10-PI support Max Gurtovoy
2020-03-27 17:15 ` [PATCH 1/1] nvme-cli/fabrics: Add pi_enable param to connect cmd Max Gurtovoy
2020-03-27 17:15 ` [PATCH 01/17] nvme: introduce namespace features flag Max Gurtovoy
2020-04-21 11:59 ` Christoph Hellwig [this message]
2020-04-21 15:53 ` James Smart
2020-04-21 18:11 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 02/17] nvme: Add has_md field to the nvme_req structure Max Gurtovoy
2020-04-21 11:59 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 03/17] nvme: Enforce extended LBA format for fabrics metadata Max Gurtovoy
2020-04-21 12:08 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 04/17] nvme: introduce max_integrity_segments ctrl attribute Max Gurtovoy
2020-04-21 12:09 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 05/17] nvme-fabrics: Allow user enabling metadata/T10-PI support Max Gurtovoy
2020-04-21 12:12 ` Christoph Hellwig
2020-04-21 15:17 ` Christoph Hellwig
2020-04-22 22:07 ` Max Gurtovoy
2020-04-22 22:24 ` James Smart
2020-04-22 22:39 ` Max Gurtovoy
2020-04-23 5:54 ` Christoph Hellwig
2020-04-23 7:30 ` Max Gurtovoy
2020-04-24 7:06 ` Christoph Hellwig
2020-04-26 9:48 ` Max Gurtovoy
2020-04-27 6:04 ` Christoph Hellwig
2020-04-27 13:52 ` Max Gurtovoy
2020-04-27 13:54 ` Christoph Hellwig
2020-04-28 9:18 ` Max Gurtovoy
2020-04-23 5:53 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 06/17] nvme: introduce NVME_INLINE_MD_SG_CNT Max Gurtovoy
2020-04-21 12:12 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 07/17] nvme-rdma: Introduce nvme_rdma_sgl structure Max Gurtovoy
2020-04-21 12:13 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 08/17] nvme-rdma: add metadata/T10-PI support Max Gurtovoy
2020-04-21 12:20 ` Christoph Hellwig
2020-04-23 9:22 ` Max Gurtovoy
2020-04-24 7:09 ` Christoph Hellwig
2020-04-26 10:04 ` Max Gurtovoy
2020-03-27 17:15 ` [PATCH 09/17] nvmet: prepare metadata request Max Gurtovoy
2020-04-21 15:21 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 10/17] nvmet: add metadata characteristics for a namespace Max Gurtovoy
2020-04-21 15:23 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 11/17] nvmet: Rename nvmet_rw_len to nvmet_rw_data_len Max Gurtovoy
2020-03-27 17:15 ` [PATCH 12/17] nvmet: Rename nvmet_check_data_len to nvmet_check_transfer_len Max Gurtovoy
2020-03-27 17:15 ` [PATCH 13/17] nvme: Add Metadata Capabilities enumerations Max Gurtovoy
2020-04-21 15:24 ` Christoph Hellwig
2020-04-23 12:09 ` Max Gurtovoy
2020-04-24 7:12 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 14/17] nvmet: Add metadata/T10-PI support Max Gurtovoy
2020-04-21 15:30 ` Christoph Hellwig
2020-04-23 12:39 ` Max Gurtovoy
2020-04-24 7:14 ` Christoph Hellwig
2020-04-26 10:50 ` Max Gurtovoy
2020-04-27 6:06 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 15/17] nvmet: Add metadata support for block devices Max Gurtovoy
2020-04-21 15:33 ` Christoph Hellwig
2020-04-23 17:25 ` Max Gurtovoy
2020-04-24 7:54 ` Christoph Hellwig
2020-03-27 17:15 ` [PATCH 16/17] RDMA/rw: Expose maximal page list for a device per 1 MR Max Gurtovoy
2020-03-27 17:15 ` [PATCH 17/17] nvmet-rdma: Add metadata/T10-PI support Max Gurtovoy
2020-04-21 15:37 ` 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=20200421115912.GB26432@lst.de \
--to=hch@lst.de \
--cc=axboe@kernel.dk \
--cc=idanb@mellanox.com \
--cc=israelr@mellanox.com \
--cc=jgg@mellanox.com \
--cc=jsmart2021@gmail.com \
--cc=kbusch@kernel.org \
--cc=linux-nvme@lists.infradead.org \
--cc=linux-rdma@vger.kernel.org \
--cc=martin.petersen@oracle.com \
--cc=maxg@mellanox.com \
--cc=oren@mellanox.com \
--cc=sagi@grimberg.me \
--cc=shlomin@mellanox.com \
--cc=vladimirk@mellanox.com \
/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;
as well as URLs for NNTP newsgroup(s).