Re: [PATCH RFC v3 08/12] hw/block/nvme: end-to-end data protection

[Klaus Jensen <its@irrelevant.dk>]

[-- Attachment #1: Type: text/plain, Size: 1445 bytes --]

On Feb 16 15:08, Keith Busch wrote:
> On Mon, Feb 15, 2021 at 12:02:36AM +0100, Klaus Jensen wrote:
> > From: Klaus Jensen <k.jensen@samsung.com>
> > 
> > Add support for namespaces formatted with protection information. The
> > type of end-to-end data protection (i.e. Type 1, Type 2 or Type 3) is
> > selected with the `pi` nvme-ns device parameter. If the number of
> > metadata bytes is larger than 8, the `pil` nvme-ns device parameter may
> > be used to control the location of the 8-byte DIF tuple. The default
> > `pil` value of '0', causes the DIF tuple to be transferred as the last
> > 8 bytes of the metadata. Set to 1 to store this in the first eight bytes
> > instead.
> 
> 
> This file is getting quite large. I think this feature can have the bulk
> of the implementation in a separate file. For ex, nvme-dif.c.

Yes, makes sense to split it off. I think moving[1] the device to hw/nvme
first would be good.

> But like the linux implementation this is based on, it isn't really
> nvme specific, so even better if t10 dif is implemented in a generic
> location with an API for nvme and others.

That is true, but in the absence of any interest from other subsystems
to implement this, I think we can keep it local for now? I keep a pretty
close eye on qemu-block, so if other subsystems should care about this,
I promise that I will pitch in :)


  [1]: <20210209110826.585987-1-its@irrelevant.dk>

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 488 bytes --]