Re: [PATCH v4 12/16] libnvdimm, nfit: enable support for volatile ranges

[Dan Williams]

On Thu, Jun 29, 2017 at 3:35 PM, Linda Knippers <linda.knippers@xxxxxxx> wrote:
> On 06/29/2017 06:28 PM, Dan Williams wrote:
>> On Thu, Jun 29, 2017 at 3:12 PM, Linda Knippers <linda.knippers@xxxxxxx> wrote:
>> [..]
>>>> The /dev/pmem
>>>> device name just tells you that your block device is hosted by a
>>>> driver that knows how to handle persistent memory constraints, but any
>>>> other details about the nature of the address range need to come from
>>>> other sources of information, and potentially information sources that
>>>> the kernel does not know about.
>>>
>>>
>>> I'm asking about the other source of information in this specific case
>>> where we're exposing pmem devices that will never ever be persistent.
>>> Before we add these devices, I think we should be able to tell the user
>>> how they can know the properties of the underlying device.
>>
>> The only way I can think to indicate this is with a platform + device
>> whitelist in a tool like ndctl. Where the tool says "yes, these
>> xyz-vendor DIMMs on this abc-vendor platform with this 123-version
>> BIOS" is a known good persistent configuration.
>
> Doesn't the kernel know that something will never ever be persistent
> because the NFIT type says NFIT_SPA_VDISK, NFIT_SPA_VCD, or NFIT_SPA_VOLATILE?
> That's the case I'm asking about here.   In this patch, you're adding support
> for creating /dev/pmem devices for those address ranges.  My question is
> how the admin/user knows that those devices will never ever be persistent.

The parent region of the namespace will have a 'volatile' type:

# cat /sys/bus/nd/devices/region0/devtype
nd_volatile

> I don't think we need ndctl to know which vendors' hardware/firmware
> actually works as advertised.

Certification stickers is what I was thinking, but I was missing your
direction question.