Re: [PATCH] nvme: make providing NGUID as UUID usage less scary

[AlanCui4080 <me@alancui.cc>]

On Monday, 11 May 2026 16:59，I wrote:
> For NVMe spec revision 1.3, there is no UUID yet.
> 

> If it's mandatory for userspace, why the nvme2n1 has not, if it's not
> mandatory, why should we create the UUID sysfd attr. then put NGUID in it?
> But for a NVMe 1.3 device there is both no UUID and no NGUID, while kernel
> is not reporting warnings at all?

Sorry, I checked the 1.3 specification and found that there is
NGUID and UUID in it.

The only difference between revision 1.3 and 1.4 is Revision 1.4
states that UUIDs should be reported when neither EUI64 nor NGUID is supported.

It seems that for an NVMe 1.4 device, its UUID is mandatory when EUI64/NGUID
is not supported. However, the kernel chooses to expose both the UUID and
NGUID of a device with an NGUID but no UUID to sysfs, then warns us every
time we read the device that an NGUID is being used to impersonate a UUID.

But for an NVMe 1.3 device, its UUID is optional when EUI64/NGUID is not
present. However, the kernel chooses not to expose either the UUID or NGUID
of a device with an NGUID but no UUID to sysfs.

This is incredibly strange.

Alan.