Re: Where did the nvd devices go?

On Wed, Jun 21, 2023, 6:22 PM Kevin Oberman <rkoberman@gmail.com> wrote:

> Well, they are still around, but not functional. They are symlinks to nda
> devices, but the symlinks don't work well.
>

They work for filesystem access.


I have no idea when the symlink of nvd to nda happened, but after updating
> my system to main-n263630-ab3e6234ab6e, at least geom related commands no
> longer function using nvd0p?. I hit this when trying to use gpart and geli.
> gpart claims "gpart: No such geom: /dev/nvd0." geli responds (after
> entering a passphrase) "geli: Provider not found: "/dev/nvd0p7"My previous
> system version was main-n262908-c16e08e5f324.
>

These will work with nda. They should likely work with the nvd aliases, but
don't it seems (though you don't need the /dev/ for geom commands).

Was this just a failure of muscle memory, or was there persistent config
that failed?

Was this intentional? If so, why was this change made?  If not, could it be
> fixed? Since I usually use geli with the /dev/gpt devices, I didn't notice
> it right away, but it could certainly  surprise many users.
>

All these questions are answered in the UPDATING entry from when I switched
the default:

20230612:
	Belatedly switch the default nvme block device on x86 from nvd to nda.
	nda created nvd compatibility links by default, so this should be a
	nop. If this causes problems for your application, set hw.nvme.use_nvd=1
	in your loader.conf or add `options NVME_USE_NVD=1` to your kernel
	config. To disable the nvd compatibility aliases, add
	kern.cam.nda.nvd_compat=0 to loader.conf.  The default has been nda on
	all non-x86 platforms for some time now. If you need to fall back,
	please email imp@freebsd.org about why.


-- 
> Kevin Oberman, Part time kid herder and retired Network Engineer
> E-mail: rkoberman@gmail.com
> PGP Fingerprint: D03FB98AFA78E3B78C1694B318AB39EF1B055683
>