Why not just name the cam-ata devices the same as the old names?

[Pan Tsu]

"Jason J. Hellenthal" <jhell at DataIX.net> writes:

> I do not know if this was summed up in a easy way by Jeremy's nice
> message below but in short a summary can be made here to clear that up.
>
> /dev/gptid/* /dev/gpt/*
> 	* These survive its raw partition being newfs'd
> 	* Are only created for disks that are partitioned
> 	  and contain a GPT table as can be seen with gpart
> 	  show
> 	* Operations on these or the raw partition will not remove them.

Not sure if we have support for labels based on disk serial number
similar to /dev/serno/* from DragonFlyBSD but

/dev/serno/*
	* no extra step to setup, e.g. `gpart create' or `newfs'
	* survive wiping entire disk, no metadata stored on-disk
	* available on every ata disk

As CAM_ATA transition goes such labels could be a better choice as
they'd lessen the pain when compat naming is gone, e.g. in 10-CURRENT.

> The best possible thing you could use here is a GPT scheme for the disks
> to remain consistent across newfs's. But relying on GPT for all disks
> will not always work in situations where the disk also involves a
> operating system that does not support booting off of a GPT disk, like
> all of Windows XP and then Win7 for non-Itanium based architectures. Yes
> Win7 last checked was said to only support booting GPT schemes on
> Itanium systems, so this leaves a lot of systems to only rely on
> /dev/ufs*/ labels or generic labels.