Re: get rid of MEMORY_DISK_IS_ROOT?

["Toru Nishimura" <locore64%alkyltechnology.com@localhost>]

||| Why not getting rid of md_attach_hook and use the configured ram disk
||| for the first read of an unconfigured md(4)?
|| 
|| It would be much cleaner if we wouldn't instantiate a fixed
|| number of md(4) devices in the pseudo-device attach function,
|| but make it dynamic, like eg vnd(4).
|
| Yes, but it's a slightly different issue.
|
|| I'm a bit wary here because md seems to be used by embedded
|| ports in more creative ways.
|
| I'm with you there;  maybe someone can enlighten us about those creative
| ways.

(Yet another pulp fuctions and pipe dreams)

Instantiate (materialize) at md creation from particular backing store, say,
from kernel data[] section (INSTALL kernel we are using), from "special
storage" like a section of bare NAND and/or NOR, even from HTTP down
loading during the booting stage.  If we allow the md change snapshot'ed
away to backing store given any moment, that'd be a knockdown win over
initrd.

Toru Nishimura / ALKYL Technolog