Re: Can't upgrade RAIDframe system -- won't boot from CD

[Alan Barrett <apb%cequrux.com@localhost>]

On Wed, 27 Feb 2013, Martin Husemann wrote:
> On Wed, Feb 27, 2013 at 12:32:28PM +0200, Alan Barrett wrote:
> > We already have MEMORY_DISK_IS_ROOT.  Perhaps BOOT_DISK_IS_ROOT would
> > be a reasonable name for a new option.
>
> When I filed PR 44774 (suggesting a /boot.cfg statement instead of a
> kernel option) a lengthy discussion starting here:
>
> http://mail-index.netbsd.org/tech-kern/2011/04/18/msg010381.html

Yes, a boothowto option would be better than a kernel build option.

I mostly agree with what kre said in 
<http://mail-index.netbsd.org/tech-kern/2009/11/11/msg006485.html>:

" It should always be
"        1. what the user explicitly asks for
"        2. what the kernel has built in
"        3. hacks like "[raidctl] -A root" (or perhaps similar things
"           for cgd etc)
"        4. where I think I came from
"
" in that order to find the root device.

except that a way to suppress some of the steps is needed.

Your suggestion in
<http://mail-index.netbsd.org/tech-kern/2011/04/18/msg010381.html> 
was for a boothowto option to suppress step 3.  I think that a 
boothowto option to suppress both 2 and 3 would be useful, leaving 
us with:

   1. what the user or boot loader explicitly asks for
      (via command line, multiboot, RB_MINIROOT, RB_ASKNAME,
      RB_STRING, or any other mechanism);
   2. the root device that the kernel has built in,
      provided this is not suppressed by a boothowto option;
   3. devices that explicitly label themselves for use as the root
      file system (via "raidctl -A root" or similar mechanisms),
      provided this is not suppressed by a boothowto option;
   4. the device from which the kernel was booted, if that
      is known;
   5. as a last resort, set RB_ASKNAME and try again.

I think that the same boothowto option could suppress both steps 
2 and 3, essentially making root on the boot device be the second 
highest priority, just below root on the device explicitly 
requested by the user.

We should also address the problem reported by Klaus Heinz in 
<http://mail-index.netbsd.org/tech-kern/2009/10/31/msg006410.html>, 
ensuring that a root device explicitly specified on the 
multiboot command line takes precedence over a device with 
"raidctl -A root".

--apb (Alan Barrett)