Boot Images & "diskboot"

Hiya,

I’m not a big fan of plug & pray, automatically detecting new hardware or
partitions that mount themselves (I end up having to find them, umount 'em
then mount 'em where I want 'em). Therefore, I was wondering what would
happen if I eschewed the modern RtP build file with diskboot in it for a
more “traditional” if you will Nto 2.0 style build file where if I don’t
specify it, it don’t happen. I understand that I’ll miss out on having
diskboot fire off a shell which runs a script that does the enumeration then
runs /etc/config/rc.whatever which checks and runs a bunch of other stuff
and so on ad infinitum (or so it seems, sometimes :slight_smile:. If I want to keep it
simple where I name the disk driver and the partitions I want mounted and
spin off a script (yes, just one script, not 48 of 'em) to do the startup
as I see fit, am I going to break something that will be relied upon further
on down the line? Are there any global envirovars that I will need to set
up to keep things happy? Thoughts on why this might be a bad idea? Thanks
in advance for the input.

-Warren “I’ll do it my way” Peece

Warren Peece <Warren@nospam.com> wrote:

If I want to keep it
simple where I name the disk driver and the partitions I want mounted and
spin off a script (yes, just one script, not 48 of 'em) to do the startup
as I see fit, am I going to break something that will be relied upon further
on down the line? Are there any global envirovars that I will need to set
up to keep things happy? Thoughts on why this might be a bad idea? Thanks
in advance for the input.

The short answer is yes, you can do this. Remember that at it’s heart RtP
is just Neutrino, with all the core components able to work in a closed
embedded system. We always need to keep the ability for people to just
manually start what they need to start. Longer answer is you might run
into some glitches and might miss some things from the patches if you
make such a major fiddle with the boot up sequence.