[OLPC-devel] Re: just tested new buildrom with linuxbios version 2376

[Jim Gettys]

On Tue, 2006-08-15 at 16:33 -0600, Jordan Crouse wrote:
> On 15/08/06 14:49 -0500, Richard Smith wrote:
> > >
> > >The graphical boot isn't persistant.  It goes away when the timeout
> > >expires or when a menu option is selected.  Do we seriously want a
> shell
> > >always available on the production ROM?  I say no.
> > >
> > 
> > I say yes.
> > 
> > You are going to blow the whole thing away when you kexec into the
> new
> > kernel so this would only be active if the boot failed.  Exactly
> when
> > you need a shell.
> 
> Ok, fine. You can come up with a 1000 strawman arguments for a shell,
> and I'm not equiped to refute them all. 

I agree with Richard on this one.

And I don't want to close the door on the smart kids; a shell should be
possible to get to, even if we want to make sure in a normal hands off
boot sequence you never see it.

> 
> Regardless, the menu and shell running together right now probably
> won't
> have the effect that you imagine it will.  To help Ron's problem, we
> can
> add new shell scripts to boot the NAND and USB, so one could type
> boot_nand
> or boot_usb at the shell to continue on with the boot.  That should do
> what
> Ron needs with minimal effort.
> 
That seems like exactly the right way to go.
                   - Jim

-- 
Jim Gettys
One Laptop Per Child