Re: initrd & /linuxrc behavior

[H. Peter Anvin]

Followup to:  <[e-mail]>
By author:    =?iso-8859-1?Q?S=E9bastien=20C=F4t=E9?= <[e-mail]>
In newsgroup: linux.dev.kernel
>
> Hi,
> 
> When you boot linux using initrd, /linuxrc gets started if, and only if,
> the root= partion != the current root partition (/dev/ram0).  IMHO this
> is a bug, initrd's behavior should be consistent in each boot
> situation.  If there's a /linuxrc, run it each and every time !
> 
> Maybe there's a reason for such a behavior and in that case, I'd really
> like to know it!  In my case, I use /linuxrc to create /dev/console
> before /sbin/init is started so I can see the output messages.  So it's
> an interesting feature (I know I could just replace init by a script
> that would create /dev/console and then call the real init but since
> /linuxrc exist, why not use it ?)
> 

init=/linuxrc

Seriously, the whole linuxrc-versus-init system is pretty broken.  As
Werner pointed out in his OLS talk, pivot_root() or overmounting is so
much cleaner, without relying on hacks.

No need to have the kernel know about two names for init.

	-hpa
-- 
<[e-mail]> at work, <[e-mail]> in private!
"Unix gives you enough rope to shoot yourself in the foot."
http://www.zytor.com/~hpa/puzzle.txt