trouble with init in EABI toolchain

Brian Austin brian.austin at cirrus.com
Wed May 9 12:58:53 UTC 2007


On Wed, 2007-05-09 at 20:46 +0200, Denis Vlasenko wrote:
> On Wednesday 09 May 2007 14:15, Brian Austin wrote:
> > On Wed, 2007-05-09 at 11:03 -0700, Larry Brigman wrote:
> > > On 5/9/07, Brian Austin <brian.austin at cirrus.com> wrote:
> > > > On Wed, 2007-05-09 at 12:30 -0400, Mike Frysinger wrote:
> > > > > On Wednesday 09 May 2007, Brian Austin wrote:
> > > > > > anyone have issues with init in EABI busybox?
> > > > >
> > > > > most likely an EABI issue rather than a busybox issue ... as such, you should
> > > > > ask on the arm mailing lists as they prob have a lot more experience in
> > > > > common user gotchas
> > > > > -mike
> > > >
> > > > I can see a kernel panic attempting to kill init when I set
> > > > init=/bin/busybox on the kernel command line.
> > > 
> > > Because busybox will exit unless you give it a command like:
> > > /bin/busybox sh
> > > 
> > 
> > correct, but when I do init=/bin/busybox sh nothing happens.
> 
> It means "init=/bin/busybox" "sh". sh is not going into value
> of init, it becomes another kernel prarmeter!
> 
> make a symlink /bin/sh -> /bin/busybox and boot with init=/bin/sh
> --
> vda
> 

I can take out all refs to init in init/main.c in the kernel and just
have it try /bin/sh, it still just sits after freeing init...





More information about the busybox mailing list