[prev in list] [next in list] [prev in thread] [next in thread] 

List:       busybox
Subject:    Re: init
From:       "hinko.kocevar () cetrtapot ! si" <hinko ! kocevar () cetrtapot ! si>
Date:       2005-11-30 9:30:20
Message-ID: 438D712C.9000507 () cetrtapot ! si
[Download RAW message or body]

Rob Landley wrote:
> On Tuesday 29 November 2005 09:44, hinko.kocevar@cetrtapot.si wrote:
> 
> 
>>My inittab:
>># This is run first except when booting in single-user mode.
>>#
>>#::sysinit:/etc/init.d/rcS
> 
> 
> That would be a commented out line, yes?

Sorry, I've just pasted last working config from the target and missed 
the comment there.

> 
> 
>>The problem is - when I try to run /etc/init.d/rcS (sysinit line) the
>>system executes rcS and then hangs there
> 
> 
> Does it?  Even though it's commented out?

Yes, it just sits there (when the rcS line is not commented of course 
and rcS gets executed).

> 
> Does your rcS have the executable bit set?

Yes. Otherwise init complained about permisions for rcS.

> 
> Could you give us something we might be able to reproduce in a finite amount 
> of time?

Well, I've copy pasted the rcS, inittab and linuxrc as I use it on 
target platform in my previous email. Now you can try with those files - 
first uncomment the rcS line in the inittab.
This is console output when rcS line is commented out:
...
/linuxrc: Mounting /mnt/flash... Using fallback suid method
done
/linuxrc: Starting init.
init started:  BusyBox v1.00 (2005.10.26-09:59+0000) multi-call binary
Starting pid 662, console /dev/ttyS0: '/bin/sh'
#

And this this is console output with rcS being called:
/linuxrc: Mounting /mnt/flash... Using fallback suid method
done
/linuxrc: Starting init.
init started:  BusyBox v1.00 (2005.10.26-09:59+0000) multi-call binary
Starting pid 663, console /dev/ttyS0: '/etc/init.d/rcS'

<HANG>

I'm starting to think this is cris architecture issue - as I've used 
busybox's init on ARM platform with no sweat.

> 
> 
> I've abused the heck out of init in the past, but these days there's initramfs 
> which is designed for this sort of thing.

Our current layout is one as introduced by the buildroot enviroment from 
axis. It uses cramfs for kernel and binaries, jffs2 for etc and tmpfs 
for var and tmp directories. We do not use any kernel modules here - as 
this is what initramfs is meant for, right - preloading kernel modules 
and later mounting the real rootfs?!

regards,
hinko
_______________________________________________
busybox mailing list
busybox@busybox.net
http://busybox.net/cgi-bin/mailman/listinfo/busybox
[prev in list] [next in list] [prev in thread] [next in thread]