[busybox] No such device tty2, tty3, tty4 on embedded system

[Denys Vlasenko]

On Wednesday 21 May 2008 18:59, Roberto A. Foglietta wrote:
> >>  Sorry, what appear strangely arbitrary? and change what? I always use
> >> 1.0.x and 1.2.x version of busybox and this is the first time I am
> >> using "new" busybox. I think is natural having to familiarize with new
> >> bb and its options.
> >
> > I mean - this "log to vc #5" behavior seems to be arbitrary.
> > Why vc #5 and not #6? Why vc and not serial line? and so on.
> >
> > You just got bitten by this arbitrariness. Good moment to think
> > what can be a better behavior.
> >
> > Would you like init to just log to console - say, fd #2 which
> > is passed to init by the kernel?
> 
>  IMHO passing arbitrary arguments to kernel in not standard board
> could not an easy thing.

What arguments?

>  I would prefer choose by:
> 
>  1) config in which console logs fall in case syslog is not initialized.
>  2) previous one choose could be overwritten by a special inittab line
>  3) previous twos could be overwritten by kernel line arguments with
> syslog=ttyN or syslog=ttySn
> 
>  and fd #2 kernel argument seem too less auto-explicative.

Well, it's sort of standard behavior of unix programs to use standard
output/standard error fds for logging. init also has such descriptors open.
They are open to console by the kernel before init is executed.

Added bonus: almost everybody have console configured so that output
can me made visible for humans (this is needed to debug kernel
boot problems), so automatically init's messages will also be visible.

The downside is that these messages may intermigle with getty/shell/etc
output if console is also used for login.
--
vda