[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: U-Boot, d-i and selecting the console device

On Sun, Jun 21, 2015 at 09:17:12AM -0700, Vagrant Cascadian wrote:

> On 2015-06-21, Karsten Merker wrote:
> > If we don't pass a console variable to the kernel, d-i starts
> > on the console device defined by the platform's stdout-path
> > property, i.e. usually on a serial console - again bad luck
> > for people having only a framebuffer console.
> >
> > If we set the console variable to tty0, d-i starts on the
> > framebuffer console if it is available, or on a dummy console,
> > if no framebuffer console is available. The dummy console is
> > invisible, i.e. a serial-console-only user doesn't get any
> > output, so bad luck for the people using a serial console
> > in this case.
> What about adding support in debian-installer to dsisplay on both the
> serial console and framebuffer?

My memory on that topic is unfortunately a bit shady, but IIRC
this was proposed some time ago already and was dismissed due to
technical problems, but I cannot remember any details.
Kibi, do you by chance have some information in this regard?

> > AFAICS there is no way to determine the "active" console in
> > u-boot and set either the u-boot console variable or the
> > stdout-path property accordingly.  Even if there was such a
> > mechanism, the question remains: what is the "active" console? 
> > With our d-i SD-card images, the system autoboots directly into
> > d-i without any user interaction required, to there is no way for
> > u-boot to determine which console the user is intending to use.
> Failing automatic selection, we could use u-boot's extlinux.conf support
> to present the user with a menu to select between unspecified console,
> console=${console}, and console=tty0...  Although I'm not sure if
> variables are suported... hrm. Not a pretty user-interface, either.

At least the u-boot extlinux.conf docs and the bootloader spec
which it refers to do not mention variable support.  I'll try to
test whether it works nonetheless when I find time in the next

Another problem which we still have with regard to extlinux.conf
is the baudrate mismatch issue on i.mx6.  Our current workaround
for that requires a real boot script and does not work with
extlinux.conf, so unless that gets fixed upstream, we cannot
change from boot.scr to extlinux.conf or we lose proper i.mx6
serial console support.

> Other options that come to mind are adding the ability to read a file on
> the boot media and use that to determine which console to set... such as
> uEnv.txt implemented in some boards. Or even just checking for the
> presence of certain files on the boot media...

That is of course possible, but requires manual intervention by
the user and is probably more complicated than running "setenv
console tty0" with the current setup.

Gem. Par. 28 Abs. 4 Bundesdatenschutzgesetz widerspreche ich der Nutzung
sowie der Weitergabe meiner personenbezogenen Daten für Zwecke der
Werbung sowie der Markt- oder Meinungsforschung.

Reply to: