Bug#762634: initramfs-tools: [armhf] mounting rootfs on USB disk fails / some USB host controller drivers missing in initramfs

[Karsten Merker <merker@debian.org>]

On Thu, Sep 25, 2014 at 01:31:16PM +0200, Michael Prokop wrote:
> * Cyril Brulebois [Thu Sep 25, 2014 at 12:32:22PM +0200]:
> > Michael Prokop <mika@debian.org> (2014-09-25):
> > > * Karsten Merker [Thu Sep 25, 2014 at 08:15:48AM +0200]:
 
> > > > Inside the /target chroot created by d-i, update-initramfs is by default
> > > > configured to run with "MODULES=dep" and gives only the following
> > > > modules on my armhf/sunxi test system:
> 
> > > Hmpf, I so much am not a fan of this default "MODULES=dep" behaviour
> > > of d-i...
> 
> > Can you please clarify? I see this in base-installer:
> > |                 if db_get base-installer/initramfs-tools/driver-policy && \
> > |                    [ -z "$RET" ]; then
> > |                         # Get default for architecture
                                            ^^^^^^^^^^^^^^^^
[snip] 

> > Granted, no coffee yet. But I seem to recall that people having issues
> > with MODULES="dep" are those who actually selected it manually (e.g.
> > through expert install).
> 
> I'm just not a friend of "MODULES=dep" as a default behaviour, good
> to know that d-i uses a sane default here. :) Thanks for verifying!
> 
> > My local test confirms that a basic installation (netboot-gtk image on
> > amd64 using udebs from sid) leads to MODULES="most" in /target. I didn't
> > go further than the popcon prompt though.

Hello,

I have just run another test on armhf with today's daily d-i
build (dated 25-Sep-2014 05:17) and default debconf priority
(at which no debconf prompt regarding initramfs-tools gets
displayed). This results in:

  ~ # cat /target/etc/initramfs-tools/conf.d/driver-policy
  # Driver inclusion policy selected during installation
  # Note: this setting overrides the value set in the file
  # /etc/initramfs-tools/initramfs.conf
  MODULES=dep

and the short module list in the initramfs confirms that it was
really built with MODULES=dep. The comment marked above suggests
that this is an architecture-dependent default, so we could
possibly both be right ;-). If yes, we will have to find a
solution regarding the inclusion of the phy-sun4i-usb module
when building the initramfs with MODULES=dep, though.

Regards,
Karsten
-- 
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.