Re: Bug#290952: Should be run before cryptdisks
retitle 290952 usb keyboard driver should be loaded earlier
reassign 290952 initrd-tools
tag 290952 -wontfix
thanks
On Tue, Jan 18, 2005 at 10:50:27AM +0100, Marco d'Itri wrote:
> severity 290952 wishlist
> tag 290952 wontfix
> thanks
>
> On Jan 17, Emidio Planamente <eplanamente@gmx.ch> wrote:
>
> > I have a crypted partition on my laptop. This has been created with
> > dm-crypt.
> > During the boot process, the /etc/init.d/cryptdisks script is called
> > (in /etc/rcS.d/S28cryptdisks) and I'm asked for the passphrase, but I
> > cannot enter it because my keyboard doesn't seem to react at this point.
>
> > I have discovered that my keyboard starts working only later, after
> > hotplug is called. Strange, because I thought to have a ps2 keyboard,
> > but in fact I'm not sure about this. Anyway, calling hotplug
> You probably have an USB keyboard.
>
> I know that it's annoying, but your configuration is currently not
> supported by hotplug, sorry.
> Starting hotplug before /usr has been mounted is not possible for some
> devices, so I'm not sure if this will be changed even in the future.
> As a workaround you can add to /etc/modules the drivers you need to
> operate your keyboard and disks. Probably this should be documented in
> the cryptsetup package.
Yes, this is also a problem if you have an initrd which mounts cryptoroot.
A better solution than /etc/modules is put a symlink in
/lib/modules/<kernelversion>/boot to usbhib.ko. This way the module gets
loaded during the initial ram disk phase (you will have to re-run mkinitrd).
I have moved this bug report b/c Marco is right that this needs to be
documented/fixed elsewhere. Actually, I think this should be done in
mkinitrd. Even if you aren't using cryptsetup, when the initial ram disk
drops you into interactive mode (on failure), you need keyboard control.
Right now, usb keyboard users just get screwed. =)
I propose: always add usbhib to the list of modules mkinitrd includes.
I'm not certain whether this works w/o the usb host driver..?
--
Wesley W. Terpstra
Reply to: