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

Re: getting espeakup working with LVM encryption on boot using cryptsetup



Hi, What is wrong with a process from the initial ramdisk image running in the main system as me being totally blind I would have thought that this issue was taken care of when the speakup_soft module was put into the initial ramdisk image by the debian-installer?  If the espeakup process was just started on boot as early as possible and then not started on the running system would this cause many problems as from the users point of view we would have speech as early in the boot process as possible which I thought was the point?  Is there no guide for modifying the initial ramdisk image to at least cause speech to be started as early as possible or just modify it so the pass phrase was automatically entered at boot as yes normally this would not be an issue but I would at least like this as an option if I was away and needed to reboot this system remotely and yes I know methods exist to put a ssh server in the initial ramdisk image but not sure if that supports wireless networking?  Nick Gawronski

On 2/6/2024 10:16 AM, Sam Hartman wrote:
"Nick" == Nick Gawronski <nick@nickgawronski.com> writes:
     Nick> Hi, I used the option during the debian installation to do
     Nick> full encryption and the encryption unlocking is done I think
     Nick> in the initial ramdisk image and the speakup_soft module does
     Nick> exist in there but I do not get any speech prompting me for
     Nick> the pass phrase and so have to guess as to when exactly to
     Nick> enter it

I think the big issue here is that you do not want any processes
lingering from the initrd to the main system.
speakup_soft calls out to espeakup; speakup_soft is in the kernel, and
espeakup is a user land process.

And the issue is that you would need to kill off the espeakup from the
initrd and let it restart from the rootfs.
If you are a developer, you probably could create an initramfs hook to
do that, and submit a patch to the espeakup package.
You would also need to mess with setting sound volume etc in the
initramfs.
All that is possible, it just takes a fair bit of fiddly work.

I've never been particularly bothered by needing to wait/guess when to
type the password, so I'm not particularly motivated to do the work.
I would kind of like to have speech access on the console when things go
wrong at initramfs time, but again, that has never risen to be a high
enough priority that it was worth the time investment.



Reply to: