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

Re: Alternate approaches to signed module packaging



On Tue, 2016-06-07 at 19:06 +0100, Ben Hutchings wrote:
> On Sun, 2016-05-29 at 13:38 +0100, Ben Hutchings wrote:
> [...]
> > So I think I have to abandon my current approach and instead do one of:
> > 
> > 1. Attach module signatures at installation time, in a subdirectory.
> >    Change kmod to prefer this subdirectory (this is purely a
> >    configuration change).  It would also be possible to check during
> >    installation that signatures match the installed unsigned modules,
> >    and if not then abort and leave any older signed modules in place.
> > 
> > 2. Attach module signatures at package build time, making the
> >    linux-image-signed packages provide/conflict/replace the
> >    corresponding linux-image packages.  For architectures with
> >    signed modules, udebs would be built from linux-signed and not
> >    from linux.
> [...]
> 
> I'm now implementing the second approach above.

That's now done, in version 1.1, uploaded to unstable today.

[...]
> I think there are three steps left::
> 
> 1. Move udeb generation for configurations with module signing enabled
>    from linux to linux-signed.  (This is in progress.)

Now implemented on the benh/udebsig branches.

> 2. (Optional) Remove the '-signed' suffix from signed packages and add
>    a '-unsigned' suffix to unsigned linux-image packages built with
>    module signing enabled.  Adjust the Conflicts/Replaces/Provides
>    fields accordingly.

I intend to implement steps 1 and 2 in the next upload to experimental.

> 3. Change the signing script to use an HSM.

This is still TBD.

Ben.

-- 
Ben Hutchings
Any smoothly functioning technology is indistinguishable from a rigged
demo.

Attachment: signature.asc
Description: This is a digitally signed message part


Reply to: