On Thu, Jun 9, 2022 at 8:21 AM <sartr...@email.ecagroup.com> wrote:
>
> Hello,
>
> I've noticed that on kirkstone branch, the call to objcopy 
> --add-gnu-debuglink, made by splitdebuginfo() function, on a signed kernel 
> module is removing the signature.
> An easy fix is to set  INHIBIT_PACKAGE_DEBUG_SPLIT to "1" in the kernel 
> recipe, this way kernel module signatures are not removed.
> Maybe I can submitt this simple fix to meta-security layer.
> But maybe it is better to handle it in oe-core as it was done in runstrip() 
> function in lib/oe/package.py, there is already a is_kernel_module_signed() 
> helper that detect if a kernel module is signed or not.
> What do you think?

It would be worth checking with Saul/Joshua to make sure that SBOM or
licensing isn't hooked into the debug split packages. I vaguely recall
Saul requiring the debug split, or at least debug info, to track down
the kernel source for a given module, so that it could be part of
SBOM.

Anything we did for this signed issue, would have be be sure to not break that.

Bruce

>
> Regards
> --
> Leo
> 
>


--
- Thou shalt not follow the NULL pointer, for chaos and madness await
thee at its end
- "Use the force Harry" - Gandalf, Star Trek II
-=-=-=-=-=-=-=-=-=-=-=-
Links: You receive all messages sent to this group.
View/Reply Online (#166780): 
https://lists.openembedded.org/g/openembedded-core/message/166780
Mute This Topic: https://lists.openembedded.org/mt/91643928/21656
Group Owner: openembedded-core+ow...@lists.openembedded.org
Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub 
[arch...@mail-archive.com]
-=-=-=-=-=-=-=-=-=-=-=-

Reply via email to