On Thu, Oct 02, 2025 at 02:55:07PM -0700, Dave Hansen wrote:

On 10/2/25 14:31, Daniel Walker (danielwa) wrote:
...

quoted

quoted

BTW, your series looks like a *really* good idea. Please don't let it
die. But you might want to trim it down a bit. I'd probably remove the
tests and the 'insert-sys-cert' changes to make it more approachable to
folks.

Since x86 is asking for it I think it would trim it down to just
what is needed for x86. If I don't trim down the architectures it
ropes in too many people anyway.

That's not a bad idea. Or, even if you can pick two amenable
architectures to start with it will make it really obvious that this is
useful. Two architectures means a *lot*, IMNHO. Two is a billion times
better than one.

ARM64 has also request this series in the past, but I don't know what their
current code looks like since my last submissions.

quoted

The biggest issue is that libstub would need to be modified, but
I've never had any luck getting the libstub maintainer to review
anything. I suspect he would ignore private email too, particularly
from people he's doesn't know.

Are you talking about Ard?

	EXTENSIBLE FIRMWARE INTERFACE (EFI)
	M:      Ard Biesheuvel [off-list ref]
	L:      linux-efi@vger.kernel.org
	S:      Maintained
	...
	F:      drivers/firmware/efi/
	F:      include/linux/efi*.h

Yes I think so.

He's a pretty nice guy and has been active in this thread, so I'm kinda
surprised to hear you're having a hard time there. I'd just try asking
nicely. I'm pretty sure he's "ardb" on the usual IRC networks. IRC is a
great alternative when you're having problems getting your emails seen
in the normal email flood.

That was the first response from him over many submissions. Which channels are
popular on IRC ?

BTW, reading the changelog for libstub, it wasn't clear to me that
changes there were _required_ for the series to go forward. For
instance, is the x86 patch useful without libstub changes?

I think it is required because x86 can be compiled with libstub and the 
changes take effect for the whole kernel when compiled for that architecture.
libstub uses the Kconfig options directly which I modify in the series, so there
would need to be some compatibility changes.

Daniel