1 |
On Mon, Jun 27, 2022 at 3:42 PM Georgy Yakovlev <gyakovlev@g.o> wrote: |
2 |
> |
3 |
> On Mon, 2022-06-27 at 14:56 -0400, Mike Gilbert wrote: |
4 |
> > On Mon, Jun 27, 2022 at 2:35 PM Kenton Groombridge |
5 |
> > <concord@g.o> wrote: |
6 |
> > > > so looks like we need to combine both methods and do the |
7 |
> > > > following: |
8 |
> > > > - if signing requested without compression - sign in |
9 |
> > > > pkg_preinst. |
10 |
> > > > - if signing requested with compression - sign in src_install |
11 |
> > > > |
12 |
> > > |
13 |
> > > Why can't we do both in pkg_preinst? I am thinking it would be best |
14 |
> > > if |
15 |
> > > we drop the current compression implementation and rework your old |
16 |
> > > code |
17 |
> > > to handle both compression and signing since the signing code is |
18 |
> > > more or |
19 |
> > > less already complete. |
20 |
> > |
21 |
> > Signing modules in pkg_preinst seems like a bad idea to me. That |
22 |
> > means |
23 |
> > you need to copy your private keys around to every host where the |
24 |
> > package might be installed. |
25 |
> > |
26 |
> > If you sign in src_compile or src_install, you only need private keys |
27 |
> > on the system building your binpkg. |
28 |
> > |
29 |
> |
30 |
> unfortunately portage will unconditionally strip .ko objects, rendering |
31 |
> modules unloadable by stripping signature, unless we do dostrip -x |
32 |
> (requires EAPI7+, which should not be a problem nowadays, but was a |
33 |
> problem back in 2018), which can be quite unfortunate on debug enabled |
34 |
> kernels. |
35 |
|
36 |
Sounds like something to fix/change in Portage. It could probably be |
37 |
updated to not strip the signature. However, I would guess the |
38 |
signature needs to be updated after the binary is modified in any |
39 |
case. |
40 |
|
41 |
Or as a workaround you could disable automatic striping via dostrip -x |
42 |
and run the proper commands to strip the modules in src_install as |
43 |
well. |