[LTP] [PATCH v2 1/4] Fallback landlock network support

[Li Wang]

On Tue, Nov 5, 2024 at 9:15 PM Cyril Hrubis <chrubis@suse.cz> wrote:

> Hi!
> > In this way it could work, but we loose guarded buffers which are passed
> > to the syscall and might be useful during debugging. In this case we
> > should use tst_buffers_alloc(). @Cyril what do you think?
>
> That woudln't work either, since we cannot allocate "half a structure"
> can we?
>
> Unfortunatelly I think that having a per API version structures is the
> cleanest solution. Because:
>

Okay, I agree to go this way, unless we find a better solution, and it can
be replaced anytime if we have it in the future.

Andrea, feel free to fix the tiny problems and add my RBT.
Reviewed-by: Li Wang <liwang@redhat.com>



>
> - our tests should run everywhere they can, that means that we have to
>   use the minimal ABI that is required for the test
>
> - important part of the backward compatibilty testing is that there are
>   no accesses past the declared ruleset size, which can be easily done
>   by the guarded buffers provided that we allocate exactly the size
>   needed
>
> And backward compatibility also means that we have to properly handle
> the case when we need newer ABI than currently supported, so the
> verify_landlock_enabled() needs a size parameter so that we can check
> that the ABI is >= than the minimal ABI the test needs...
>
> --
> Cyril Hrubis
> chrubis@suse.cz
>
>

-- 
Regards,
Li Wang