On Fri, May 28, 2021 at 10:43 AM Daniel Borkmann [off-list ref] wrote:

On 5/28/21 3:42 PM, Ondrej Mosnacek wrote:

quoted

(I'm off work today and plan to reply also to Paul's comments next
week, but for now let me at least share a couple quick thoughts on
Daniel's patch.)

Oooh, I sense some disagreement brewing :)

quoted

On Fri, May 28, 2021 at 11:56 AM Daniel Borkmann [off-list ref] wrote:

quoted

On 5/28/21 9:09 AM, Daniel Borkmann wrote:

quoted

On 5/28/21 3:37 AM, Paul Moore wrote:

quoted

On Mon, May 17, 2021 at 5:22 AM Ondrej Mosnacek [off-list ref] wrote:

...

quoted

quoted

Ondrej / Paul / Jiri: at least for the BPF tracing case specifically (I haven't looked
at the rest but it's also kind of independent), the attached fix should address both
reported issues, please take a look & test.

Thanks, I like this solution, although there are a few gotchas:

1. This patch creates a slight "regression" in that if someone flips
the Lockdown LSM into lockdown mode on runtime, existing (already
loaded) BPF programs will still be able to call the
confidentiality-breaching helpers, while before the lockdown would
apply also to them. Personally, I don't think it's a big deal (and I
bet there are other existing cases where some handle kept from before
lockdown could leak data), but I wanted to mention it in case someone
thinks the opposite.

Yes, right, though this is nothing new either in the sense that there are
plenty of other cases with security_locked_down() that operate this way
e.g. take the open_kcore() for /proc/kcore access or the module_sig_check()
for mod signatures just to pick some random ones, same approach where the
enforcement is happen at open/load time.

Another, yes, this is not really a good thing to do.  Also, just
because there are other places that don't really do The Right Thing
doesn't mean that it is okay to also not do The Right Thing here.
It's basically the two-wrongs-don't-make-a-right issue applied to
kernel code.

-- 
paul moore
www.paul-moore.com