lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAHk-=wjbahY4JjLCXNT4o3xVq2cejqBG69z+MNfHsN9FQBnbOw@mail.gmail.com>
Date: Tue, 25 Mar 2025 16:02:13 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: Paul Moore <paul@...l-moore.com>
Cc: selinux@...r.kernel.org, linux-security-module@...r.kernel.org, 
	linux-kernel@...r.kernel.org
Subject: Re: [GIT PULL] selinux/selinux-pr-20250323

On Sun, 23 Mar 2025 at 12:39, Paul Moore <paul@...l-moore.com> wrote:
>
> - Add additional SELinux access controls for kernel file reads/loads
>
>   The SELinux kernel file read/load access controls were never updated
>   beyond the initial kernel module support, this pull request adds
>   support for firmware, kexec, policies, and x.509 certificates.

Honestly, is there a *reason* for this, or is this just some misguided
"for completeness" issue?

Because dammit, adding more complexity to the security rules isn't a
feature, and isn't security. It's just theater.

And it adds completely pointless extra cases making the different
cases use artificially different code.

The commit message doesn't actually make any mention of *why* any of
this would be a good idea.

I've pulled this, but honestly, I think all those new cases should be
removed, and if people object to us having "LOADING_MODULE" for other
kinds of reads, then I think the fix should be to just rename it to
"KERNEL_LOAD" or whatever.

Because dammit, this "make security interfaces more complicated just
because" needs to STOP.

We are literally wasting enormous amounts of time in the security
layers - I regularly see the lsm and selinux layers spending *more*
time on security than the actual operation takes because the security
people have written random code without taking performance into
account - and I need to see *reasons* for adding more crap in this
area, not "let's expand" with no actual reason given.

So I really think that commit needs to be either reverted or explained
very clearly why user policy _needs_ to care for the difference
between a module load and a firmware image load.

Because I think any such explanation is likely complete BS. The
difference between loading modules and loading firmware usually would
boil down to "different hardware does things differently". There's no
obvious reason why one should be considered different from another
from a security standpoint.

               Linus

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ