netdev - Re: 6.17 crashes in ipv6 code when booted fips=1 [was: [GIT PULL] Crypto Update for 6.17]

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] [day] [month] [year] [list]

Message-ID: <CAHk-=whTKNg8+F5EUP2oxcfr14P7geOOpaPBwhxF7a0jjBm2GA@mail.gmail.com>
Date: Wed, 8 Oct 2025 09:15:49 -0700
From: Linus Torvalds <torvalds@...ux-foundation.org>
To: "Theodore Ts'o" <tytso@....edu>
Cc: Simo Sorce <simo@...hat.com>, Eric Biggers <ebiggers@...nel.org>, 
	Vegard Nossum <vegard.nossum@...cle.com>, Jiri Slaby <jirislaby@...nel.org>, 
	Herbert Xu <herbert@...dor.apana.org.au>, "David S. Miller" <davem@...emloft.net>, 
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>, 
	Linux Crypto Mailing List <linux-crypto@...r.kernel.org>, netdev@...r.kernel.org, 
	Jakub Kicinski <kuba@...nel.org>, "nstange@...e.de" <nstange@...e.de>, "Wang, Jay" <wanjay@...zon.com>
Subject: Re: 6.17 crashes in ipv6 code when booted fips=1 [was: [GIT PULL]
 Crypto Update for 6.17]

On Wed, 8 Oct 2025 at 05:13, Theodore Ts'o <tytso@....edu> wrote:
>
> If there is something beyond hard-disabling CONFIG_CRYPTO_SHA1 which
> all distributions could agree with --- what would that set of patches
> look like, and would it be evenly vaguely upstream acceptable.  It
> could even hidden behind CONFIG_BROKEN.  :-)

Maybe just

 (a) make it a runtime flag - so that it can't mess up the boot, at least

 (b) make it ternary so that you get a "warn vs turn off"

 (c) and allow people to clear it too - so that you can sanely *test*
it without forcing a possibly unusable machine

and then

 (d) make the clearing be dependent on that 'lockdown' set that nobody
remotely normal uses anyway

would make this thing a whole lot more testable, and a whole lot less abrupt.

Christ, if even FIPS went "we know this is a big thing, we'll give you
a decade to sort things out", then the kernel damn well shouldn't make
it some black-and-white sudden flag.

So we should *NOT* say "FIPS says turn it off eventually, so we should
turn it off". No. That was very very wrong.

We should say "FIPS says turn it off eventually, so we should give
users simple tools to find problem spots".

And that "simple tools" very much is about not making it some kind of
"Oh, what happens is that the machine is unusable". It should be that
"Oh, look, now it gives a warning that I would have a problem in XYZ
if it wasn't available".

                   Linus