linux-kernel - Re: d7be102f29 ("cfg80211: initialize regulatory keys/database later"): kernel BUG at crypto/asymmetric_keys/public

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: <1511819541.5456.19.camel@sipsolutions.net>
Date:   Mon, 27 Nov 2017 22:52:21 +0100
From:   Johannes Berg <johannes@...solutions.net>
To:     Linus Torvalds <torvalds@...ux-foundation.org>,
        "Wu, Fengguang" <fengguang.wu@...el.com>,
        David Howells <dhowells@...hat.com>,
        Kees Cook <keescook@...omium.org>,
        Josh Boyer <jwboyer@...hat.com>
Cc:     "Ye, Xiaolong" <xiaolong.ye@...el.com>,
        LKML <linux-kernel@...r.kernel.org>, LKP <lkp@...org>
Subject: Re: d7be102f29 ("cfg80211: initialize regulatory keys/database
 later"): kernel BUG at crypto/asymmetric_keys/public_key.c:80!

On Mon, 2017-11-27 at 21:46 +0000, Linus Torvalds wrote:
> On Sat, Nov 25, 2017 at 7:07 PM, Fengguang Wu <fengguang.wu@...el.com> wrote:
> > FYI, we noticed the following commit (built with gcc-4.8):
> > 
> > commit: d7be102f2945a626f55e0501e52bb31ba3e77b81 ("cfg80211: initialize regulatory keys/database later")
> 
> The attached 'dmesg.xz' doesn't actually match the kernel or the
> report. Very odd.
> 
> > [    8.602885] kernel BUG at crypto/asymmetric_keys/public_key.c:80!
> 
> In that reported kernel, that is
> 
>     80          BUG_ON(!sig->digest);
> 
> so the public key signature has no digest.

Yeah. I believe this was fixed by commit 01a95b2141e3 ("cfg80211:
select CRYPTO_SHA256 if needed"), and since that was already on the way
I didn't reply to this report specifically.

> I'm not seeing why it would ever be ok to do BUG_ON() instead of just
> returning an error, though.

That's kinda a good question - here we might fail to load this
certificate, print a key, and then never be able to load the signed
file - but at least we wouldn't crash immediately :-)

johannes