mailing list archives
Re: CVE request: crypt_blowfish 8-bit character mishandling
From: Solar Designer <solar () openwall com>
Date: Fri, 8 Jul 2011 03:54:56 +0400
On Thu, Jul 07, 2011 at 04:41:17PM +0200, Ludwig Nussel wrote:
So here is a first draft of a pam_unix2 patch for discussion. The
password function of pam_unix2 already reads /etc/default/passwd
and /etc/login.defs, now the auth function also does.
The new option BLOWFISH_2a2x=yes/no toggles the compat mode to treat
2a as 2x as discussed before.
I've also added a second option BLOWFISH_2y=yes/no (default yes) as
safeguard. When set to 'no' new passwords are still stored as 2a
instead of 2y. I've been thinking about networked environments where
not all systems might get patched at the same time. If some user on
a patched system changes the password and gets 2y he can not log in
on other systems anymore, even if no 8bit characters are used. So if
that use case is important it may be better to risk locking out a
few users with umlauts in the password rather than having 2y not
accepted by unpatched systems.
That's a nasty problem I had not thought of. I am unhappy about your
workaround for it. With such a workaround, we'll have more uncertainty
of what $2a$ hashes are on SUSE systems. (Before this point, you could
have more confidence that all of those were computed with the sign
extension bug, so treating them as $2x$ would allow all users to log in
with no issues if that's what's desired despite of the security risk.)
Also, it brings up the question: why merely use $2a$ running the new
code rather than fully emulate the bug even for newly set passwords,
which would make all passwords work, even on other networked machines?
Sure, that would be even nastier for security, so maybe you managed to
strike a balance well. But nevertheless the question is there. One of
your options results in full backwards compatibility at a security cost
(for the local system), but the other somehow chooses to strike a
balance between compatibility and security without achieving either of
these fully (for a network of systems).
Maybe you can afford to drop BLOWFISH_2y to avoid those inconsistencies?
I imagine that people won't know to enable this option unless/until they
have already run into an issue anyway (that is, someone is already
unable to log in). At this point, they could likely upgrade the rest of
their networked systems as well... or downgrade this one. ;-(
BTW, this problem is a reason for us to possibly use the magic salt
substring approach instead of $2y$, although it has its drawbacks too.
+ /* if compat mode is turned on treat all 2a hashes as affected by
+ signess bug, ie 2x */
+ if (options->blowfish_2a2x && !strncmp(hash, "$2a$", 4))
+ char* h = strdupa(hash);
+ h = 'x';
+ hash = h;
+ return (strcmp (hash, crypt_r (pass, hash, &output)) == 0);
You probably want to wipe and free your altered copy of the hash here
(after strcmp(), but before return).