phc-discussions - Re: [PHC] "Why I Don't Recommend Scrypt"

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 for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <op.xco7o3w1yldrnw@laptop-air>
Date: Thu, 13 Mar 2014 23:16:05 -0700
From: "Jeremy Spilman" <jeremy@...link.co>
To: "discussions@...sword-hashing.net" <discussions@...sword-hashing.net>,
 "Tony Arcieri" <bascule@...il.com>
Subject: Re: [PHC] "Why I Don't Recommend Scrypt"

On Thu, 13 Mar 2014 11:45:39 -0700, Tony Arcieri <bascule@...il.com> wrote:

> Account name? If you don't compute a password hash even for invalid  
> account names, you're vulnerable to user enumeration attacks.

I supposed you could try to make all attempts return after X ms,  
regardless of whether any hashing was actually done. But if you're to the  
point of hashing the password for an invalid username, then maybe next  
you're worried about side-channels where an attacker can [remotely]  
measure the load on your login servers, and be able to tell when you're  
just sleeping instead of hashing.

User enumeration is probably one of those battles which a sufficiently  
motivated attacker is not going to lose. There are almost certainly going  
to be many, many code paths which can be exploited as oracles for user  
existence beyond just the login API.

It's probably so easy to enumerate users on a system through timing, it's  
almost worth giving up on the premise, and just provide a descriptive  
login error message.
Content of type "text/html" skipped