[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20261107044646.GC23438@brightrain.aerifal.cx>
Date: Fri, 6 Nov 2026 23:46:47 -0500
From: Rich Felker <dalias@...c.org>
To: Yao Zi <me@...ao.cc>
Cc: musl@...ts.openwall.com
Subject: Re: OpenSSH password auth bypass analysis [was: Re:
Advisory for OpenSSH+musl users]
On Sat, Nov 07, 2026 at 03:57:55AM +0000, Yao Zi wrote:
> On Fri, Nov 06, 2026 at 04:39:24PM -0500, Jeffrey Walton wrote:
> > Hi Rich,
> >
> > One comment on Item (5) below.
> >
> > On Fri, Nov 6, 2026 at 1:55 PM Rich Felker <dalias@...c.org> wrote:
> > >
> > > [...]
> > > 5. Having a failure token not match any password is contingent on the
> > > caller of crypt passing the correct hash as the settign argument to
> > > crypt. While it's guaranteed that:
> > >
> > > crypt(pass,failure_token) != failure_token
> > >
> > > it is impossible to guarantee that:
> > >
> > > crypt(pass,some_other_hash) != failure_token
> >
> > I think you can bind the username _and_ (salt,password) to the hash,
> > and that will ensure the conditions above behave as expected. That
> > is, instead of storing the hash of (salt,password), instead store
> > (username,salt,password). That should ensure a negligible probability
> > in cases where two users happen to use the same password, like
> > P@...ord1.
>
> This couldn't be achieved in the current way that crypt() is commonly
> used, right?
Two users happening to use the same password is not even the issue at
hand. I don't understand the motivation for that suggestion.
Rich
Powered by blists - more mailing lists
Confused about mailing lists and their use?
Read about mailing lists on Wikipedia
and check out these
guidelines on proper formatting of your messages.
