ssh server vulnerable to regreSSHion?

[Brian Inglis]

On 2024-07-04 09:31, Tom Kent via Cygwin wrote:
> For anyone not aware, a major, remotely exploitable, vulnerability has been
> found in OpenSSH servers.
> 
> It has been assigned CVE-2024-6387 [1] and titled "regreSSHion" [2] because
> it is actually a regression of a pair of early 2000s bugs:
> CVE-2006-5051 and CVE-2008-4109.
> 
> The vulnerability is a race condition related to its interaction with
> glibc. Because of the way cygwin is built, it isn't clear to me if this is
> something that could possibly be impacting or not, thus I wanted to see if
> smarter heads could identify if this is a potential (or actual) issue.
> 
> Either way, it might be nice to get a determination posted somewhere for
> people to find, as I expect there will be more out there wondering about
> this in the next days/weeks.

If you subscribed to Cygwin Announce mailing list

	https://cygwin.com/mailman/listinfo/cygwin-announce

	https://inbox.sourceware.org/cygwin-announce/

you would have seen the openssh 9.8p1-1 upgrade announcement

	https://cygwin.com/pipermail/cygwin-announce/2024-July/011846.html

https://inbox.sourceware.org/cygwin-announce/20240702194232.2039121-1-corinna-cygwin@cygwin.com

which should take care of any potential issues whether vulnerable or not.

The Cygwin OpenSSH maintainer was also involved in pre-release testing:

	https://marc.info/?l=openssh-unix-dev&m=171956630724852&w=2

validated the release, and caught an out-of-tree build test bug, so they are 
taking care on Cygwin, as Cygwin developers and package maintainers are likely 
to be dependent on OpenSSH servers and clients.

The regression issues are dependent on how certain libc functions are 
implemented and used, in Cygwin's case by newlib and/or Cygwin functions.
Other newlib and other libc, like musl, hosted implementations may have similar 
or independent issues.
Certainly Ubuntu and Debian (both 32 bit) have similar issues with significant 
differences.
As the OpenSSH announcement included above says:
"Exploitation on 64-bit systems is believed to be possible but has not been 
demonstrated at this time."
It requires weak ALSR applied to sshd and async-signal-unsafe syslog() calling 
malloc() allowing it to be be vulnerable to a race condition exploitable by 
SIGALARM, for the demonstrated vulnerability.

The ObscureKeystrokeTiming password timing attack is assigned as:

	https://www.cve.org/CVERecord?id=CVE-2024-39894

> [1] https://www.cve.org/CVERecord?id=CVE-2024-6387
> [2]
> https://blog.qualys.com/vulnerabilities-threat-research/2024/07/01/regresshion-remote-unauthenticated-code-execution-vulnerability-in-openssh-server
-- 
Take care. Thanks, Brian Inglis              Calgary, Alberta, Canada

La perfection est atteinte                   Perfection is achieved
non pas lorsqu'il n'y a plus rien à ajouter  not when there is no more to add
mais lorsqu'il n'y a plus rien à retirer     but when there is no more to cut
                                 -- Antoine de Saint-Exupéry