[Rio-pm] Fwd: [oss-security] CVE Request -- perl v5.8.* -- stack overflow by processing certain regex (Gentoo BTS#313565 / RH BZ#580605)

[Oscar Bacelar]

Para conhecimento.

---------- Forwarded message ----------
From: Jan Lieskovsky <jlieskov em redhat.com>
Date: 2010/4/8
Subject: [oss-security] CVE Request -- perl v5.8.* -- stack overflow by
processing certain regex (Gentoo BTS#313565 / RH BZ#580605)
To: "Steven M. Christey" <coley em linus.mitre.org>
Cc: oss-security <oss-security em lists.openwall.com>


Hi Steve, vendors,

 1, wouldn't like to open a can of worms,
 2, but for purpose of properly tracking it, requesting a CVE id for the
    following Perl regular expression engine issue:

Bruce Merry reported:
 [1] http://bugs.gentoo.org/show_bug.cgi?id=313565

an integer overflow, leading to stack overflow in the way
Perl regular expression engine processed certain regular
expression(s). Remote attacker could use this flaw to cause
a denial of service (crash of an application, using the
Perl regular expression engine).

Public PoC from [1]:
--------------------
 perl -e 'if ((("a " x 100000) . "a\n") =~ /\A\S+(?: \S+)*\n\z/) {}'

References:
 [2] http://bugs.gentoo.org/show_bug.cgi?id=313565
 [3] https://bugzilla.redhat.com/show_bug.cgi?id=580605

Affected Perl versions:
 Issue tested and confirmed in Perl of versions v5.8.*.
 Versions of Perl v5.10.* are not affected by this.

Steve, what's the Mitre's opinion on cases like this --
denial of service reachable via certain regular expression.

Should we track them on per issue basis? Or only for cases,
where more than a DoS is possible? (doesn't seem to be
this case though).

Thanks && Regards, Jan.
--
Jan iankko Lieskovsky / Red Hat Security Response Team
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://mail.pm.org/pipermail/rio-pm/attachments/20100408/284ad365/attachment.html>