linux-kernel - Re: [PATCH 1/2] audit: fix NUL handling in untrusted strings

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: Windows password security audit tool. GUI, reports in PDF.

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

Message-Id: <1221143113.2992.9.camel@localhost.localdomain>
Date:	Thu, 11 Sep 2008 10:25:13 -0400
From:	Eric Paris <eparis@...hat.com>
To:	Miloslav Trmač <mitr@...hat.com>
Cc:	viro@...iv.linux.org.uk, linux-audit <linux-audit@...hat.com>,
	linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/2] audit: fix NUL handling in untrusted strings

On Thu, 2008-09-11 at 00:23 +0200, Miloslav Trmač wrote:
> From: Miloslav Trmac <mitr@...hat.com>
> 
> audit_string_contains_control() stops checking at the first NUL byte.
> If audit_string_contains_control() returns FALSE,
> audit_log_n_untrustedstring() submits the complete string - including
> the NUL byte and all following bytes, up to the specified maximum length
> - to audit_log_n_string(), which copies the data unchanged into the
> audit record.
> 
> The audit record can thus contain a NUL byte (and some unchecked data
> after that).  Because the user-space audit daemon treats audit records
> as NUL-terminated strings, an untrusted string that is shorter than the
> specified maximum length effectively terminates the audit record.
> 
> This patch modifies audit_log_n_untrustedstring() to only log the data
> before the first NUL byte, if any.

I'm going to have to say NAK on this patch.

It's still not right looking at the other user,
audit_log_single_execve_arg().  An execve arg with a NULL could loose
the stuff after the NULL (not break the record like audit_tty) since the
execve uses %s rather than calling trusted string.

How about we change the meaning of audit_string_contains_control()
return values?  If it returns positive that is the number of bytes in a
legitimate string up to the first null.  -1 means it is hex.  That
eliminates your code duplication and allows us to do the right thing in
both the generic untrusted_string code and the execve code.....

-Eric

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/