Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20160725080331.GB24232@suse.de>
Date: Mon, 25 Jul 2016 10:03:31 +0200
From: Sebastian Krahmer <krahmer@...e.com>
To: "Eric W. Biederman" <ebiederm@...ssion.com>,
	oss-security@...ts.openwall.com,
	pkg-shadow-devel@...ts.alioth.debian.org
Subject: Re: Re: [Pkg-shadow-devel] subuid security patches
 for shadow package

On Wed, Jul 20, 2016 at 11:48:52PM +0200, Nicolas François wrote:
> Hi,
> 
> The first point looks like a non issue to me.
> 
> getlogin() is used to differentiate users with the same UID.
> The result of getlogin() is checked: if it returns a username that do not
> have the UID returned by getuid(), it will be ignored.
> 
> 
> Best Regards,
> -- 
> Nekral

I agree that its not a severe issue. But its dubious code at best.
I couldnt even imagine someone would have usernames with different UID's?
Maybe such configs should not be encouraged and potential issues with
that discussed.

My understanding of secure coding is that getlogin() should not
be trusted. Having same username with multiple UIDs is also to be avoided
IMHO, since its asking for trouble (I dont know if thats some requirement
of LSB or POSIX or so?)

So, I am open for discussion about this point, as removing getlogin()
should not break valid configs. If it can be removed without breakage,
it should.

Sebastian

-- 

~ perl self.pl
~ $_='print"\$_=\47$_\47;eval"';eval
~ krahmer@...e.com - SuSE Security Team

Powered by blists - more mailing lists

Please check out the Open Source Software Security Wiki, which is counterpart to this mailing list.

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.