linux-kernel - Re: [PATCH] user_ns: use correct check for single-threadedness

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 for Android: free password hash cracker in your pocket

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

Message-ID: <87614tr2jd.fsf@x220.int.ebiederm.org>
Date:	Wed, 05 Aug 2015 13:52:22 -0500
From:	ebiederm@...ssion.com (Eric W. Biederman)
To:	Oleg Nesterov <oleg@...hat.com>
Cc:	"Kirill A. Shutemov" <kirill@...temov.name>,
	Andrew Morton <akpm@...ux-foundation.org>,
	Kees Cook <keescook@...omium.org>,
	David Howells <dhowells@...hat.com>,
	linux-kernel@...r.kernel.org,
	Peter Zijlstra <peterz@...radead.org>,
	Ingo Molnar <mingo@...nel.org>,
	"Kirill A. Shutemov" <kirill.shutemov@...ux.intel.com>,
	Rik van Riel <riel@...hat.com>,
	Vladimir Davydov <vdavydov@...allels.com>,
	Ricky Zhou <rickyz@...omium.org>,
	Julien Tinnes <jln@...gle.com>
Subject: Re: [PATCH] user_ns: use correct check for single-threadedness

Hmm.

On closer inspection this patch touches on a greater inconsistency then
the test to see if the task is the only task using the mm_struct.

We currently allow tasks created with clone to have a different user
namespace and to share a mm_struct, and I don't think that is wrong.

What we actually care about are the uid and gid values that show up in
signals that are reported to a process, and for that what we care about
is the question do the tasks share signal handling state, which is
controlled by the flags CLONE_SIGHAND and CLONE_THREAD.

As such current_is_single_threaded() is wrong because it tests to see if
there is someone else sharing an mm_struct.

So I have to ask.  Is it possible to rework these checks such that we
look at the sighand struct and signal sharing handling sharing instead
of the count on the mm_struct?

I suspect we could do that more cheaply, as well as making the code more
correct.

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/