linux-kernel - Re: [PATCH RESEND2] ptrace: make ptrace() fail if the tracee changed its pid unexpectedly

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]

Date:   Wed, 12 May 2021 09:56:49 -0400 (EDT)
From:   Mathieu Desnoyers <mathieu.desnoyers@...icios.com>
To:     Oleg Nesterov <oleg@...hat.com>
Cc:     Linus Torvalds <torvalds@...ux-foundation.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        "Eric W. Biederman" <ebiederm@...ssion.com>,
        Eugene Syromiatnikov <esyr@...hat.com>,
        Jan Kratochvil <jan.kratochvil@...hat.com>,
        Michael Kerrisk <mtk.manpages@...il.com>,
        Pedro Alves <palves@...hat.com>,
        Simon Marchi <simon.marchi@...icios.com>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH RESEND2] ptrace: make ptrace() fail if the tracee
 changed its pid unexpectedly

----- On May 12, 2021, at 9:36 AM, Oleg Nesterov oleg@...hat.com wrote:

> On 05/11, Oleg Nesterov wrote:
>>
>> On 05/11, Oleg Nesterov wrote:
>> >
>> > On 05/11, Linus Torvalds wrote:
>> >
>> > > That said, why this:
>> > >
>> > > > +       rcu_read_lock();
>> > > > +       pid = task_pid_nr_ns(task, task_active_pid_ns(task->parent));
>> > > > +       rcu_read_unlock();
>> > >
>> > > I don't see why the RCU read lock would be needed? task_pid_nr_ns()
>> > > does any required locking itself, afaik.
>> > >
>> > > And even if it wasn't, this all happens with siglock held, can
>> > > anything actually change.
>> >
>> > ... and with tasklist_lock held.
>> >
>> > Hmm. Linus, I am shy to admit I can't answer immediately, I'll recheck
>> > tomorrow after sleep. But it seems you are right.
>>
>> most probably to protect task->parent, not sure, this was 6 month ago...
>> but in this case we can use "current". I'll recheck.
> 
> Of course you are right, rcu_read_lock() is not needed. Plus we can use
> task_pid_vnr() rather than task_pid_nr_ns(). I've sent v2.

Out of curiosity: what makes it OK to use either the current task or its
parent's pid namespace in this specific case ? What happens if they are
in different pid namespaces ?

Thanks,

Mathieu


-- 
Mathieu Desnoyers
EfficiOS Inc.
http://www.efficios.com