linux-kernel - Re: [PATCH] seccomp: Fix tracer exit notifications during fatal signals

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: <CAP045ArzsETMp0-1Yi=vu89pqR=08FWGkKRD9q4ZNMsXW=-3jQ@mail.gmail.com>
Date:   Mon, 22 Aug 2016 20:27:03 -0700
From:   Kyle Huey <me@...ehuey.com>
To:     Kees Cook <keescook@...omium.org>
Cc:     Oleg Nesterov <oleg@...hat.com>,
        LKML <linux-kernel@...r.kernel.org>,
        Kyle Huey <khuey@...ehuey.com>,
        "Robert O'Callahan" <robert@...llahan.org>,
        Andy Lutomirski <luto@...capital.net>,
        Will Drewry <wad@...omium.org>
Subject: Re: [PATCH] seccomp: Fix tracer exit notifications during fatal signals

On Thu, Aug 11, 2016 at 11:18 AM, Kees Cook <keescook@...omium.org> wrote:
> On Thu, Aug 11, 2016 at 8:12 AM, Oleg Nesterov <oleg@...hat.com> wrote:
>> On 08/10, Kees Cook wrote:
>>>
>>> This fixes a ptrace vs fatal pending signals bug as manifested in seccomp
>>> now that ptrace was reordered to happen after ptrace. The short version is
>>> that seccomp should not attempt to call do_exit() while fatal signals are
>>> pending under a tracer. This was needlessly paranoid. Instead, the syscall
>>> can just be skipped and normal signal handling, tracer notification, and
>>> process death can happen.
>>
>> ACK.
>>
>> I think this change is fine in any case, but...
>>
>>> The bug happens because when __seccomp_filter() detects
>>> fatal_signal_pending(), it calls do_exit() without dequeuing the fatal
>>> signal. When do_exit() sends the PTRACE_EVENT_EXIT
>>
>> I _never_ understood what PTRACE_EVENT_EXIT should actually do. I mean,
>> when it should actually stop. This was never defined.
>
> Yeah, agreed. I spent some time reading through what should happen to
> __TASK_TRACED during exit and my head spun. :)
>
>>> notification and
>>> that task is descheduled, __schedule() notices that there is a fatal
>>> signal pending and changes its state from TASK_TRACED to TASK_RUNNING.
>>
>> And this can happen anyway, with or without this change, with or without
>> seccomp. Because another fatal signal can be pending. So PTRACE_EVENT_EXIT
>> actually depends on /dev/random.
>>
>> Perhaps we should finally define what it should do. Say, it should only
>> stop if SIGKILL was sent "implicitely" by exit/exec. But as for exec,
>> there are more (off-topic) complications, not sure we actually want this...
>>
>> Nevermind, the main problem is that _any_ change in this area can break
>> something. This code is sooooooo old.
>>
>> But let me repeat, I think this change is fine anyway.
>>
>> Acked-by: Oleg Nesterov <oleg@...hat.com>
>
> Awesome, thanks!

Hi folks,

Can't help but notice this didn't make it into rc3.  Not sure if it's
bubbling up somewhere I can't see, but we'd really like this to get
into 4.8 so we don't have to work around the regression.

Thanks!

- Kyle