MIME-Version: 1.0
References: <20211101034147.6203-1-khuey@kylehuey.com> <877ddqabvs.fsf@disp2133>
 <CAP045AqJVXA60R9RF8Gb2PWGBsK6bZ7tVBkdCcPYYrp6rOkG-Q@mail.gmail.com>
 <87fsse8maf.fsf@disp2133> <CAP045ApAX725ZfujaK-jJNkfCo5s+oVFpBvNfPJk+DKY8K7d=Q@mail.gmail.com>
 <CAP045AqsstnxfTyXhhCGDSucqGN7BTtfHJ5s6ZxUQC5K-JU56A@mail.gmail.com>
 <87bl2kekig.fsf_-_@email.froward.int.ebiederm.org> <CAP045AqSKv8hAz79ntbj6ZoGiO-v7M0UBuQF2hwJdXqqL7oxWA@mail.gmail.com>
 <87y25m9154.fsf@email.froward.int.ebiederm.org>
In-Reply-To: <87y25m9154.fsf@email.froward.int.ebiederm.org>
From:   =?UTF-8?B?TWFya28gTcOka2Vsw6Q=?= <marko.makela@mariadb.com>
Date:   Thu, 18 Nov 2021 08:12:24 +0200
Message-ID: <CAJuX1hyqRzOkYhoVLwUORTOn2ncJxiru8JfM9cpZt8uXigff2g@mail.gmail.com>
Subject: Re: [PATCH 0/3] signal: requeuing undeliverable signals
To:     "Eric W. Biederman" <ebiederm@xmission.com>
Cc:     Kyle Huey <me@kylehuey.com>,
        open list <linux-kernel@vger.kernel.org>,
        Jens Axboe <axboe@kernel.dk>,
        Peter Zijlstra <peterz@infradead.org>,
        Marco Elver <elver@google.com>,
        Oleg Nesterov <oleg@redhat.com>,
        Thomas Gleixner <tglx@linutronix.de>,
        Peter Collingbourne <pcc@google.com>,
        Alexey Gladkov <legion@kernel.org>,
        "Robert O'Callahan" <rocallahan@gmail.com>,
        Linux API <linux-api@vger.kernel.org>,
        Al Viro <viro@zeniv.linux.org.uk>,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Kees Cook <keescook@chromium.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Nov 17, 2021 at 6:51 PM Eric W. Biederman <ebiederm@xmission.com> w=
rote:
>
> Kyle Huey <me@kylehuey.com> writes:
>
> > On Mon, Nov 15, 2021 at 9:31 PM Eric W. Biederman <ebiederm@xmission.co=
m> wrote:
> >>
> >>
> >> Kyle Huey recently reported[1] that rr gets confused if SIGKILL preven=
ts
> >> ptrace_signal from delivering a signal, as the kernel setups up a sign=
al
> >> frame for a signal that rr did not have a chance to observe with ptrac=
e.
> >>
> >> In looking into it I found a couple of bugs and a quality of
> >> implementation issue.
> >>
> >> - The test for signal_group_exit should be inside the for loop in get_=
signal.
> >> - Signals should be requeued on the same queue they were dequeued from=
.
> >> - When a fatal signal is pending ptrace_signal should not return anoth=
er
> >>   signal for delivery.
> >>
> >> Kyle Huey has verified[2] an earlier version of this change.
> >>
> >> I have reworked things one more time to completely fix the issues
> >> raised, and to keep the code maintainable long term.
> >>
> >> I have smoke tested this code and combined with a careful review I
> >> expect this code to work fine.  Kyle if you can double check that
> >> my last round of changes still works for rr I would appreciate it.
> >
> > This still fixes the race we reported.
>
> >
> > Tested-by: Kyle Huey <khuey@kylehuey.com>
>
> Thank you very much for retesting.
>
> Eric

Thank you, Kyle and Eric, for reporting and fixing the root cause of this r=
ace.

Meanwhile, I followed Kyle's suggestion and will disable the crash
handlers in the tracee whenever it is being traced.

Marko
--=20
Marko M=C3=A4kel=C3=A4, Lead Developer InnoDB
MariaDB Corporation