Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   ebiederm@xmission.com (Eric W. Biederman)
To:     Kees Cook <keescook@chromium.org>
Cc:     linux-kernel@vger.kernel.org, linux-arch@vger.kernel.org,
        Linus Torvalds <torvalds@linux-foundation.org>,
        Oleg Nesterov <oleg@redhat.com>,
        Al Viro <viro@zeniv.linux.org.uk>,
        David Miller <davem@davemloft.net>, sparclinux@vger.kernel.org
References: <87y26nmwkb.fsf@disp2133>
        <20211020174406.17889-15-ebiederm@xmission.com>
        <202110210927.D0B4B0342@keescook>
Date:   Thu, 21 Oct 2021 11:56:57 -0500
In-Reply-To: <202110210927.D0B4B0342@keescook> (Kees Cook's message of "Thu,
        21 Oct 2021 09:34:31 -0700")
Message-ID: <87k0i69uzq.fsf@disp2133>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Subject: Re: [PATCH 15/20] signal/sparc32: Exit with a fatal signal when try_to_clear_window_buffer fails
Precedence: bulk

Kees Cook <keescook@chromium.org> writes:

> On Wed, Oct 20, 2021 at 12:44:01PM -0500, Eric W. Biederman wrote:
>> The function try_to_clear_window_buffer is only called from
>> rtrap_32.c.  After it is called the signal pending state is retested,
>
> nit: rtrap_32.S
>
>> and signals are handled if TIF_SIGPENDING is set.  This allows
>> try_to_clear_window_buffer to call force_fatal_signal and then rely on
>> the signal being delivered to kill the process, without any danger of
>> returning to userspace, or otherwise using possible corrupt state on
>> failure.
>
> The TIF_SIGPENDING test happens in do_notify_resume(), though I see
> other code before that:
>
> ...
>         call    try_to_clear_window_buffer
>         add    %sp, STACKFRAME_SZ, %o0
>
>         b       signal_p
> ...
> signal_p:
>         andcc   %g2, _TIF_DO_NOTIFY_RESUME_MASK, %g0
>         bz,a    ret_trap_continue
>         ld     [%sp + STACKFRAME_SZ + PT_PSR], %t_psr
>
>         mov     %g2, %o2
>         mov     %l6, %o1
>         call    do_notify_resume
>
> Will the ret_trap_continue always be skipped?

The ret_trap_continue is the break out of the loop.  So unless the code
is not properly setting the signal to be pending the code should be good.

> Also I see the "tp->w_saved = 0" never happens due to the "return" in
> try_to_clear_window_buffer. Is that okay?

It should be.  As you point out the w_saved value is only used in
generating signal frames.  The code in get_signal should never
return and should call do_group_exit which calls do_exit, so building
signal frames that happens after get_signal returns should never be
reached.

Further this is the same way the code makes it to do_exit today.

Also looking at it I think the logic is that w_saved == 0
says that the register windows have been saved on the user mode stack,
and that clearly has not happened so I think it would in general
be a bug to clear w_saved on failure.

> Only synchronize_user_stack()
> uses it, and that could be called in do_sigreturn(). Should the "return"
> be removed?

Of course I could be wrong, if David or someone else who knows sparc32
better than me wants to set me straight I would really appreciate it.

Eric