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;
MIME-Version: 1.0
References: <20210811155914.19550-1-quic_eberman@quicinc.com>
In-Reply-To: <20210811155914.19550-1-quic_eberman@quicinc.com>
From:   Sami Tolvanen <samitolvanen@google.com>
Date:   Wed, 11 Aug 2021 09:12:11 -0700
Message-ID: <CABCJKueDLEqxu1OVM4bWGmeX1aQQXVKsnr6_jCXcPpcW3H_A3g@mail.gmail.com>
Subject: Re: [PATCH 1/1] cfi: Use rcu_read_{un}lock_sched_notrace
To:     Elliot Berman <quic_eberman@quicinc.com>
Cc:     Kees Cook <keescook@chromium.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Jinlong Mao <jinlmao@codeaurora.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi Elliot,

On Wed, Aug 11, 2021 at 9:00 AM Elliot Berman <quic_eberman@quicinc.com> wrote:
>
> If rcu_read_lock_sched tracing is enabled, the tracing subsystem can
> perform a jump which needs to be checked by CFI. For example, stm_ftrace
> source is enabled as a module and hooks into enabled ftrace events. This
> can cause an recursive loop where find_shadow_check_fn ->
> rcu_read_lock_sched -> (call to stm_ftrace generates cfi slowpath) ->
> find_shadow_check_fn -> rcu_read_lock_sched -> ...
>
> To avoid the recursion, either the ftrace codes needs to be marked with
> __no_cfi or CFI should not trace. Use the "_notrace" in CFI to avoid
> tracing so that CFI can guard ftrace.
>
> Signed-off-by: Elliot Berman <quic_eberman@quicinc.com>
> ---
>  kernel/cfi.c | 8 ++++----
>  1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/kernel/cfi.c b/kernel/cfi.c
> index e17a56639766..9594cfd1cf2c 100644
> --- a/kernel/cfi.c
> +++ b/kernel/cfi.c
> @@ -248,9 +248,9 @@ static inline cfi_check_fn find_shadow_check_fn(unsigned long ptr)
>  {
>         cfi_check_fn fn;
>
> -       rcu_read_lock_sched();
> +       rcu_read_lock_sched_notrace();
>         fn = ptr_to_check_fn(rcu_dereference_sched(cfi_shadow), ptr);
> -       rcu_read_unlock_sched();
> +       rcu_read_unlock_sched_notrace();
>
>         return fn;
>  }
> @@ -269,11 +269,11 @@ static inline cfi_check_fn find_module_check_fn(unsigned long ptr)
>         cfi_check_fn fn = NULL;
>         struct module *mod;
>
> -       rcu_read_lock_sched();
> +       rcu_read_lock_sched_notrace();
>         mod = __module_address(ptr);
>         if (mod)
>                 fn = mod->cfi_check;
> -       rcu_read_unlock_sched();
> +       rcu_read_unlock_sched_notrace();
>
>         return fn;
>  }

Thanks for the patch! I agree, this looks like a better solution than
littering tracing code with __nocfi annotations.

Reviewed-by: Sami Tolvanen <samitolvanen@google.com>

Sami