Received-SPF: softfail (google.com: domain of transitioning linux-kernel-owner@vger.kernel.org does not designate 23.128.96.19 as permitted sender) client-ip=23.128.96.19;
MIME-Version: 1.0
References: <20220304064324.331217-1-hasegawa-hitomi@fujitsu.com> <20220304064324.331217-2-hasegawa-hitomi@fujitsu.com>
In-Reply-To: <20220304064324.331217-2-hasegawa-hitomi@fujitsu.com>
From:   Doug Anderson <dianders@chromium.org>
Date:   Fri, 4 Mar 2022 10:05:45 -0800
Message-ID: <CAD=FV=Udf=MzyPa_o=vz=nc7ZVXBuuVNqw-VOSfrShuv0hN64Q@mail.gmail.com>
Subject: Re: [PATCH v2 1/2] tty/sysrq: Make sysrq handler NMI aware
To:     Hitomi Hasegawa <hasegawa-hitomi@fujitsu.com>
Cc:     Linux ARM <linux-arm-kernel@lists.infradead.org>,
        SoC Team <soc@kernel.org>, linux-serial@vger.kernel.org,
        Sumit Garg <sumit.garg@linaro.org>,
        Arnd Bergmann <arnd@arndb.de>, Olof Johansson <olof@lixom.net>,
        Catalin Marinas <catalin.marinas@arm.com>,
        Will Deacon <will@kernel.org>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Jiri Slaby <jirislaby@kernel.org>,
        Jason Wessel <jason.wessel@windriver.com>,
        Daniel Thompson <daniel.thompson@linaro.org>,
        LKML <linux-kernel@vger.kernel.org>,
        kgdb-bugreport@lists.sourceforge.net,
        Peter Zijlstra <peterz@infradead.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

Hi,

On Thu, Mar 3, 2022 at 10:45 PM Hitomi Hasegawa
<hasegawa-hitomi@fujitsu.com> wrote:
>
>  void __handle_sysrq(int key, bool check_mask)
>  {
>         const struct sysrq_key_op *op_p;
> @@ -573,6 +606,10 @@ void __handle_sysrq(int key, bool check_mask)
>         int orig_suppress_printk;
>         int i;
>
> +       /* Skip sysrq handling if one already in progress */
> +       if (sysrq_nmi_key != -1)
> +               return;

Should this give a warning?

Also, can you remind me why this is safe if two CPUs both call
handle_sysrq() at the same time? Can't both of them make it past this?
That doesn't seem so great.


> @@ -596,7 +633,13 @@ void __handle_sysrq(int key, bool check_mask)
>                 if (!check_mask || sysrq_on_mask(op_p->enable_mask)) {
>                         pr_info("%s\n", op_p->action_msg);
>                         console_loglevel = orig_log_level;
> -                       op_p->handler(key);
> +
> +                       if (in_nmi() && !op_p->nmi_safe) {
> +                               sysrq_nmi_key = key;
> +                               irq_work_queue(&sysrq_irq_work);

It looks like irq_work_queue() returns false if it fails to queue.
Maybe it's worth checking and setting "sysrq_nmi_key" back to -1 if it
fails?

-Doug