linux-kernel - Re: [PATCH] watchdog: sbsa_gwdt: disable watchdog when system panic was trigged by signal WS0

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] [day] [month] [year] [list]

Date:   Mon, 2 Mar 2020 06:19:44 -0800
From:   Guenter Roeck <linux@...ck-us.net>
To:     luanshi <zhangliguang@...ux.alibaba.com>,
        Wim Van Sebroeck <wim@...ux-watchdog.org>
Cc:     linux-watchdog@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] watchdog: sbsa_gwdt: disable watchdog when system panic
 was trigged by signal WS0

On 3/2/20 12:46 AM, luanshi wrote:
> When ARM SBSA Generic Watchdog worked in the two stages mode by setting
> module parameter action to 1, when the timeout is reached, the first
> signal (WS0) will trigger panic. Before panic, the value of register
> SBSA_GWDT_WCS was 0x0, after panic and system startup, the value of
> register SBSA_GWDT_WCS was 0x7, status bits SBSA_GWDT_WCS_EN
> SBSA_GWDT_WCS_WS0 and SBSA_GWDT_WCS_WS1 were set, this will increase the
> refcnt of module sbsa_gwdt by function watchdog_cdev_register because
> flag WDOG_HW_RUNNING was set, so we cannot unload the module again. To be
> consistent with reboot, watchdog should be disabled when system panic was
> trigged by signal(WS0).
> 
> Signed-off-by: Liguang Zhang <zhangliguang@...ux.alibaba.com>
> ---
>  drivers/watchdog/sbsa_gwdt.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/drivers/watchdog/sbsa_gwdt.c b/drivers/watchdog/sbsa_gwdt.c
> index f0f1e3b..6bee5bb 100644
> --- a/drivers/watchdog/sbsa_gwdt.c
> +++ b/drivers/watchdog/sbsa_gwdt.c
> @@ -194,6 +194,12 @@ static int sbsa_gwdt_stop(struct watchdog_device *wdd)
>  
>  static irqreturn_t sbsa_gwdt_interrupt(int irq, void *dev_id)
>  {
> +	struct sbsa_gwdt *gwdt = (struct sbsa_gwdt *)dev_id;
> +	struct watchdog_device *wdd = &gwdt->wdd;
> +
> +	if (wdd->ops->stop)
> +		wdd->ops->stop(wdd);
> +
>  	panic(WATCHDOG_NAME " timeout");
>  
>  	return IRQ_HANDLED;
> 

This prevents the 2nd stage (the actual system reset) from working if the panic
call gets stuck and doesn't result in a reboot. I don't think it is a good idea
to do that. It effectively disables the second stage.

Ultimately, this is a bios problem; the bios/rommon should have stopped
the watchdog when rebooting. If you don't want the watchdog to run on startup,
you have to add a module parameter to disable the watchdog if it is running
at probe time.

Guenter