linux-kernel - RE: [PATCH 2/2] watchdog: da9062: reset board on watchdog timeout

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <DB9PR10MB4652529FAC6253C266B172EF80639@DB9PR10MB4652.EURPRD10.PROD.OUTLOOK.COM>
Date:   Fri, 26 Nov 2021 13:28:28 +0000
From:   Adam Thomson <Adam.Thomson.Opensource@...semi.com>
To:     Andrej Picej <andrej.picej@...ik.com>,
        Support Opensource <Support.Opensource@...semi.com>,
        "wim@...ux-watchdog.org" <wim@...ux-watchdog.org>,
        "linux@...ck-us.net" <linux@...ck-us.net>,
        "linux-watchdog@...r.kernel.org" <linux-watchdog@...r.kernel.org>
CC:     "y.bas@...tec.de" <y.bas@...tec.de>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: RE: [PATCH 2/2] watchdog: da9062: reset board on watchdog timeout

On 24 November 2021 08:33, Andrej Picej wrote:

> From: Stefan Christ <s.christ@...tec.de>
>
> In the default fuse configuration the watchdog poweroffs the system
> instead of resetting it on a watchdog timeout. This patch changes the
> behavior. Now the board is reseted and reboots.
>
> Note: This patch requires that the config register CONFIG_I is
> configured as writable in the da9062 multi function device.

I'm a little concerned about forcing this change in the driver. There may be
platforms which don't want the PMIC to perform a full reset through OTP re-read
and if we hard code this change then that's impacting those platforms. If we
want/need this then I think it should probably be a DT binding for da9061/2
which then indicates the behaviour we want.

NRES_MODE bit also plays a part here as it controls whether or not the nRESET
line state is changed as part of the power-down/up process. I'm assuming for
your setup this bit is 0?

>
> Signed-off-by: Stefan Christ <s.christ@...tec.de>
> Signed-off-by: Andrej Picej <andrej.picej@...ik.com>
> ---
>  drivers/watchdog/da9062_wdt.c | 14 ++++++++++++++
>  1 file changed, 14 insertions(+)
>
> diff --git a/drivers/watchdog/da9062_wdt.c b/drivers/watchdog/da9062_wdt.c
> index f02cbd530538..a04613e68f4a 100644
> --- a/drivers/watchdog/da9062_wdt.c
> +++ b/drivers/watchdog/da9062_wdt.c
> @@ -87,6 +87,20 @@ static int da9062_wdt_start(struct watchdog_device *wdd)
>  	unsigned int selector;
>  	int ret;
>
> +	/*
> +	 * Use da9062's SHUTDOWN mode instead of POWERDOWN for
> watchdog reset.
> +	 * On timeout the PMIC should reset the system, not powering it
> +	 * off.
> +	 */
> +	ret = regmap_update_bits(wdt->hw->regmap,
> +				 DA9062AA_CONFIG_I,
> +				 DA9062AA_WATCHDOG_SD_MASK,
> +				 DA9062AA_WATCHDOG_SD_MASK);
> +	if (ret)
> +		dev_err(wdt->hw->dev,
> +			"failed to set wdt reset mode. Expect poweroff on
> watchdog reset: %d\n",
> +			ret);
> +
>  	selector = da9062_wdt_timeout_to_sel(wdt->wdtdev.timeout);
>  	ret = da9062_wdt_update_timeout_register(wdt, selector);
>  	if (ret)
> --
> 2.25.1