linux-kernel - Re: [PATCH] watchdog: arm_smc_wdt: get wdt status through SMCWD_GET

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: <df2124f7-8df9-4fb3-b687-5968805c668a@roeck-us.net>
Date: Mon, 19 May 2025 10:58:29 -0700
From: Guenter Roeck <linux@...ck-us.net>
To: Antonio Borneo <antonio.borneo@...s.st.com>,
 Julius Werner <jwerner@...omium.org>, Evan Benn <evanbenn@...omium.org>,
 Wim Van Sebroeck <wim@...ux-watchdog.org>, linux-watchdog@...r.kernel.org,
 linux-kernel@...r.kernel.org
Cc: linux-stm32@...md-mailman.stormreply.com
Subject: Re: [PATCH] watchdog: arm_smc_wdt: get wdt status through
 SMCWD_GET_TIMELEFT

On 5/19/25 10:00, Antonio Borneo wrote:
> The optional SMCWD_GET_TIMELEFT command can be used to detect if
> the watchdog has already been started.
> See the implementation in OP-TEE secure OS [1].
> 
> If CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED is set, at probe time check
> if the watchdog is already started and then set WDOG_HW_RUNNING in
> the watchdog status. This will cause the watchdog framework to
> ping the watchdog until a userspace watchdog daemon takes over the
> control.
> 
> Link: https://github.com/OP-TEE/optee_os/commit/a7f2d4bd8632 [1]
> 
> Signed-off-by: Antonio Borneo <antonio.borneo@...s.st.com>
> ---
>   drivers/watchdog/arm_smc_wdt.c | 18 +++++++++++++++---
>   1 file changed, 15 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/watchdog/arm_smc_wdt.c b/drivers/watchdog/arm_smc_wdt.c
> index 8f3d0c3a005fb..f1268f43327ea 100644
> --- a/drivers/watchdog/arm_smc_wdt.c
> +++ b/drivers/watchdog/arm_smc_wdt.c
> @@ -46,6 +46,8 @@ static int smcwd_call(struct watchdog_device *wdd, enum smcwd_call call,
>   		return -ENODEV;
>   	if (res->a0 == PSCI_RET_INVALID_PARAMS)
>   		return -EINVAL;
> +	if (res->a0 == PSCI_RET_DISABLED)
> +		return -ENODATA;
>   	if (res->a0 != PSCI_RET_SUCCESS)
>   		return -EIO;
>   	return 0;
> @@ -131,10 +133,20 @@ static int smcwd_probe(struct platform_device *pdev)
>   
>   	wdd->info = &smcwd_info;
>   	/* get_timeleft is optional */
> -	if (smcwd_call(wdd, SMCWD_GET_TIMELEFT, 0, NULL))
> -		wdd->ops = &smcwd_ops;
> -	else
> +	err = smcwd_call(wdd, SMCWD_GET_TIMELEFT, 0, NULL);
> +	switch (err) {
> +	case 0:
> +		if (IS_ENABLED(CONFIG_WATCHDOG_HANDLE_BOOT_ENABLED))
> +			set_bit(WDOG_HW_RUNNING, &wdd->status);

This is the wrong use of this configuration option. It is only needed
in a driver if the watchdog status can not be read from hardware.
That is not the case here. Worse, using it in a driver like this
overrides the watchdog core module parameter "handle_boot_enabled".

Guenter

> +		fallthrough;
> +	case -ENODATA:
>   		wdd->ops = &smcwd_timeleft_ops;
> +		break;
> +	default:
> +		wdd->ops = &smcwd_ops;
> +		break;
> +	}
> +
>   	wdd->timeout = res.a2;
>   	wdd->max_timeout = res.a2;
>   	wdd->min_timeout = res.a1;
> 
> base-commit: a5806cd506af5a7c19bcd596e4708b5c464bfd21