linux-kernel - Re: [v4, 2/2] drivers/watchdog: ASPEED reference dev tree properties for config

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: <20170708145530.GA3436@roeck-us.net>
Date:   Sat, 8 Jul 2017 07:55:30 -0700
From:   Guenter Roeck <linux@...ck-us.net>
To:     Christopher Bostic <cbostic@...ux.vnet.ibm.com>
Cc:     wim@...ana.be, robh+dt@...nel.org, mark.rutland@....com,
        joel@....id.au, linux-watchdog@...r.kernel.org,
        devicetree@...r.kernel.org, linux-kernel@...r.kernel.org,
        openbmc@...ts.ozlabs.org
Subject: Re: [v4, 2/2] drivers/watchdog: ASPEED reference dev tree properties
 for config

On Thu, Jul 06, 2017 at 07:49:00PM -0500, Christopher Bostic wrote:
> Reference the system device tree when configuring the watchdog
> engines. If property 'aspeed,reset_type' is present then set
> reset behavior based on the specified value.  This can be one of
> three different mutually exclusive values
>   * cpu - Reset CPU only on watchdog timeout
>   * soc - Reset System on Chip
>   * system - Full system reset
> 
> No reset can also be specified by indicating:
>   * none - No reset, assumes another watchdog is responsible for
>            this.
> 
> Add optional property 'aspeed,external-signal'. If present then
> configure to generate external signal on watchdog timeout.
> 
> Signed-off-by: Christopher Bostic <cbostic@...ux.vnet.ibm.com>
> ---
> v4 - Change the three reset type parameters to a new property
>      'aspeed,reset_type' and check assignment for one of four
>      different values, cpu, soc, system, none
> v3 - Invert the logic for system reset dev tree property to
>      preserve backwards compatibility. If not specified the
>      default is to configure for system reset
>    - Add check for 'aspeed,no-soc-reset' property and only if
>      not present is SOC reset to be configured.  This preserves
>      backwards compatibility.
> v2 - Change of_get_property() to of_property_read_bool()
>    - Remove redundant check for NULL struct device_node pointer
>    - Optional property names now start with prefix 'aspeed,'
> ---
>  drivers/watchdog/aspeed_wdt.c | 27 ++++++++++++++++++++++-----
>  1 file changed, 22 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/watchdog/aspeed_wdt.c b/drivers/watchdog/aspeed_wdt.c
> index 1c65258..3ca79565 100644
> --- a/drivers/watchdog/aspeed_wdt.c
> +++ b/drivers/watchdog/aspeed_wdt.c
> @@ -36,6 +36,7 @@ struct aspeed_wdt {
>  #define WDT_CTRL		0x0C
>  #define   WDT_CTRL_RESET_MODE_SOC	(0x00 << 5)
>  #define   WDT_CTRL_RESET_MODE_FULL_CHIP	(0x01 << 5)
> +#define   WDT_CTRL_RESET_MODE_ARM_CPU	(0x10 << 5)
>  #define   WDT_CTRL_1MHZ_CLK		BIT(4)
>  #define   WDT_CTRL_WDT_EXT		BIT(3)
>  #define   WDT_CTRL_WDT_INTR		BIT(2)
> @@ -140,6 +141,8 @@ static int aspeed_wdt_probe(struct platform_device *pdev)
>  {
>  	struct aspeed_wdt *wdt;
>  	struct resource *res;
> +	struct device_node *np;
> +	const char *reset_type;
>  	int ret;
>  
>  	wdt = devm_kzalloc(&pdev->dev, sizeof(*wdt), GFP_KERNEL);
> @@ -164,14 +167,28 @@ static int aspeed_wdt_probe(struct platform_device *pdev)
>  	wdt->wdd.timeout = WDT_DEFAULT_TIMEOUT;
>  	watchdog_init_timeout(&wdt->wdd, 0, &pdev->dev);
>  
> +	wdt->ctrl = WDT_CTRL_1MHZ_CLK;
> +
>  	/*
>  	 * Control reset on a per-device basis to ensure the
> -	 * host is not affected by a BMC reboot, so only reset
> -	 * the SOC and not the full chip
> +	 * host is not affected by a BMC reboot
>  	 */
> -	wdt->ctrl = WDT_CTRL_RESET_MODE_SOC |
> -		WDT_CTRL_1MHZ_CLK |
> -		WDT_CTRL_RESET_SYSTEM;
> +	np = pdev->dev.of_node;
> +	ret = of_property_read_string(np, "aspeed,reset-type", &reset_type);
> +	if (ret) {
> +		wdt->ctrl |= WDT_CTRL_RESET_SYSTEM;
> +	} else {
> +		if (!strcmp(reset_type, "cpu"))
> +			wdt->ctrl |= WDT_CTRL_RESET_MODE_ARM_CPU;
> +		else if (!strcmp(reset_type, "soc"))
> +			wdt->ctrl |= WDT_CTRL_RESET_MODE_SOC; 
> +		else if (!strcmp(reset_type, "system"))
> +			wdt->ctrl |= WDT_CTRL_RESET_SYSTEM;

This silently accepts 'aspeed,reset-type="junk"' as "none".
I think it would be better to explicitly check for "none".

Also, if I remember the datasheet correctly, bit 1 (WDT_CTRL_RESET_SYSTEM)
enables the reset in general, and bit 5/6 specify the reset type.
With that in mind, I don't think the above will work as expected.
I think it would have to be

	WDT_CTRL_RESET_MODE_FULL_CHIP | WDT_CTRL_RESET_SYSTEM
	WDT_CTRL_RESET_MODE_SOC | WDT_CTRL_RESET_SYSTEM
	WDT_CTRL_RESET_MODE_ARM_CPU | WDT_CTRL_RESET_SYSTEM

To match the original code, the default should be

	WDT_CTRL_RESET_MODE_SOC | WDT_CTRL_RESET_SYSTEM

> +	}
> +	if (of_property_read_bool(np, "aspeed,external-signal"))
> +		wdt->ctrl |= WDT_CTRL_WDT_EXT;
> +
No plan to support the other new configuration flags ?

> +	writel(wdt->ctrl, wdt->base + WDT_CTRL);
>  
>  	if (readl(wdt->base + WDT_CTRL) & WDT_CTRL_ENABLE)  {
>  		aspeed_wdt_start(&wdt->wdd);