Re: [PATCH v10 5/5] Watchdog: ARM SBSA Generic Watchdog half timeout panic support

[Mathieu Poirier]

On 3 February 2016 at 10:18,  <fu.wei@xxxxxxxxxx> wrote:
> From: Fu Wei <fu.wei@xxxxxxxxxx>
>
> This patch registers the WS0 interrupt routine to trigger panic,
> when the watchdog reachs the first stage (the half timeout).
> This function can help administrator to backup the system context
> info by panic console output or kdump (if supported), once system
> goes wrong (doesn't feed the watchdog in the half timeout).
>
> User also can skip panic by setting panic_enabled (module parameter) as 0
>
> Signed-off-by: Fu Wei <fu.wei@xxxxxxxxxx>
> ---
>  Documentation/watchdog/watchdog-parameters.txt |  1 +
>  drivers/watchdog/Kconfig                       | 10 +++++
>  drivers/watchdog/sbsa_gwdt.c                   | 54 +++++++++++++++++++++++---
>  3 files changed, 60 insertions(+), 5 deletions(-)
>
> diff --git a/Documentation/watchdog/watchdog-parameters.txt b/Documentation/watchdog/watchdog-parameters.txt
> index 300eb4d..31641e2 100644
> --- a/Documentation/watchdog/watchdog-parameters.txt
> +++ b/Documentation/watchdog/watchdog-parameters.txt
> @@ -286,6 +286,7 @@ nowayout: Watchdog cannot be stopped once started
>  -------------------------------------------------
>  sbsa_gwdt:
>  timeout: Watchdog timeout in seconds. (default 20s)
> +panic_enabled: Enable panic at half timeout. (default=true)
>  nowayout: Watchdog cannot be stopped once started
>         (default=kernel config parameter)
>  -------------------------------------------------
> diff --git a/drivers/watchdog/Kconfig b/drivers/watchdog/Kconfig
> index 4ab1b05..42adfdf 100644
> --- a/drivers/watchdog/Kconfig
> +++ b/drivers/watchdog/Kconfig
> @@ -218,6 +218,16 @@ config ARM_SBSA_WATCHDOG
>           To compile this driver as module, choose M here: The module
>           will be called sbsa_gwdt.
>
> +config ARM_SBSA_WATCHDOG_PANIC
> +       bool "ARM SBSA Generic Watchdog triggers panic at the half timeout"
> +       depends on ARM_SBSA_WATCHDOG
> +       help
> +         ARM SBSA Generic Watchdog will trigger panic in the first signal
> +         (WS0) interrupt routine when the half timeout is reached.
> +         This function can help administrator to backup the system context
> +         info by panic console output or kdump (if supported).
> +         But user can skip panic by setting moduleparam panic_enabled as 0.
> +
>  config ASM9260_WATCHDOG
>         tristate "Alphascale ASM9260 watchdog"
>         depends on MACH_ASM9260
> diff --git a/drivers/watchdog/sbsa_gwdt.c b/drivers/watchdog/sbsa_gwdt.c
> index 5a2dba3..d18cf37 100644
> --- a/drivers/watchdog/sbsa_gwdt.c
> +++ b/drivers/watchdog/sbsa_gwdt.c
> @@ -16,18 +16,22 @@
>   * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>   * GNU General Public License for more details.
>   *
> - * This SBSA Generic watchdog driver is a single stage timeout version.
> + * This SBSA Generic watchdog driver is a two stages version.
>   * Since this watchdog timer has two stages, and each stage is determined
>   * by WOR. So the timeout is (WOR * 2).
> - * When first timeout is reached, WS0 is triggered, the interrupt
> - * triggered by WS0 will be ignored, then the second watch period starts;
> - * when second timeout is reached, then WS1 is triggered, system reset.
> + * When the first stage(the half timeout) is reached, WS0 interrupt is
> + * triggered, at this moment the second watch period starts;
> + * In the WS0 interrupt routine, panic will be triggered for saving the
> + * system context.
> + * If the system is getting into trouble and cannot be reset by panic or
> + * restart properly by the kdump kernel(if supported), then the second
> + * stage (the timeout) will be reached, system will be reset by WS1.
>   *
>   * More details about the hardware specification of this device:
>   * ARM DEN0029B - Server Base System Architecture (SBSA)
>   *
>   * SBSA GWDT: |--------WOR-------WS0--------WOR-------WS1
> - *            |----------------timeout----------------reset
> + *            |--half_timeout--(panic)--half_timeout--reset
>   *
>   */
>
> @@ -84,6 +88,13 @@ MODULE_PARM_DESC(timeout,
>                  "Watchdog timeout in seconds. (>=0, default="
>                  __MODULE_STRING(DEFAULT_TIMEOUT) ")");
>
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +static bool panic_enabled = true;
> +module_param(panic_enabled, bool, 0);
> +MODULE_PARM_DESC(panic_enabled,
> +                "enable panic at half timeout. (default=true)");
> +#endif
> +
>  static bool nowayout = WATCHDOG_NOWAYOUT;
>  module_param(nowayout, bool, S_IRUGO);
>  MODULE_PARM_DESC(nowayout,
> @@ -159,6 +170,16 @@ static int sbsa_gwdt_stop(struct watchdog_device *wdd)
>         return 0;
>  }
>
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +static irqreturn_t sbsa_gwdt_interrupt(int irq, void *dev_id)
> +{
> +       if (panic_enabled)
> +               panic("SBSA Watchdog half timeout");
> +
> +       return IRQ_HANDLED;
> +}
> +#endif
> +
>  static struct watchdog_info sbsa_gwdt_info = {
>         .identity       = "SBSA Generic Watchdog",
>         .options        = WDIOF_SETTIMEOUT |
> @@ -186,6 +207,9 @@ static int sbsa_gwdt_probe(struct platform_device *pdev)
>         struct resource *res;
>         u32 status;
>         int ret;
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +       int irq;
> +#endif
>
>         gwdt = devm_kzalloc(dev, sizeof(*gwdt), GFP_KERNEL);
>         if (!gwdt)
> @@ -202,6 +226,14 @@ static int sbsa_gwdt_probe(struct platform_device *pdev)
>         if (IS_ERR(rf_base))
>                 return PTR_ERR(rf_base);
>
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +       irq = platform_get_irq(pdev, 0);
> +       if (irq < 0) {
> +               dev_err(dev, "unable to get ws0 interrupt.\n");
> +               return irq;
> +       }
> +#endif
> +

Can't the driver revert to single stage mode if platform_get_irq()
fails?  That way the value of 'irq' can be tested throughout the
_probe() function and the #ifdefs removed.

Thanks,
Mathieu

>         /*
>          * Get the frequency of system counter from the cp15 interface of ARM
>          * Generic timer. We don't need to check it, because if it returns "0",
> @@ -228,6 +260,14 @@ static int sbsa_gwdt_probe(struct platform_device *pdev)
>                 dev_warn(dev, "System reset by WDT.\n");
>                 wdd->bootstatus |= WDIOF_CARDRESET;
>         }
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +       ret = devm_request_irq(dev, irq, sbsa_gwdt_interrupt, 0,
> +                              pdev->name, gwdt);
> +       if (ret) {
> +               dev_err(dev, "unable to request IRQ %d\n", irq);
> +               return ret;
> +       }
> +#endif
>
>         ret = watchdog_register_device(wdd);
>         if (ret)
> @@ -242,6 +282,10 @@ static int sbsa_gwdt_probe(struct platform_device *pdev)
>
>         dev_info(dev, "Initialized with %ds timeout @ %u Hz%s\n", wdd->timeout,
>                  gwdt->clk, status & SBSA_GWDT_WCS_EN ? " [enabled]" : "");
> +#ifdef CONFIG_ARM_SBSA_WATCHDOG_PANIC
> +       dev_info(dev, "Half timeout panic %s.\n",
> +                panic_enabled ? "enabled" : "disabled");
> +#endif
>
>         return 0;
>  }
> --
> 2.5.0
>