Linux-Watchdog Archive on lore.kernel.org
 help / color / Atom feed
From: Guenter Roeck <linux@roeck-us.net>
To: Woody Lin <woodylin@google.com>,
	Wim Van Sebroeck <wim@linux-watchdog.org>
Cc: linux-watchdog@vger.kernel.org
Subject: Re: [PATCH v3] softdog: Add options 'soft_reboot_cmd' and 'soft_active_on_boot'
Date: Wed, 1 Jul 2020 07:10:35 -0700
Message-ID: <d2ea870d-a9d0-fe8f-c11b-aa54750f1d0d@roeck-us.net> (raw)
In-Reply-To: <20200701110339.2833449-1-woodylin@google.com>

On 7/1/20 4:03 AM, Woody Lin wrote:
> Add module parameters 'soft_reboot_cmd' and 'soft_active_on_boot' for
> customizing softdog configuration; config reboot command by assigning
> soft_reboot_cmd, and set soft_active_on_boot to start up softdog
> timer at module initialization stage.
> 
> Signed-off-by: Woody Lin <woodylin@google.com>

Sigh. Now I'll have to look up old versions and compare to figure out
changes from v2. Why do people always believe that kernel maintainers
have endless amounts of time available ?

[ Sorry, yes, I know, we are supposed to remain friendly all the time.
  It is just _so_ frustrating to have to deal with this all the time. ]

Guenter

> ---
>  drivers/watchdog/softdog.c | 56 ++++++++++++++++++++++++++++++++++++++
>  1 file changed, 56 insertions(+)
> 
> diff --git a/drivers/watchdog/softdog.c b/drivers/watchdog/softdog.c
> index 3e4885c1545e..8c8d214b6aa7 100644
> --- a/drivers/watchdog/softdog.c
> +++ b/drivers/watchdog/softdog.c
> @@ -20,11 +20,13 @@
>  #include <linux/hrtimer.h>
>  #include <linux/init.h>
>  #include <linux/kernel.h>
> +#include <linux/kthread.h>
>  #include <linux/module.h>
>  #include <linux/moduleparam.h>
>  #include <linux/reboot.h>
>  #include <linux/types.h>
>  #include <linux/watchdog.h>
> +#include <linux/workqueue.h>
>  
>  #define TIMER_MARGIN	60		/* Default is 60 seconds */
>  static unsigned int soft_margin = TIMER_MARGIN;	/* in seconds */
> @@ -49,11 +51,33 @@ module_param(soft_panic, int, 0);
>  MODULE_PARM_DESC(soft_panic,
>  	"Softdog action, set to 1 to panic, 0 to reboot (default=0)");
>  
> +static char *soft_reboot_cmd;
> +module_param(soft_reboot_cmd, charp, 0000);
> +MODULE_PARM_DESC(soft_reboot_cmd,
> +	"Set reboot command. Emergency reboot takes place if unset");
> +
> +static bool soft_active_on_boot;
> +module_param(soft_active_on_boot, bool, 0000);
> +MODULE_PARM_DESC(soft_active_on_boot,
> +	"Set to true to active Softdog on boot (default=false)");
> +
>  static struct hrtimer softdog_ticktock;
>  static struct hrtimer softdog_preticktock;
>  
> +static int reboot_kthread_fn(void *data)
> +{
> +	kernel_restart(soft_reboot_cmd);
> +	return -EPERM; /* Should not reach here */
> +}
> +
> +static void reboot_work_fn(struct work_struct *unused)
> +{
> +	kthread_run(reboot_kthread_fn, NULL, "softdog_reboot");
> +}
> +
>  static enum hrtimer_restart softdog_fire(struct hrtimer *timer)
>  {
> +	static bool soft_reboot_fired;
>  	module_put(THIS_MODULE);
>  	if (soft_noboot) {
>  		pr_crit("Triggered - Reboot ignored\n");
> @@ -62,6 +86,33 @@ static enum hrtimer_restart softdog_fire(struct hrtimer *timer)
>  		panic("Software Watchdog Timer expired");
>  	} else {
>  		pr_crit("Initiating system reboot\n");
> +		if (!soft_reboot_fired && soft_reboot_cmd != NULL) {
> +			static DECLARE_WORK(reboot_work, reboot_work_fn);
> +			/*
> +			 * The 'kernel_restart' is a 'might-sleep' operation.
> +			 * Also, executing it in system-wide workqueues blocks
> +			 * any driver from using the same workqueue in its
> +			 * shutdown callback function. Thus, we should execute
> +			 * the 'kernel_restart' in a standalone kernel thread.
> +			 * But since starting a kernel thread is also a
> +			 * 'might-sleep' operation, so the 'reboot_work' is
> +			 * required as a launcher of the kernel thread.
> +			 *
> +			 * After request the reboot, restart the timer to
> +			 * schedule an 'emergency_restart' reboot after
> +			 * 'TIMER_MARGIN' seconds. It's because if the softdog
> +			 * hangs, it might be because of scheduling issues. And
> +			 * if that is the case, both 'schedule_work' and
> +			 * 'kernel_restart' may possibly be malfunctional at the
> +			 * same time.
> +			 */
> +			soft_reboot_fired = true;
> +			schedule_work(&reboot_work);
> +			hrtimer_add_expires_ns(timer,
> +					(u64)TIMER_MARGIN * NSEC_PER_SEC);
> +
> +			return HRTIMER_RESTART;
> +		}
>  		emergency_restart();
>  		pr_crit("Reboot didn't ?????\n");
>  	}
> @@ -145,12 +196,17 @@ static int __init softdog_init(void)
>  		softdog_preticktock.function = softdog_pretimeout;
>  	}
>  
> +	if (soft_active_on_boot)
> +		softdog_ping(&softdog_dev);
> +
>  	ret = watchdog_register_device(&softdog_dev);
>  	if (ret)
>  		return ret;
>  
>  	pr_info("initialized. soft_noboot=%d soft_margin=%d sec soft_panic=%d (nowayout=%d)\n",
>  		soft_noboot, softdog_dev.timeout, soft_panic, nowayout);
> +	pr_info("             soft_reboot_cmd=%s soft_active_on_boot=%d\n",
> +		soft_reboot_cmd, soft_active_on_boot);
>  
>  	return 0;
>  }
> 


  reply index

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-07-01 11:03 Woody Lin
2020-07-01 14:10 ` Guenter Roeck [this message]
2020-07-01 14:32   ` Woody Lin
2020-07-02 15:07     ` Guenter Roeck
2020-07-07  4:03 ` Guenter Roeck
2020-07-07  8:46   ` Woody Lin
2020-07-07 11:47     ` Guenter Roeck
2020-07-08  7:17       ` Woody Lin

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=d2ea870d-a9d0-fe8f-c11b-aa54750f1d0d@roeck-us.net \
    --to=linux@roeck-us.net \
    --cc=linux-watchdog@vger.kernel.org \
    --cc=wim@linux-watchdog.org \
    --cc=woodylin@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

Linux-Watchdog Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/linux-watchdog/0 linux-watchdog/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 linux-watchdog linux-watchdog/ https://lore.kernel.org/linux-watchdog \
		linux-watchdog@vger.kernel.org
	public-inbox-index linux-watchdog

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.linux-watchdog


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git