diff mbox series

[04/12] watchdog: Allow pretimeout governor setting to be accessed from modules

Message ID 20190819203711.32599-5-minyard@acm.org (mailing list archive)
State Changes Requested
Headers show
Series [01/12] watchdog: NULL the default governor if it is unregistered | expand

Commit Message

Corey Minyard Aug. 19, 2019, 8:37 p.m. UTC
From: Corey Minyard <cminyard@mvista.com>

This is so watchdog driver (like IPMI) can set it.

Signed-off-by: Corey Minyard <cminyard@mvista.com>
---
 drivers/watchdog/watchdog_pretimeout.c | 1 +
 1 file changed, 1 insertion(+)

Comments

Guenter Roeck Aug. 19, 2019, 9:49 p.m. UTC | #1
On Mon, Aug 19, 2019 at 03:37:03PM -0500, minyard@acm.org wrote:
> From: Corey Minyard <cminyard@mvista.com>
> 
> This is so watchdog driver (like IPMI) can set it.
> 
> Signed-off-by: Corey Minyard <cminyard@mvista.com>
> ---
>  drivers/watchdog/watchdog_pretimeout.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/watchdog/watchdog_pretimeout.c b/drivers/watchdog/watchdog_pretimeout.c
> index b45041b0ef39..270baf7b3fa0 100644
> --- a/drivers/watchdog/watchdog_pretimeout.c
> +++ b/drivers/watchdog/watchdog_pretimeout.c
> @@ -95,6 +95,7 @@ int watchdog_pretimeout_governor_set(struct watchdog_device *wdd,
>  
>  	return 0;
>  }
> +EXPORT_SYMBOL_GPL(watchdog_pretimeout_governor_set);
>  

I don't think that is a good idea. The whole point of pretimeout governor
selection was to be able to configure it from userspace.

>  void watchdog_notify_pretimeout(struct watchdog_device *wdd)
>  {
> -- 
> 2.17.1
>
Corey Minyard Aug. 20, 2019, 12:24 a.m. UTC | #2
On Mon, Aug 19, 2019 at 02:49:53PM -0700, Guenter Roeck wrote:
> On Mon, Aug 19, 2019 at 03:37:03PM -0500, minyard@acm.org wrote:
> > From: Corey Minyard <cminyard@mvista.com>
> > 
> > This is so watchdog driver (like IPMI) can set it.
> > 
> > Signed-off-by: Corey Minyard <cminyard@mvista.com>
> > ---
> >  drivers/watchdog/watchdog_pretimeout.c | 1 +
> >  1 file changed, 1 insertion(+)
> > 
> > diff --git a/drivers/watchdog/watchdog_pretimeout.c b/drivers/watchdog/watchdog_pretimeout.c
> > index b45041b0ef39..270baf7b3fa0 100644
> > --- a/drivers/watchdog/watchdog_pretimeout.c
> > +++ b/drivers/watchdog/watchdog_pretimeout.c
> > @@ -95,6 +95,7 @@ int watchdog_pretimeout_governor_set(struct watchdog_device *wdd,
> >  
> >  	return 0;
> >  }
> > +EXPORT_SYMBOL_GPL(watchdog_pretimeout_governor_set);
> >  
> 
> I don't think that is a good idea. The whole point of pretimeout governor
> selection was to be able to configure it from userspace.

Yeah, this is really just a temporary thing.  There is a module parameter in
the IPMI watchdog that does basically the same thing as the device parameter
for setting the watchdog governor.  This is so that code can perform its
function.

I would expect this to go away eventually.

-corey

> 
> >  void watchdog_notify_pretimeout(struct watchdog_device *wdd)
> >  {
> > -- 
> > 2.17.1
> >
diff mbox series

Patch

diff --git a/drivers/watchdog/watchdog_pretimeout.c b/drivers/watchdog/watchdog_pretimeout.c
index b45041b0ef39..270baf7b3fa0 100644
--- a/drivers/watchdog/watchdog_pretimeout.c
+++ b/drivers/watchdog/watchdog_pretimeout.c
@@ -95,6 +95,7 @@  int watchdog_pretimeout_governor_set(struct watchdog_device *wdd,
 
 	return 0;
 }
+EXPORT_SYMBOL_GPL(watchdog_pretimeout_governor_set);
 
 void watchdog_notify_pretimeout(struct watchdog_device *wdd)
 {