[PATCH] watchdog: ebc-c384_wdt: Simplify ebc_c384_wdt_start timeout computation

[PATCH] watchdog: ebc-c384_wdt: Simplify ebc_c384_wdt_start timeout computation

[William Breathitt Gray]

The ebc_c384_wdt_set_timeout function already rounds up timeout values
greater than 255 to a multiple of 60, so there is no need to round again
in the ebc_c384_wdt_start function; a timeout value greater than 255 in
ebc_c384_wdt_start will divide cleanly by 60.

Signed-off-by: William Breathitt Gray <vilhelm.gray@gmail.com>
---
 drivers/watchdog/ebc-c384_wdt.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/watchdog/ebc-c384_wdt.c b/drivers/watchdog/ebc-c384_wdt.c
index 77fda0b..948e687 100644
--- a/drivers/watchdog/ebc-c384_wdt.c
+++ b/drivers/watchdog/ebc-c384_wdt.c
@@ -51,7 +51,7 @@ static int ebc_c384_wdt_start(struct watchdog_device *wdev)
 
 	/* resolution is in minutes for timeouts greater than 255 seconds */
 	if (t > 255)
-		t = DIV_ROUND_UP(t, 60);
+		t /= 60;
 
 	outb(t, PET_ADDR);
 
-- 
2.7.3

Re: [PATCH] watchdog: ebc-c384_wdt: Simplify ebc_c384_wdt_start timeout computation

[Guenter Roeck]

On Wed, May 11, 2016 at 09:05:51AM -0400, William Breathitt Gray wrote:
> The ebc_c384_wdt_set_timeout function already rounds up timeout values
> greater than 255 to a multiple of 60, so there is no need to round again
> in the ebc_c384_wdt_start function; a timeout value greater than 255 in
> ebc_c384_wdt_start will divide cleanly by 60.
> 
... except if the timeout is set with the module parameter, which is now
rounded down, which would be a problem since it would time out earlier than
expected (a timeout of 299 seconds, set with the module parameter, would
time out after 240 seconds). So you would have to do some calculations
on the timeout value set with the module parameter to ensure that it is
set to a correct value.

Guenter

> Signed-off-by: William Breathitt Gray <vilhelm.gray@gmail.com>
> ---
>  drivers/watchdog/ebc-c384_wdt.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/watchdog/ebc-c384_wdt.c b/drivers/watchdog/ebc-c384_wdt.c
> index 77fda0b..948e687 100644
> --- a/drivers/watchdog/ebc-c384_wdt.c
> +++ b/drivers/watchdog/ebc-c384_wdt.c
> @@ -51,7 +51,7 @@ static int ebc_c384_wdt_start(struct watchdog_device *wdev)
>  
>  	/* resolution is in minutes for timeouts greater than 255 seconds */
>  	if (t > 255)
> -		t = DIV_ROUND_UP(t, 60);
> +		t /= 60;
>  
>  	outb(t, PET_ADDR);
>  
> -- 
> 2.7.3
> 

Re: [PATCH] watchdog: ebc-c384_wdt: Simplify ebc_c384_wdt_start timeout computation

[William Breathitt Gray]

On Wed, May 11, 2016 at 06:43:27AM -0700, Guenter Roeck wrote:
>On Wed, May 11, 2016 at 09:05:51AM -0400, William Breathitt Gray wrote:
>> The ebc_c384_wdt_set_timeout function already rounds up timeout values
>> greater than 255 to a multiple of 60, so there is no need to round again
>> in the ebc_c384_wdt_start function; a timeout value greater than 255 in
>> ebc_c384_wdt_start will divide cleanly by 60.
>> 
>... except if the timeout is set with the module parameter, which is now
>rounded down, which would be a problem since it would time out earlier than
>expected (a timeout of 299 seconds, set with the module parameter, would
>time out after 240 seconds). So you would have to do some calculations
>on the timeout value set with the module parameter to ensure that it is
>set to a correct value.
>
>Guenter

Oops, I overlooked that scenario. Please ignore this patch as your are
correct that a timeout set via module parameter would not necessarily be
a multiple of 60; timeout in the start callback should indeed be
rounded.

William Breathitt Gray