[PATCH] watchdog: imx2_wdt: Fix ioctl() results

[PATCH] watchdog: imx2_wdt: Fix ioctl() results

[Markus Rinne]

WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT returned the initial timeout and
not the one that was last set.  Fix this by updating struct
watchdog_device member 'timeout'.  This is how it's supposed to be done
according to Documentation/watchdog/watchdog-kernel-api.txt.

This is the test case I used:

#include <fcntl.h>
#include <linux/watchdog.h>
#include <sys/stat.h>
#include <sys/types.h>

#include <assert.h>
#include <stdlib.h>

static const int TIMEOUT = 127;

int main(void)
{
	int fd = open("/dev/watchdog", O_WRONLY);
	if (fd == -1)
		return EXIT_FAILURE;

	int timeout = TIMEOUT;
	ioctl(fd, WDIOC_SETTIMEOUT, &timeout);
	assert(timeout == TIMEOUT);

	close(fd);

	return EXIT_SUCCESS;
}

Signed-off-by: Markus Rinne <markus.rinne@vincit.fi>
---
 drivers/watchdog/imx2_wdt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/watchdog/imx2_wdt.c b/drivers/watchdog/imx2_wdt.c
index 5e6d808..b636799 100644
--- a/drivers/watchdog/imx2_wdt.c
+++ b/drivers/watchdog/imx2_wdt.c
@@ -168,6 +168,7 @@ static int imx2_wdt_set_timeout(struct watchdog_device *wdog,
 
 	regmap_update_bits(wdev->regmap, IMX2_WDT_WCR, IMX2_WDT_WCR_WT,
 			   WDOG_SEC_TO_COUNT(new_timeout));
+	wdog->timeout = new_timeout;
 	return 0;
 }
 
-- 
1.8.4

Re: [PATCH] watchdog: imx2_wdt: Fix ioctl() results

[Guenter Roeck]

On 06/05/2015 05:41 AM, Markus Rinne wrote:
> WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT returned the initial timeout and
> not the one that was last set.  Fix this by updating struct
> watchdog_device member 'timeout'.  This is how it's supposed to be done
> according to Documentation/watchdog/watchdog-kernel-api.txt.
>
> This is the test case I used:
>
> #include <fcntl.h>
> #include <linux/watchdog.h>
> #include <sys/stat.h>
> #include <sys/types.h>
>
> #include <assert.h>
> #include <stdlib.h>
>
> static const int TIMEOUT = 127;
>
> int main(void)
> {
> 	int fd = open("/dev/watchdog", O_WRONLY);
> 	if (fd == -1)
> 		return EXIT_FAILURE;
>
> 	int timeout = TIMEOUT;
> 	ioctl(fd, WDIOC_SETTIMEOUT, &timeout);
> 	assert(timeout == TIMEOUT);
>
> 	close(fd);
>
> 	return EXIT_SUCCESS;
> }
>
> Signed-off-by: Markus Rinne <markus.rinne@vincit.fi>
> ---

Good catch.

Reviewed-by: Guenter Roeck <linux@roeck-us.net>

Re: [PATCH] watchdog: imx2_wdt: Fix ioctl() results

[Vladimir Zapolskiy]

Hello Markus,

On 05.06.2015 15:41, Markus Rinne wrote:
> WDIOC_SETTIMEOUT and WDIOC_GETTIMEOUT returned the initial timeout and
> not the one that was last set.  Fix this by updating struct
> watchdog_device member 'timeout'.  This is how it's supposed to be done
> according to Documentation/watchdog/watchdog-kernel-api.txt.
> 
> This is the test case I used:
> 
> #include <fcntl.h>
> #include <linux/watchdog.h>
> #include <sys/stat.h>
> #include <sys/types.h>
> 
> #include <assert.h>
> #include <stdlib.h>
> 
> static const int TIMEOUT = 127;
> 
> int main(void)
> {
> 	int fd = open("/dev/watchdog", O_WRONLY);
> 	if (fd == -1)
> 		return EXIT_FAILURE;
> 
> 	int timeout = TIMEOUT;
> 	ioctl(fd, WDIOC_SETTIMEOUT, &timeout);
> 	assert(timeout == TIMEOUT);
> 
> 	close(fd);
> 
> 	return EXIT_SUCCESS;
> }
> 
> Signed-off-by: Markus Rinne <markus.rinne@vincit.fi>
> ---
>  drivers/watchdog/imx2_wdt.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/drivers/watchdog/imx2_wdt.c b/drivers/watchdog/imx2_wdt.c
> index 5e6d808..b636799 100644
> --- a/drivers/watchdog/imx2_wdt.c
> +++ b/drivers/watchdog/imx2_wdt.c
> @@ -168,6 +168,7 @@ static int imx2_wdt_set_timeout(struct watchdog_device *wdog,
>  
>  	regmap_update_bits(wdev->regmap, IMX2_WDT_WCR, IMX2_WDT_WCR_WT,
>  			   WDOG_SEC_TO_COUNT(new_timeout));
> +	wdog->timeout = new_timeout;
>  	return 0;
>  }
>  
> 

this change repeats the change sent by Michael Grzeschik one month ago
-- http://www.spinics.net/lists/linux-watchdog/msg06296.html

--
With best wishes,
Vladimir

Re: [PATCH] watchdog: imx2_wdt: Fix ioctl() results

[Guenter Roeck]

On 06/05/2015 07:14 AM, Vladimir Zapolskiy wrote:
> Hello Markus,
>
[ ... ]
>
> this change repeats the change sent by Michael Grzeschik one month ago
> -- http://www.spinics.net/lists/linux-watchdog/msg06296.html
>

... and I even have it in my list of patches.

Thanks a lot for noticing.

Guenter