On 12/4/18 4:01 AM, Wolfram Sang wrote:
After discussing this mail thread [1] again, we concluded that giving
userspace enough time to prepare is our favourite option. So, do not
keep the time value when suspended but reset it when resuming.
[1] https://patchwork.kernel.org/patch/10252209/
Signed-off-by: Wolfram Sang <wsa+renesas@xxxxxxxxxxxxxxxxxxxx>
Above exchange says it all, no need to repeat.
Reviewed-by: Guenter Roeck <linux@xxxxxxxxxxxx>
---
Fabrizio: can you agree to that? The R-Car BSP team and we (the R-Car upstream
team) would prefer it this way (knowing it is also not perfect).
drivers/watchdog/renesas_wdt.c | 11 ++++-------
1 file changed, 4 insertions(+), 7 deletions(-)
diff --git a/drivers/watchdog/renesas_wdt.c b/drivers/watchdog/renesas_wdt.c
index b570962e84f3..9f2307bf727b 100644
--- a/drivers/watchdog/renesas_wdt.c
+++ b/drivers/watchdog/renesas_wdt.c
@@ -48,7 +48,6 @@ struct rwdt_priv {
void __iomem *base;
struct watchdog_device wdev;
unsigned long clk_rate;
- u16 time_left;
u8 cks;
};
@@ -263,10 +262,9 @@ static int __maybe_unused rwdt_suspend(struct device *dev)
{
struct rwdt_priv *priv = dev_get_drvdata(dev);
- if (watchdog_active(&priv->wdev)) {
- priv->time_left = readw(priv->base + RWTCNT);
+ if (watchdog_active(&priv->wdev))
rwdt_stop(&priv->wdev);
- }
+
return 0;
}
@@ -274,10 +272,9 @@ static int __maybe_unused rwdt_resume(struct device *dev)
{
struct rwdt_priv *priv = dev_get_drvdata(dev);
- if (watchdog_active(&priv->wdev)) {
+ if (watchdog_active(&priv->wdev))
rwdt_start(&priv->wdev);
- rwdt_write(priv, priv->time_left, RWTCNT);
- }
+
return 0;
}
