Call runtime_pm_put*() if watchdog is not already started during probe and re
enable it in watchdog start as required.
On K3 SoCs, watchdogs and their corresponding CPUs are under same
power-domain, so if the reference count of unused watchdogs aren't
dropped, it will lead to CPU hotplug failures as Device Management
firmware won't allow to turn off the power-domain due to dangling
reference count.
Fixes: 2d63908bdbfb ("watchdog: Add K3 RTI watchdog support")
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Tested-by: Manorit Chawdhry <m-chawdhry@ti.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Link: https://lore.kernel.org/r/20231213140110.938129-1-vigneshr@ti.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Wim Van Sebroeck <wim@linux-watchdog.org>
 {
        u32 timer_margin;
        struct rti_wdt_device *wdt = watchdog_get_drvdata(wdd);
+       int ret;
+
+       ret = pm_runtime_resume_and_get(wdd->parent);
+       if (ret)
+               return ret;
 
        /* set timeout period */
        timer_margin = (u64)wdd->timeout * wdt->freq;
        if (last_ping)
                watchdog_set_last_hw_keepalive(wdd, last_ping);
 
+       if (!watchdog_hw_running(wdd))
+               pm_runtime_put_sync(&pdev->dev);
+
        return 0;
 
 err_iomap:
        struct rti_wdt_device *wdt = platform_get_drvdata(pdev);
 
        watchdog_unregister_device(&wdt->wdd);
-       pm_runtime_put(&pdev->dev);
+
+       if (!pm_runtime_suspended(&pdev->dev))
+               pm_runtime_put(&pdev->dev);
+
        pm_runtime_disable(&pdev->dev);
 }