On Thu, Dec 22, 2022 at 11:43:31PM +0000, Thomas Weißschuh wrote:
> As per the specification the action QUERY_COUNTDOWN_PERIOD is optional.
> If the action is not implemented by the physical device the driver would
> always report "0" from get_timeleft().
> Avoid confusing userspace by only providing get_timeleft() when
> implemented by the hardware.
>
> Signed-off-by: Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
Assuming that there can be only one of those watchdogs in the system,
Reviewed-by: Guenter Roeck <linux@xxxxxxxxxxxx>
Guenter
> ---
> drivers/watchdog/wdat_wdt.c | 6 ++++--
> 1 file changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/watchdog/wdat_wdt.c b/drivers/watchdog/wdat_wdt.c
> index ce7a4a9e4b03..0ba99bed59fc 100644
> --- a/drivers/watchdog/wdat_wdt.c
> +++ b/drivers/watchdog/wdat_wdt.c
> @@ -301,13 +301,12 @@ static const struct watchdog_info wdat_wdt_info = {
> .identity = "wdat_wdt",
> };
>
> -static const struct watchdog_ops wdat_wdt_ops = {
> +static struct watchdog_ops wdat_wdt_ops = {
> .owner = THIS_MODULE,
> .start = wdat_wdt_start,
> .stop = wdat_wdt_stop,
> .ping = wdat_wdt_ping,
> .set_timeout = wdat_wdt_set_timeout,
> - .get_timeleft = wdat_wdt_get_timeleft,
> };
>
> static int wdat_wdt_probe(struct platform_device *pdev)
> @@ -436,6 +435,9 @@ static int wdat_wdt_probe(struct platform_device *pdev)
> list_add_tail(&instr->node, instructions);
> }
>
> + if (wdat->instructions[ACPI_WDAT_GET_CURRENT_COUNTDOWN])
> + wdat_wdt_ops.get_timeleft = wdat_wdt_get_timeleft;
> +
> wdat_wdt_boot_status(wdat);
> wdat_wdt_set_running(wdat);
>
>
> ---
> base-commit: 0a924817d2ed9396401e0557c6134276d2e26382
> change-id: 20221221-wdat_wdt-timeleft-b38d0db39d9f
>
> Best regards,
> --
> Thomas Weißschuh <linux@xxxxxxxxxxxxxx>
