On Tue, Jun 01, 2021 at 11:43:27AM +0200, nicolas saenz julienne wrote:

On Sun, 2021-05-30 at 11:26 +0200, Stefan Wahren wrote:

quoted

Until now all Raspberry Pi boards used the power off function of the SoC.
But the Raspberry Pi 400 uses gpio-poweroff for the whole board which
possibly cannot register the poweroff handler because the it's
already registered by this watchdog driver. So consider the
system-power-controller property for registering, which is already
defined in soc/bcm/brcm,bcm2835-pm.txt .

Signed-off-by: Stefan Wahren <redacted>
---
 drivers/watchdog/bcm2835_wdt.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/watchdog/bcm2835_wdt.c b/drivers/watchdog/bcm2835_wdt.c
index dec6ca0..9490717 100644
--- a/drivers/watchdog/bcm2835_wdt.c
+++ b/drivers/watchdog/bcm2835_wdt.c

@@ -205,9 +205,13 @@ static int bcm2835_wdt_probe(struct platform_device *pdev)

 	if (err)
 		return err;
 

-	if (pm_power_off == NULL) {
-		pm_power_off = bcm2835_power_off;
-		bcm2835_power_off_wdt = wdt;
+	if (of_device_is_system_power_controller(pdev->dev.parent->of_node)) {
+		if (!pm_power_off) {
+			pm_power_off = bcm2835_power_off;
+			bcm2835_power_off_wdt = wdt;
+		} else {
+			dev_info(dev, "Poweroff handler already present!\n");

nit: shoudln't this be a warning?

No. If anything, it should not be there in the first place, but I didn't want
to argue about it. It is perfectly valid for a system to have more than one
means to reset it, meaning this message is just noise. Making it a warning
would just make it worse.

Guenter

Other than that,

Reviewed-by: Nicolas Saenz Julienne <nsaenz@kernel.org>

Regards,
Nicolas

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel