Hi Florian, Am 04.07.24 um 16:12 schrieb Florian Fainelli:
On 6/30/2024 5:53 PM, Stefan Wahren wrote:This adds suspend/resume support for the 8250_bcm2835aux driver to provide power management support on attached devices. Signed-off-by: Stefan Wahren <wahrenst@xxxxxxx> --- Since i don't have a RS485 setup, any test feedback would be great. drivers/tty/serial/8250/8250_bcm2835aux.c | 23 +++++++++++++++++++++++ 1 file changed, 23 insertions(+) diff --git a/drivers/tty/serial/8250/8250_bcm2835aux.c b/drivers/tty/serial/8250/8250_bcm2835aux.c index 121a5ce86050..cccd2a09cb6f 100644 --- a/drivers/tty/serial/8250/8250_bcm2835aux.c +++ b/drivers/tty/serial/8250/8250_bcm2835aux.c @@ -213,11 +213,34 @@ static const struct acpi_device_id bcm2835aux_serial_acpi_match[] = { }; MODULE_DEVICE_TABLE(acpi, bcm2835aux_serial_acpi_match); +static int __maybe_unused bcm2835aux_suspend(struct device *dev) +{ + struct bcm2835aux_data *data = dev_get_drvdata(dev); + + serial8250_suspend_port(data->line);Don't you also need to disable the clock here, unless the device is a wake-up source, and conversely re-enable the clock upon resume?
at first I experiment with the pm implementation from 8250_uniphier.c, but this didn't work as soon as I drop "no_console_suspend" from the Kernel cmdline. Maybe that's the wrong pattern.
