On Tue, Feb 13, 2024 at 11:00:43AM +0100, Javier Carrasco wrote:
> On 06.02.24 19:40, Matthias Kaehlcke wrote:
> > On Tue, Feb 06, 2024 at 02:59:30PM +0100, Javier Carrasco wrote:
> >> Most of the functionality this driver provides can be used by non-hub
> >> devices as well.
> >>
> >> To account for the hub-specific code, add a flag to the device data
> >> structure and check its value for hub-specific code.
> >>
> >> Signed-off-by: Javier Carrasco <javier.carrasco@xxxxxxxxxxxxxx>
> >> ---
> >> drivers/usb/misc/onboard_usb_dev.c | 3 +++
> >> drivers/usb/misc/onboard_usb_dev.h | 10 ++++++++++
> >> 2 files changed, 13 insertions(+)
> >>
> >> diff --git a/drivers/usb/misc/onboard_usb_dev.c b/drivers/usb/misc/onboard_usb_dev.c
> >> index e2e1e1e30c1e..3ac21ec38ac0 100644
> >> --- a/drivers/usb/misc/onboard_usb_dev.c
> >> +++ b/drivers/usb/misc/onboard_usb_dev.c
> >> @@ -123,6 +123,9 @@ static int __maybe_unused onboard_dev_suspend(struct device *dev)
> >> if (onboard_dev->always_powered_in_suspend)
> >> return 0;
> >>
> >> + if (!onboard_dev->pdata->is_hub)
> >> + return onboard_dev_power_off(onboard_dev);
> >
> > Why turn the device always off when it isn't a hub? It could be a device
> > with wakeup support.
> >
> > I really regret making 'off in suspend' the default :(
> >
>
>
> The power management seems to be a critical point to consider.
>
> Maybe we keep the current implementation and add support to non-hub
> devices by simply adding a check to set power_off to false:
>
> static int __maybe_unused onboard_dev_suspend(struct device *dev)
>
> {
>
> struct onboard_dev *onboard_dev = dev_get_drvdata(dev);
>
> struct usbdev_node *node;
>
> bool power_off = true;
>
>
>
> if (onboard_dev->always_powered_in_suspend)
>
> return 0;
>
>
>
> mutex_lock(&onboard_dev->lock);
>
>
>
> list_for_each_entry(node, &onboard_dev->udev_list, list) {
>
> if (!device_may_wakeup(node->udev->bus->controller))
>
> continue;
>
>
>
> ~ if (usb_wakeup_enabled_descendants(node->udev) ||
>
> + !onboard_dev->pdata->is_hub) {
>
> power_off = false;
>
> break;
>
> }
It isn't really necessary to perform this check in the loop, it isn't
dependent on any properties of the USB devices, right? It could be
combined with the check of 'always_powered_in_suspend' above.
