On 28.02.24 19:10, Matthias Kaehlcke wrote: > On Wed, Feb 28, 2024 at 02:51:33PM +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. >> >> The 'always_powered_in_supend' attribute is only available for hub >> devices, keeping the driver's default behavior for non-hub devices (keep >> on in suspend). >> >> Signed-off-by: Javier Carrasco <javier.carrasco@xxxxxxxxxxxxxx> >> --- >> drivers/usb/misc/onboard_usb_dev.c | 25 +++++++++++++++++++++++-- >> drivers/usb/misc/onboard_usb_dev.h | 10 ++++++++++ >> 2 files changed, 33 insertions(+), 2 deletions(-) >> >> diff --git a/drivers/usb/misc/onboard_usb_dev.c b/drivers/usb/misc/onboard_usb_dev.c >> index e1779bd2d126..df0ed172c7ec 100644 >> --- a/drivers/usb/misc/onboard_usb_dev.c >> +++ b/drivers/usb/misc/onboard_usb_dev.c >> @@ -132,7 +132,8 @@ static int __maybe_unused onboard_dev_suspend(struct device *dev) >> struct usbdev_node *node; >> bool power_off = true; >> >> - if (onboard_dev->always_powered_in_suspend) >> + if (onboard_dev->always_powered_in_suspend && >> + !onboard_dev->pdata->is_hub) >> return 0; > > With this non-hub devices would always be powered down, since > 'always_powerd_in_suspend' is not set for them. This should be: > May I ask you what you meant in v4 with this comment? > Even without the sysfs attribute the field 'always_powered_in_suspend' > could > be set to true by probe() for non-hub devices. > if (!onboard_dev->pdata->is_hub || > onboard_dev->always_powered_in_suspend) > > Checking for the (non-)hub status first is clearer IMO, also it avoids > an unneccessary check of 'always_powered' for non-hub devices. > That makes sense and will be fixed. > Without code context: for hubs there can be multiple device tree nodes > for the same physical hub chip (e.g. one for the USB2 and another for > the USB3 part). I suppose this could also be the case for non-hub > devices. For hubs there is the 'peer-hub' device tree property to > establish a link between the two USB devices, as a result the onboard > driver only creates a single platform device (which is desired, > otherwise two platform devices would be in charge for power sequencing > the same phyiscal device. For non-hub devices there is currently no such > link. In many cases I expect there will be just one DT entry even though > the device has multiple USB interfaces, but it could happen and would > actually be a more accurate representation. > > General support is already there (the code dealing with 'peer-hub'), but > we'd have to come up with a suitable name. 'peer-device' is the first > thing that comes to my mind, but there might be better options. If such > a generic property is added then we should deprecate 'peer-hub', but > maintain backwards compatibility. I have nothing against that, but the first non-hub device that will be added does not have multiple DT nodes, so I have nothing to test that extension with real hardware. That could be added in the future, though, if the need ever arises. Best regards, Javier Carrasco