Re: [PATCH v13 2/6] usb: dwc3: core: Host wake up support from system suspend

[Matthias Kaehlcke <mka@chromium.org>]

On Tue, Apr 12, 2022 at 12:08:02PM +0530, Sandeep Maheswaram (Temp) wrote:
> Hi Pavan,
> 
> On 4/12/2022 10:30 AM, Pavan Kondeti wrote:
> > Hi Sandeep,
> > 
> > On Tue, Apr 12, 2022 at 10:16:39AM +0530, Sandeep Maheswaram (Temp) wrote:
> > > Hi Matthias,
> > > 
> > > On 4/12/2022 2:24 AM, Matthias Kaehlcke wrote:
> > > > On Tue, Apr 12, 2022 at 12:46:50AM +0530, Sandeep Maheswaram wrote:
> > > > > During suspend read the status of all port and set hs phy mode
> > > > > based on current speed. Use this hs phy mode to configure wakeup
> > > > > interrupts in qcom glue driver.
> > > > > 
> > > > > Check wakep-source property for dwc3 core node to set the
> > > > s/wakep/wakeup/
> > > Okay. Will update in next version.
> > > > > wakeup capability. Drop the device_init_wakeup call from
> > > > > runtime suspend and resume.
> > > > > 
> > > > > Also check during suspend if any wakeup capable devices are
> > > > > connected to the controller (directly or through hubs), if there
> > > > > are none set a flag to indicate that the PHY is powered
> > > > > down during suspend.
> > > > > 
> > > > > Signed-off-by: Sandeep Maheswaram <quic_c_sanm@quicinc.com>
> > > > > ---
> > > > A per-patch change log would be really helpful for reviewers, even
> > > > if it doesn't include older versions.
> > > Okay. Will update in next version.
> > > > >   drivers/usb/dwc3/core.c | 33 ++++++++++++++++++++-------------
> > > > >   drivers/usb/dwc3/core.h |  4 ++++
> > > > >   drivers/usb/dwc3/host.c | 25 +++++++++++++++++++++++++
> > > > >   3 files changed, 49 insertions(+), 13 deletions(-)
> > > > > 
> > > > > diff --git a/drivers/usb/dwc3/core.c b/drivers/usb/dwc3/core.c
> > > > > index 1170b80..effaa43 100644
> > > > > --- a/drivers/usb/dwc3/core.c
> > > > > +++ b/drivers/usb/dwc3/core.c
> > > > > @@ -32,6 +32,7 @@
> > > > >   #include <linux/usb/gadget.h>
> > > > >   #include <linux/usb/of.h>
> > > > >   #include <linux/usb/otg.h>
> > > > > +#include <linux/usb/hcd.h>
> > > > >   #include "core.h"
> > > > >   #include "gadget.h"
> > > > > @@ -1723,6 +1724,7 @@ static int dwc3_probe(struct platform_device *pdev)
> > > > >   	platform_set_drvdata(pdev, dwc);
> > > > >   	dwc3_cache_hwparams(dwc);
> > > > > +	device_init_wakeup(&pdev->dev, of_property_read_bool(dev->of_node, "wakeup-source"));
> > > > >   	spin_lock_init(&dwc->lock);
> > > > >   	mutex_init(&dwc->mutex);
> > > > > @@ -1865,6 +1867,7 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
> > > > >   {
> > > > >   	unsigned long	flags;
> > > > >   	u32 reg;
> > > > > +	struct usb_hcd  *hcd = platform_get_drvdata(dwc->xhci);
> > > > >   	switch (dwc->current_dr_role) {
> > > > >   	case DWC3_GCTL_PRTCAP_DEVICE:
> > > > > @@ -1877,10 +1880,7 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
> > > > >   		dwc3_core_exit(dwc);
> > > > >   		break;
> > > > >   	case DWC3_GCTL_PRTCAP_HOST:
> > > > > -		if (!PMSG_IS_AUTO(msg)) {
> > > > > -			dwc3_core_exit(dwc);
> > > > > -			break;
> > > > > -		}
> > > > > +		dwc3_check_phy_speed_mode(dwc);
> > > > >   		/* Let controller to suspend HSPHY before PHY driver suspends */
> > > > >   		if (dwc->dis_u2_susphy_quirk ||
> > > > > @@ -1896,6 +1896,16 @@ static int dwc3_suspend_common(struct dwc3 *dwc, pm_message_t msg)
> > > > >   		phy_pm_runtime_put_sync(dwc->usb2_generic_phy);
> > > > >   		phy_pm_runtime_put_sync(dwc->usb3_generic_phy);
> > > > > +
> > > > > +		if (!PMSG_IS_AUTO(msg)) {
> > > > > +			if (device_may_wakeup(dwc->dev) &&
> > > > > +			    usb_wakeup_enabled_descendants(hcd->self.root_hub)) {
> > > > You did not answer my question on v12, reposting it:
> > > > 
> > > >    Did you ever try whether you could use device_children_wakeup_capable() from
> > > >    [1] instead of usb_wakeup_enabled_descendants()?
> > > > 
> > > >    [1] https://patchwork.kernel.org/project/linux-usb/patch/1635753224-23975-2-git-send-email-quic_c_sanm@quicinc.com/#24566065
> > > Sorry ..I have replied in mail yesterday but it is not showing up in
> > > patchwork link.
> > > 
> > > Tried with  device_children_wakeup_capable(dwc->dev) instead of
> > > usb_wakeup_enabled_descendants and it always returns true even
> > > 
> > > when no devices are connected.
> > > 
> > What do you mean by when no devices are connected? There is always
> > root hub connected and we should not power down the DWC3 here even
> > when remote wakeup for root hub is enabled. Essentially
> > usb_wakeup_enabled_descendants() returns true even without any
> > physical devices connected.
> > 
> > What does device_children_wakeup_capable() do? Sorry, I could not
> > find this function definition.
> > 
> > Thanks,
> > Pavan
> 
> usb_wakeup_enabled_descendants() doesn't consider hubs. It only returns true if any devices
> are connected with wakeup capability apart from hubs.

Actually it considers hubs:

unsigned usb_wakeup_enabled_descendants(struct usb_device *udev)
{
	struct usb_hub *hub = usb_hub_to_struct_hub(udev);

	return udev->do_remote_wakeup +
		(hub ? hub->wakeup_enabled_descendants : 0);
}

'udev' may or may not be a hub, if 'do_remote_wakeup' is set then the
device is considered a wakeup enabled descendant.

And for system suspebd 'do_remote_wakeup' is set based on the wakeup
config of the device:

static void choose_wakeup(struct usb_device *udev, pm_message_t msg)
{
	...
	w = device_may_wakeup(&udev->dev);
	...
	udev->do_remote_wakeup = w;
}

I checked on three systems with different Linux distributions, on all of
the wakeup flag of a connected hub is 'disabled'. Wakeup still works, so
apparently that flag doesn't really have an impact for child ports.

> If we consider hubs also dwc3 core exit and phy exit will never be called.
> 
> device_children_wakeup_capable() implementation was shared by Matthias in below thread
> https://patchwork.kernel.org/project/linux-usb/patch/1635753224-23975-2-git-send-email-quic_c_sanm@quicinc.com/#24566065
> 
> Probably device_children_wakeup_capable() is returning true because it considers hubs also.

I thought I did a basic test when I sent the patch, I did another (?) one
with v13 of your patch set. In this tests with a hub connected the
function returns true when an HID device is connected, and false when
nothing is connected. The wakeup flag of the hub is disabled (default).

Sandeep, are the wakeup flags of the child hub(s) set to 'enabled' on
the system you tested on?