Re: [PATCH v3 2/3] dt-bindings: usb: snps,dwc3: Add allow-rtsusp-on-u3 property

[Thinh Nguyen]

On Wed, Jul 12, 2023, Roger Quadros wrote:
> +Rob
> 
> On 12/07/2023 00:56, Thinh Nguyen wrote:
> > On Tue, Jul 11, 2023, Elson Roy Serrao wrote:
> >> This property allows dwc3 runtime suspend when bus suspend interrupt
> >> is received even with cable connected. This would allow the dwc3
> >> controller to enter low power mode during bus suspend scenario.
> >>
> >> This property would particularly benefit dwc3 IPs where hibernation is
> >> not enabled and the dwc3 low power mode entry/exit is handled by the
> >> glue driver. The assumption here is that the platform using this dt
> >> property is capable of detecting resume events to bring the controller
> >> out of suspend.
> >>
> >> Signed-off-by: Elson Roy Serrao <quic_eserrao@xxxxxxxxxxx>
> >> ---
> >>  Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 5 +++++
> >>  1 file changed, 5 insertions(+)
> >>
> >> diff --git a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
> >> index a696f23730d3..18ad99a26dd9 100644
> >> --- a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
> >> +++ b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
> >> @@ -403,6 +403,11 @@ properties:
> >>      description:
> >>        Enable USB remote wakeup.
> >>  
> >> +  snps,allow-rtsusp-on-u3:
> > 
> > Please spell out the whole thing as "rtsusp" isn't clear. Also, it's not
> > just for U3 right? For highspeed, it's L2.
> > 
> > How about the name that Roger use: "snps,gadget-keep-connect-sys-sleep"
> 
> That property was meant to keep the USB device controller connected
> during system sleep. So that name may not be appropriate here as this
> is about allowing controller runtime suspend during USB suspend.
> 

Ah... you're right. Don't know what I was thinking. They're not the
same, and they may need to be separated. I'm not great with the naming,
but it can be something like this: "snps,runtime-suspend-on-usb-suspend"

BTW, I assume that your platform (both from Roger and Elson) does not
turn off the bus that handle DMA of events on suspend? (e.g. AXI bus)
Looking at the change, it looks like the events are still written in the
event buffer.

Thanks,
Thinh