Hi, On 12/10/2020 16:22, Serge Semin wrote: > On Mon, Oct 12, 2020 at 09:54:25AM +0200, Neil Armstrong wrote: >> Hi, >> >> On 11/10/2020 00:41, Serge Semin wrote: >>> An empty snps,quirk-frame-length-adjustment won't cause any change >>> performed by the driver. Moreover the DT schema validation will fail, >>> since it expects the property being assigned with some value. So just >>> discard the property declaration then from the example. >>> >>> Signed-off-by: Serge Semin <Sergey.Semin@xxxxxxxxxxxxxxxxxxxx> >>> >>> --- >>> >>> Note the same problem is in the DT source file >>> arch/arm64/boot/dts/amlogic/meson-g12-common.dtsi . >>> --- >>> .../devicetree/bindings/usb/amlogic,meson-g12a-usb-ctrl.yaml | 1 - >>> 1 file changed, 1 deletion(-) >>> >>> diff --git a/Documentation/devicetree/bindings/usb/amlogic,meson-g12a-usb-ctrl.yaml b/Documentation/devicetree/bindings/usb/amlogic,meson-g12a-usb-ctrl.yaml >>> index 5b04a7dfa018..88184d7e26cc 100644 >>> --- a/Documentation/devicetree/bindings/usb/amlogic,meson-g12a-usb-ctrl.yaml >>> +++ b/Documentation/devicetree/bindings/usb/amlogic,meson-g12a-usb-ctrl.yaml >>> @@ -209,6 +209,5 @@ examples: >>> interrupts = <30>; >>> dr_mode = "host"; >>> snps,dis_u2_susphy_quirk; >>> - snps,quirk-frame-length-adjustment; >>> }; >>> }; >>> >> > >> Thanks for reporting this, actually the fladj must be 0x20 on this hw, >> but we do set this on the PHY side, so we can let the dwc3 side 0 here. > > I can convert this patch to initializing the "snps,quirk-frame-length-adjustment" > property with 0x20 value instead. Since most likely I'll have to send a v2/v3/etc > of this patchset, that modification won't be too much work to do. What do you think? Yes, do this please, anyway it's only an example so it's ok. > > Anyway please note, that I've fixed the improper property usage in the DT schema > example only. "snps,quirk-frame-length-adjustment" defined as boolean still > persists in the DTS file: arch/arm64/boot/dts/amlogic/meson-g12-common.dtsi . > So if you ever try to validate that dts file with "make dtbs_check" scenario, it > will fail. Yes, I'll push a fix to pass the dtbs_check when this is merged. Thanks, Neil > > -Sergey > >> >> Acked-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx> >> >> Neil >>
