Hi Geert-san, Thank you for your review! > From: Geert Uytterhoeven, Sent: Wednesday, April 15, 2020 10:56 PM <snip> > > --- /dev/null > > +++ b/Documentation/devicetree/bindings/dma/renesas,usb-dmac.yaml > > @@ -0,0 +1,99 @@ > > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) > > +%YAML 1.2 > > +--- > > +$id: http://devicetree.org/schemas/dma/renesas,usb-dmac.yaml# > > +$schema: http://devicetree.org/meta-schemas/core.yaml# > > + > > +title: Renesas USB DMA Controller > > + > > +maintainers: > > + - Yoshihiro Shimoda <yoshihiro.shimoda.uh@xxxxxxxxxxx> > > + > > +allOf: > > + - $ref: "dma-controller.yaml#" > > + > > +properties: > > + compatible: > > + items: > > + - enum: > > + - renesas,r8a7743-usb-dmac # RZ/G1M > > + - renesas,r8a7744-usb-dmac # RZ/G1N > > + - renesas,r8a7745-usb-dmac # RZ/G1E > > + - renesas,r8a77470-usb-dmac # RZ/G1C > > + - renesas,r8a774a1-usb-dmac # RZ/G2M > > + - renesas,r8a774b1-usb-dmac # RZ/G2N > > + - renesas,r8a774c0-usb-dmac # RZ/G2E > > + - renesas,r8a7790-usb-dmac # R-Car H2 > > + - renesas,r8a7791-usb-dmac # R-Car M2-W > > + - renesas,r8a7793-usb-dmac # R-Car M2-N > > + - renesas,r8a7794-usb-dmac # R-Car E2 > > + - renesas,r8a7795-usb-dmac # R-Car H3 > > + - renesas,r8a7796-usb-dmac # R-Car M3-W > > + - renesas,r8a77961-usb-dmac # R-Car M3-W+ > > + - renesas,r8a77965-usb-dmac # R-Car M3-N > > + - renesas,r8a77990-usb-dmac # R-Car E3 > > + - renesas,r8a77995-usb-dmac # R-Car D3 > > + - const: renesas,usb-dmac > > + > > + reg: > > + maxItems: 1 > > + > > + interrupts: > > + maxItems: 2 > > Is there a use case for specifying a single interrupt? No. All USB-DMAC of R-Car Gen2/3 and RZ/Gn has 2 channels. In case of R-Car Gen3, please refer to the Figure 75.1 USB-DMAC Block Diagram. These USB-DMACn have CH0 and CH1 > > + > > + interrupt-names: > > + maxItems: 2 > > + items: > > + - pattern: "^ch[0-1]$" > > + - pattern: "^ch[0-1]$" > > Would it make sense to list the (two) actual channel names instead? I agree. Just using ch0 and ch1 is simpler. > > + > > + clocks: > > + maxItems: 1 > > + > > + '#dma-cells': > > + const: 1 > > + description: > > + The cell specifies the channel number of the DMAC port connected to > > + the DMA client. > > + > > + dma-channels: > > + maximum: 2 > > Is there a use case for specifying a single channel? > > > + > > + iommus: > > + maxItems: 2 > > Likewise? As I mentioned above, there is not a use case for specifying a single channel. > > + > > + power-domains: > > + maxItems: 1 > > + > > + resets: > > + maxItems: 1 > > + > > +required: > > + - compatible > > + - reg > > + - interrupts > > + - interrupt-names > > + - clocks > > + - '#dma-cells' > > + - dma-channels > > Shouldn't "power-domains" and "resets" be mandatory, too? > All covered SoCS have them. Oops. I'll add these properties as required. Best regards, Yoshihiro Shimoda > Gr{oetje,eeting}s, > > Geert > > -- > Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx > > In personal conversations with technical people, I call myself a hacker. But > when I'm talking to journalists I just say "programmer" or something like that. > -- Linus Torvalds
