On Sun, 17 Jan 2021 22:28:47 -0600 Samuel Holland <samuel@xxxxxxxxxxxx> wrote: Hi, > On 1/17/21 8:08 PM, Andre Przywara wrote: > > Add simple "allwinner,sun50i-h616-xxx" compatible names to existing > > bindings, and pair them with an existing fallback compatible string, > > as the devices are compatible. > > This covers I2C, infrared, RTC and SPI. > > > > Use enums to group all compatible devices together. > > > > Signed-off-by: Andre Przywara <andre.przywara@xxxxxxx> > > Acked-by: Rob Herring <robh@xxxxxxxxxx> > > Acked-by: Wolfram Sang <wsa@xxxxxxxxxx> # for I2C > > --- > > .../bindings/i2c/marvell,mv64xxx-i2c.yaml | 21 +++++++------------ > > .../media/allwinner,sun4i-a10-ir.yaml | 16 ++++++-------- > > .../bindings/rtc/allwinner,sun6i-a31-rtc.yaml | 3 +++ > > .../bindings/spi/allwinner,sun6i-a31-spi.yaml | 1 + > > 4 files changed, 17 insertions(+), 24 deletions(-) > > > > diff --git a/Documentation/devicetree/bindings/i2c/marvell,mv64xxx-i2c.yaml b/Documentation/devicetree/bindings/i2c/marvell,mv64xxx-i2c.yaml > > index 5b5ae402f97a..eb72dd571def 100644 > > --- a/Documentation/devicetree/bindings/i2c/marvell,mv64xxx-i2c.yaml > > +++ b/Documentation/devicetree/bindings/i2c/marvell,mv64xxx-i2c.yaml > > @@ -18,21 +18,14 @@ properties: > > - const: allwinner,sun4i-a10-i2c > > - const: allwinner,sun6i-a31-i2c > > - items: > > - - const: allwinner,sun8i-a23-i2c > > + - enum: > > + - allwinner,sun8i-a23-i2c > > + - allwinner,sun8i-a83t-i2c > > + - allwinner,sun50i-a64-i2c > > + - allwinner,sun50i-a100-i2c > > + - allwinner,sun50i-h6-i2c > > + - allwinner,sun50i-h616-i2c > > - const: allwinner,sun6i-a31-i2c > > - - items: > > - - const: allwinner,sun8i-a83t-i2c > > - - const: allwinner,sun6i-a31-i2c > > - - items: > > - - const: allwinner,sun50i-a64-i2c > > - - const: allwinner,sun6i-a31-i2c > > - - items: > > - - const: allwinner,sun50i-a100-i2c > > - - const: allwinner,sun6i-a31-i2c > > - - items: > > - - const: allwinner,sun50i-h6-i2c > > - - const: allwinner,sun6i-a31-i2c > > - > > - const: marvell,mv64xxx-i2c > > - const: marvell,mv78230-i2c > > - const: marvell,mv78230-a0-i2c > > diff --git a/Documentation/devicetree/bindings/media/allwinner,sun4i-a10-ir.yaml b/Documentation/devicetree/bindings/media/allwinner,sun4i-a10-ir.yaml > > index 5fa19d4aeaf3..6d8395d6bca0 100644 > > --- a/Documentation/devicetree/bindings/media/allwinner,sun4i-a10-ir.yaml > > +++ b/Documentation/devicetree/bindings/media/allwinner,sun4i-a10-ir.yaml > > @@ -20,16 +20,12 @@ properties: > > - const: allwinner,sun5i-a13-ir > > - const: allwinner,sun6i-a31-ir > > - items: > > - - const: allwinner,sun8i-a83t-ir > > - - const: allwinner,sun6i-a31-ir > > - - items: > > - - const: allwinner,sun8i-r40-ir > > - - const: allwinner,sun6i-a31-ir > > - - items: > > - - const: allwinner,sun50i-a64-ir > > - - const: allwinner,sun6i-a31-ir > > - - items: > > - - const: allwinner,sun50i-h6-ir > > + - enum: > > + - allwinner,sun8i-a83t-ir > > + - allwinner,sun8i-r40-ir > > + - allwinner,sun50i-a64-ir > > + - allwinner,sun50i-h6-ir > > + - allwinner,sun50i-h616-ir > > - const: allwinner,sun6i-a31-ir > > > > reg: > > diff --git a/Documentation/devicetree/bindings/rtc/allwinner,sun6i-a31-rtc.yaml b/Documentation/devicetree/bindings/rtc/allwinner,sun6i-a31-rtc.yaml > > index 37c2a601c3fa..97928efd2bc9 100644 > > --- a/Documentation/devicetree/bindings/rtc/allwinner,sun6i-a31-rtc.yaml > > +++ b/Documentation/devicetree/bindings/rtc/allwinner,sun6i-a31-rtc.yaml > > @@ -26,6 +26,9 @@ properties: > > - const: allwinner,sun50i-a64-rtc > > - const: allwinner,sun8i-h3-rtc > > - const: allwinner,sun50i-h6-rtc > > + - items: > > + - const: allwinner,sun50i-h616-rtc > > + - const: allwinner,sun50i-h6-rtc > > Since H6, the RTC manages the 24 MHz DCXO, so it provides a fourth clock > output. If this is easy to change later, then it is fine for now, but > maybe it is better to get the H616 binding correct from the beginning? So you mean that RTC register +0x160 controls the system HOSC clock, so the main input clock for all the PLLs and other clocks? And by clearing bit 1 in there we can stop this? And if that is the case, do you suggest that we should model this in the DT, so that the fixed-clock "<&osc24M>" should be replaced with "<&rtc 3>"? So from a "the DT describes the hardware" point of view that would probably the right way, but not sure if Linux is happy about that. At the very least that would mean to extend the RTC driver to export a fourth clock, and all devices would now depend on the RTC (also for probing!). And Linux can realistically never turn that clock off anyway (without grinding to a halt), so this register is more useful for ARISC firmware? So I am somewhat undecided: changing this for the H6 would make newer DTs unusable on older kernels, without anything we really gain. When we really want this, we should indeed use the opportunity to introduce this at least for the H616 from day one, to avoid this situation here. But this requires more changes than just the binding, doesn't it? Cheers, Andre > > > reg: > > maxItems: 1 > > diff --git a/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml b/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml > > index 7866a655d81c..908248260afa 100644 > > --- a/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml > > +++ b/Documentation/devicetree/bindings/spi/allwinner,sun6i-a31-spi.yaml > > @@ -25,6 +25,7 @@ properties: > > - enum: > > - allwinner,sun8i-r40-spi > > - allwinner,sun50i-h6-spi > > + - allwinner,sun50i-h616-spi > > - const: allwinner,sun8i-h3-spi > > > > reg: > > >