On 26 January 2017 at 15:39, Rob Herring <r...@kernel.org> wrote: > On Mon, Jan 23, 2017 at 11:56 AM, Chris Brandt <chris.bra...@renesas.com> > wrote: >> Hello Rob, >> >> >> On Monday, January 23, 2017, Rob Herring wrote: >>> > --- a/Documentation/devicetree/bindings/mmc/tmio_mmc.txt >>> > +++ b/Documentation/devicetree/bindings/mmc/tmio_mmc.txt >>> > @@ -25,8 +25,32 @@ Required properties: >>> > "renesas,sdhi-r8a7795" - SDHI IP on R8A7795 SoC >>> > "renesas,sdhi-r8a7796" - SDHI IP on R8A7796 SoC >>> > >>> > +- clocks: Most controllers only have 1 clock source per channel. >>> However, on >>> > + some variations of this controller, the internal card detection >>> >>> This should be explicit as to which compatible strings have 2 clocks and >>> which have 1 clock. >> >> OK. I'll make a list like I did for sh_mmcif: >> >> https://patchwork.kernel.org/patch/9512091/ >> >> >> >>> > + >>> > +Example showing 2 clocks: >>> > + sdhi0: sd@e804e000 { >>> >>> mmc@... >> >> I'm confused. I see that for all SDHI controllers, it either "sd@" or >> "sdhci@". >> >> $ grep sdhi $(find arch/arm/boot/dts -name "*.dtsi") >> >> $ grep sdhci $(find arch/arm/boot/dts -name "*.dtsi") > > Yes, there's lots of variation. Node names are supposed to be generic > for their class of device (e.g. ethernet, pci, usb, > interrupt-controller, etc.). I'd be fine with "sd", but think "mmc" is > more common. Either way, we should pick one moving forward. "sdhci" > should not be used as that's a specific implementation.
I think we discussed this earlier. The problem with these devices is that it covers several type of devices, so it's hard to find a generic node name. The devices are: MMC, eMMC, SD, SDIO. I don't have strong opinions on whether to chose a generic node name, and perhaps "mmc" is the best we can get? Whatever we decide, we should update all mmc DTS docs, to avoid future confusion. Kind regards Uffe
