[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <2038198.JSHZf9jCco@diego>
Date: Tue, 14 Apr 2020 14:01:19 +0200
From: Heiko Stübner <heiko@...ech.de>
To: Johan Jonker <jbx6244@...il.com>
Cc: Robin Murphy <robin.murphy@....com>, devicetree@...r.kernel.org,
robh+dt@...nel.org, linux-kernel@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org,
linux-rockchip@...ts.infradead.org
Subject: Re: [PATCH] arm64: dts: rockchip: remove bus-width from mmc nodes in rk3308.dtsi
Hi Johan,
Am Dienstag, 14. April 2020, 13:45:00 CEST schrieb Johan Jonker:
> Hi Robin, Heiko,
>
> If the Rockchip DT maintainers(= Heiko) agree that the new line for the
> 'bus-width' properties is that it should be placed in dtsi I'll produce
> a version 2. Please advise what should be done with the other Rockchip
> SoCs. Change them too?
(1) as Robin pointed out bus-width and pinctrl containing the bus-pins
should be in the same file, as they describe parts of the same property
(2) essentially it is ok for pinctrl-defaults to live in the dtsi, when there
are no pin variants ... (like the uartX_mY pin variants), so if you enable
a node and only have essentially one pin variant to enable, this should
live in the soc dtsi (like essentially all boards using 4-pin sdmmc
and 8-pin emmc)
(3) Fixing other devicetrees is optional, so I won't oppose it of course
but it's also not something "that must be done" ;-)
Heiko
> On 4/14/20 12:02 PM, Robin Murphy wrote:
> > On 2020-04-13 8:36 pm, Johan Jonker wrote:
> >> The 'bus-width' property for mmc nodes is defined both in
> >> 'rk3308.dtsi' and 'rk3308-roc-cc.dts'.
> >> In line with the other Rockchip SoCs define that in a user dts only,
> >> so remove all entries from mmc nodes in 'rk3308.dtsi'.
> >
> > Judging by the pinctrl entries, these represent the number of pins
> > provided by the SoC itself. Obviously boards need to override that if
> > for some reason they don't wire up all the available data lines, but it
> > seems backwards to have every board restate the SoC's default value.
> >
> > In fact, having brought it up, for this particular case the pinctrl
> > setting is inherently related to the bus width, so having one without
> > the other in either place doesn't smell right.
> >
> > Robin.
> >
> >> Signed-off-by: Johan Jonker <jbx6244@...il.com>
> >> ---
> >> arch/arm64/boot/dts/rockchip/rk3308.dtsi | 3 ---
> >> 1 file changed, 3 deletions(-)
> >>
> >> diff --git a/arch/arm64/boot/dts/rockchip/rk3308.dtsi
> >> b/arch/arm64/boot/dts/rockchip/rk3308.dtsi
> >> index a9b98555d..130771ede 100644
> >> --- a/arch/arm64/boot/dts/rockchip/rk3308.dtsi
> >> +++ b/arch/arm64/boot/dts/rockchip/rk3308.dtsi
> >> @@ -587,7 +587,6 @@
> >> compatible = "rockchip,rk3308-dw-mshc",
> >> "rockchip,rk3288-dw-mshc";
> >> reg = <0x0 0xff480000 0x0 0x4000>;
> >> interrupts = <GIC_SPI 76 IRQ_TYPE_LEVEL_HIGH>;
> >> - bus-width = <4>;
> >> clocks = <&cru HCLK_SDMMC>, <&cru SCLK_SDMMC>,
> >> <&cru SCLK_SDMMC_DRV>, <&cru SCLK_SDMMC_SAMPLE>;
> >> clock-names = "biu", "ciu", "ciu-drive", "ciu-sample";
> >> @@ -602,7 +601,6 @@
> >> compatible = "rockchip,rk3308-dw-mshc",
> >> "rockchip,rk3288-dw-mshc";
> >> reg = <0x0 0xff490000 0x0 0x4000>;
> >> interrupts = <GIC_SPI 77 IRQ_TYPE_LEVEL_HIGH>;
> >> - bus-width = <8>;
> >> clocks = <&cru HCLK_EMMC>, <&cru SCLK_EMMC>,
> >> <&cru SCLK_EMMC_DRV>, <&cru SCLK_EMMC_SAMPLE>;
> >> clock-names = "biu", "ciu", "ciu-drive", "ciu-sample";
> >> @@ -615,7 +613,6 @@
> >> compatible = "rockchip,rk3308-dw-mshc",
> >> "rockchip,rk3288-dw-mshc";
> >> reg = <0x0 0xff4a0000 0x0 0x4000>;
> >> interrupts = <GIC_SPI 78 IRQ_TYPE_LEVEL_HIGH>;
> >> - bus-width = <4>;
> >> clocks = <&cru HCLK_SDIO>, <&cru SCLK_SDIO>,
> >> <&cru SCLK_SDIO_DRV>, <&cru SCLK_SDIO_SAMPLE>;
> >> clock-names = "biu", "ciu", "ciu-drive", "ciu-sample";
> >>
>
>
Powered by blists - more mailing lists