Re: [PATCH v3 2/3] ARM: dts: rockchip: add dts for Firefly Firefly-RK3288 boards

Hi,

On Mon, Jan 19, 2015 at 8:36 AM, Heiko Stübner [off-list ref] wrote:

quoted

+     ext_gmac: external-gmac-clock {
+             compatible = "fixed-clock";
+             #clock-cells = <0>;
+             clock-frequency = <125000000>;
+             clock-output-names = "phy_clkout125";

This should be named "ext_gmac". See the rockchip,rk3288-cru.txt binding
document.

I'll revert this to use "ext_gmac".

quoted

+&gmac {
+     assigned-clocks = <&cru SCLK_MAC>;
+     assigned-clock-parents = <&ext_gmac>;
+     clock_in_out = "input";
+     pinctrl-names = "default";
+     pinctrl-0 = <&rgmii_pins>, <&phy_rst>, <&phy_pmeb>, <&phy_int>;
+     phy_regulator = "vcc_lan";

This is wrong in the dwmac-rk implementation at the moment which Romain Perier
was/is trying to rectify. I.e. there is an established devicetree api for
handling regulators, thus the approach the current net-code takes is just
wrong and thus this will need to change after.

I see.

quoted

+     hym8563: hym8563@51 {
+             compatible = "haoyu,hym8563";
+             reg = <0x51>;
+             #clock-cells = <0>;
+             clock-frequency = <32768>;
+             clock-output-names = "rtc_clkout";

this should be named "xin32k" . See radxarock.dts (as a rk3188 based similar
example) and the rk3288 clock controller devicetree binding.

I'll revert this to use "xin32k".

This is due to the fact that this is also the input for the suspend clock and
the core clock code expects the specific naming according to the soc
documentation.

can I assume clock named "xin32k" is always enabled?
no need to describe consumer explicitly?

quoted

+&sdio0 {
+     broken-cd;
+     bus-width = <4>;
+     clocks = <&hym8563>;
+     clock-names = "lpo";

why are you _overriding_ the clocks of the sdio controller?

sorry, I was just wrong. I'll remove clocks and clock-names from here.

quoted

+&spi2 {
+     pinctrl-names = "default";
+     pinctrl-0 = <&spi2_clk>, <&spi2_cs0>;

you're enabling only the clk and chipselect, what about the data signals?

no idea, I just follow schematic. if it's really unusable on mainline,
I'll disable(remove) spi2.

quoted

+&tsadc {
+     clocks = <&hym8563>;
+     clock-names = "clkin_32k";

The tsadc driver only recognizes the clocks "tsadc", "apb_pclk" so I'm not
sure what you're doing with the clkin_32k name here? Also you're again
overwriting the existing property.

I was wrong here too. I'll remove clocks and clock-names from here.
and if it makes tsadc non-working, I'll remove tsadc too for now.

Regards,
--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html