imx6dl/imx6q fec rmii mode with external ref_clk

On Wed, Jan 29, 2014 at 10:54:29AM +0100, Philippe De Muyter wrote:

quoted

diff --git a/arch/arm/mach-imx/mach-imx6q.c

b/arch/arm/mach-imx/mach-imx6q.c
index e51e3da..be8d074 100644

--- a/arch/arm/mach-imx/mach-imx6q.c
+++ b/arch/arm/mach-imx/mach-imx6q.c

@@ -188,7 +188,7 @@ static void __init imx6q_1588_init(void)
        if (!IS_ERR(gpr))
                regmap_update_bits(gpr, IOMUXC_GPR1,
                                IMX6Q_GPR1_ENET_CLK_SEL_MASK,
-                               IMX6Q_GPR1_ENET_CLK_SEL_ANATOP);
+                               0);
        else
                pr_err("failed to find fsl,imx6q-iomux-gpr regmap\n");

Yes, it works now.  Thanks.  Is that configurable from the device tree ?

Yes, with some code change, I think that's configurable from device
tree.  But, it's not that simple.  As I explained in thread [1], the
RMII reference clock ENET_REF_CLK from pad GPIO_16 will also be used
by PTP (IEEE 1588) as sampling clock in your configuration.  Right now,
the ptp clock in fec node is defaulted to be enet_ref from IMX6 CCM
(id 190) in imx6qdl.dtsi.  You will need to correct it in your board dts
file, something like the following.

clocks {
	#address-cells = <1>;
	#size-cells = <0>;

	rmii_clk: clock at 0 {
		compatible = "fixed-clock";
		reg = <0>;
		#clock-cells = <0>;
		clock-frequency = <25000000>;  /* 25MHz for example */
	};
};

&fec {
	clocks = <&clks 117>, <&clks 117>, <&rmii_clk>;
};

I think we can even save a device tree property and ask kernel to figure
it out that bit GPR1[21] should be set or cleared.  Yes, kernel should
be able to do that by checking whether the ptp_clk in fec node points to
the enet_ref clock coming from IMX6 clock driver (ANATOP/CCM).  I will
send a patch for this soon.

Shawn

[1] http://thread.gmane.org/gmane.linux.ports.arm.kernel/269265/focus=269646