Bring imx6qp dts definitions upstream

[Luís Mendes]

Hi Fabio,

Regarding the dts... I've followed your suggestion, but I've run into some
issues, namely:
- The base imx6qdl-wandboard.dtsi file has definitions for the Gigabit
Ethernet controller that are incompatible with the new Rev. D1. I had to
remove two lines that define the ineterrupt lines, set it to disabled and
redefine the pins in imx6qdl-wandboard-revd1.dtsi for rev. D1 ethernet to
work. Thus this already implies changes to the dtsi files for revb1 and
revc1.
Can I modify them? I don't have rev B1 hardware to test, only a i.MX6Q
Wandboard revC1 and another i.MX6QP rev D1 .

- I'm having quite some difficulty making BCM4339 WiFI module to work. It
seems it doesn't want to power up and if I define a vmmc-supply entry for
the sdhc controller with a fixed regulator. I get:
"Failed getting OCR mask: 0" and lots of
"sdhci-esdhc-imx 2198000.usdhc: could not set regulator OCR (-22)"
If I try to use wlreg_on-supply property, I believe it provides no effect
for vanilla kernel drivers.
    brcmf: brcmf at 1 {
        reg = <1>;
        compatible = "brcm,bcm4329-fmac";
        interrupt-parent = <&gpio1>;
                interrupts = <29 IRQ_TYPE_LEVEL_LOW>;
        wlreg_on-supply = <&reg_wlreg_on>;
    };
Anyway the brcmfmac driver is not picking up the OF definitions above.
I'm trying to place printk in the kernel driver to trace the probe process
but is not doing any serious probing that I can see.


Regards,
Luís
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/etnaviv/attachments/20170827/ab27a91c/attachment.html>