Re: [PATCH v2 07/13] arm64: dts: Add Pensando Elba SoC support

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Rob,

On Tue, Mar 30, 2021 at 2:55 PM Rob Herring <robh@xxxxxxxxxx> wrote:
>
> On Sun, Mar 28, 2021 at 06:59:32PM -0700, Brad Larson wrote:
> > Add Pensando common and Elba SoC specific device nodes
>
> Build your dtb with W=1 and 'make dtbs_check' and fix warnings.

I'll do that thanks.

> > +
> > +             // CLUSTER 0
> > +             cpu0: cpu@0 {
> > +                     device_type = "cpu";
> > +                     compatible = "arm,cortex-a72", "arm,armv8";
>
> This should give you a warning.

I'll be sure to W=1 and make dtbs_check.  All the C++ comments are C syntax now.

> > +&qspi {
> > +     status = "okay";
> > +     flash0: mt25q@0 {
>
> flash@0

Yes, generic node name should be used.  Changed mt25q to flash.
-       flash0: mt25q@0 {
+       flash0: flash@0 {

> > +             compatible = "jdec,spi-nor";
>
> jedec,spi-nor

Changed jdec to jedec
-               compatible = "jdec,spi-nor";
+               compatible = "jedec,spi-nor";

> > +     spi@0 {
>
> 'spi@' is reserved for SPI controllers.

Changed the node name to indicate the spi bus number and chip select

&spi0 {
        num-cs = <4>;
        cs-gpios = <&spics 0 GPIO_ACTIVE_LOW>, <&spics 1 GPIO_ACTIVE_LOW>,
                   <&porta 1 GPIO_ACTIVE_LOW>, <&porta 7 GPIO_ACTIVE_LOW>;
        status = "okay";
-       spi@0 {
+       spi0_cs0@0 {
                compatible = "pensando,cpld";
                #address-cells = <1>;
                #size-cells = <1>;
                spi-max-frequency = <12000000>;
                reg = <0>;
        };
-       spi@1 {
+       spi0_cs1@1 {
        ...

> > +             compatible = "pensando,cpld";
>
> Any new compatibles need to be documented with schema.

I'll create the bindings file for the v3 patchset dependent on the
discussion revolving around adding pensando,cpld to this list in
spi/spidev.c.

static const struct of_device_id spidev_dt_ids[] = {
        { .compatible = "rohm,dh2228fv" },
        { .compatible = "lineartechnology,ltc2488" },
        { .compatible = "ge,achc" },
        { .compatible = "semtech,sx1301" },
        { .compatible = "lwn,bk4" },
        { .compatible = "dh,dhcom-board" },
        { .compatible = "menlo,m53cpld" },
        { .compatible = "pensando,cpld" },    <===
        {},
};
MODULE_DEVICE_TABLE(of, spidev_dt_ids);

> > --- /dev/null
> > +++ b/arch/arm64/boot/dts/pensando/elba-flash-parts.dtsi
> > @@ -0,0 +1,78 @@
> > +&flash0 {
> > +     partitions {
> > +             compatible = "fixed-partitions";
>
> This should just be moved into elba-asic-common.dtsi IMO unless you have
> good reasons to keep it separate.

This file is shared across platforms which have additional flash
devices and partition schemes.

> You need SPDX license tags at the top of each file. checkpatch.pl should
> tell you this.

I don't recall checkpatch.pl reporting that for device tree files.
I've added to each file now.

> > +/*
> > + * Copyright (c) 2019, Pensando Systems Inc.
> > + * This program is free software; you can redistribute it and/or modify
> > + * it under the terms of the GNU General Public License version 2 as
> > + * published by the Free Software Foundation.
> > + */
> > +
> > +#include "dt-bindings/interrupt-controller/arm-gic.h"
> > +
> > +/ {
> > +     compatible = "pensando,elba";
>
> This needs an SoC family schema.
>
> Generally this and 'model' should be in the same file (the board file).

Will add schema for pensando,elba in the v3 patchset.  The compatible
and model are now in the board file elba-asic.dts as shown below

+/ {
+       model = "Elba ASIC Board";
+       compatible = "pensando,elba";
+
+       aliases {
+               serial0 = &uart0;
+               spi0 = &spi0;
+               spi1 = &qspi;
+       };
+
+       chosen {
+               stdout-path = "serial0:115200n8";
+       };
+};

> > +
> > +     interrupt-parent = <&gic>;
> > +     #address-cells = <2>;
> > +     #size-cells = <2>;
> > +
> > +     clocks {
>
> Drop this container node.

Removed the container node.

> > +     /* Common UIO device for MSI drivers */
> > +     uio_penmsi {
> > +             compatible = "pensando,uio_penmsi";
> > +             name = "uio_penmsi";
>
> No. What's UIO?

UIO nodes are removed, only core SoC support will be included

> > +     soc: soc {
> > +             compatible = "simple-bus";
> > +             #address-cells = <2>;
> > +             #size-cells = <2>;
> > +             ranges;
> > +
> > +             gic: interrupt-controller@800000 {
>
> Order nodes by increasing unit address.

Yes, I've done this and will be part of the v3 patchset

> > +             /*
> > +              * Until we  know the interrupt domain following this, we
> > +              * are forced to use this is the place where interrupts from
> > +              * PCI converge. In the ideal case, we use one domain higher,
> > +              * where the PCI-ness has been shed.
> > +              */
> > +             pxc0_intr: intc@20102200 {
>
> interrupt-controller@...

Changed intc to interrupt-controller

> > +                     compatible = "pensando,soc-ictlr-csrintr";
> > +                     interrupt-controller;
> > +                     reg = <0x0 0x20102200 0x0 0x4>;
> > +                     #interrupt-cells = <3>;
> > +                     interrupt-parent = <&gic>;
> > +                     interrupts = <GIC_SPI 17 IRQ_TYPE_LEVEL_HIGH>;
> > +                     interrupt-names = "pxc0_intr";
> > +             };
> > +
> > +             uart0: serial@4800 {
> > +                     device_type = "serial";
>
> Drop 'device_type'. It's deprecated for all but memory and pci.
>

Removed  device_type = "serial";

> > +             gpio0: gpio@4000 {
> > +                     #address-cells = <1>;
> > +                     #size-cells = <0>;
> > +                     compatible = "snps,dw-apb-gpio";
> > +                     reg = <0x0 0x4000 0x0 0x78>;
> > +                     status = "disabled";
> > +
> > +                     porta: gpio-controller@0 {
>
> gpio@0

Changed from gpio@0 to gpio-port@0 based on additional reviewer input

-                       porta: gpio-controller@0 {
+                       porta: gpio-port@0 {

> > +
> > +             /* UIO device using interrupt line PCIEMAC */
>
> UIO is a kernel thing. It doesn't belong in DT.

UIO nodes are removed for core SoC support patchset

> > +
> > +             emmc: mmc@30440000 {
> > +                     compatible = "pensando,elba-emmc", "cdns,sd4hc";
> > +                     clocks = <&emmc_clk>;
> > +                     interrupts = <GIC_SPI 26 IRQ_TYPE_LEVEL_HIGH>;
> > +                     reg = <0x0 0x30440000 0x0 0x10000
> > +                            0x0 0x30480044 0x0 0x4>;
>
> Use <> around each entry.

The system fails to boot if I make that change, mmc probe fails and
rootfs not found

-               reg = <0x0 0x30440000 0x0 0x10000
-                      0x0 0x30480044 0x0 0x4>;
+               reg = <0x0 0x30440000 0x0 0x10000>,
+                     <0x0 0x30480044 0x0 0x4>;

Regards,
Brad



[Index of Archives]     [Linux SPI]     [Linux Kernel]     [Linux ARM (vger)]     [Linux ARM MSM]     [Linux Omap]     [Linux Arm]     [Linux Tegra]     [Fedora ARM]     [Linux for Samsung SOC]     [eCos]     [Linux Fastboot]     [Gcc Help]     [Git]     [DCCP]     [IETF Announce]     [Security]     [Linux MIPS]     [Yosemite Campsites]

  Powered by Linux