Re: [PATCH 2/3] dt-bindings: arm: amlogic: Add SoC information bindings

From: Rob Herring
Date: Wed Apr 05 2017 - 15:12:38 EST


On Tue, Apr 4, 2017 at 7:49 AM, Neil Armstrong <narmstrong@xxxxxxxxxxxx> wrote:
> On 04/04/2017 02:26 PM, Rob Herring wrote:
>> On Tue, Apr 4, 2017 at 3:51 AM, Neil Armstrong <narmstrong@xxxxxxxxxxxx> wrote:
>>> On 04/03/2017 06:34 PM, Rob Herring wrote:
>>>> On Fri, Mar 31, 2017 at 04:10:30PM +0200, Neil Armstrong wrote:
>>>>> On 03/31/2017 03:44 PM, Arnd Bergmann wrote:
>>>>>> On Fri, Mar 31, 2017 at 10:47 AM, Neil Armstrong
>>>>>> <narmstrong@xxxxxxxxxxxx> wrote:
>>>>>>> Add bindings for the SoC information register of the Amlogic SoCs.
>>>>>>>
>>>>>>> Signed-off-by: Neil Armstrong <narmstrong@xxxxxxxxxxxx>
>>>>>>> ---
>>>>>>> Documentation/devicetree/bindings/arm/amlogic.txt | 20 ++++++++++++++++++++
>>>>>>> 1 file changed, 20 insertions(+)
>>>>>>>
>>>>>>> diff --git a/Documentation/devicetree/bindings/arm/amlogic.txt b/Documentation/devicetree/bindings/arm/amlogic.txt
>>>>>>> index bfd5b55..b850985 100644
>>>>>>> --- a/Documentation/devicetree/bindings/arm/amlogic.txt
>>>>>>> +++ b/Documentation/devicetree/bindings/arm/amlogic.txt
>>>>>>> @@ -52,3 +52,23 @@ Board compatible values:
>>>>>>> - "amlogic,q201" (Meson gxm s912)
>>>>>>> - "nexbox,a95x" (Meson gxbb or Meson gxl s905x)
>>>>>>> - "nexbox,a1" (Meson gxm s912)
>>>>>>> +
>>>>>>> +Amlogic Meson GX SoCs Information
>>>>>>> +----------------------------------
>>>>>>> +
>>>>>>> +The Meson SoCs have a Product Register that allows to retrieve SoC type,
>>>>>>> +package and revision information. If present, a device node for this register
>>>>>>> +should be added.
>>>>>>> +
>>>>>>> +Required properties:
>>>>>>> + - compatible: For Meson GX SoCs, must be "amlogic,meson-gx-socinfo".
>>>>>>> + - reg: Base address and length of the register block.
>>>>>>> +
>>>>>>> +Examples
>>>>>>> +--------
>>>>>>> +
>>>>>>> + chipid@220 {
>>>>>>> + compatible = "amlogic,meson-gx-socinfo";
>>>>>>> + reg = <0x0 0x00220 0x0 0x4>;
>>>>>>> + };
>>>>>>> +
>>>>>>
>>>>>> The register location would hint that this is in the middle of some block of
>>>>>> random registers, i.e. a syscon or some unrelated device.
>>>>>>
>>>>>> Are you sure that "socinfo" is the actual name of the IP block and that
>>>>>> it only has a single 32-bit register?
>>>>>>
>>>>>> Arnd
>>>>>>
>>>>>
>>>>> Hi Arnd,
>>>>>
>>>>> I'm sorry I did not find any relevant registers in the docs or source code describing
>>>>> it in a specific block of registers, and no close enough register definitions either.
>>>>> They may be used by the secure firmware I imagine.
>>>>>
>>>>> For the register name, Amlogic refers it to "cpu_version" in their code, but it really
>>>>> gives some details on the whole SoC and package, and socinfo seems better.
>>>>
>>>> A register at address 0x220 seems a bit strange (unless there's ranges
>>>> you're not showing), but ROM code at this address would be fairly
>>>> typical. And putting version information into the ROM is also common.
>>>>
>>>> Rob
>>>>
>>>
>>> Hi Rob.
>>>
>>> Indeed it's part of a larger range :
>>> aobus: aobus@c8100000 {
>>> compatible = "simple-bus";
>>> reg = <0x0 0xc8100000 0x0 0x100000>;
>>> #address-cells = <2>;
>>> #size-cells = <2>;
>>> ranges = <0x0 0x0 0x0 0xc8100000 0x0 0x100000>;
>>>
>>>
>>> While scrubbing on the uboot source, I found a sort of block of registers dedicated to communicate with
>>> the secure firmware :
>>> AO_SEC_REG0 0x140
>>> AO_SEC_REG1 0x144
>>> AO_SEC_REG2 0x148
>>> AO_SEC_TMODE_PWD0 0x160
>>> AO_SEC_TMODE_PWD1 0x164
>>> AO_SEC_TMODE_PWD2 0x168
>>> AO_SEC_TMODE_PWD3 0x16C
>>> AO_SEC_SCRATCH 0x17C
>>> AO_SEC_JTAG_PWD0 0x180
>>> AO_SEC_JTAG_PWD1 0x184
>>> AO_SEC_JTAG_PWD2 0x188
>>> AO_SEC_JTAG_PWD3 0x18C
>>> AO_SEC_JTAG_SEC_CNTL 0x190
>>> AO_SEC_JTAG_PWD_ADDR0 0x194
>>> AO_SEC_JTAG_PWD_ADDR1 0x198
>>> AO_SEC_JTAG_PWD_ADDR2 0x19C
>>> AO_SEC_JTAG_PWD_ADDR3 0x1A0
>>> AO_SEC_SHARED_AHB_SRAM_REG0_0 0x1C0
>>> AO_SEC_SHARED_AHB_SRAM_REG0_1 0x1C4
>>> AO_SEC_SHARED_AHB_SRAM_REG0_2 0x1C8
>>> AO_SEC_SHARED_AHB_SRAM_REG1_0 0x1CC
>>> AO_SEC_SHARED_AHB_SRAM_REG1_1 0x1D0
>>> AO_SEC_SHARED_AHB_SRAM_REG1_2 0x1D4
>>> AO_SEC_SHARED_AHB_SRAM_REG2_0 0x1D8
>>> AO_SEC_SHARED_AHB_SRAM_REG2_1 0x1DC
>>> AO_SEC_SHARED_AHB_SRAM_REG2_2 0x1E0
>>> AO_SEC_SHARED_AHB_SRAM_REG3_0 0x1E4
>>> AO_SEC_SHARED_AHB_SRAM_REG3_1 0x1E8
>>> AO_SEC_SHARED_AHB_SRAM_REG3_2 0x1EC
>>> AO_SEC_AO_AHB_SRAM_REG0_0 0x1F0
>>> AO_SEC_AO_AHB_SRAM_REG0_1 0x1F4
>>> AO_SEC_AO_AHB_SRAM_REG1_0 0x1F8
>>> AO_SEC_AO_AHB_SRAM_REG1_1 0x1FC
>>> AO_SEC_SD_CFG8 0x220
>>> AO_SEC_SD_CFG9 0x224
>>> AO_SEC_SD_CFG10 0x228
>>> AO_SEC_SD_CFG11 0x22C
>>> AO_SEC_SD_CFG12 0x230
>>> AO_SEC_SD_CFG13 0x234
>>> AO_SEC_SD_CFG14 0x238
>>> AO_SEC_SD_CFG15 0x23C
>>> AO_SEC_GP_CFG0 0x240
>>> AO_SEC_GP_CFG1 0x244
>>> AO_SEC_GP_CFG2 0x248
>>> AO_SEC_GP_CFG3 0x24C
>>> AO_SEC_GP_CFG4 0x250
>>> AO_SEC_GP_CFG5 0x254
>>> AO_SEC_GP_CFG6 0x258
>>> AO_SEC_GP_CFG7 0x25C
>>> AO_SEC_GP_CFG8 0x260
>>> AO_SEC_GP_CFG9 0x264
>>> AO_SEC_GP_CFG10 0x268
>>> AO_SEC_GP_CFG11 0x26C
>>> AO_SEC_GP_CFG12 0x270
>>> AO_SEC_GP_CFG13 0x274
>>> AO_SEC_GP_CFG14 0x278
>>> AO_SEC_GP_CFG15 0x27C
>>>
>>>
>>> As you see, the register we use here is AO_SEC_SD_CFG8...
>>>
>>> Should I define all this block as simple-mfd and refer to it as a regmap ?
>>>
>>> aobus: aobus@c8100000 {
>>> compatible = "simple-bus";
>>> reg = <0x0 0xc8100000 0x0 0x100000>;
>>> #address-cells = <2>;
>>> #size-cells = <2>;
>>> ranges = <0x0 0x0 0x0 0xc8100000 0x0 0x100000>;
>>>
>>> ao_secure: ao-secure@140 {
>>> compatible = "amlogic,meson-gx-ao-secure", "simple-mfd";
>>> reg = <0x0 0x140 0x0 0x140>;
>>> };
>>> };
>>>
>>> chipid {
>>> compatible = "amlogic,meson-gx-socinfo";
>>> ao-secure = <&ao_secure>;
>>> chip-info-reg = <0xe0>;
>>
>> Why even divide it up further in DT? IMO, describing single
>> registers/address in DT is too fine grained.
>>
>> Rob
>>
>
> Rob, I don't get it.
>
> Maybe something like that ?
>
> aobus: aobus@c8100000 {
> compatible = "simple-bus";
> reg = <0x0 0xc8100000 0x0 0x100000>;
> #address-cells = <2>;
> #size-cells = <2>;
> ranges = <0x0 0x0 0x0 0xc8100000 0x0 0x100000>;
>
> ao_secure: ao-secure@140 {
> compatible = "amlogic,meson-gx-ao-secure", "simple-mfd", "simple-bus";
> reg = <0x0 0x140 0x0 0x140>;
> #address-cells = <1>;
> #size-cells = <1>;
>
> chipid@e0 {
> compatible = "amlogic,meson-gx-socinfo";
> reg = <0xe0 0x4>;
> };
> };
> };

That's somewhat better, though your addressing is wrong.

>
> Concerning the fine graining, I'm sorry but the actual information comes from a single register here...

Yes, but the only useful information here is really "0xe0". I imagine
you also want "amlogic,meson-gx-socinfo" to instantiate a driver, but
that's not a reason to put a node into DT. You can just easily have
whatever handles "amlogic,meson-gx-ao-secure" provide the version
information out of register 0xe0.

Rob