Re: [PATCH v4 2/3] arm: shmobile: Add the R9A06G032 SMP enabler driver

[Rob Herring]

On Wed, Jun 6, 2018 at 2:42 PM, Geert Uytterhoeven <geert@xxxxxxxxxxxxxx> wrote:
> Hi Rob,
>
> On Wed, Jun 6, 2018 at 9:35 PM, Rob Herring <robh+dt@xxxxxxxxxx> wrote:
>> On Wed, Jun 6, 2018 at 2:30 PM, Frank Rowand <frowand.list@xxxxxxxxx> wrote:
>>> On 06/05/18 23:36, Michel Pollet wrote:
>>>> On 05 June 2018 18:34, Frank wrote:
>>>>> On 06/05/18 04:28, Michel Pollet wrote:
>>>>>> The Renesas R9A06G032 second CA7 is parked in a ROM pen at boot time,
>>>>>> it requires a special enable method to get it started.
>>>>>>
>>>>>> Signed-off-by: Michel Pollet <michel.pollet@xxxxxxxxxxxxxx>
>
>>>>>> --- /dev/null
>>>>>> +++ b/arch/arm/mach-shmobile/smp-r9a06g032.c
>
>>>>>> +/*
>>>>>> + * The second CPU is parked in ROM at boot time. It requires waking
>>>>>> +it after
>>>>>> + * writing an address into the BOOTADDR register of sysctrl.
>>>>>> + *
>>>>>> + * So the default value of the "cpu-release-addr" corresponds to
>>>>> BOOTADDR...
>>>>>> + *
>>>>>> + * *However* the BOOTADDR register is not available when the kernel
>>>>>> + * starts in NONSEC mode.
>>>>>> + *
>>>>>> + * So for NONSEC mode, the bootloader re-parks the second CPU into a
>>>>>> +pen
>>>>>> + * in SRAM, and changes the "cpu-release-addr" of linux's DT to a
>>>>>> +SRAM address,
>>>>>> + * which is not restricted.
>>>>>
>>>>> The binding document for cpu-release-addr does not have a definition for 32
>>>>> bit arm.  The existing definition is only 64 bit arm.  Please add the definition
>>>>> for 32 bit arm to patch 1.
>>>>
>>>> Hmmm I do find a definition in
>>>> Documentation/devicetree/bindings/arm/cpus.txt -- just under where I
>>>> added my 'enable-method' -- And it is already used as 32 bits in at least
>>>> arch/arm/boot/dts/stih407-family.dtsi.
>>>
>>> From cpus.txt:
>>>
>>>         - cpu-release-addr
>>>                 Usage: required for systems that have an "enable-method"
>>>                        property value of "spin-table".
>>>                 Value type: <prop-encoded-array>
>>>                 Definition:
>>>                         # On ARM v8 64-bit systems must be a two cell
>>>                           property identifying a 64-bit zero-initialised
>>>                           memory location.
>>>
>>> The definition specifies a two cell property for 64-bit systems.
>>>
>>> Please add to the definition that cpu-release-addr is a one cell property
>>> for 32-bit systems.
>>
>> Actually, this is all already documented in the DT spec and it is
>> always 2 cells[1]. We should perhaps just remove whatever is
>> duplicated from the spec.
>>
>> Rob
>>
>> [1]
>>    ``cpu-release-addr``   | SD  | ``<u64>``        The
>> cpu-release-addr property is required for
>>                                                    cpu nodes that have
>> an enable-method property
>>                                                    value of
>> ``"spin-table"``. The value specifies the
>>                                                    physical address of
>> a spin table entry that
>>                                                    releases a
>> secondary CPU from its spin loop.
>
> Interesting. But why is this <u64>, and not just following #address-cells?

As you said in your other email, it's not the same.

> Due to the ePAPR-spec being 64-bit Power System-centric?

Other than #address-cells already having another defined purpose in
/cpus, my guess is 64-bit works for either and 32-bit SMP systems
didn't predate 64-bit (for the ePAPR author's perspective).

> There's also "initial-mapped-area", which must use 64-bit values for effective
> and physical addresses, according to ePAPR.

I would have thought that followed #{size,address}-cells being
/memory. Though, I guess the bootloader fills this in and it is much
easier to work with fixed sizes.

Rob