[U-Boot] [PATCH RFC] sandbox: Add 64-bit sandbox

[Simon Glass]

Hi Mario,

On 14 January 2018 at 23:23, Mario Six <mario.six at gdsys.cc> wrote:
> Hi Simon,
>
> On Fri, Dec 29, 2017 at 4:13 AM, Simon Glass <sjg at chromium.org> wrote:
>> Hi Mario,
>>
>> On 20 December 2017 at 07:31, Mario Six <mario.six at gdsys.cc> wrote:
>>> From: Mario Six <six at gdsys.cc>
>>>
>>> To debug device tree issues involving 32- and 64-bit platforms, it is useful to
>>> have a generic 64-bit platform available.
>>>
>>> Add a version of the sandbox that uses 64-bit integers for its physical
>>> addresses as well as a modified device tree.
>>>
>>> Signed-off-by: Mario Six <mario.six at gdsys.cc>
>>>
>>> ---
>>>  arch/sandbox/Kconfig             |   6 +
>>>  arch/sandbox/cpu/cpu.c           |   2 +-
>>>  arch/sandbox/dts/Makefile        |   4 +
>>>  arch/sandbox/dts/sandbox64.dts   | 317 +++++++++++++++++++++++++++++++++++++++
>>>  arch/sandbox/include/asm/io.h    |   6 +
>>>  arch/sandbox/include/asm/types.h |  14 +-
>>>  cmd/demo.c                       |   6 +-
>>>  configs/sandbox64_defconfig      | 200 ++++++++++++++++++++++++
>>>  drivers/demo/demo-simple.c       |   2 +-
>>>  9 files changed, 550 insertions(+), 7 deletions(-)
>>>  create mode 100644 arch/sandbox/dts/sandbox64.dts
>>>  create mode 100644 configs/sandbox64_defconfig
>>
>> Reviewed-by: Simon Glass <sjg at chromium.org>
>>
>> Can you please update the sandbox README?
>>
>
> Sure, I'll extend the README.
>
>> Also, how does this play with CONFIG_SANDBOX_64BIT and CONFIG_PHYS_64BIT ?
>>
>
> Yeah, the names of the config options are all a bit confusing (hence why I
> added the RFC to see if someone may have a better idea). Setting the
> CONFIG_PHYS_64BIT is not strictly necessary for the 64-bit sandbox, but some
> architectures use the option to decide whether they run on a 32- or 64-bit
> system, so I thought it was consistent to also set it, just in case. The
> CONFIG_SANDBOX_64BIT option causes a 64-bit sandbox *binary* to be built. This
> binary still uses 32-bit-wide types in the device tree (phys_addr_t and
> phys_size_t, mostly, so the width of the physical addresses are still 32-bit),
> so that's not what we want in this case. I decided to override the type
> definitions in arch/sandbox/include/asm/types.h based on CONFIG_SANDBOX64
> directly, since I didn't want to introduce yet another option that has the
> words sandbox and 64 in its name. But it's pretty confusing still, I admit.
>
> Maybe we could rename CONFIG_SANDBOX_64BIT to CONFIG_HOST_64BIT, and use
> CONFIG_PHYS_64BIT to control the width of the physical addresses in types.h
> (which is what it is used for in arch/arm/include/asm/types.h now that I'm
> taking a closer look); that would at least make the names congruent with their
> actual semantics. We would then have the options CONFIG_HOST_64BIT,
> CONFIG_PHYS_64BIT, and CONFIG_SANDBOX64 as the "board option".
>
> Maybe that's a cleaner solution?

Yes I think that is a good idea.

Regards,
Simon