[U-Boot] [linux-sunxi] Re: [PATCH 02/13] sunxi: add basical memory map definitions of H6 SoC

[Andre Przywara]

Hi,

On 25/06/18 13:50, Icenowy Zheng wrote:
> 
> 
> 于 2018年6月25日 GMT+08:00 下午8:33:34, Maxime Ripard <maxime.ripard at bootlin.com> 写到:
>> On Mon, Jun 25, 2018 at 06:37:12PM +0800, Icenowy Zheng wrote:
>>> The Allwinner H6 SoC come with a totally new memory map.
>>>
>>> Add basical definition of the new memory map into a header file, and
>> let
>>> the cpu.h header include it in the situation of H6.
>>>
>>> Signed-off-by: Icenowy Zheng <icenowy at aosc.io>
>>> ---
>>>  arch/arm/include/asm/arch-sunxi/cpu.h         |  2 +
>>>  .../include/asm/arch-sunxi/cpu_sun50i_h6.h    | 73
>> +++++++++++++++++++
>>>  2 files changed, 75 insertions(+)
>>>  create mode 100644 arch/arm/include/asm/arch-sunxi/cpu_sun50i_h6.h
>>>
>>> diff --git a/arch/arm/include/asm/arch-sunxi/cpu.h
>> b/arch/arm/include/asm/arch-sunxi/cpu.h
>>> index caec865264..08be963e8e 100644
>>> --- a/arch/arm/include/asm/arch-sunxi/cpu.h
>>> +++ b/arch/arm/include/asm/arch-sunxi/cpu.h
>>> @@ -9,6 +9,8 @@
>>>  
>>>  #if defined(CONFIG_MACH_SUN9I)
>>>  #include <asm/arch/cpu_sun9i.h>
>>> +#elif defined(CONFIG_MACH_SUN50I_H6)
>>> +#include <asm/arch/cpu_sun50i_h6.h>
>>>  #else
>>>  #include <asm/arch/cpu_sun4i.h>
>>>  #endif
>>> diff --git a/arch/arm/include/asm/arch-sunxi/cpu_sun50i_h6.h
>> b/arch/arm/include/asm/arch-sunxi/cpu_sun50i_h6.h
>>> new file mode 100644
>>> index 0000000000..b12f2dd1a2
>>> --- /dev/null
>>> +++ b/arch/arm/include/asm/arch-sunxi/cpu_sun50i_h6.h
>>> @@ -0,0 +1,73 @@
>>> +/*
>>> + * (C) Copyright 2017 Icenowy Zheng <icenowy at aosc.io>
>>> + *
>>> + * SPDX-License-Identifier:	GPL-2.0+
>>> + */
>>> +
>>> +#ifndef _SUNXI_CPU_SUN50I_H6_H
>>> +#define _SUNXI_CPU_SUN50I_H6_H
>>> +
>>> +#define SUNXI_SRAM_A1_BASE		0x00020000
>>
>> Isn't that the same thing than SUNXI_SRAM_ADDRESS? Having it defined
>> twice and expecting the value to be matching seems a bit weird.
> 
> This assumes SPL will be in SRAM A1, which may change.

Can it? I believe this part is fixed, because this is where the BootROM
loads the initial code to. And for the boot flow we support the SPL is
the first one to execute. Not sure if a TPL would change that, but as
Maxime said, we can address that when we really need to.

I guess we can't use that symbol in Kconfig, can we? So that we could
say: "default SUNXI_SRAM_A1_BASE if MACH_SUN50I_H6" in patch 12/13 ...

Cheers,
Andre