[PATCH 3/7] ARM: mmp: set correct uart according to board

[Eric Miao]

On Wed, Apr 13, 2011 at 5:01 AM, Russell King - ARM Linux
<linux at arm.linux.org.uk> wrote:
> On Fri, Apr 08, 2011 at 08:15:40PM +0800, Haojian Zhuang wrote:
>> UART3 is the default console in jasper, brownstone and avengers lite.
>>
>> Change-Id: Iacfe665cc0e20a47d50e0b2ad8724097f28e7d2c
>
> Please don't include Change-Ids in patch submissions.
>
>> diff --git a/arch/arm/mach-mmp/include/mach/debug-macro.S b/arch/arm/mach-mmp/include/mach/debug-macro.S
>> index 7e2ebd3..b353785 100644
>> --- a/arch/arm/mach-mmp/include/mach/debug-macro.S
>> +++ b/arch/arm/mach-mmp/include/mach/debug-macro.S
>> @@ -14,8 +14,14 @@
>>               .macro  addruart, rp, rv
>>               ldr     \rp, =APB_PHYS_BASE             @ physical
>>               ldr     \rv, =APB_VIRT_BASE             @ virtual
>> +#if defined(CONFIG_MACH_BROWNSTONE) || defined(CONFIG_MACH_MARVELL_JASPER) \
>> +     || defined(CONFIG_MACH_AVENGERS_LITE)
>> +             orr     \rp, \rp, #0x00018000
>> +             orr     \rv, \rv, #0x00018000
>> +#else
>>               orr     \rp, \rp, #0x00017000
>>               orr     \rv, \rv, #0x00017000
>> +#endif
>
> It probably makes more sense to have this in the configuration scripts:
>
> choice
>        prompt "PXA debug uart"
>        default DEBUG_PXA_UART0
>
> config DEBUG_PXA_UART0
>        bool "UART 0"
>
> config DEBUG_PXA_UART1
>        bool "UART 1"
>        default y if MACH_BROWNSTONE || MACH_MARVELL_JASPER || MACH_AVENGERS_LITE
>
> endchoice
>
> and select the debug uart address using these symbols.  That'd avoid the
> need to couple lots of platform specific stuff into the compiled files.
>
> Alternatively, you could do more in the Kconfig like this:
>
> config DEBUG_PXA_UART
>        hex
>        default 0x17000 if DEBUG_PXA_UART0
>        default 0x18000 if DEBUG_PXA_UART1
>
> and use CONFIG_DEBUG_PXA_UART in the assembly code, eliminating the
> need to change sourcecode for additional uarts.  Or maybe combine the
> two ideas.
>

This looks much cleaner. I was actually thinking if it's necessary to make
this run-time determined, when multiple boards are all selected.