Re: [PATCH] x86-64: more fixes and cleanup to AMD Fam10 MMCONF enabling

["Jan Beulich" <JBeulich@novell.com>]

>>> On 08.11.10 at 17:13, "H. Peter Anvin" <hpa@zytor.com> wrote:
> On 11/05/2010 03:59 AM, Jan Beulich wrote:
>>  
>> --- 2.6.37-rc1-x86_64-mmconf-fam10h.orig/arch/x86/kernel/mmconf-fam10h_64.c
>> +++ 2.6.37-rc1-x86_64-mmconf-fam10h/arch/x86/kernel/mmconf-fam10h_64.c
>> @@ -43,7 +43,7 @@ static int __cpuinit cmp_range(const voi
>>  	return start1 - start2;
>>  }
>>  
>> -#define UNIT (1ULL << (5 + 3 + 12))
>> +#define UNIT (1ULL << FAM10H_MMIO_CONF_BASE_SHIFT)
>>  #define MASK (~(UNIT - 1))
>>  #define SIZE (UNIT << 8)
> 
> Could we avoid macros named UNIT, MASK, and SIZE at all?  I realize
> they're already in the code, but still...

I could understand if these were definition in a header, but why
do you think we need to have unnecessarily long identifiers (e.g.
by prefixing all of the defines here with FAM10H_MMIO_CONF_BASE_)
in places like this? After all, one of the two goals of using a macro
here at all is to keep things small and simple...

But sure, if just the names hinder acceptance, I can fold this and
the original patches together and use less ambiguous names.

Jan