[OpenAFS] clarification on client caches

Horst Birthelmer horst@riback.net
Fri, 10 Sep 2004 16:35:53 +0200

On Sep 10, 2004, at 4:22 PM, Derek Atkins wrote:

> Jim Rees <rees@umich.edu> writes:
>>   Why would it need to be conditional on AFS_64BIT_ENV?  Linux 
>> supports
>>   the "long long" 64-bit integer type natively in the 32bit OS just 
>> fine.
>> I'm amazed you would say that.  You of all people should know that 
>> linux is
>> not the only OS in the world.  It's not even the only OS that OpenAFS 
>> runs
>> on.
> Sorry, *BSD supports it, Solaris supports it, MacOS supports it, I'm
> fairly sure windows supports it, too..  So, MANY of the OSes out there
> support it...  I can't claim that ALL the supported OSes have a native
> 64-bit integer type, but I would bet a dollar to a dime that a
> majority of them do.
> My confusion is about what AFS_64BIT_ENV actually signifies.  Does it
> mean "we have a 64-bit integer type"?  Or "we support 64-bit file
> lengths"?  Or "this is a native 64-bit OS?"  Or does it mean something
> completely different?

 From the code ... it has nothing to do with large files because it was 
there before the large file support.
In the beginning I think it was there for 64 bit environments :-)) like 
OSes on 64 bit hardware.
That changed a little bit since almost every platform has a 64 bit 
derivative ;-).
I'm speaking from experience from the PPC64 port.