[PATCH v2] arch: Fix image and bitmap byte order for ppc64le

[Alan Coopersmith]

On 02/24/14 08:19 AM, Guillem Jover wrote:
> On Mon, 2014-02-24 at 08:09:33 -0800, Keith Packard wrote:
>> walter harms <wharms at bfs.de> writes:
>>> Assuming that __LITTLE_ENDIAN__ is of general interest,
>>> i would avoid defined(__ppc__) and friends otherwise we will need to define
>>> this for every achitecture.
>>
>> That would risk breaking little endian platforms that failed to define
>> this. Safest to limit the scope of the change the PPC platforms, where
>> __LITTLE_ENDIAN__ is a new option, and hence likely to be set correctly.
>>
>> I would sure like to be able to detect the difference between 'this
>> option is set to false' and 'this option isn't defined'; is there
>> something that PPC defines in both cases, and sets to TRUE in one case
>> and FALSE in the other?
>
> I'm probably missing something (maybe this header cannot include a
> dix-config.h or equivalent), but why not use AC_C_BIGENDIAN?

Because tradition!  And because Imake didn't work that way!

But mostly because no one has had the patience to change it since,
and to make a servermd.h.in that configure can generate servermd.h
from to install for use by driver builds that may not have the exact
same macros set by AC_C_BIGENDIAN in their builds.

-- 
	-Alan Coopersmith-              alan.coopersmith at oracle.com
	 Oracle Solaris Engineering - http://blogs.oracle.com/alanc