[PATCHv3][Ping] ldbl-128ibm-compat: Add tests for IBM long double functions

[Paul E Murphy]

And committed.

On 3/3/20 2:17 PM, Paul E Murphy wrote:
> I will commit this friday if there are no more requests for improvement.
> 
> On 2/28/20 9:33 AM, Paul E Murphy wrote:
>> Ping.
>>
>> On 2/18/20 4:00 PM, Paul E. Murphy wrote:
>>> From: Rajalakshmi Srinivasaraghavan <raji@linux.vnet.ibm.com>
>>>
>>> Simplify this further.  As correctly pointed out by Joseph, compiler
>>> _Float128 compiler flags should always be present where needed.  Thus,
>>> I can greatly simplify this patch.  It had collected too much cruft from
>>> my constant rebasing.
>>>
>>> Likewise, defer the ppc64le makefile changes.  They really shouldn't be
>>> here, and contained harmful rebasing noise.  They have been squashed
>>> into later, more appropriate patches.
>>>
>>> Similarly, the hack for gamma is removed.  `make regen-ulps` behaves
>>> as expected and fills the missing entries correctly.  The ULP changes
>>> will be included with the enablement patch.
>>>
>>> Tested on x86-64 and ppc64le.
>>>
>>> ---8<---
>>>
>>> This patch creates test-ibm128* tests from the long double function 
>>> tests.
>>> In order to explicitly test IBM long double functions 
>>> -mabi=ibmlongdouble is
>>> added to CFLAGS.
>>>
>>> Likewise, update the test headers to correct choose ULPs when redirects
>>> are enabled.
>>>
>>> Co-Authored-By: Tulio Magno Quites Machado Filho  <tuliom@linux.ibm.com>
>>> Co-Authored-By: Paul E. Murphy  <murphyp@linux.vnet.ibm.com>