[5/5] math: Sync atanh with CORE-MATH

[Adhemerval Zanella Netto]

I am not sure why the CI is saying that the *atanh* patch it the one that
has failed, the patch [1] itself is for log2p1f.

[1] https://www.delorie.com/trybots/32bit/57055/the-patch.txt

On 02/03/26 09:49, Paul Zimmermann wrote:
> ok, but why does it break the atanh patch and not the log2p1f one?
> 
> Paul
> 
>> Date: Mon, 2 Mar 2026 09:42:14 -0300
>> From: Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>
>> Organization: Linaro
>>
>>
>>
>> On 02/03/26 07:07, Paul Zimmermann wrote:
>>>        Hi,
>>>
>>> this patch fails one bot test:
>>>
>>> https://www.delorie.com/trybots/32bit/57055/
>>>
>>> However I don't see any relation with atanh,
>>> since the failures are in log2p1f:
>>>
>>> https://www.delorie.com/trybots/32bit/57055/math-test-float-log2p1.out
>>>
>>> Paul
>>
>> I think this is due a missing fix that I did on CORE-MATH and forgot
>> to actually do in glibc [1].
>>
>> [1] https://sourceware.org/pipermail/libc-alpha/2026-January/174670.html
>>