Re: [PATCH] build-many-glibcs.py: Add some s390x glibc variants.

[Adhemerval Zanella via Libc-alpha <libc-alpha@sourceware.org>]

On 05/08/2020 12:39, Florian Weimer via Libc-alpha wrote:
> * Stefan Liebler via Libc-alpha:
> 
>> There is a s390x configure check which checks the architecture level set.
>>
>> This ALS influences e.g. which ifunc variants are needed or which one is
>> the default variant or if the symbol will be an ifunc-symbol at all.
>>
>> The ALS also enables to use the gcc builtins for e.g. the round function
>> in libm and others.
>>
>> Therefore this patch adds some glibc variants which are using different
>> architecture level sets for s390x.
> 
> Do these additional build targets actually result in build breakage?
> 
> I'm not sure if anyone does run-time testing on build-many-glibcs.py
> output.> 
> I would rather suggest adding an -O3 targets (maybe s390x due to its
> inliner differences, and x86-64), where we know we have occasional build
> breakage.

Such build permutations stress how s390 port is organized internally,
where it adds an optimization to avoid build some ifunc variations
if the minimum defined architecture avoids selecting it (different
tha x86_64 that always build all ifunc variant regardless of the
target chip/ABI). Another build permutation is whether it will use
the compiler builtins on some match function or whether it will use
the generic implementation.

I don't have a strong opinion on a '-O3' variant, although currently
build-many-glibcs.py usually focus on ABIs variants that might
trigger different code path within glibc itself (although it does
exercise the compiler itself).