Ilya Leoshkevich <[email protected]> writes:

> On 1/22/26 17:43, Alex Bennée wrote:
>> Ilya Leoshkevich <[email protected]> writes:
>>
>>> Add a test to prevent regressions. Data is generated using a
>>> libFuzzer-based fuzzer and hopefully covers all the important corner
>>> cases.
>>>
>>> Signed-off-by: Ilya Leoshkevich <[email protected]>
>>> ---
>>>   tests/tcg/s390x/Makefile.target     |   3 +
>>>   tests/tcg/s390x/divide-to-integer.c | 215 ++++++++++++++++++++++++++++
>>>   2 files changed, 218 insertions(+)
>>>   create mode 100644 tests/tcg/s390x/divide-to-integer.c
>>>
>>> diff --git a/tests/tcg/s390x/Makefile.target 
>>> b/tests/tcg/s390x/Makefile.target
>>> index da5fe71a407..d5ec01d04fd 100644
>>> --- a/tests/tcg/s390x/Makefile.target
>>> +++ b/tests/tcg/s390x/Makefile.target
>>> @@ -49,14 +49,17 @@ TESTS+=cvd
>>>   TESTS+=cvb
>>>   TESTS+=ts
>>>   TESTS+=ex-smc
>>> +TESTS+=divide-to-integer
>>>     cdsg: CFLAGS+=-pthread
>>>   cdsg: LDFLAGS+=-pthread
>>>     rxsbg: CFLAGS+=-O2
>>> +divide-to-integer: CFLAGS+=-O2
>> As we generally compile -O0 to make life easier for people debugging
>> behaviour via gdbstub could we have an explanation of why -O2 is needed
>> here? Is it the same reason as rxsbg?
>>
>> <snip>
>
> Yes, this is because inlining is mandatory for the mask argument, and
> -O0 does not honor always_inline.

Could we add a comment to that effect in the Makefile. Otherwise looks
fine to me:

Acked-by: Alex Bennée <[email protected]>


-- 
Alex Bennée
Virtualisation Tech Lead @ Linaro

Reply via email to