[RFC] Extend ipa-bitwise-cp with pointer alignment propagation

Prathamesh Kulkarni prathamesh.kulkarni@linaro.org
Thu Oct 6 20:55:00 GMT 2016


On 6 October 2016 at 18:51, Jan Hubicka <hubicka@ucw.cz> wrote:
>> >
>> > What do you mean by "for instance?"  What are the other cases when it
>> > happens?
>> Well ipa_get_type() returned NULL for 481.wrf, and I assumed it was a
>> fortran-only
>> code-base but apparently it's a mix of C and fortran.
>
> Yep, I also have expreinece that the K&R style declarations are more common
> than one would expect 30 years after the language was obsoletted.
>> ipa_get_type() returned NULL for param 'n' in
>> get_initial_data_value_() which is defined
>> as K&R style function in wrf_num_bytes_between.c, so it seems K&R is the
>> only case where ipa_get_type() returns NULL.
>>
>> I have a question though on how to reduce fortran test-cases or a mix
>> of fortran/C
>> like 481.wrf ? For C/C++ only tests-cases, I use creduce.
>
> I use delta/multidelta
> https://gcc.gnu.org/wiki/A_guide_to_testcase_reduction
>
> It works pretty well and one can reduce first list of files and then file at a
> time.
Thanks! I will try to reduce 481.wrf assert ICE for practice -;)
Is the alignment propagation patch OK to commit ?

Thanks,
Prathamesh
>
> Honza



More information about the Gcc-patches mailing list