This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH, build] Fix PR 37137
Adam Nemet <anemet@caviumnetworks.com> writes:
> Adam Nemet <anemet@caviumnetworks.com> writes:
>> Paolo Bonzini <bonzini@gnu.org> writes:
>>> I think the problem is that those host libraries should be built with
>>> CFLAGS, not LIBCFLAGS. LIBCFLAGS is a legacy that should be superseded
>>> by CFLAGS_FOR_TARGET, and it should be useless to use it in libiberty.
>>>
>>> In other words, does the attached patch work?
>
> Bootstrapping on mips64octeon-linux-gnu has already completed. Testing is
> still ongoing, I will let you know. My script will bootstrap and test
> x86_64-linux next so by tomorrow morning I should have that as well.
Everything's finished now. Results look good.
Adam
>> I will try it but are you sure this does the right thing WRT libiberty as a
>> target library. Specifically, would target-libiberty be compiled with
>> optimization in stage1? I am thinking of the original bug, PR/32161 (stage1
>> libgcc is built unoptimized)?
>
> To answer my own question, target-libiberty is not bootstrapped so I guess the
> issue does not currently arise.
>
> Can you please add the comments from your explanation to the STAGE*_LIBCFLAGS
> variables or whereever you think it's most helpful. My patch was trying to
> also improve the code and explain these variables.
>
> Adam