This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH v6 00/13]port C-SKY to glibc
On Thu, Nov 29, 2018 at 05:16:58PM +0000, Joseph Myers wrote:
> (There are other ways of ensuring GCC uses initfini-array support
> unconditionally for a target - look at what's done for Arm EABI, which
> also uses initfini-array only. But in the first instance, fixing the
> configure test is the natural thing to do to make this work better for all
> ELF architectures. In any case, this should be fixed so that GCC gets it
> right and no special use of --enable-initfini-array is needed in
> build-many-glibcs.py.)
>
We trend to solve this issue like what RISC-V did, force
gcc_cv_initfini_array=yes in config.gcc when glibc is built.
The first way may take quite a long time to find out what compilation test
is required to get things right. We'v got some question about what compilation
test should be like for C-SKY. We'v got several other C library do not use
init-array at present. When gcc is built first time we are not able
to find out whether the C library want init-array or not, if gcc is not
rebuilt for C++, init/fini will never be called. Even if gcc is rebuilt,
we can not execute test program if it is cross-compiled. We may try to make
init-array check in configure work later when we find ways to solve the
issues above.
Best Regards,
Mao Han