Re: RISC-V: GCC crti.S vs. Newlib __libc_init_array()

On Fri, Jul 27, 2018 at 12:51 AM, Sebastian Huber
<[email protected]> wrote:
> glibc has a more sophisticated configuration machinery than Newlib. I think
> in Newlib there exists currently no generic mechanism for machines
> (architectures) to define options for the generic code. They can only
> overwrite/replace files via the build system. Since machine defines are
> already used throughout the Newlib code base, I will prepare a patch using
> this approach. I think a configuration script detection is less obvious and
> harder to understand.

I already briefly pointed out how to fix it.  If you look at
newlib/configure.host, you will see that there are a bunch of shell
variables that can be set in a target dependent fashion.  We just need
one more to indicate whether _init/_fini are supported.  This is on by
default, and turned off for RISC-V.  Then in configure this shell
variable is used to define a macro that can be tested in
newlib/libc/misc/init.c and fini.c.  Or if that is too complicated, we
could just define a macro in newlib_cflags for RISC-V.

There should be no explicit test for RISC-V in the init.c and fini.c
files.  This is not a RISC-V specific feature.  _init and _fini are
deprecated for all targets.  Most targets just haven't explicitly
dropped support for them yet.

Jim