Richard Henderson <r...@redhat.com> wrote:
> It does raise the question of whether it's worth while to move these
> sequences inline in the compiler? We'd need to keep the symbols from
> linux-atomic.S of course, but there's no reason we couldn't satisfy them with
> now-trivial C code.
The same question was raised to me when writing sh/linux-atomic.S.
I was not sure whether the instruction sequences depending on
a specific kernel feature is good or not and the library functions
looked to be enough at that time. The situation has changed though.
Regards,
kaz
