This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: inlining default
Jason Merrill wrote:-
> >>>>> "Neil" == Neil Booth <neil@daikokuya.demon.co.uk> writes:
>
> > Jason Merrill wrote:-
> >> It looks like if (!optimize && flag_instrument_function_entry_exit), we
> >> will end up trying to do RTL inlining and fail.
>
> > Could you elaborate?
>
> Sorry, it would also need flag_no_inline = 0. Then your changed code
> doesn't ever set it again, because flag_no_inline is not 1, and f_i_f_e_e
> is not 0.
Hmmm. Still not clear.
You're saying that
-finline -O0 -finstrument-functions
should set flag_no_inline to 1 despite the -finline, otherwise it will
"fail". I agree that the current code keeps it at 0. Why is that
bad? Why will it fail? (I don't know the RTL part of the compiler
well.)
Neil.