This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] target/37283: Fix -fno-unit-at-a-time and -fno-toplevel-reorder handling
>>>>> "Andrew" == Andrew Pinski <pinskia@gmail.com> writes:
Andrew> I think the arm back-end should do the same thing as the
Andrew> rs6000 back-end does with respect of -fsection-anchors.
It does mostly the same thing (set flag_section_anchors to 2 instead
of 1), but I do not know whether this makes sense not to set the flag
in C++ and Objective-C (what about Ada and other front-ends?). My
patch errs on the safe side and is a minimal change.
Incidentally, I wonder whether the rs6000 handling may be bogus too:
it checks flags_toplevel_reorder, but OPTIMIZATION_OPTIONS is called
*before* handle_options, so flags_toplevel_reorder still holds its
default value, not the user-chosen one. The bug is harmless though as
flags_section_anchors will be reset properly if needed (with my patch
applied) but the test "flag_toplevel_reorder != 0" could be removed
from rs6000.c.
Sam
--
Samuel Tardieu -- sam@rfc1149.net -- http://www.rfc1149.net/