This is the mail archive of the
gcc-bugs@gcc.gnu.org
mailing list for the GCC project.
[Bug target/52555] [4.6/4.7/4.8 Regression] ICE unrecognizable insn with -ffast-math and __attribute__((optimize(xx)))
- From: "jakub at gcc dot gnu.org" <gcc-bugzilla at gcc dot gnu dot org>
- To: gcc-bugs at gcc dot gnu dot org
- Date: Wed, 15 Aug 2012 14:21:54 +0000
- Subject: [Bug target/52555] [4.6/4.7/4.8 Regression] ICE unrecognizable insn with -ffast-math and __attribute__((optimize(xx)))
- Auto-submitted: auto-generated
- References: <bug-52555-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52555
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rth at gcc dot gnu.org
--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2012-08-15 14:21:54 UTC ---
Another option is to switch this_target_optabs not just for SWITCHABLE_TARGETs,
but also when switching between cfun with different optimize/target attributes
(at the time when the optimization/target node is created for a function it
would try to init_all_optabs into a newly allocated array, and if that differed
from the default, it would stick it into the node and switch at set_cfun time).
I actually wonder how could target attribute work the way it is implemented
right now so far.