------- Comment #3 from jakub at gcc dot gnu dot org 2010-03-02 08:04 ------- Unfortunately, the patch didn't bootstrap. One issue is loc_cmp, which asserts everywhere the modes are the same. Which isn't necessarily true, we can end up comparing e.g.: (subreg:QI (value/s/u:DI 28:4092 @0x4828ae8/0x47ad420) 0) and (subreg:QI (value/s/u:SI 94:208 @0x4781938/0x47825f0) 0) and ICE on it.
-- http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43177