http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49678
Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target| |arm
Status|UNCONFIRMED |NEW
Last reconfirmed| |2011.07.08 16:47:18
CC| |rearnsha at gcc dot gnu.org
Ever Confirmed|0 |1
--- Comment #1 from Richard Earnshaw <rearnsha at gcc dot gnu.org> 2011-07-08
16:47:18 UTC ---
I think the problem here is that sched1 is re-arranging the insns in a way that
makes register allocation without using a mov pretty much impossible. Running
the test at -Os seems to work because the code reorg doesn't happen.
However, looking at the code generated after sched1 has run doesn't seem to be
any better than before (indeed it often looks worse); but I'm not sure yet why
this happens.
