On Tue, 4 Oct 2011, Andi Kleen wrote: > On Tue, Oct 04, 2011 at 03:08:02PM +0200, Richard Guenther wrote > > > Sure, the question is if "-flto" counts as magic and thus > > "don't do it when it hurts" ;)) I suppose with -flto-partition=none > > (or 1to1) it would be reasonable to make -fno-toplevel-reorder work > > (and thus maybe -fno-toplevel-reorder should simply force 1to1 > > partitioning). > > At least =none is incredible slow for larger programs. I would really > prefer normal whopr, otherwise at some point LTO becomes unpracticable > due to the extreme compile time penalty. I can use it right now as > a workaround, but it's not a long term solution. > > What I really want is not full -fno-toplevel-reorder anyways, but just > marking a few special variables with a special attribute to not > reorder. But this patch is still needed even for that and likely a good > idea anyways for the other reasons.
Yes, of course - I was just arguing about the supposed final state we want to arrive at with respect to -fno-toplevel-reorder and LTO. The patch is ok (it probably will also help testcase reductions). Thanks, Richard.
