On 07/22/2011 01:33 PM, Paolo Bonzini wrote: > On Fri, Jul 22, 2011 at 02:09, Richard Henderson <r...@redhat.com> wrote: >> I should mention here that I suspect the Cleanest solution is >> to make DF_DEFER_INSN_RESCAN imply that we also shouldn't grab >> data from the blocks those new insns are in either. In this >> way one can create the block, link it up properly, and have >> everything Just Work later, when deferred insns are processed. > > Yes, this also seems like the right approach. I'll try to find some > time. I guess with this in place the rest of the patch should also > Just Work? :)
I should think so, with the caveat that there's a trivial crash with -O0 -freorder-blocks-and-partition in that in that case the df data structures have not yet been allocated at all. The right solution would seem to be to test optimize in gate_handle_partition_blocks, since we also test optimize in gate_handle_reorder_blocks. I'll probably commit a patch to this effect later today; I'm currently testing another patch and do not want to disturb that. r~
