On 11/15/11 01:43, Richard Henderson wrote: > On 11/14/2011 11:56 AM, Alan Modra wrote: >> * function.c (thread_prologue_and_epilogue_insns): Guard >> emitting return with single_succ_p test. > > Hmm. This looks plausible too. > > Bernd's patch made sure that cfglayout didn't do something impossible. > Of course, it's possible that his patch should merely be an assert, > and the correct fix goes here.
I haven't really looked at Alan's patch; I'm thinking there were probably two bugs - the one I found being latent before. Note that the problematic CFG exists way before thread_prologue_and_epilogue even runs, so an assert would still trigger. Bernd