https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103267
--- Comment #9 from hubicka at kam dot mff.cuni.cz --- > @@ -1,4 +1,3 @@ > -static int > __attribute__ ((noinline,const)) > infinite (int p) > { Just for a record, it crahes with or without static int here for me :) I run across it because the code tracking must access in ipa-sra is IMO conceptually wrong. I noticed that because ipa-modref solves similar problem for kills (both need to verify that given access will always happen). The post-dominance check is not enough to verify that because earlier function calls can do things like EH. I failed to construct an actual testcase because on interesting stuff like EH we punt for other reasons (missed fnspec annotations on EH builtins). I will play with it more today.