[Bug rtl-optimization/127606] [17 Regression] wrong code at -O{2,3} with "-fno-guess-branch-probability -fno-tree-loop-im -fno-tree-pre -fno-forward-propagate -fno-ipa-ra -fno-tree-ch" on x86_64-linux-gnu

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127606

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Sam James from comment #3)
> I can't repro SIGFPE but I do see this under Valgrind with the long option
> list only:
> 
> ==1848749== Use of uninitialised value of size 8
> ==1848749==    at 0x4001248: cm (a.c:26)
> ==1848749==    by 0x4001285: cq (a.c:32)
> ==1848749==    by 0x4001FC0: cp (a.c:181)
> ==1848749==    by 0x40021E7: dd (a.c:144)
> ==1848749==    by 0x40021E7: main (a.c:268)
> ==1848749==
> ==1848749== Use of uninitialised value of size 8
> ==1848749==    at 0x4001255: cm (a.c:26)
> ==1848749==    by 0x4001285: cq (a.c:32)
> ==1848749==    by 0x4001FC0: cp (a.c:181)
> ==1848749==    by 0x40021E7: dd (a.c:144)
> ==1848749==    by 0x40021E7: main (a.c:268)

Ah, I can reproduce that.  Gemini is having a look.