https://gcc.gnu.org/bugzilla/show_bug.cgi?id=71779
--- Comment #17 from Bernd Edlinger <bernd.edlinger at hotmail dot de> ---
And here an alternative patch for the other paradoxical subreg bug:
Index: cse.c
===================================================================
--- cse.c (revision 238891)
+++ cse.c (working copy)
@@ -4716,10 +4716,6 @@ cse_insn (rtx_insn *insn)
}
}
-#if 0
- /* It is no longer clear why we used to do this, but it doesn't
- appear to still be needed. So let's try without it since this
- code hurts cse'ing widened ops. */
/* If source is a paradoxical subreg (such as QI treated as an SI),
treat it as volatile. It may do the work of an SI in one context
where the extra bits are not being used, but cannot replace an SI
@@ -4726,7 +4722,6 @@ cse_insn (rtx_insn *insn)
in general. */
if (paradoxical_subreg_p (src))
sets[i].src_volatile = 1;
-#endif
/* Locate all possible equivalent forms for SRC. Try to replace
SRC in the insn with each cheaper equivalent.
I'd be happy if you could try a boot-strap with this patch too.
Thanks
Bernd.
