https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84660
--- Comment #4 from Kito Cheng <npickito at gmail dot com> ---
Hi Jim:
We've written a patch which is based on your patch for gcc 7, and we've tested
with GCC testsuite and several testsuite.
diff --git a/gcc/combine.c b/gcc/combine.c
index ff250aa..033387d 100644
--- a/gcc/combine.c
+++ b/gcc/combine.c
@@ -8770,7 +8770,10 @@ force_to_mode (rtx x, machine_mode mode, unsigned
HOST_WIDE_INT mask,
&& INTVAL (XEXP (x, 1)) < GET_MODE_PRECISION (mode))
&& ! (GET_MODE (XEXP (x, 1)) != VOIDmode
&& (nonzero_bits (XEXP (x, 1), GET_MODE (XEXP (x, 1)))
- < (unsigned HOST_WIDE_INT) GET_MODE_PRECISION (mode))))
+ < (unsigned HOST_WIDE_INT) GET_MODE_PRECISION (mode))
+ && (SHIFT_COUNT_TRUNCATED
+ && nonzero_bits (XEXP (x, 1), GET_MODE (XEXP (x, 1)))
+ < (unsigned HOST_WIDE_INT) GET_MODE_PRECISION (GET_MODE
(x)))))
break;
/* If the shift count is a constant and we can do arithmetic in
