The comment says this test is supposed to prevent "a narrower
operation than requested", but it actually only allows a larger
subreg, not one the same size. Fix that.
Bootstrapped and regression tested powerpc64-linux. OK for stage1?
Note that this bug was found when investigating why gcc-6 does not
suffer from pr69548, ie. this bug was masking a powerpc backend bug.
* combine.c (simplify_set): Correct WORD_REGISTER_OPERATIONS test.
diff --git a/gcc/combine.c b/gcc/combine.c
index 858552d..9f284a7 100644
--- a/gcc/combine.c
+++ b/gcc/combine.c
@@ -6736,7 +6736,7 @@ simplify_set (rtx x)
+ (UNITS_PER_WORD - 1)) / UNITS_PER_WORD))
&& (WORD_REGISTER_OPERATIONS
|| (GET_MODE_SIZE (GET_MODE (src))
- < GET_MODE_SIZE (GET_MODE (SUBREG_REG (src)))))
+ <= GET_MODE_SIZE (GET_MODE (SUBREG_REG (src)))))
#ifdef CANNOT_CHANGE_MODE_CLASS
&& ! (REG_P (dest) && REGNO (dest) < FIRST_PSEUDO_REGISTER
&& REG_CANNOT_CHANGE_MODE_P (REGNO (dest),
--
Alan Modra
Australia Development Lab, IBM
