Since r230164, in PR68330 combine ends up with a sign_extend of an
lshiftrt by some constant, and it does not know to morph that into a
zero_extract (the extend will always extend with zeroes). I think
it is best to let simplify-rtx always replace such a sign_extend by
a zero_extend, after which everything works as expected.
Bootstrapped and tested on powerpc64-linux. Is this okay for trunk?
Segher
2015-11-15 Segher Boessenkool <seg...@kernel.crashing.org>
PR rtl-optimization/68330
* simplify-rtx.c (simplify_unary_operation_1): Simplify SIGN_EXTEND
of LSHIFTRT by a non-zero constant integer.
---
gcc/simplify-rtx.c | 7 +++++++
1 file changed, 7 insertions(+)
diff --git a/gcc/simplify-rtx.c b/gcc/simplify-rtx.c
index c4fc42a..413d61b 100644
--- a/gcc/simplify-rtx.c
+++ b/gcc/simplify-rtx.c
@@ -1462,6 +1462,13 @@ simplify_unary_operation_1 (enum rtx_code code,
machine_mode mode, rtx op)
}
}
+ /* (sign_extend:M (lshiftrt:N <X> (const_int I))) is better as
+ (zero_extend:M (lshiftrt:N <X> (const_int I))) if I is not 0. */
+ if (GET_CODE (op) == LSHIFTRT
+ && CONST_INT_P (XEXP (op, 1))
+ && XEXP (op, 1) != const0_rtx)
+ return simplify_gen_unary (ZERO_EXTEND, mode, op, GET_MODE (op));
+
#if defined(POINTERS_EXTEND_UNSIGNED)
/* As we do not know which address space the pointer is referring to,
we can do this only if the target does not support different pointer
--
1.9.3
