So here's the next chunk of conditional move work from Shreya.
It's been a long standing wart that the conditional move expander does
not support sub-word operands in the comparison. Particularly since we
have support routines to handle the necessary extensions for that case.
This patch adjusts the expander to use riscv_extend_comparands rather
than fail for that case. I've built spec2017 before/after this and we
definitely get more conditional moves and they look sensible from a
performance standpoint. None are likely hitting terribly hot code, so I
wouldn't expect any performance jumps.
Waiting on pre-commit testing to do its thing.
Jeff
* config/riscv/riscv.cc (riscv_expand_conditional_move): Use
riscv_extend_comparands to extend sub-word comparison arguments.
diff --git a/gcc/config/riscv/riscv.cc b/gcc/config/riscv/riscv.cc
index 99eeba64b6f..dd29059412b 100644
--- a/gcc/config/riscv/riscv.cc
+++ b/gcc/config/riscv/riscv.cc
@@ -5436,13 +5436,18 @@ riscv_expand_conditional_move (rtx dest, rtx op, rtx
cons, rtx alt)
machine_mode mode0 = GET_MODE (op0);
machine_mode mode1 = GET_MODE (op1);
- /* An integer comparison must be comparing WORD_MODE objects. We
- must enforce that so that we don't strip away a sign_extension
- thinking it is unnecessary. We might consider using
- riscv_extend_operands if they are not already properly extended. */
+ /* An integer comparison must be comparing WORD_MODE objects.
+ Extend the comparison arguments as necessary. */
if ((INTEGRAL_MODE_P (mode0) && mode0 != word_mode)
|| (INTEGRAL_MODE_P (mode1) && mode1 != word_mode))
- return false;
+ riscv_extend_comparands (code, &op0, &op1);
+
+ /* We might have been handed back a SUBREG. Just to make things
+ easy, force it into a REG. */
+ if (!REG_P (op0) && !CONST_INT_P (op0))
+ op0 = force_reg (word_mode, op0);
+ if (!REG_P (op1) && !CONST_INT_P (op1))
+ op1 = force_reg (word_mode, op1);
/* In the fallback generic case use MODE rather than WORD_MODE for
the output of the SCC instruction, to match the mode of the NEG
