From: Andrew Pinski <quic_apinski@quicinc.com>
Date: Tue, 11 Feb 2025 23:13:07 +0000 (-0800)
Subject: ifcvt: Don't speculation move inline-asm [PR102150]
X-Git-Tag: basepoints/gcc-16~2058
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=aa972d027437784686dfc66180dc1b640e7dbb39;p=thirdparty%2Fgcc.git

ifcvt: Don't speculation move inline-asm [PR102150]

So unlike loop invariant motion, moving an inline-asm out of an
if is not always profitable and the cost estimate for the instruction
inside inline-asm is unknown.

This is a regression from GCC 4.6 which didn't speculatively move inline-asm
as far as I can tell.
Bootstrapped and tested on x86_64-linux-gnu.

	PR rtl-optimization/102150
gcc/ChangeLog:

	* ifcvt.cc (cheap_bb_rtx_cost_p): Return false if the insn
	has an inline-asm in it.

Signed-off-by: Andrew Pinski <quic_apinski@quicinc.com>
---

diff --git a/gcc/ifcvt.cc b/gcc/ifcvt.cc
index cb5597bc171..6d67bef41c8 100644
--- a/gcc/ifcvt.cc
+++ b/gcc/ifcvt.cc
@@ -166,6 +166,13 @@ cheap_bb_rtx_cost_p (const_basic_block bb,
     {
       if (NONJUMP_INSN_P (insn))
 	{
+	  /* Inline-asm's cost is not very estimatable.
+	     It could be a costly instruction but the
+	     estimate would be the same as a non costly
+	     instruction.  */
+	  if (asm_noperands (PATTERN (insn)) >= 0)
+	    return false;
+
 	  int cost = insn_cost (insn, speed) * REG_BR_PROB_BASE;
 	  if (cost == 0)
 	    return false;