[gcc r17-3532] [committed] Fix minor arc testsuite regression

https://gcc.gnu.org/g:b236deea5082eb240da32096845193ec163666ba

commit r17-3532-gb236deea5082eb240da32096845193ec163666ba
Author: Jeff Law <[email protected]>
Date:   Sat Aug 22 11:05:59 2026 -0600

    [committed] Fix minor arc testsuite regression
    
    This change:
    
    > commit aac02c9666184d19334a0527ec444b5e55fbd113 (HEAD)
    > Author: Konstantinos Eleftheriou <[email protected]>
    > Date:   Thu Jul 9 02:22:54 2026 -0700
    >
    >     widening_mul: Lower long-multiply chains to inline longhand
    >
    >     The forwprop long-multiply recognizer canonicalizes longhand
    >     high-part multiplies into
    >
    >       lhs = (N) (((2N) op1 * (2N) op2) >> N)
    [ ... ]
    
    Cause a minor testsuite regression on the arc port.  Basically it wants to see
    a all to muldi, but instead we component SI multiplies.  AFAICT the patch is
    behaving per design, so I've just twiddled the testsuite by adding
    -fno-expensive-optimizations  which restores the older code generation and the
    intent of the test.
    
    gcc/testsuite
            * gcc.target/arc/mulsi3_highpart-2.c: Disable -fexpensive-optimizations.

Diff:
---
 gcc/testsuite/gcc.target/arc/mulsi3_highpart-2.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gcc/testsuite/gcc.target/arc/mulsi3_highpart-2.c b/gcc/testsuite/gcc.target/arc/mulsi3_highpart-2.c
index 22b28cf25070..d478ad43810e 100644
--- a/gcc/testsuite/gcc.target/arc/mulsi3_highpart-2.c
+++ b/gcc/testsuite/gcc.target/arc/mulsi3_highpart-2.c
@@ -1,7 +1,7 @@
 /* { dg-do run } */
 /* { dg-skip-if "ARC700 always has mpy option on" { arc700 } } */
 /* { dg-skip-if "ARC600 doesn't have mpy instruction" { arc6xx } } */
-/* { dg-options "-O2 -mmpy-option=0 -w -save-temps" } */
+/* { dg-options "-O2 -mmpy-option=0 -w -save-temps -fno-expensive-optimizations" } */
 
 #include <stdlib.h>