[PATCH] s390: Fix up *cmp_and_trap_unsigned_int<mode> constraints [PR104775]

[PATCH] s390: Fix up *cmp_and_trap_unsigned_int<mode> constraints [PR104775]

[Jakub Jelinek]

Hi!

The following testcase fails to assemble due to clgte %r6,0(%r1,%r10)
insn not being accepted by assembler.
My rough understanding is that in the RSY-b insn format the spot
in other formats used for index registers is used instead for M3 what
kind of comparison it is, so this patch follows what other similar
instructions use for constraint (i.e. one without index register).

Bootstrapped on s390x-linux, regtest there still pending, ok for
trunk if it passes it?

2022-03-05  Jakub Jelinek  <jakub@redhat.com>

	PR target/104775
	* config/s390/s390.md (*cmp_and_trap_unsigned_int<mode>): Use
	S constraint instead of T in the last alternative.

	* gcc.target/s390/pr104775.c: New test.

--- gcc/config/s390/s390.md.jj	2022-02-08 20:08:13.873404137 +0100
+++ gcc/config/s390/s390.md	2022-03-04 14:38:23.252988476 +0100
@@ -9578,7 +9578,7 @@ (define_insn "*cmp_and_trap_signed_int<m
 (define_insn "*cmp_and_trap_unsigned_int<mode>"
   [(trap_if (match_operator 0 "s390_unsigned_integer_comparison"
 	       [(match_operand:GPR 1 "register_operand" "d,d,d")
-		(match_operand:GPR 2 "general_operand"  "d,D,T")])
+		(match_operand:GPR 2 "general_operand"  "d,D,S")])
 	    (const_int 0))]
   "TARGET_Z10"
   "@
--- gcc/testsuite/gcc.target/s390/pr104775.c.jj	2022-03-04 14:49:58.190134898 +0100
+++ gcc/testsuite/gcc.target/s390/pr104775.c	2022-03-04 14:49:42.845352647 +0100
@@ -0,0 +1,14 @@
+/* PR target/104775 */
+/* { dg-do assemble { target s390_zEC12_hw } } */
+/* { dg-options "-O2 -march=zEC12" } */
+
+long a[64];
+void bar (void);
+
+void
+foo (int x, int y)
+{
+  if (x != a[y])
+    bar ();
+  __builtin_trap ();
+}

	Jakub

Re: [PATCH] s390: Fix up *cmp_and_trap_unsigned_int<mode> constraints [PR104775]

[Jakub Jelinek]

On Sat, Mar 05, 2022 at 09:33:58AM +0100, Jakub Jelinek via Gcc-patches wrote:
> Hi!
> 
> The following testcase fails to assemble due to clgte %r6,0(%r1,%r10)
> insn not being accepted by assembler.
> My rough understanding is that in the RSY-b insn format the spot
> in other formats used for index registers is used instead for M3 what
> kind of comparison it is, so this patch follows what other similar
> instructions use for constraint (i.e. one without index register).
> 
> Bootstrapped on s390x-linux, regtest there still pending, ok for
> trunk if it passes it?

Regression testing passed too.

> 2022-03-05  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR target/104775
> 	* config/s390/s390.md (*cmp_and_trap_unsigned_int<mode>): Use
> 	S constraint instead of T in the last alternative.
> 
> 	* gcc.target/s390/pr104775.c: New test.

	Jakub

Re: [PATCH] s390: Fix up *cmp_and_trap_unsigned_int<mode> constraints [PR104775]

[Andreas Krebbel]

On 3/5/22 09:33, Jakub Jelinek wrote:
> Hi!
> 
> The following testcase fails to assemble due to clgte %r6,0(%r1,%r10)
> insn not being accepted by assembler.
> My rough understanding is that in the RSY-b insn format the spot
> in other formats used for index registers is used instead for M3 what
> kind of comparison it is, so this patch follows what other similar
> instructions use for constraint (i.e. one without index register).
> 
> Bootstrapped on s390x-linux, regtest there still pending, ok for
> trunk if it passes it?
> 
> 2022-03-05  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR target/104775
> 	* config/s390/s390.md (*cmp_and_trap_unsigned_int<mode>): Use
> 	S constraint instead of T in the last alternative.
> 
> 	* gcc.target/s390/pr104775.c: New test.

Ok. Thanks for the fix!

Bye,

Andreas