[PATCH] IBM Z: Fix load-and-test peephole2 condition

[PATCH] IBM Z: Fix load-and-test peephole2 condition

[Stefan Schulze Frielinghaus]

For a peephole2 condition variable insn points to the first matched
insn.  In order to refer to the second matched insn use
peep2_next_insn(1) instead.

Bootstrapped and regtested on IBM Z.  Ok for mainline and gcc-{11,10,9}?

gcc/ChangeLog:

	* config/s390/s390.md (define_peephole2): Variable insn points
	to the first matched insn.  Use peep2_next_insn(1) to refer to
	the second matched insn.
---
 gcc/config/s390/s390.md | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/gcc/config/s390/s390.md b/gcc/config/s390/s390.md
index 4debdcd1247..c4f92bde061 100644
--- a/gcc/config/s390/s390.md
+++ b/gcc/config/s390/s390.md
@@ -1003,7 +1003,7 @@
 	(match_operand:GPR 2 "memory_operand"))
    (set (reg CC_REGNUM)
 	(compare (match_dup 0) (match_operand:GPR 1 "const0_operand")))]
-  "s390_match_ccmode(insn, CCSmode) && TARGET_EXTIMM
+  "s390_match_ccmode (peep2_next_insn (1), CCSmode) && TARGET_EXTIMM
    && GENERAL_REG_P (operands[0])
    && satisfies_constraint_T (operands[2])
    && !contains_constant_pool_address_p (operands[2])"
-- 
2.31.1

[PATCH] IBM Z: Fix load-and-test peephole2 condition

[Stefan Schulze Frielinghaus]

For a peephole2 condition variable insn points to the first matched
insn.  In order to refer to the second matched insn use
peep2_next_insn(1) instead.

Update: Added a test case

Bootstrapped and regtested on IBM Z.  Ok for mainline and gcc-{11,10,9}?

gcc/ChangeLog:

	* config/s390/s390.md (define_peephole2): Variable insn points
	to the first matched insn.  Use peep2_next_insn(1) to refer to
	the second matched insn.

gcc/testsuite/ChangeLog:

	* gcc.target/s390/20211119.c: New test.
---
 gcc/config/s390/s390.md                  |  2 +-
 gcc/testsuite/gcc.target/s390/20211119.c | 12 ++++++++++++
 2 files changed, 13 insertions(+), 1 deletion(-)
 create mode 100644 gcc/testsuite/gcc.target/s390/20211119.c

diff --git a/gcc/config/s390/s390.md b/gcc/config/s390/s390.md
index 4debdcd1247..c4f92bde061 100644
--- a/gcc/config/s390/s390.md
+++ b/gcc/config/s390/s390.md
@@ -1003,7 +1003,7 @@
 	(match_operand:GPR 2 "memory_operand"))
    (set (reg CC_REGNUM)
 	(compare (match_dup 0) (match_operand:GPR 1 "const0_operand")))]
-  "s390_match_ccmode(insn, CCSmode) && TARGET_EXTIMM
+  "s390_match_ccmode (peep2_next_insn (1), CCSmode) && TARGET_EXTIMM
    && GENERAL_REG_P (operands[0])
    && satisfies_constraint_T (operands[2])
    && !contains_constant_pool_address_p (operands[2])"
diff --git a/gcc/testsuite/gcc.target/s390/20211119.c b/gcc/testsuite/gcc.target/s390/20211119.c
new file mode 100644
index 00000000000..1a6df4f4b89
--- /dev/null
+++ b/gcc/testsuite/gcc.target/s390/20211119.c
@@ -0,0 +1,12 @@
+/* { dg-do run } */
+/* { dg-options "-Os -march=z10" } */
+signed char a;
+int b = -925974181, c;
+unsigned *d = &b;
+int *e = &c;
+int main() {
+  *e = ((217 ^ a) > 585) < *d;
+  if (c != 1)
+    __builtin_abort();
+  return 0;
+}
-- 
2.31.1

Re: [PATCH] IBM Z: Fix load-and-test peephole2 condition

[Andreas Krebbel]

On 11/19/21 10:45, Stefan Schulze Frielinghaus wrote:
...
> diff --git a/gcc/testsuite/gcc.target/s390/20211119.c b/gcc/testsuite/gcc.target/s390/20211119.c
> new file mode 100644
> index 00000000000..1a6df4f4b89
> --- /dev/null
> +++ b/gcc/testsuite/gcc.target/s390/20211119.c
> @@ -0,0 +1,12 @@
> +/* { dg-do run } */
> +/* { dg-options "-Os -march=z10" } */

Although z10 is pretty old we will need an effective target check here. Ok with that change.

Thanks!

Andreas