[PATCH] s390: Fix s390_const_int_pool_entry_p and movdi peephole2 [PR114605]

[PATCH] s390: Fix s390_const_int_pool_entry_p and movdi peephole2 [PR114605]

[Jakub Jelinek]

Hi!

The following testcase is miscompiled, because we have initially
a movti which loads the 0x3f8000003f800000ULL TImode constant
from constant pool.  Later on we split it into a pair of DImode
loads.  Now, for the first load (why just that?, though not stage4
material) we trigger the peephole2 which uses s390_const_int_pool_entry_p.
That function doesn't check at all the constant pool mode though, sees
the constant pool at that address has a CONST_INT value and just assumes
that is the value to return, which is especially wrong for big-endian,
if it is a DImode load from offset 0, it should be loading 0 rather than
0x3f8000003f800000ULL.
The following patch adds checks if we are extracing a MODE_INT mode,
if the constant pool has MODE_INT mode as well, punts if constant pool
has smaller mode size than the extraction one (then it would be UB),
if it has the same mode as before keeps using what it did before,
if constant pool has a larger mode than the one being extracted, uses
simplify_subreg.  I'd have used avoid_constant_pool_reference
instead which can handle also offsets into the constant pool constants,
but it can't handle UNSPEC_LTREF.

Another thing is that once that is fixed, we ICE when we extract constant
like 0, ior insn predicate require non-0 constant.  So, the patch also
fixes the peephole2 so that if either 32-bit half is zero, it uses a mere
load of the constant into register rather than a pair of such load and ior.

Bootstrapped/regtested on s390x-linux, ok for trunk?

2024-04-06  Jakub Jelinek  <jakub@redhat.com>

	PR target/114605
	* config/s390/s390.cc (s390_const_int_pool_entry_p): Punt
	if mem doesn't have MODE_INT mode, or pool constant doesn't
	have MODE_INT mode, or if pool constant mode is smaller than
	mem mode.  If mem mode is different from pool constant mode,
	try to simplify subreg.  If that doesn't work, punt, if it
	does, use the simplified constant instead of the constant pool
	constant.
	* config/s390/s390.md (movdi from const pool peephole): If
	either low or high 32-bit part is zero, just emit move insn
	instead of move + ior.

	* gcc.dg/pr114605.c: New test.

--- gcc/config/s390/s390.cc.jj	2024-03-14 14:07:34.088426911 +0100
+++ gcc/config/s390/s390.cc	2024-04-05 15:58:57.757057420 +0200
@@ -9984,7 +9984,7 @@ s390_const_int_pool_entry_p (rtx mem, HO
      - (mem (unspec [(symbol_ref) (reg)] UNSPEC_LTREF)).
      - (mem (symbol_ref)).  */
 
-  if (!MEM_P (mem))
+  if (!MEM_P (mem) || GET_MODE_CLASS (GET_MODE (mem)) != MODE_INT)
     return false;
 
   rtx addr = XEXP (mem, 0);
@@ -9998,9 +9998,19 @@ s390_const_int_pool_entry_p (rtx mem, HO
     return false;
 
   rtx val_rtx = get_pool_constant (sym);
-  if (!CONST_INT_P (val_rtx))
+  machine_mode mode = get_pool_mode (sym);
+  if (!CONST_INT_P (val_rtx)
+      || GET_MODE_CLASS (mode) != MODE_INT
+      || GET_MODE_SIZE (mode) < GET_MODE_SIZE (GET_MODE (mem)))
     return false;
 
+  if (mode != GET_MODE (mem))
+    {
+      val_rtx = simplify_subreg (GET_MODE (mem), val_rtx, mode, 0);
+      if (val_rtx == NULL_RTX || !CONST_INT_P (val_rtx))
+	return false;
+    }
+
   if (val != nullptr)
     *val = INTVAL (val_rtx);
   return true;
--- gcc/config/s390/s390.md.jj	2024-01-03 11:51:54.638410489 +0100
+++ gcc/config/s390/s390.md	2024-04-05 16:17:17.322234553 +0200
@@ -2152,6 +2152,16 @@ (define_peephole2
   gcc_assert (ok);
   operands[2] = GEN_INT (val & 0xFFFFFFFF00000000ULL);
   operands[3] = GEN_INT (val & 0x00000000FFFFFFFFULL);
+  if (operands[2] == const0_rtx)
+    {
+      emit_move_insn (operands[0], operands[3]);
+      DONE;
+    }
+  else if (operands[3] == const0_rtx)
+    {
+      emit_move_insn (operands[0], operands[2]);
+      DONE;
+    }
 })
 
 ;
--- gcc/testsuite/gcc.dg/pr114605.c.jj	2024-04-05 16:25:34.678505438 +0200
+++ gcc/testsuite/gcc.dg/pr114605.c	2024-04-05 16:25:10.388834268 +0200
@@ -0,0 +1,37 @@
+/* PR target/114605 */
+/* { dg-do run } */
+/* { dg-options "-O0" } */
+
+typedef struct { const float *a; int b, c; float *d; } S;
+
+__attribute__((noipa)) void
+bar (void)
+{
+}
+
+__attribute__((noinline, optimize (2))) static void
+foo (S *e)
+{
+  const float *f;
+  float *g;
+  float h[4] = { 0.0, 0.0, 1.0, 1.0 };
+  if (!e->b)
+    f = h;
+  else
+    f = e->a;
+  g = &e->d[0];
+  __builtin_memcpy (g, f, sizeof (float) * 4);
+  bar ();
+  if (!e->b)
+    if (g[0] != 0.0 || g[1] != 0.0 || g[2] != 1.0 || g[3] != 1.0)
+      __builtin_abort ();
+}
+
+int
+main ()
+{
+  float d[4];
+  S e = { .d = d };
+  foo (&e);
+  return 0;
+}

	Jakub

Re: [PATCH] s390: Fix s390_const_int_pool_entry_p and movdi peephole2 [PR114605]

[Ilya Leoshkevich]

On Sat, 2024-04-06 at 18:58 +0200, Jakub Jelinek wrote:
> Hi!
> 
> The following testcase is miscompiled, because we have initially
> a movti which loads the 0x3f8000003f800000ULL TImode constant
> from constant pool.  Later on we split it into a pair of DImode
> loads.  Now, for the first load (why just that?, though not stage4
> material) we trigger the peephole2 which uses
> s390_const_int_pool_entry_p.
> That function doesn't check at all the constant pool mode though,
> sees
> the constant pool at that address has a CONST_INT value and just
> assumes
> that is the value to return, which is especially wrong for big-
> endian,
> if it is a DImode load from offset 0, it should be loading 0 rather
> than
> 0x3f8000003f800000ULL.
> The following patch adds checks if we are extracing a MODE_INT mode,
> if the constant pool has MODE_INT mode as well, punts if constant
> pool
> has smaller mode size than the extraction one (then it would be UB),
> if it has the same mode as before keeps using what it did before,
> if constant pool has a larger mode than the one being extracted, uses
> simplify_subreg.  I'd have used avoid_constant_pool_reference
> instead which can handle also offsets into the constant pool
> constants,
> but it can't handle UNSPEC_LTREF.
> 
> Another thing is that once that is fixed, we ICE when we extract
> constant
> like 0, ior insn predicate require non-0 constant.  So, the patch
> also
> fixes the peephole2 so that if either 32-bit half is zero, it uses a
> mere
> load of the constant into register rather than a pair of such load
> and ior.
> 
> Bootstrapped/regtested on s390x-linux, ok for trunk?

Hi Jakub, thanks for the patch, it looks good to me.
Since I'm not a maintainer, we need to wait for Andreas' opinion.

> 
> 2024-04-06  Jakub Jelinek  <jakub@redhat.com>
> 
> 	PR target/114605
> 	* config/s390/s390.cc (s390_const_int_pool_entry_p): Punt
> 	if mem doesn't have MODE_INT mode, or pool constant doesn't
> 	have MODE_INT mode, or if pool constant mode is smaller than
> 	mem mode.  If mem mode is different from pool constant mode,
> 	try to simplify subreg.  If that doesn't work, punt, if it
> 	does, use the simplified constant instead of the constant
> pool
> 	constant.
> 	* config/s390/s390.md (movdi from const pool peephole): If
> 	either low or high 32-bit part is zero, just emit move insn
> 	instead of move + ior.
> 
> 	* gcc.dg/pr114605.c: New test.

Re: [PATCH] s390: Fix s390_const_int_pool_entry_p and movdi peephole2 [PR114605]

[Andreas Krebbel]

On 4/8/24 13:43, Ilya Leoshkevich wrote:
> On Sat, 2024-04-06 at 18:58 +0200, Jakub Jelinek wrote:
>> Hi!
>>
>> The following testcase is miscompiled, because we have initially
>> a movti which loads the 0x3f8000003f800000ULL TImode constant
>> from constant pool.  Later on we split it into a pair of DImode
>> loads.  Now, for the first load (why just that?, though not stage4
>> material) we trigger the peephole2 which uses
>> s390_const_int_pool_entry_p.
>> That function doesn't check at all the constant pool mode though,
>> sees
>> the constant pool at that address has a CONST_INT value and just
>> assumes
>> that is the value to return, which is especially wrong for big-
>> endian,
>> if it is a DImode load from offset 0, it should be loading 0 rather
>> than
>> 0x3f8000003f800000ULL.
>> The following patch adds checks if we are extracing a MODE_INT mode,
>> if the constant pool has MODE_INT mode as well, punts if constant
>> pool
>> has smaller mode size than the extraction one (then it would be UB),
>> if it has the same mode as before keeps using what it did before,
>> if constant pool has a larger mode than the one being extracted, uses
>> simplify_subreg.  I'd have used avoid_constant_pool_reference
>> instead which can handle also offsets into the constant pool
>> constants,
>> but it can't handle UNSPEC_LTREF.
>>
>> Another thing is that once that is fixed, we ICE when we extract
>> constant
>> like 0, ior insn predicate require non-0 constant.  So, the patch
>> also
>> fixes the peephole2 so that if either 32-bit half is zero, it uses a
>> mere
>> load of the constant into register rather than a pair of such load
>> and ior.
>>
>> Bootstrapped/regtested on s390x-linux, ok for trunk?
> 
> Hi Jakub, thanks for the patch, it looks good to me.
> Since I'm not a maintainer, we need to wait for Andreas' opinion.

Ok. Thank you very much Jakub for fixing this!

Andreas

> 
>>
>> 2024-04-06  Jakub Jelinek  <jakub@redhat.com>
>>
>> 	PR target/114605
>> 	* config/s390/s390.cc (s390_const_int_pool_entry_p): Punt
>> 	if mem doesn't have MODE_INT mode, or pool constant doesn't
>> 	have MODE_INT mode, or if pool constant mode is smaller than
>> 	mem mode.  If mem mode is different from pool constant mode,
>> 	try to simplify subreg.  If that doesn't work, punt, if it
>> 	does, use the simplified constant instead of the constant
>> pool
>> 	constant.
>> 	* config/s390/s390.md (movdi from const pool peephole): If
>> 	either low or high 32-bit part is zero, just emit move insn
>> 	instead of move + ior.
>>
>> 	* gcc.dg/pr114605.c: New test.