[PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[Jakub Jelinek]

Hi!

This simple patch improves the f3 function in the testcase below,
a parameter with TYPE_RESTRICT REFERENCE_TYPE IMHO can be safely treated
like the DECL_BY_REFERENCE case where the source actually didn't contain
a reference, but compiler created it anyway.  If source contains &__restrict
parameter, the parameter again points to a chunk of (for the function
global) restrict memory that nothing inside of the function should access
otherwise than through this parameter or pointers derived from it.

Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?

2011-09-23  Jakub Jelinek  <jakub@redhat.com>

	* tree-ssa-structalias.c (intra_create_variable_infos): Treat
	TYPE_RESTRICT REFERENCE_TYPE parameters like restricted
	DECL_BY_REFERENCE parameters.

	* g++.dg/tree-ssa/restrict2.C: New test.

--- gcc/tree-ssa-structalias.c.jj	2011-09-15 12:18:37.000000000 +0200
+++ gcc/tree-ssa-structalias.c	2011-09-23 15:36:23.000000000 +0200
@@ -5589,10 +5589,11 @@ intra_create_variable_infos (void)
       varinfo_t p;
 
       /* For restrict qualified pointers to objects passed by
-         reference build a real representative for the pointed-to object.  */
-      if (DECL_BY_REFERENCE (t)
-	  && POINTER_TYPE_P (TREE_TYPE (t))
-	  && TYPE_RESTRICT (TREE_TYPE (t)))
+         reference build a real representative for the pointed-to object.
+	 Treat restrict qualified references the same.  */
+      if (TYPE_RESTRICT (TREE_TYPE (t))
+	  && ((DECL_BY_REFERENCE (t) && POINTER_TYPE_P (TREE_TYPE (t)))
+	      || TREE_CODE (TREE_TYPE (t)) == REFERENCE_TYPE))
 	{
 	  struct constraint_expr lhsc, rhsc;
 	  varinfo_t vi;
--- gcc/testsuite/g++.dg/tree-ssa/restrict2.C.jj	2011-09-23 16:11:27.000000000 +0200
+++ gcc/testsuite/g++.dg/tree-ssa/restrict2.C	2011-09-23 16:10:28.000000000 +0200
@@ -0,0 +1,62 @@
+// { dg-do compile }
+// { dg-options "-O2 -fdump-tree-optimized" }
+
+struct S { int *__restrict p; int q; };
+S s;
+
+int
+f1 (S x, S y)
+{
+  x.p[0] = 1;
+  y.p[0] = 0;
+// { dg-final { scan-tree-dump-times "return 1" 1 "optimized" } }
+  return x.p[0];
+}
+
+int
+f2 (S x)
+{
+  x.p[0] = 2;
+  s.p[0] = 0;
+// { dg-final { scan-tree-dump-times "return 2" 1 "optimized" } }
+  return x.p[0];
+}
+
+int
+f3 (S &__restrict x, S &__restrict y)
+{
+  x.p[0] = 3;
+  y.p[0] = 0;
+// { dg-final { scan-tree-dump-times "return 3" 1 "optimized" } }
+  return x.p[0];
+}
+
+int
+f4 (S &x, S &y)
+{
+  x.p[0] = 4;
+  y.p[0] = 0;
+// { dg-final { scan-tree-dump-times "return 4" 0 "optimized" } }
+  return x.p[0];
+}
+
+int
+f5 (S *__restrict x, S *__restrict y)
+{
+  x->p[0] = 5;
+  y->p[0] = 0;
+// We might handle this some day
+// { dg-final { scan-tree-dump-times "return 5" 0 "optimized" } }
+  return x->p[0];
+}
+
+int
+f6 (S *x, S *y)
+{
+  x->p[0] = 6;
+  y->p[0] = 0;
+// { dg-final { scan-tree-dump-times "return 6" 0 "optimized" } }
+  return x->p[0];
+}
+
+// { dg-final { cleanup-tree-dump "optimized" } }

	Jakub

Re: [PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[Richard Guenther]

On Fri, Sep 23, 2011 at 7:06 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> Hi!
>
> This simple patch improves the f3 function in the testcase below,
> a parameter with TYPE_RESTRICT REFERENCE_TYPE IMHO can be safely treated
> like the DECL_BY_REFERENCE case where the source actually didn't contain
> a reference, but compiler created it anyway.  If source contains &__restrict
> parameter, the parameter again points to a chunk of (for the function
> global) restrict memory that nothing inside of the function should access
> otherwise than through this parameter or pointers derived from it.
>
> Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
>
> 2011-09-23  Jakub Jelinek  <jakub@redhat.com>
>
>        * tree-ssa-structalias.c (intra_create_variable_infos): Treat
>        TYPE_RESTRICT REFERENCE_TYPE parameters like restricted
>        DECL_BY_REFERENCE parameters.
>
>        * g++.dg/tree-ssa/restrict2.C: New test.
>
> --- gcc/tree-ssa-structalias.c.jj       2011-09-15 12:18:37.000000000 +0200
> +++ gcc/tree-ssa-structalias.c  2011-09-23 15:36:23.000000000 +0200
> @@ -5589,10 +5589,11 @@ intra_create_variable_infos (void)
>       varinfo_t p;
>
>       /* For restrict qualified pointers to objects passed by
> -         reference build a real representative for the pointed-to object.  */
> -      if (DECL_BY_REFERENCE (t)
> -         && POINTER_TYPE_P (TREE_TYPE (t))
> -         && TYPE_RESTRICT (TREE_TYPE (t)))
> +         reference build a real representative for the pointed-to object.
> +        Treat restrict qualified references the same.  */
> +      if (TYPE_RESTRICT (TREE_TYPE (t))
> +         && ((DECL_BY_REFERENCE (t) && POINTER_TYPE_P (TREE_TYPE (t)))
> +             || TREE_CODE (TREE_TYPE (t)) == REFERENCE_TYPE))
>        {
>          struct constraint_expr lhsc, rhsc;
>          varinfo_t vi;

I don't see why

  f4 (s, s)

would be invalid.  But you would miscompile it.
(I'm not sure that a restrict qualified component is properly defined
by the C standard - we're just making this extension in a very constrained
case to allow Fortran array descriptors to work).

Richard.

> +
> +int
> +f2 (S x)
> +{
> +  x.p[0] = 2;
> +  s.p[0] = 0;
> +// { dg-final { scan-tree-dump-times "return 2" 1 "optimized" } }
> +  return x.p[0];
> +}
> +
> +int
> +f3 (S &__restrict x, S &__restrict y)
> +{
> +  x.p[0] = 3;
> +  y.p[0] = 0;
> +// { dg-final { scan-tree-dump-times "return 3" 1 "optimized" } }
> +  return x.p[0];
> +}
> +
> +int
> +f4 (S &x, S &y)
> +{
> +  x.p[0] = 4;
> +  y.p[0] = 0;
> +// { dg-final { scan-tree-dump-times "return 4" 0 "optimized" } }
> +  return x.p[0];
> +}
> +
> +int
> +f5 (S *__restrict x, S *__restrict y)
> +{
> +  x->p[0] = 5;
> +  y->p[0] = 0;
> +// We might handle this some day
> +// { dg-final { scan-tree-dump-times "return 5" 0 "optimized" } }
> +  return x->p[0];
> +}
> +
> +int
> +f6 (S *x, S *y)
> +{
> +  x->p[0] = 6;
> +  y->p[0] = 0;
> +// { dg-final { scan-tree-dump-times "return 6" 0 "optimized" } }
> +  return x->p[0];
> +}
> +
> +// { dg-final { cleanup-tree-dump "optimized" } }
>
>        Jakub
>

Re: [PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[Jakub Jelinek]

On Sat, Sep 24, 2011 at 01:26:36PM +0200, Richard Guenther wrote:
> > +int
> > +f3 (S &__restrict x, S &__restrict y)
> > +{
> > +  x.p[0] = 3;
> > +  y.p[0] = 0;
> > +// { dg-final { scan-tree-dump-times "return 3" 1 "optimized" } }
> > +  return x.p[0];
> > +}
> > +
> > +int
> > +f4 (S &x, S &y)
> > +{
> > +  x.p[0] = 4;
> > +  y.p[0] = 0;
> > +// { dg-final { scan-tree-dump-times "return 4" 0 "optimized" } }
> > +  return x.p[0];
> > +}

> I don't see why
> 
>   f4 (s, s)
> 
> would be invalid.  But you would miscompile it.

f3 (s, s) you mean?  I believe it is invalid.  For f4 it would be valid
and not optimized out.

> (I'm not sure that a restrict qualified component is properly defined
> by the C standard - we're just making this extension in a very constrained
> case to allow Fortran array descriptors to work).

Well, C standard doesn't have references, and C++ doesn't have restrict.
So it is all about extensions.
But what else would be & __restrict for than similar to *__restrict
to say that the pointed (resp. referenced) object must not be accessed
through other means than the reference or references/pointers derived from
it, in the spirit of ISO C99 6.7.3.1.
So, before jumping to __restrict fields, consider
int a[10], b[10];
int *
f8 (S &__restrict x, S &__restrict y)
{
  x.p = a;
  y.p = b;
  return x.p;
}
which we already optimize even before the patch.
It is certainly invalid to call f8 (s, s).

And the restricted fields, it is a straightforward extension to the restrict
definition of ISO C99.  We don't use it just for Fortran descriptors, but
e.g. std::valarray uses __restrict fields too and has that backed up by the
C++ standard requirements.  Two different std::valarray objects will have
different pointers inside of the structure.

My intent currently is to be able to vectorize:
#include <valarray>

std::valarray<int>
f9 (std::valarray<int> a, std::valarray<int> b, std::valarray<int> c, int z)
{
  int i;
  for (i = 0; i < z; i++)
    {
      a[i] = b[i] + c[i];
      a[i] += b[i] * c[i];
    }
  return a;
}

void
f10 (std::valarray<int> &__restrict a, std::valarray<int> &__restrict b, std::valarray<int> &__restrict c, int z)
{
  int i;
  for (i = 0; i < z; i++)
    {
      a[i] = b[i] + c[i];
      a[i] += b[i] * c[i];
    }
}

In f9 we currently handle it differently than in f10, while IMHO it should
be the same thing, a is guaranteed in both cases not to alias b nor c and b
is guaranteed not to alias c, furthermore, a._M_data[0] through a._M_data[z-1]
is guaranteed not to alias b._M_data[0] through b._M_data[z-1] and c._M_data[0]
through c._M_data[z-1] and similarly for b vs. c.  The __restrict on the
_M_data field in std::valarray is a hint that different std::valarray
objects will have different pointers.

In f9 we have:
  size_tD.1850 D.53593;
  intD.9 * restrict D.53592;
  intD.9 & D.53591;
...
  D.53592_7 = MEM[(struct valarrayD.50086 *)aD.50087_6(D) + 8B];
  D.53593_42 = D.53456_5 * 4;
  # PT = nonlocal escaped { D.53660 } (restr)
  D.53591_43 = D.53592_7 + D.53593_42;
...
  *D.53591_43 = D.53462_19;
and while PTA computes the restricted property here, we unfortunately still
don't use it, because D.53591 (which comes from all the inlined wrappers)
isn't TYPE_RESTRICT.  Shouldn't we propagate that property to either
SSA_NAMEs initialized from restricted pointers resp. POINTER_PLUS_EXPRs,
or if it is common to all VAR_DECLs underlying such SSA_NAMEs, to the
VAR_DECLs?

But in f10 we don't get even that far, the a._M_data (which is actually
a->_M_data, since a is a (restricted) reference) load is already itself
not considered as restricted by PTA.

It is nice that we optimize Fortran arrays well, but it would be nice if we
did the same for C++ too.

	Jakub

Re: [PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[Richard Guenther]

On Sat, Sep 24, 2011 at 2:15 PM, Jakub Jelinek <jakub@redhat.com> wrote:
> On Sat, Sep 24, 2011 at 01:26:36PM +0200, Richard Guenther wrote:
>> > +int
>> > +f3 (S &__restrict x, S &__restrict y)
>> > +{
>> > +  x.p[0] = 3;
>> > +  y.p[0] = 0;
>> > +// { dg-final { scan-tree-dump-times "return 3" 1 "optimized" } }
>> > +  return x.p[0];
>> > +}
>> > +
>> > +int
>> > +f4 (S &x, S &y)
>> > +{
>> > +  x.p[0] = 4;
>> > +  y.p[0] = 0;
>> > +// { dg-final { scan-tree-dump-times "return 4" 0 "optimized" } }
>> > +  return x.p[0];
>> > +}
>
>> I don't see why
>>
>>   f4 (s, s)
>>
>> would be invalid.  But you would miscompile it.
>
> f3 (s, s) you mean?  I believe it is invalid.  For f4 it would be valid
> and not optimized out.

Ah, I misread the dump-test.

>> (I'm not sure that a restrict qualified component is properly defined
>> by the C standard - we're just making this extension in a very constrained
>> case to allow Fortran array descriptors to work).
>
> Well, C standard doesn't have references, and C++ doesn't have restrict.
> So it is all about extensions.
> But what else would be & __restrict for than similar to *__restrict
> to say that the pointed (resp. referenced) object must not be accessed
> through other means than the reference or references/pointers derived from
> it, in the spirit of ISO C99 6.7.3.1.
> So, before jumping to __restrict fields, consider
> int a[10], b[10];
> int *
> f8 (S &__restrict x, S &__restrict y)
> {
>  x.p = a;
>  y.p = b;
>  return x.p;
> }
> which we already optimize even before the patch.
> It is certainly invalid to call f8 (s, s).
>
> And the restricted fields, it is a straightforward extension to the restrict
> definition of ISO C99.  We don't use it just for Fortran descriptors, but
> e.g. std::valarray uses __restrict fields too and has that backed up by the
> C++ standard requirements.  Two different std::valarray objects will have
> different pointers inside of the structure.
>
> My intent currently is to be able to vectorize:
> #include <valarray>
>
> std::valarray<int>
> f9 (std::valarray<int> a, std::valarray<int> b, std::valarray<int> c, int z)
> {
>  int i;
>  for (i = 0; i < z; i++)
>    {
>      a[i] = b[i] + c[i];
>      a[i] += b[i] * c[i];
>    }
>  return a;
> }
>
> void
> f10 (std::valarray<int> &__restrict a, std::valarray<int> &__restrict b, std::valarray<int> &__restrict c, int z)
> {
>  int i;
>  for (i = 0; i < z; i++)
>    {
>      a[i] = b[i] + c[i];
>      a[i] += b[i] * c[i];
>    }
> }
>
> In f9 we currently handle it differently than in f10, while IMHO it should
> be the same thing, a is guaranteed in both cases not to alias b nor c and b
> is guaranteed not to alias c, furthermore, a._M_data[0] through a._M_data[z-1]
> is guaranteed not to alias b._M_data[0] through b._M_data[z-1] and c._M_data[0]
> through c._M_data[z-1] and similarly for b vs. c.  The __restrict on the
> _M_data field in std::valarray is a hint that different std::valarray
> objects will have different pointers.

Ok, I'm just worried that people get bitten by this (given the two existing
wrong-code issues we still have with restrict tracking and inlining).

But yes, your patch looks safe to me.  Maybe we can document how GCC
treats restrict qualification of structure members?

>
> In f9 we have:
>  size_tD.1850 D.53593;
>  intD.9 * restrict D.53592;
>  intD.9 & D.53591;
> ...
>  D.53592_7 = MEM[(struct valarrayD.50086 *)aD.50087_6(D) + 8B];
>  D.53593_42 = D.53456_5 * 4;
>  # PT = nonlocal escaped { D.53660 } (restr)
>  D.53591_43 = D.53592_7 + D.53593_42;
> ...
>  *D.53591_43 = D.53462_19;
> and while PTA computes the restricted property here, we unfortunately still
> don't use it, because D.53591 (which comes from all the inlined wrappers)
> isn't TYPE_RESTRICT.

Yeah, that's extra safety checks because we ignore nonlocal/escaped when
applying the restrict tag match.  Otherwise the restrict tags are prone to leak
into other variables solutions.  Maybe finally fixing that two
wrong-code restrict
bugs will solve this issue though.

>  Shouldn't we propagate that property to either
> SSA_NAMEs initialized from restricted pointers resp. POINTER_PLUS_EXPRs,
> or if it is common to all VAR_DECLs underlying such SSA_NAMEs, to the
> VAR_DECLs?

I'm not sure where to best do that.  In principle we shouldn't need to check
TYPE_RESTRICT at all, but that requires some thoughts.

> But in f10 we don't get even that far, the a._M_data (which is actually
> a->_M_data, since a is a (restricted) reference) load is already itself
> not considered as restricted by PTA.
>
> It is nice that we optimize Fortran arrays well, but it would be nice if we
> did the same for C++ too.

Yeah, I agree.

Your patch is ok.

Thanks,
Richard.

>        Jakub
>

Re: [PATCH] Handle &__restrict parameters in tree-ssa-structalias.c like DECL_BY_REFERENCE parameters

[Jason Merrill]

On 09/24/2011 07:26 AM, Richard Guenther wrote:
> I don't see why
>
>    f4 (s, s)
>
> would be invalid.  But you would miscompile it.

>> +int
>> +f4 (S&x, S&y)
>> +{
>> +  x.p[0] = 4;
>> +  y.p[0] = 0; // { dg-final { scan-tree-dump-times "return 4" 0 "optimized" } }
>> +  return x.p[0];
>> +}

It looks to me like the testcase is testing that we *don't* optimize f4, 
which I think is the correct result.

> +// We might handle this some day
> +// { dg-final { scan-tree-dump-times "return 5" 0 "optimized" } }

But we could optimize f5, so I don't think we want to test for not 
optimizing.  Better would be to test for the optimization, but mark it 
as xfail.

Jason